INFORMATION TO USERS This manuscript has been reproduced from the microfilm master. UMI films the text directly from the original or copy submitted. Thus, some thesis and dissertation copies are in typewriter face, while others may be from any type of computer printer. The quality of this reproduction is dependent upon the quality of the copy submitted. Broken or indistinct print, colored or poor quality illustrations and photographs, print bleedthrough, substandard margins, and improper alignment can adversely affect reproduction. In the unlikely event that the author did not send UMI a complete manuscript and there are missing pages, these will be noted. Also, if unauthorized copyright material had to be removed, a note will indicate the deletion. Oversize materials (e.g., maps, drawings, charts) are reproduced by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each original is also photographed in one exposure and is included in reduced form at the back of the book. Photographs included in the original manuscript have been reproduced xerographically in this copy. Higher quality 6" x 9" black and white photographic prints are available for any photographs or illustrations appearing in this copy for an additional charge. Contact UMI directly to order. U niversity M icrofilm s International A B ell & H ow ell Inform ation C o m p a n y 3 0 0 N orth Z e e b R o a d . Ann Arbor. Ml 4 8 1 0 6 - 1 3 4 6 U SA 3 1 3 /7 6 1 - 4 7 0 0 8 0 0 /5 2 1 -0 6 0 0 O rd er N u m b e r 9 4 0 6 5 4 7 R esearch u n iv ersity fa cu lty and th e influ ence o f gender: A n e x a m in a tio n o f g en d er’s sign ifican ce in th e p rofession al sa tisfa c tio n s, a llo c a tio n o f w ork effort, and geograph ic m ob ility stra te g ies o f M ich igan S ta te U n iv e r sity ’s facu lty R oels, Shirley Jean, P h .D . Michigan State University, 1993 C o p y righ t © 1 9 9 3 b y R o e ls, S h irley J e a n . A ll rig h ts reserved . UMI 300 N. Zeeb Rd. Ann Arbor, MI 48106 RESEARCH UNIVERSITY FACULTY AND TH E IN FLU EN CE OF GENDER: AN EXAM INATION OF GENDER’S SIGNIFICANCE IN TH E PROFESSIONAL SATISFACTIONS, ALLOCATION OF WORK EFFORT, AND GEOGRAPHIC MOBILITY STRATEGIES OF MICHIGAN STATE UNIVERSITY’S FACULTY By Shirley Jean Roels A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Educational Administration 1993 ABSTRACT RESEARCH UNIVERSITY FACULTY AND THE INFLUENCE OF GENDER: AN EXAM INATION OF G END ER’S SIGNIFICANCE IN THE PROFESSIONAL SATISFACTIONS, ALLOCATION OF WORK EFFORT, AND GEOGRAPHIC MOBILITY STRATEGIES OF M ICHIGAN STATE UNIVERSITY’S FACULTY By Shirley Jean Roels This study investigates whether shifts in the gender balance among faculty in a research university could significantly alter the activity and culture of the research university. It analyzes whether male and female faculty at Michigan State University, a Research I institution, are significantly different from each other in the areas o f professional satisfaction and the allocation o f effort among teaching, research, and other academic activities. The study also tries to assess whether gender is related to perceptions of geographic mobility. To investigate these topics a subset o f data generated by a survey of Michigan State U niversity’s faculty was used. This subset of data was consolidated into a smaller group of dependent variables which were evaluated in relation to independent variables representing sex, age, rank, marital status, the presence o f family children, and university colleges. Two statistical methodologies, ANOVA and regression analysis, were then used to evaluate faculty responses. and geographic mobility. university’s support for it. Female faculty reported less satisfaction with their w ork and the They were also less satisfied with opportunities for professional growth and development than their male counterparts. Women faculty, particularly in the lower ranks, spent a substantially smaller percentage o f their time in research and larger percentage o f time in teaching than did their male colleagues. The difference in teaching effort was directly explained by differences in age and rank, not by gender. However, difference in research effort was directly attributable to gender. Yet female faculty members expressed stronger desires than male faculty to restructure their professional efforts, desiring more restructuring for research than did the men. M ale and female faculty did not differ significantly in the relative importance they assigned to teaching, research, and service activities for tenure, promotion, or merit salary increases. Concerning perceived geographic mobility females reported feeling slightly more constrained than did males. Copyright by SHIRLEY JEAN ROELS 1993 V Dedication To John Roels, my husband Daniel and Steven Roels, my children Gertrude and Herbert Wolthuis, my parents for believing that my education could make a Christian difference vii ACKNOWLEDGMENTS I would like to acknowledge the special support and assistance o f Dr. Kathryn Moore, my dissertation committee chairperson. She was a helpful knowledge source, a patient advisor, an insightful critic, an evaluator with high standards, and a supportive friend in this process. I would also like to thank Dr. Stanley Kaplowitz for responding to the many questions I posed to him about statistical methodology and the mysteries of SPSS. I would also like to express my appreciation for Dorothy Britton, my word processing specialist. She turned rough text into beautiful form with speed and accuracy. The knowledge, skill, and dedication o f these three people helped immensely in turning raw data into final form. TABLE OF CONTENTS List o f Tables ............................................................................................................................. ix Chapter 1 - The Research P ro b le m ....................................................................................... 1 Chapter 2 - The Status o f Women Faculty in Higher E d u c a tio n ....................................11 Chapter 3 - Research Methodology ......................................................................................86 Chapter 4 - Statistical R e s u lts ................................................................................................123 Chapter 5 - Discussion of the R e s u l ts ................................................................................. 181 Appendix A - Faculty Mobility S u r v e y .............................................................................205 Bibliography ............................................................................................................................. 221 LIST OF TABLES Chapter 1 Table 1 - Chapter Table 2 Table 3 Table 4 Table 5 2 - Chapter 3 Table 6 Table 7 Table 8 Table 9 Table 10Table 11Table 12Table 13Table 14Table 15- Chapter 4 Table 16Table 17Table 18Table 19- AGE DISTRIBUTION OF REGULAR FULL-TIM E FACULTY, BY GENDER - FALL 1987 ...................................................... 3 PERCENTAGE OF FEMALES BY INSTITUTIONAL TYPE . . . . 14 PERCENTAGE OF WOMEN W ITHIN ACADEMIC RANKS . . . 17 WOMEN PROFESSORS IN 20 RESEARCH UNIVERSITIES . . . 18 PERCENTAGE OF TIM E SPENT ON VARIOUS ACTIVITIES BY FULL-TIM E REGULAR FACULTY, BY TYPE AND CONTROL OF INSTITUTION - 1987 ........................ 82 COLLEGE DISTRIBUTION OF UNIVERSITY FACULTY AND THEIR SURVEY RESPONSE RATES BY GENDER . . . COMPARATIVE RESPONSE RATE BY ACADEMIC RANK . . SURVEY RESPONDENTS BY UNIT OF APPOINTM ENT, RANK, AND GENDER ............................................................ MEAN AGE OF SURVEY RESPONDENTS BY RANK AND S E X .............................. .. ...................................................... USE OF SURVEY Q U E S T IO N S ............................................................ CROSSTABULATION OF RANK AND TENURE STATUS OF SURVEY R E SP O N D E N T S......................................................... CONSOLIDATION OF SELECTED RESEARCH VARIABLES . OTHER SURVEY VARIABLES USED ............................................. FEMALES AS A PROPORTION OF MSU COLLEGE FACULTIES ................................................................................. PERCENTAGE OF TIM E SPENT ON RESEARCH (ROC) FOR MAJORITY SEX AMONG FACULTY BY COLLEGE . . RESPONDENT RESULTS - ONE-WAY A N O V A ........................... ANOVA ON PROFESSIONAL SATISFACTION VARIABLES . . ANOVA ON PROFESSIONAL SATISFACTION VARIABLES - PART II ............................................................ ANOVA ON IM PORTANCE OF PROFESSIONAL SATISFACTION VARIABLES ................................................ . 90 . 91 92 93 104 106 108 110 114 116 124 130 132 134 X Table 20Table 21Table 22Table 23Table 24Table 25Table 26Table 27Table 28Table 29Table 30- Table 31- Table 32- Table 33- Table 34- Table 35- Table 36- Table 37- ANOVA ON ALLOCATION OF WORK TIME VARIABLES ....................................................................................136 ANOVA ON PREFERRED ALLOCATION OF WORK TIM E VARIABLES ........................................................................138 ANOVA ON IM PORTANCE OF TASKS FOR TENURE, RANK, AND MERIT INCREASES .......................................... 140 ANOVA ON GEOGRAPHIC MOBILITY VARIABLES .................... 142 DEPENDENT VARIABLES FOR WHICH NULL HYPOTHESIS CAN BE REJECTED .....................................................................144 DEPENDENT VARIABLE MEANS FOR SIGNIFICANT TWOWAY INTERACTIONS BETWEEN SEX AND MARRY . 155 DEPENDENT VARIABLE MEANS FOR SIGNIFICANT TWOWAY INTERACTIONS BETWEEN SEX AND NCHILD . 156 DEPENDENT VARIABLE MEANS FOR SIGNIFICANT TWOWAY INTERACTIONS BETWEEN MARRY AND NCHILD57 DEPENDENT VARIABLE MEANS FOR SIGNIFICANT TWOWAY INTERACTIONS BETWEEN RANK AND SEX . . 158 DEPENDENT VARIABLE MEANS FOR SIGNIFICANT TWOWAY INTERACTIONS BETWEEN RANK AND NCHILD 160 DEPENDENT VARIABLE, RSCP4, MEANS FOR SIGNIFICANT THREE-W AY INTERACTIONS AMONG SEX, MARRY, AND N C H I L D .................................................................................162 DEPENDENT VARIABLE, RERTRM, MEANS FOR SIGNIFICANT THREE-WAY INTERACTIONS AMONG RANK, MARRY, AND SEX .................................... 163 DEPENDENT VARIABLE, RERTRM, MEANS FOR SIGNIFICANT THREE-WAY INTERACTIONS AMONG RANK, MARRY, AND NCHILD ........................... 164 DEPENDENT VARIABLE, PSEX, MEANS FOR SIGNIFICANT THREE-W AY INTERACTIONS AMONG RANK, MARRY, AND N C H IL D ...............................................................164 DEPENDENT VARIABLE, PTEACH, MEANS FOR SIGNIFICANT THREE-WAY INTERACTIONS AMONG RANK, NCHILD, AND S E X .................................... 165 DEPENDENT VARIABLE, PPDT, MEANS FOR SIGNIFICANT THREE-W AY INTERACTIONS AMONG RANK, N CHILD, AND SEX .....................................................................166 DEPENDENT VARIABLE, PSEX, MEANS FOR SIGNIFICANT THREE-W AY INTERACTIONS AMONG RANK, NCHILD, AND SEX .....................................................................166 DEPENDENT VARIABLE, RETTRM, MEANS FOR SIGNIFICANT THREE-WAY INTERACTIONS AMONG RANK, NCHILD, AND S E X .................................... 167 xi Table 38- Table 39Table 40Table 41Table 42Table 43Table 44Table 45- D EPENDENT VARIABLE, RSCP4, MEANS FOR SIGNIFICANT FOUR-WAY INTERACTIONS AMONG RANK, SEX, MARRY, AND N C H I L D ........................................ 168 RESULTS OF TESTS ON REGRESSION ASSUMPTIONS . . . . 170 ZERO ORDER CORRELATIONS FOR INDEPENDENT VARIABLES ....................................................................................172 REGRESSION RESULTS ON PTEACH - CURRENT PERCENTAGE OF TIME SPENT IN T E A C H IN G .....174 REGRESSION RESULTS ON PRSP - CURRENT PERCENTAGE OF TIM E SPENT IN R E S E A R C H ...................................175 REGRESSION RESULTS ON RSCP1 - DESIRE TO RESTRUCTURE AMOUNT OF TEACHING TIM E . . . . 175 REGRESSION RESULTS ON RSCP2 - DESIRE TO RESTRUCTURE AMOUNT OF RESEARCH TIM E . . . . 176 SUMMARY - REJECTION OF NULL HYPOTHESIS ALLOWED .......................................................................................178 CHAPTER ONE TH E RESEARCH PROBLEM During the next 25 years three quarters o f current higher education faculty will need to be replaced nationwide (Bowen and Schuster, 1986). W hile these replacements will occur across a broad spectrum o f institutions, many new faculty members will be needed in the major research universities o f this nation. As the predominantly male faculty cohort hired during the expansive 1960’s and 1970’s retires from these universities in the years ahead, they could be replaced by a group o f faculty which includes a higher percentage o f women. Could this shift in gender balance among faculty significantly alter the activity and the culture of the research university? A shift in the gender balance o f faculty will affect the research university culture only if female faculty as a group are significantly different from male faculty and these differences can be attributed to gender. If there are differences between male and female faculty but they can be explained by other factors such as rank, age, or chosen discipline, it is possible that the presence o f more female faculty per se will not alter the core o f the research university. In such cases, factors other than gender will be much more significant in any noticeable changes. Thus three significant questions must be addressed to understand whether potential shifts in the gender balance o f faculty will affect research universities in a major way. They are as follows: 1 1. Are male and female faculty in the research university different from each other in key areas which affect university activity and culture? 2. If differences between male and female faculty are significant, can these differences be attributed prim arily and/or directly to gender? 3. If these differences cannot be attributed to gender, what other factors contribute to differences among these men and women? Understanding the national context for these questions is important. W hile female faculty have been involved in higher education in the United States for over 100 years, historically there have been several key differences between the roles and careers o f male and female faculty. Nationwide women faculty are still much more concentrated in parttime positions and in smaller two and four-year institutions o f higher education. Even within doctoral universities women have consistently been concentrated in the lower academic ranks. They have had lower salaries, even when rank was held constant (Dwyer, 1991). However, numerous changes in female socialization patterns, social structures, and institutional policy have begun to change these traditional patterns for academic women. Three basic factors will favor greater employment o f academic women in research universities. First, the supply o f women with doctorates has improved substantially since the 1960’s and early 1970’s, the last substantial hiring period for higher education. By 1986-87 women were awarded 35% o f all doctoral degrees and were increasing their credentials in many fields previously dominated by men (Dwyer, 1991) This shift in supply contrasts sharply with the trends from 1930-1960, when the pool o f academic women with PhD ’s was continually declining (Bernard, 1964). Second, the pool o f female faculty already in the higher education system, from which research universities could draw , is on average younger than the male faculty pool. A comparison o f male and female faculty by age in all non-proprietary accredited U.S. post-secondary institutions that grant a two-year or higher degree is shown in Table 1 to illustrate the differences. TABLE 1 - AGE DISTRIBUTION OF REGULAR FULL-TIM E FA CU LTY, BY GENDER - FALL 1987 Age group Total Under 30 30-44 45-54 55-59 60-64 65 or older Percentage Male 100 1 36 35 14 10 4 Source: SRI International. For the National Center for Education Statistics, 1990 Female 100 3 49 31 8 6 3 4 Thus some women faculty already in the system, who have appropriate professional backgrounds and experiences, may still have several years o f professional effort ahead, years which could be used in research universities. Third, labor market studies indicate that employment fields with growth achieve gender balance more rapidly than occupational fields in decline (Jolly, 1990). The demand of research universities for PhD faculty within the next ten years will be stronger than it has been in two decades (Bowen and Schuster, 1986). In private research institutions 32% o f their faculty expect to retire by the year 2000 and 42% expect to stop teaching. In public research universities, 44% o f their faculty expect to stop teaching by the year 2000 (SRI International, 1990). W hile higher education em ployment as a whole may not grow in the next decade, strong replacement needs within research universities will likely have gender-balancing effects similar to those o f high growth occupational fields. Replacement demand in research universities coupled with a more highly qualified and larger female applicant pool could create a different faculty gender balance. However, a shift in gender balance in research universities will not happen automatically. Older women pioneers may not keep their positions until age 65. The 1988 National Survey o f Post-secondary Faculty found that women tended toward earlier ages of retirement and cessation of teaching (SRI International, 1990). W hile they might be replaceable by younger women academics, this same survey found that younger faculty were more likely to leave academia than older faculty (SRI International, 1990). If younger cohorts o f women faculty have lower retention rates in the profession and older women faculty tend to retire early, there is no guarantee that the supply o f female faculty available to research universities will increase. Questions also remain about their fit with the current configuration o f research institutions. Much o f the literature to date suggests that female faculty have a stronger allegiance to teaching and public service than do male faculty. Jesse Bernard observed this propensity for teaching already in the early 1960’s, noting that women seemed to be particularly effective with poorer students (Bernard, 1964). Boyer’s 1989 survey for the Carnegie Commission indicated that women’s interest in teaching and belief in its value has persisted. Fifteen percent more women than men faculty agreed that "teaching effectiveness should be the prim ary criterion for promotion o f faculty". Eleven percent more women than men faculty agreed that their interests leaned toward or were primarily in teaching rather than research (Boyer, 1989). Aisenberg, in her interviews with women faculty also noted a strong interest in both teaching and public service. (Aisenberg, 1988). Typically, when these general studies o f female faculty find gender-based differences, early socialization, graduate school experience, degree, chosen discipline, age, rank, or relative mobility are cited as possible reasons for differences between male and female academics. Research related to the effects o f socialization, graduate school, degree, and chosen discipline has been described by Aisenberg(1988), Bemard(1964), Dwyer(1991), Feldman(1974), Finkelstein(1984), Simeone(1987), and other researchers. However, little has been done to include institutional type in the analysis. These more general studies do not differentiate research universities from other types o f 4-year institutions o f higher education. Thus, these general inclinations among women faculty cannot be assumed for women in research universities. The findings may simply indicate personal preference; or they may be clouded by the effects o f wom en’s locations in schools focused on teaching and community service and in non-tenured positions within research universities. Research to date does not allow the conclusion that tenured and tenure track women faculty in research universities would necessarily respond with a strong preference for teaching. Little is actually known about whether gender itself is a significant factor for these women faculty. Self-selection and institutional screening may insure that the sources of career satisfaction, w ork priorities, and values among tenure track faculty in research universities, both male and female, are more alike than different. Together these faculty may be necessarily different from those in other types o f higher education institutions. Olsen, in one of the few studies comparing white male and white female faculty members at Indiana University, a research university, concluded that "All three subsamples,{the third subsample being all minority faculty}...,clearly recognize the centrality o f the research mission o f the university ".(Olsen, 1990, p. 20). Further analysis on her part showed only a borderline negative relationship between gender and the allocation o f faculty time. In contrast, rank was a significant positive predictor of the percentage o f time spent on teaching or service and the intrinsic satisfactions which faculty received from their work. (Olsen, 1991). In short, Olsen’s w ork indicates that gender is a factor at work, but may be a much less important characteristic than several others in a research university context. Yet because o f the dearth o f research on women faculty in research universities, the perception lingers that women’s preferences will conflict with the mission o f the research university. In addition, the relative geographic mobility o f women academics, whether by personal choice or social factors, also could have implications for joining and staying in research universities. If women faculty outside research institutions perceive they are less geographically mobile than male faculty, that fact alone could foster different career development strategies among these women. Women with less mobility may look for varied sources o f career satisfaction and allocate their work efforts differently. They may become more interested in teaching, interdisciplinary initiatives, regional public service, or institutional governance. It is also possible that lack o f geographic mobility will constrain opportunities for young women faculty just beginning careers in research universities. While gender and geographical mobility may be connected, little research has been done on this factor’s effect in the research university context. The question o f female faculty’s fit and career development strategies within research universities is further complicated by ongoing shifts in the nature o f these institutions themselves. In The Academic Revolution (1968) Christopher Jencks and David Riesman chronicled the rise o f the 20th century paradigm for research universities within the 8 United States. As the 21st century approaches, the mission and organizational culture o f research universities are under review. In both The Conditions o f the Professoriate (1989) and Scholarship Reconsidered (1990), Dr. Ernest Boyer argues that research universities must change the balance o f their foci to effectively serve the future of society. Research universities must instruct graduate students in the art o f teaching and must also pay more attention to the quality o f learning in their own undergraduate programs. They must also raise the priority given to public service as society reckons with some o f its major social ills. Boyer suggests that even within their research agendas, such schools must recognize and support more diverse research activities. If such changes are adopted the current culture within many U .S. research universities will also be transformed. Thus male/female differences in professional satisfaction and priorities could be further complicated by the changes advocated for the character of research universities themselves. The fit o f women faculty with these institutions will be a dynamic and iterative process as these universities reconfigure their own futures. To add to the knowledge about gender and its impact within research universities, this dissertation explores whether male and female faculty within a particular research university differ in significant areas of professional responsibility. It also examines potential reasons for any differences that are found. To explore these issues a survey o f faculty at Michigan State University, a school classified as a Research I university by the Carnegie Commission, was conducted. The survey was administered by the Collegiate Employment Research Institute, an independent institute housed on Michigan State University’s campus in the early spring o f 1991. All Michigan State University faculty who were either tenured or on tenuretrack appointments, approximately 2100 o f them, were asked to complete this survey. The resulting data base, based on responses from 50.5% o f this faculty population, is used to investigate the three questions posed at the beginning o f this chapter for the faculty of Michigan State University. W hile many areas o f potential differences could be explored, this dissertation will focus on possible differences in the sources and levels o f professional satisfaction, patterns of work load allocation, and geographic mobility plans o f male and female faculty. The environment of Michigan State University is still predominantly one o f tenured white male faculty. The overall percentage o f faculty tenured is 80%. Only 11% o f faculty identify themselves as members of racial and ethnic minorities. Women represent 22% o f the faculty. Within this context are the tenure track and tenured women who are now moving into and progressing within Michigan State University similar to male faculty in their work satisfactions, allocation of work effort, and geographic mobility strategies? When the sample is controlled for age or other relevant variables do they still differ? This study investigates these questions, trying to determine whether, on this particular university 10 campus, gender is a significant source o f differences when compared to the effects o f other variables. In summary, first o f all, the survey o f Michigan State U niversity’s faculty is used to compare characteristics o f women faculty with those o f male faculty occupying similar faculty positions at Michigan State University, a Research I university, investigating professional satisfactions, allocation o f w ork effort, and geographic mobility. Second, statistical models then examine whether those differences are prim arily attributable to gender or to other significant variables in the academic work environment. The second chapter of this dissertation reviews the current literature in areas essential to a further understanding o f the research problem. Chapter three describes both the survey and statistical methodologies used to investigate the research problems that have been posed. Chapter four presents the results o f the study and the analysis that has been undertaken. Chapter five comments on the analysis, noting both the implications o f this study and the limitations o f its findings. CHAPTER 2 TH E STATUS OF W OM EN FACULTY IN HIGHER EDUCATION Since 1960 a diverse body o f literature has developed about the involvement o f women in higher education. This literature can be divided into three significant categories focusing on women as students, wom en’s scholarship in the disciplines, and women professionals within higher education. W hile the first two areas o f study are connected to the study o f women professionals in higher education, the third area o f study is the prim ary focus o f this dissertation. The role o f women professionals in higher education is gradually being documented. Most o f the literature concerns the relative progress and impact o f women as faculty members. Four areas o f questions repeatedly surface in the discussion o f women working as faculty within higher education. They are as follows: 1. W hat are the relative amounts and allocation o f women faculty within the higher education system including distribution by type o f institution, discipline, rank and salary? How are they different than the male distributions? 2. Are female faculty different than male faculty in work-related satisfactions and the allocation o f work effort to teaching, research, and service? 11 12 3. How are career development approaches influenced by graduate school experiences, initial institutional placement, advancement, and geographic mobility? 4. If women faculty differ in their distribution, approaches to their work, or career development why do they differ? Are these differences generated primarily by individual preferences or by institutional/societal settings? W hat predominant factors shape choice? W hile much o f the literature does not isolate the effects o f the research university setting, probing these questions in general provides an important context for understanding this study o f the possible effects o f gender on faculty within a research university. Each o f these questions is addressed below. TH E DISTRIBUTION OF W OM EN FACULTY The numbers o f women faculty within the higher education system can be evaluated in several ways. First, trends in total numbers over the past decades will be examined. Second, distribution by type o f school, discipline, rank, and salary will in turn be reviewed. Some significant changes in the available numbers o f women faculty have occurred in the past century. Prior to 1900 women faculty were often lone pioneers. M aria Sanford, the first woman professor in the 1860’s through about 1900, experienced mixed success and failure. Ellen Richards, the first married women faculty member, taught sanitary 13 chemistry and home economics at MIT for 27 years. Florence Sabin taught in Johns Hopkins Medical School. Vida Schudder worked at W ellesley. Yet these women were often surrounded only by male faculty or a very small cohort o f other women faculty found primarily in the women’s colleges o f the late 1800’s. In 1875 there may have been only four women PhDs in the world (Bernard, 1964). It is not until 1900 that rapid growth in the numbers o f women faculty begins. Many of these women were added because o f their service in colleges o f home economics that grew as the land grant institutions developed. In the decades since 1900 there has been considerable fluctuation in the supply o f women faculty relative to demand. By the late 1930’s women as a proportion o f the total faculty peaked at 27.7% (Bernard, 1964). Women were receiving approximately 13-15% o f all doctorates (Fox, 1989). After the 1930’s the percentage o f women faculty began to decline. By 1950 they were receiving only 10% o f the all doctorates (Fox, 1989); and by 1960 they represented only 22.1% of the faculty of higher education (Bernard, 1964). After 1960 the percentage o f women doctorates began to rise. By 1970 it was 14%; by 1980, 28% ; by 1986, 35% (Fox, 1989). Distribution by Type o f Institution While the available supply o f women faculty has improved since 1960, this supply is not evenly disbursed among higher education institutions. Feldm an’s research documented that women, even expecting a PhD, were more likely to be employed in junior college 14 teaching than men with similar credentials. When research orientations were equivalent between men and women, men still opted for university teaching positions in greater proportions than similar women (Feldman, 1974). Finkelstein (1984) notes that women are disproportionately employed in lower strata institutions, in a concentrated set o f academic disciplines, in lower academic ranks, and in lesser paid positions even after controlling for institutional type, discipline, and rank. His findings are borne out by the data supplied by the institutions themselves. A 1989 study by the Carnegie Foundation for Advancement o f Teaching found that the percentage o f faculty who are women within each type o f institution was as follows: TABLE 2 - PERCENTAGE OF FEMALES BY INSTITUTIONAL TYPE Two year institutions Liberal Arts colleges Comprehensive institutions Doctoral II institutions Doctoral I institutions 40.9% 39.4% 31.4% 22.3% 21.7% Source: Boyer, 1989 These data show some changes over earlier statistics which indicated that women comprised 25.6% o f faculty in two year colleges, 22.7% in four year colleges, and 14.8% in universities in the early 1970’s (Rossi, 1973). However, from this data it is clear that women faculty are more strongly represented in institutions which offer 15 associate and bachelor degrees, not in those that offer graduate work and advanced degrees. It is in undergraduate institutions that women academics experience the greatest change in the gender balance o f the institutions. By comparison women are farther below parity in numbers in larger universities. Between 1978 and 1987 there was only a 1 % increase in women faculty at state and land grant institutions, most o f which offer advanced degrees (Dwyer, 1991). In Table 4, a 1989 study o f 20 leading U .S. research universities shows that only 16.2% o f their faculty members were women (AAUP, 1989). Distribution by Discipline The distribution o f women professionals by discipline is also skewed. Based on data from a 1969 Carnegie Commission Study on higher education, Feldman documented the unequal representation by field. The percentage o f female graduate students in a field was correlated with the sex stereotyping o f the field. If a field was stereotyped as a "female field" it was typically a field that was lower in prestige, economic rewards, and power. The majority o f women concentrated their efforts in such fields. Few women were found as a percentage o f higher prestige fields such as math, computer science, and business. Furtherm ore the female attrition ratios in such traditionally "masculine" fields were typically higher than attrition rates in the fields classed as "feminine" (Feldman, 1974). 16 By 1986 women earned 35% o f all doctorates. Yet o f those doctorates, they earned 32% o f them in education, 13% o f them in psychology, and 9.5% o f them in biology. Only three disciplinary areas accounted for over half o f the doctorates earned by women in 1986. In that year women also outnumber men in doctorates earned by 5 to 1 in home economics, 3 to 1 in languages, and 2 to 1 in communications, education, health, and psychology. In contrast men outnumbered women in doctorates earned by 7 to 1 in engineering, 5 to 1 in theology, and 3 to 1 in agriculture, information sciences, and math (Fox, 1989). Even within a discipline where women doctorates outnumbered those awarded to men, women are often concentrated in a limited number o f specialties. The clustering pattern within various fields that Rossi noted already in the early 1970’s has continued (Rossi, 1973). For example, in psychology women are concentrated in the study of developmental, educational, and social psychology. Few women study cognitive, experimental, industrial, or organizational psychology (Chamberlain, 1988). It is comforting to note that the number o f women in science and engineering increased by 200% from 1972-1982 compared to a 40% increase for males and that engineering and computer science employment at the doctoral level rose from 100 to 700 women in each o f these fields from 1973 to 1981 (Simeone, 1987). Yet despite these shifts, unless there are more radical changes in the ratios, women faculty will continue to be concentrated in certain disciplines, while playing a very small role in others. 17 Distribution bv Rank How are women distributed by rank? Women were concentrated in the lower ranks of the profession in earlier decades. In a 1973 Carnegie Commission report women were 17.5% o f associate professors, but only 8.6% o f full professors. At that time there was a significant dropoff between the associate and full professor ranks among women involved (Fleetham, 1991). Fox (1989) notes that in 1986 women as a percentage o f the total faculty employed in each rank were distributed as follows: TABLE 3 - PERCENTAGE OF W OM EN W ITHIN ACADEMIC RANKS Professor Associate Professor Assistant Professor Instructor Lecturer 12.3% 24.6% 38.4% 53.3% 49.5% Source: Fox, 1989 Going up from the instructor rank their percentage declines about 12-15% with each increase in rank. If one reviews the sample in another way, the Carnegie Foundation for Advancement o f Teaching in a 1989 study (Dwyer, 1991) noted that o f all women 48.3% were in nontenured ranks o f assistant professor, instructor, or lecturer; 33.3% were associate professors, and 19.4% were full professors. Women at associate and full 18 professor ranks together constituted about 17% o f all faculty in those ranks (Dwyer, 1991). When focusing more specifically on research universities, the 1989 AAUP study o f 20 leading research universities confirms the general findings o f the Carnegie Foundation (AAUP, 1989). The AAUP results shown in Table 4 below indicate the paucity o f women faculty in research institutions in general and particularly in the full professor rank. TABLE 4 W OMEN PROFESSORS IN 20 RESEARCH UNIVERSITIES All faculty Women Number % 23,482 3808 16.2 Full Professors Men % 11,524 49.1 Full Professors W omen % 1056 4.5 ’•‘Universities included were UC Berkeley, Stanford, Harvard, Yale, M IT, Princeton, U o f Chicago, UCLA, U o f Michigan, U o f W isconsin, Columbia, U of Illinois, U o f Pennsylvania, Cal Tech, U o f M innesota, U o f Texas, U o f North Carolina, Northwestern, U o f Washington, U o f Arizona Source: AAUP, 1989 Not only is the absolute number o f women very low, the percentage o f those at full professor rank is even smaller. Based on this study, M oore and Sagaria calculated that women are only 8% o f all full professors in these 20 universities which represent a sample o f U .S. research universities (Moore and Sagaria, 1990). 19 To some degree this disparity within research universities reflects differences in professional credentials. The NSOPF-88 faculty survey found that among doctoral institutions 92% o f male faculty had a doctorate or first professional degree and only 76% o f female faculty had such credentials (SRI International, 1990). It is important to note that the aggregate data on rank does not tell the whole story. The magnitude o f the differences in rank distribution varies significantly by academic discipline. Already in the 1970’s disciplines differed in terms o f their distributions of women. Controlling for type and level o f academic degree, the proportion o f women compared to men who were full professors in the field o f physics was far smaller than the proportion in modem languages. O f all the disciplines studied sociology was the greatest laggard in this index o f parity (Rossi, 1973). Distribution bv Salary Do women differ significantly from men in their salary distribution within academia? In 1959-60 Bernard indicated that women professors were paid about $1000.00 less than men in terms o f their median salaries, and that the differential was smaller for instructors than for professors. She foresaw less disparity in salaries by gender because o f the tight academic labor market o f the 1960’s (Bernard, 1964). 20 However her vision did not materialize. According to Dwyer (1991), commenting on Finkelstein’s review o f studies conducted during the 1970’s on salary disparities, "women were paid less than men even after controlling for rank, institutional type, and discipline. Astin and Bayer’s studies cited in Rossi (1973) confirm the same. They likewise note that women are paid less after controlling for rank, background, achievement, and work setting. Finkelstein found that this compensation disparity grew with increased academic rank across the length o f female faculty careers. The differentials were also higher at doctoral granting universities and lower at the liberal arts and community colleges. According to the National Center for Education Statistics (1989), an 18% disparity between full-time female and male faculty compensation existed in 1972 which grew to a 19% disparity in 1982. Female faculty fared worse in the traditional arts and sciences than in the professions although an increased proportion of women in a discipline was not necessarily proven to be related to the pay disparity (Dwyer, 1991). Disparity in salaries by gender did not lessen during the 1970’s. W hat happened in the 1980’s? Ferber and Green had hypothesized in a 1978 study that academic women’s compensation was more related to performance while m en’s was more related to longevity and the nature o f departmental differentials (in Dwyer, 1991). Their 1982 performance-based study four years later showed differences between male and female compensation after degree, field, honors, and number o f publications were held constant (in Pfeffer, 1990). Finkelstein(1984) argues that salaries o f academic women are less predictable than academic men, that there are more intangibles at work, not less, 21 which are unrelated to perform ance and that compensation practices still seem to be defined in terms o f male strengths (Finkelstein, 1984). Yet while gender may still be a factor in salary differences, it may not be as pronounced o r obvious as in earlier decades. Astin and Snyder’s 1982 study concluded that with controls for degree, field, publication record, and type o f institution, salaries for women had gotten closer to those o f men when compared with the findings in an Astin and Bayer study eight years earlier (Simeone, 1987). Another study o f land grant universities by DeReim er et al, also in 1982, found that male and female assistant professors were just as likely to attempt negotiations for higher entry level salaries and subsequent raises (Simeone, 1987). Thus, while these studies disagree on the bases for awarding compensation to women, each of these studies confirms that the bases for awarding compensation are influenced by gender, though perhaps gender is less o f a factor than in previous decades. D w yer comments further on studies conducted during the 1980’s. A study by Kelley(1989) used salary as the dependent variable and controlled for both experience levels and academic rank o f faculty members. In this study " a change from female to male status would increase one’s base salary by approximately $3600, all other variables in the equation being equal" (Dwyer, 1991, p. 21). Dwyer goes on to cite additional studies also indicating salary disparities by gender even after controlling for other relevant variables. Results indicated disparities ranging from 9.7% to 28% o f the mean 22 male salary. W hile the reasons for these differences may be multiple, the gap which Bernard first noted in the 1960’s has neither disappeared nor narrowed since that time (Dwyer, 1991). In summary, while the percentage o f women faculty in the higher education system has finally returned to the levels of the late 1930’s, differences in their distribution within the system are striking. To date their place in terms o f types o f schools, disciplines, rank, and salary has been markedly different from that o f the male faculty within the academy. Reasons for these differences in distribution, often called a two-tiered system, will be addressed below. However, it is evident that such differences exist and in several ways have influenced female faculty who become members o f research universities. FEM A LE FACULTY: TH E NATURE OF THEIR WORK In addition to the issue of numerical distribution it is important to determine whether female faculty are different than male faculty in their work-related satisfactions and allocation o f work effort. In the existing literature how do the two groups compare in this area? The literature does not crisply divide the discussion o f career satisfaction and allocation o f work effort. It seems to assume that allocation o f work effort reflects relative 23 satisfaction levels but also that sources o f satisfaction influence the allocation o f effort. Thus, the two are treated in an interrelated, often indistinguishable manner. In addition, the current tri-partite definition o f faculty work includes effort devoted to teaching, scholarship, and service. This has not always been the model for the work of faculty in higher education. Based on the English college model, the U .S. higher education community o f the 1600’s to the early 1800’s embraced a model which emphasized teaching and faculty as examples for the developing moral character o f their students. In the mid to late 1800’s with the opening o f many land grant institutions after the M orrill Act, the prevailing model balanced teaching and public service, though public service was often linked to applied research. (Rudolph, 1962; Finkelstein, 1984) A significant change in the institutionally preferred model for faculty work occurred in the 20th century. Both Jencks and Riesman (1968) and Finkelstein (1984) document the shift which expanded the weight given to research efforts. Through both the rise in importance o f the PhD degree as the preferred academic credential and the rise in funded research opportunities, the broad consensus in higher education on the best model for faculty efforts was fractured. Four-year liberal arts institutions, and later the emerging community colleges retained the earlier model emphasizing teaching and service. Universities, particularly those identified as doctoral institutions, chose a model which placed much more emphasis on research and the production o f original scholarship. (Finkelstein, 1984; Baldwin, 1985). 24 With these differentiations in mind, consider the literature on the sources o f faculty satisfaction and the ways in which women faculty use their time. Teaching W ithout distinguishing by type o f higher education institution the predom inant theme in the literature is that women place much more emphasis on teaching than scholarship. Bernard (1964) argued that women focused on stable subjects which they could teach at the elementary collegiate levels. They saw themselves as transmitters o f knowledge, not as social critics. W omen were competent in teaching standard noncontroversial bodies of knowledge but were handicapped in handling controversial materials because they were seen as less authoritative. W omen were also judged to be more successful in achieving value in personal conferences with students and were more helpful to poorer students. Related data indicated that their classes showed less variance in results of teaching than those of their male counterparts. Finkelstein (1984) concurs that women in the 1970’s endorsed teaching effectiveness as their first priority. They thought it should be the prim ary basis for promotion. In his studies women were less likely to be involved in off-campus professional activities, less likely to have a broad network of scholarly collaboration, and less likely to engage in paid consulting. 25 W ith slight variations in their viewpoints, Simeone (1987) and Aisenberg (1988) also concur that women focus primarily on teaching. Simeone argues that women choose their disciplines based on their teaching desires and people orientation. Aisenberg argues similarly, though she parts company with Bernard regarding wom en’s goals in teaching. Aisenberg suggests that women are interested in more than the transference of knowledge. They are interested in student transformations. Thus they prefer a pedagogy emphasizing questioning, engagement, and discussion instead o f straight lecturing. Boyer’s survey for the Carnegie Commission published in 1989 confirms Simeone and Aisenberg’s observations about women faculty’s priorities in the 1980’s compared to those o f men faculty. 78% of female faculty compared to 67% o f male faculty indicated their interests either lean to teaching or are primarily in teaching as contrasted with research. Similarly 73% o f women faculty surveyed felt that teaching effectiveness should be the primary criterion for promotion compared with 58% o f male faculty. Female faculty’s priority for teaching seems to be bom out in other areas o f his survey as well. When asked about student characteristics, such as academic effort and standards, cheating, or substance abuse, women faculty seemed both more knowledgeable about and more sensitive to these characteristics. The frequency o f their "don’t know" responses was much smaller than those o f the males surveyed.(Boyer, 1989). Unfortunately the Carnegie Commission’s published survey results did not break down the male/female responses by institutional type. Thus it is not possible to determine 26 whether or to what degree these responses are the result o f women’s locations primarily in colleges and universities not oriented toward research. Some evidence exists about the effect o f institutional type on reported satisfactions and effort allocation. M orlock, writing in Rossi (1973), notes that, "...D ifferences between time spent in research and teaching are greater in universities than in two-year or fouryear colleges...W om en are most likely to spend less time in research than their male colleagues in precisely those institutions with the best research facilities."(p. 283). Yet a further difficulty with these findings over three decades are that few studies distinguished responses by discipline or marital status. significant. The differences can be Astin’s 1969 research cited in Rossi (1973) indicates that women in the social sciences, arts, and humanities spent more time teaching while women in the natural sciences spent more time in research. Another Astin study o f women who received doctorates between 1955 and 1960 concluded that marital status was a significant factor. Single women doctorates spent more time in teaching and administration while married women doctorates spent more time in research. M ore recently in her research on career satisfaction, Dwyer notes that a 1989 study by Newell and Kuh found women faculty’s satisfaction centered on teaching, while a comparative group o f m en’s centered on salary. However, Dwyer also comments on a study by Ethington, Smart, and Zeltmann (1989) indicating that the greatest female 27 satisfactions were in liberal arts colleges and research I institutions. This last study would seem to indicate a dichotomy between two types o f women, those with strong preferences for teaching and those with strong preferences for research. From the Ethington et al data it may be possible that there is a group o f women for whom research is a greater source o f satisfaction than teaching. On the basis o f these studies the presumption that women necessarily gravitate first to teaching is not completely supported. While there is evidence for that as an aggregate pattern, when other defining variables are included in studies, the pattern appears to be breaking down. Scholarship Beyond the studies that have been done about w om en’s emphasis on teaching, what do these and other studies indicate about women as scholars? Reviewing the evidence without regard to institutional type, it appears that women on the whole are less productive than men. Women simply publish a smaller volume o f work as is evident in both Finkelstein’s (1984) and D w yer’s (1991) reviews o f the literature on this topic. Looking at cumulative production and publication rates male academics were as much as three times more productive as researchers in the recent past. 28 Dwyer and Finkelstein’s conclusions about the past are verified by other studies. Simeone notes that in a 1977 study by Freem an, within research universities, publication rates for women were lower. Seventeen percent more women faculty than men had published no books. Twenty-three percent more women faculty than men had published no articles. (Simeone, 1987). In studying publication o f scientific papers Cole (1979; 1987) concludes that even when women and men are matched by field, professional age, and source o f PhD, men published more and that these differences widened as careers lengthened. The differences though are more subtle than is suggested by a focus on volume. Bernard (1964) critiques that focus on volume which ignores creativity. Simple counts of products ignore weight, quality , or the nature o f the discipline involved. Bernard notes that often in studies o f research productivity, degree status, rank and discipline are not controlled. Even if scholarship by females was less characteristic o f the past, by the 1980’s researchers suggested that the source o f w om en’s satisfactions was changing. Chamberlain (1988) notes that in 1972 38% o f women and 62% o f men had some research support. By 1982 53% o f women and 69% o f men had similar support. Simeone argues that based on her interviews o f 20 women at a large prestigious northeastern research university, women academics were now committed to the intellectual challenges o f the university and not just the social and emotional development 29 o f students. She believes that women are now engaging the world o f theory, methodology, and interpretation, creating new knowledge, new data, and new paradigms (Simeone, 1987). Unfortunately her interview sampling method could well have reinforced all the biases which she wished to find since her subjects were located in a research university. If she had argued that women within research universities were engaging this new agenda, her argument would have been more credible than to generalize to the goals o f all women academics regardless o f location. Aisenberg also suggests that women academics find satisfactions in both teaching and scholarship similar to men, and that there are great variations in the balance sought among women just as among men. However, Aisenberg goes on to comment that regardless o f the relative balance, the nature o f wom en’ scholarship is different than that o f men’s. W omen gravitate toward scholarship which studies human nature, not that which studies groups, social systems, organizations o f power, or nonhuman subjects. They also tend to focus on interdisciplinary work that combines theory and reality, seeming to have little preference for research that is highly abstracted and rationalistic. Aisenberg summarizes the differences by stating that "overall, women scholars are heavily engaged in integrating knowledge. Their work combines disciplines, combines theory and reality, combines a commitment to change with a commitment to humane study. Its approach to knowledge is inclusive to the point o f ambiguity, rather than exclusive to the point of certainty. Its social vision is o f an integrated whole with the 30 characteristics and interests o f diverse groups honored and supported" (Aisenberg, 1988, p. 105). W hile Aisenberg makes some thought-provoking points about the nature o f wom en’s scholarship, again, her research on the relative satisfactions and allocation o f work effort among women academics is not based on a broad sample. Aisenberg selected her interviewees from among women who had been unable to obtain or maintain a full-time academic position in the vicinity o f a large north-eastern city. Because of the characteristics o f her sample, any generalizations about women faculty must be treated with caution. Finkelstein concurs that rank, location, and discipline are overlooked as control variables. He notes three interesting characteristics. First, the disparity o f productivity between male and female academics narrows as women approach their 20th year after PhD receipt. Second, gender differences also lessen as women move up in institutional strata. Third, by discipline, the research women do publish tends to be less pure or basic, though women in the social and natural sciences are fairly prolific publishers. Perhaps that finding is also reflected in Boyer’s survey which found women faculty as tied to their academic disciplines as men, but less tied to national and international societies of the discipline (Boyer, 1989). This tie to less pure research could also be tied to lower female computer and statistical skills in many fields. White males were much more satisfied with their skills in this area than white females, according to Olsen (1990). 31 In summary the review o f women faculty’s allocation o f effort to scholarship shows very mixed results. Studies have frequently lacked control o f rank, discipline, degree status, institutional location, and the nature o f the output. W hile the past decades document less scholarly productivity by women in the aggregate, the present interest and effort of women in scholarship, particularly as one controls for non-gender variables and research university setting, is unclear. Service Little research has been done about relative attention to service activities as related to work effort and satisfaction. A well-constructed definition o f service in higher education research on faculty effort is missing. W ithout such, there is little consistent evidence of patterns for women. Only Olsen and Boyer provide small insights into the current situation. Using a broad definition o f work-related service Boyer’s 1989 studies indicate that women do less consulting than men but that women participate more in campus wide faculty committee activity and in the faculty senate meetings than do their male counterparts. Yet Boyer’s work does not differentiate by institutional type. Olsen, commenting specifically on the research university, suggests that service for women is "in large part the product o f an institutional desire to have a diverse set o f viewpoints represented on decision-making bodies...w om en’s greater willingness to recognize the value o f service makes their participation in the work o f university committees important to the day to day functioning as well as to insuring a diversity o f perspectives. Overall, 32 white men appear to participate most heavily in the types o f service closely allied with individual research" (Olsen, 1990, p. 15). Thus it would appear that the catalyst for greater on-campus service by women faculty at least in the research university is frequently the university’s need. W omen are willing participants; however such service may only reflect their recognition o f immediate university needs, rather than their preferred long run allocation o f effort. The Distribution o f W ork Effort and Type o f Institution Three decades o f research on women faculty seem to indicate that the predominant faculty model for teaching, scholarship, and service used by most female faculty differs from that o f male faculty. However, Olsen’s 1990 study of women academics within the context o f the research university indicates unexpected similarities and also some subtle areas o f difference by gender in that context. In her study o f 146 cases by personal interview and a related questionnaire, many factors influenced the relative allocation of work within a research university for women. Within her sampled research university Olsen noted that 62% o f white males and 56% of white females saw their interests as heavily related to research or leaning toward research (Olsen, 1990). The gap between these two groups is less substantial than that of Boyer’s survey encompassing a broad mix of institutional types. Olsen also notes that her findings indicated no difference in over all job satisfaction between white males, white females, and black faculty. All indicated moderate levels o f satisfaction, which would seem to indicate that all were relatively satisfied concerning their match with a research oriented institution. In fact a higher proportion o f white women in her sample group believed they were successful at research than did their white male colleagues. Olsen did find that the males developed their career goals more clearly, at an early age, and presumed no career interruptions when compared with the female academics. They also were least likely to see their teaching as contributing a great deal to their own professional development. However white male faculty found balancing work and personal life as stressful as the white women surveyed. In a related statistical study, Olsen also tested the significance o f gender, rank, and discipline in determining faculty time allocations within the research university. Olsen concludes that rank and discipline are much more important than gender in time allocations in that setting. (Olsen, 1991). Olsen writes: "Existing research suggests a higher level o f commitment to teaching and service among women and minorities and a lower rate of research productivity than found among white male faculty. The issue is more complex than this, however. Women and minority faculty are often asked to carry a heavier course load, teaching more introductory courses and are assigned fewer graduate courses...G ender-related productivity differences have also been shown to disappear when "type o f institution" and discipline are controlled for" (Olsen, 1990, p. 34 In summary, Olsen did not find that within the research university white males and white females differed substantially in their job satisfaction or their allocation of effort, though they did differ on other aspects o f their career orientation. Yet her study refutes the assumption that women necessarily tend to prefer teaching over research. Instead her study adds weight to the belief that institutional location is a significant factor in any faculty m em ber’s satisfaction o r work effort allocation. The full picture of women faculty and their preferred work allocation is not yet pieced together. M any o f the findings are o f limited use because they do not distinguish by type o f institution. Only a few studies have been done on women within the research university setting, distinguishing them from other groups o f female faculty. When the institutional setting is controlled it remains to be determined through further study whether these women prefer teaching to scholarship and service as is the case in two and four-year colleges. It is not clear whether the character o f wom en’s work is by nature unique when compared with that o f males. Analyzing the results within a research university may be an important factor in understanding the interaction between institutional type and the work o f women faculty. CAREER DEVELOPM ENT APPROACHES A final area o f possible difference between male and female faculty is found in the literature on career development approaches. How do women’s preparation for the 35 academic profession, initial entry into academia, and career strategies for advancement compare with those o f males? Are their approaches to issues such as geographic mobility substantially different from those o f male faculty? Career entry Aisenberg (1988) writes extensively about the entry o f women into the academic profession. She suggests that women do not typically take a linear path, straight from undergraduate schools to graduate schools, and then on to employment in major colleges and universities. Instead Aisenberg hypothesizes that the catalyst for many potential women faculty is a desire to break out o f traditional marriage molds. She suggests that this quest is not well defined, however. Thus women often drift and experiment, a combination she calls "veering and tacking". In her research that quest eventually led the women in her sample to graduate school. Yet the graduate school experience was often first o f all a source of personal empowerment and transformation. Frequently the women in her sample confused such personal growth and intellectual engagement with the external requirements o f the profession. They were naive about the need to find a mentor, attend extra conferences and seminars in their discipline, gain the benefits of teaching or research assistance, or network with established professionals in their field. 36 Aisenberg’s findings confirm earlier evidence o f such problems noted by Feldman (1974). His research also found that marriage had a great effect on the intellectual motivations o f women. Married women more frequently enrolled as part-time graduate students and often were seen as the least productive by fellow students. By contrast divorced females seemed to have academic success rates about equal to those o f married males. H e concluded that it appears that marriage reduced conflict for men but increased conflict for women. Those conflicts often led women, particularly married women, to make suboptimal professional choices. Similarly, Clark and Corcoran in their 1986 study documented reasons why women had not enrolled in the best graduate programs. Either because o f their own ill-defined directions or subtle discrimination in their discipline, they were not chosen as proteges, lacked research resources, and were frequently marginalized in graduate school. This is the likely beginning of cumulative disadvantages which influence academic women’s careers as they continue to develop(in Dwyer, 1991) and frequently prevents them from being prepared for positions in major universities. As Caplow and McGee stated already in their 1958 studies the link between one’s doctoral university and initial employment is substantial. In their words: "The initial choice o f a graduate school sets an indelible mark on the student’s career. In many disciplines men trained in minor universities have virtually no chance o f achieving em inence...The handicap of initial identification with a department o f low prestige is 37 hardly ever completely overcome. The system works both in channeling students into graduate school and then in channeling them out into jobs. Thus it affects where students come from and where they g o ." (p. 225 o f The Academic M arketplace, as quoted in Bernard, 1964, p. 87). Breneman and Youn confirmed that this pattern has continued into the 1980’s. In their studies one’s entry position was explained by the prestige o f one’s PhD institution, not by predoctoral productivity. The pedigree o f one’s graduate institution affected one’s later opportunities. There was some variation by discipline. Prestige affected the locus o f career entry more heavily in social sciences than in the humanities. factor operated differently in biology than it did in mathematics. The prestige Still entry and later mobility were greatly affected in all fields by one’s doctoral institution. (Breneman and Youn, 1988). Thus wom en’s enrollment in second-tier graduate programs, whether through their own choice or active discrimination, affects their later ability to attain positions, particularly in research universities, which are highly selective in their sources o f new faculty. Instead, it is more likely that even if the supply o f women faculty is increasing they are employed in two year and four year institutions not oriented toward research. 38 Initial Placement However, Bernard noted an even more unusual problem in the link between graduate education and entry level positions. She reported at that time "women receive their doctorates from universities as good as those from which men receive theirs. Berelson reported that about the same proportion o f women students (47 percent) as men (43 percent) received their doctorates from the Top Twelve universities." (Bernard, 1964, p. 87). From among those top twelve schools the Phd graduates in the sciences were tracked into employment. The women much more frequently were placed in colleges than in universities, despite equivalent credentials. Thus she concludes that even education in comparable high quality universities did not have the same consequences for women as men. Astin (Rossi, 1973) argues similarly that initial placement is critically important. In her studies she found that frequently post-undergraduate educational experiences were similar for males and females in terms of degrees, financial aid, and post doctoral study but the career development which followed was different in both compensation and promotion. She agreed with M orlock that the greatest barrier was not graduate school but finding the gateway to a standard position within higher education at a level comparable to that of males. W omen were frequently hired at a lower rank even when their credentials were comparable. (Rossi, 1973). 39 Aisenberg and Dwyer confirm that problems in the transition from graduate school to employment have continued. A isenberg’s entire interview sample consisted o f female P hD ’s who were either unemployed or only marginally employed in colleges and universities. Dwyer also notes that even by the late 1980’s women P hD ’s were more likely than males to end up unemployed, not participating in the labor force, or as parttime faculty. (Dwyer, 1991). Thus, both wom en’s preparation as academic professionals and their transition from graduate school to initial career placement appear to be characteristically different from those o f males. Career Advancement and Mobility If women faculty do transcend the barriers to initial career employment, what characterizes their patterns of advancement and the constraints on their advancement? Initially women may need to make more adjustments than men. In a study contrasting male and female junior faculty at an elite research university, women interviewees reported more changes in their perceptions and expectations for themselves and others than did the men (Reynolds, 1989). Beyond initial adjustments what occurs? Finkelstein suggests that faculty advancement involves five factors that are rewarded. They include longevity, seniority, terminal degree, research productivity, and institutional service (Finkelstein , 1984). However, Finkelstein’s advancement factors could play out very differently in the lives o f male and 40 female academics. Even when degree credentials are comparable Dwyer (1991) comments extensively on findings which substantiate female versions o f academic career development which are very different from those o f males. Female academics progress through the academic ranks more slowly, often with more undergraduate teaching and heavier teaching loads. This has been the pattern since at least the late 1960’s (Rossi, 1973). Female faculty may also experience higher levels o f stress than males (Connors, 1991). Sometimes career development varies because o f personal strategies chosen by women themselves. M arried women employ "hidden passages" to accommodate competing demands o f family and career. These are often part-time work, leaves o f absence, or voluntary delays in the timing of the tenure decision. For some women academics longevity and seniority did not operate in tandem. Differences in research productivity have been documented. Institutional service by women occurs but differs from the type offered by males (Dwyer, 1991). Career development can also be affected by institutional climate. A recent study (Brown, 1990) suggested that career advancement for women, particularly in research universities, is undergoing a transformation. In her interviews with 32 female faculty at three research universities, Brown concluded that academic departments are in different stages o f gender equity development. While in stage one women are not hired at all, stage two is a revolving door. Women enter but do not stay because o f an unsuitable environment. 41 In stage three women sometimes gain tenure but do not become full professors. Only in the fourth stage has a completely accepting environment developed where men and women are treated fairly for tenure and promotion. Geographic mobility also plays a significant role in career development. D w yer’s review continues by examining the effects o f geographic mobility on career advancement, a factor which Finkelstein him self calls the missing but potential "ace in the hole" for women (Finkelstein, 1984, p. 212). Are women academics that different in their relative mobility? If so, how does it effect their advancement? Bernard (1964) noted that women, both married and single, were more tied to the geography o f their birthplace than men. Both social factors and personal preferences seemed to discourage moving from one institution to another. Yet the effects at that time were minimal. Finkelstein found that geographic mobility was not a critical factor in advancement for most faculty during the 1960’s and 1970’s. He documents that few faculty, male or female, moved during these two decades; and when they did, they were responding to the optimization their interests and values, not to prestige or necessarily advancement (Finkelstein, 1984). Differences between males and female academics were not affected greatly by the issue o f geographic mobility. Yet the more constrained mobility o f women was beginning to be recognized as an advancement factor that could significantly affect the future. Astin’s studies o f women 42 with doctorates obtained between 1960 and 1970 (Rossi, 1973) indicated that these women were much more likely to pursue employment in higher education rather than industry, government, or not-for-profits. These women were much less likely than men with comparable credentials to shift between the academic and nonacademic worlds. They also struggled with the issue o f spouse mobility. When surveyed in the 1968 a sample o f these women listed husband’s mobility as the third greatest barrier to career development. Studies at that time also indicated that geographic mobility often accounted for a higher salary, although not necessarily for higher rank or tenure (Rossi, 1973). By the 1980’s the academic job market was an even more national one and advancement often depended more heavily on one’s ability to relocate. The power to relocate also provided negotiating clout with a faculty mem ber’s present institution. Dwyer confirms that since the 1970’s geographic mobility has become much more related to advancement. She writes that "academics enjoy a national job market and especially at certain points in an academic career, advancement may depend on relocation. Advancement may also depend on being perceived as mobile when negotiating with present and future employing institutions."(Dwyer, 1991, p. 29) . Dwyer cites several studies by M arwell, Rosenfeld & Spilerman (1979), Tuckman and Tuckman (1981),and Rosenfeld and Jones, (1986, 1987) all o f which document the differential position o f women academics related to geographic mobility. According to Dwyer, "...w om en academics are less likely than males to leave their geographic location 43 when changing jobs and are more likely to locate in large urban areas which provide for broader but more competitive labor markets" (Dwyer, 1991, p. 30). The net effect is that the women studied did move but often stayed in the same rank and geographic location. Even single women less tied to family considerations now than in Bernard’s earlier studies are also geographically constrained. Often because o f their lower status in the academic ranks and their concentration on teaching they too have been less mobile than male faculty. By comparison men were much more likely to change their status with a change in geography. It appeal's that "...w om en do not seem to choose location for maximizing career prospects in the same way that men d o ..." (Dwyer, 1991, p .30). Again, however, these studies o f female geographic mobility related to advancement do not isolate the women faculty who staff research universities. The desire to be mobile may represent attempts to move up through the perceived hierarchy o f higher education institutions to a research university position. For example, a position in a smaller comprehensive school may only be a stepping stone to an offer from a larger research university. However, once in a research university do women faculty perceive their mobility differently from male faculty? Only Olsen’s (1990) study focuses on women who are already employed at research universities. What does she conclude about relative geographic mobility differences? Olsen concludes that, "... the number o f faculty describing themselves as ’somewhat’ or ’very’ likely to seek a new position in the next year was relatively large with about 36% 44 o f white men and minorities and 44% o f white women falling into these categories. For the most part, faculty would be seeking positions at other academic institutions." (p. 17). Given the location o f her study at Indiana University it seems safe to assume that seeking another academic position would require geographic relocation. If O lsen’s study is representative o f intentions among research university faculty to move, then at least white women expect to move even more frequently than white men. W ithin the research university context the contention that women are more geographically bound is not supported by her evidence. Instead they see themselves as more likely to move. The reasons for moving may be multiple. They may be tied either to advancement or satisfaction with organizational fit, but the perception o f capacity to move does not appear to be constrained. However, Olsen’s results should be confirmed or questioned by additional study of the research university environment. Drawing broad conclusions from one study o f limited scope and methodology would be inappropriate. To summarize the literature to date most researchers have argued that female academics choose and develop their academic careers differently from their male counterparts and are often more geographically constrained. In broad studies across different types o f institutions, this appears to be the case. However in the one study focused specifically on the research university context, there is some evidence to the contrary. It is possible that the women who populate research universities may be substantially different from women in other types o f academic institutions in their ideas about their careers and paths to advancement. 45 EXPLAINING DOCUM ENTED DIFFERENCES The sections above document the literature on possible areas o f difference between female and male faculty in higher education by commenting on their total numbers, their relative distribution in the system, their sources o f satisfaction and allocation o f work effort, and their career strategies. However, it does not explain why these differences may have occurred. This section will examine various reasons for differences in the distribution, work effort, and career development o f academic men and women. Multiple, multifaceted reasons are contributed from a variety o f knowledge fields. Scholars in anthropology, history, biology, psychology, sociology, and labor economics, as well as higher education specialists consolidating contributions from these several fields all contribute their commentary. However the most developed sources o f explanation appear to be psychology(both developmental and organizational), sociology, and labor economics. Yet researchers in each o f these fields give different emphases in the balance o f their explanations. These varied emphases reflect differing assumptions about the relationship of individual choice and societal structure. The issue is whether occupational placement prim arily results from individual preference or whether individual preferences are shaped prim arily by occupational placement (Smart, 1991). 46 For some, explanations reside in individual preferences resulting in certain career choices. They align themselves most closely with human capital theorists such as Bowen (1977) and with a functionalist’s approach to individual contributions (Davis and M oore in Tum in, 1970). Other researchers place more emphasis on social structures which affect the systematic choice set and possible preferences that are available. They align themselves with either the Weberian conflict theorists which emphasize the dominance o f the majority group and the emerging conflict as the majority/minority balance changes; or they align themselves with more Marxist educational theorists, such as Bowles and Gintis (1972) or Karabel and Halsey (1977), who argue that education is reproductive of the existing class structure and reinforces capitalist traditions, supplying the labor pool to meet its economic functions. The range o f explanations along this continuum will be developed, beginning with those explanations most emphasizing individual preference and ending with those explanations which most emphasize the social structure as the determinant o f individual choice. Nature As The Explanation For Individual Preferences One argument forwarded for the differences between male and female academics is that women are inherently different in their biological and psychological make-up. This approach argues that women faculty will make different choices simply because they are uniquely different creatures. Radical feminist theorists place great credence in the belief 47 that the special reproductive biology o f women (Stromquist,1989). results in a unique destiny They argue that this unique destiny has unfortunately not been honored in the higher education system because males serve as the gatekeepers to the system, controlling the channels of hiring, promotion, and organizational mission. The fundamental thesis is that women are inherently different in their professional orientation toward higher education’s missions and strategies. These differences thus result in satisfaction, allocation o f work effort, and career development strategies which vary from those o f males. Likewise, some educational researchers begin with this same thesis in their analysis of women faculty. Dwyer (1991) comments that H olland’s Vocational Type Theory places great emphasis on individual career choices which parallel one’s personality type and personal orientation. These factors later interact with family background and school experience to form actual choices made. However one’s inherent personal being affects the choices significantly. Dwyer suggests that Cole (1979) argues in similar tones when he states that women fare less well in academia as productive researchers because they inherently enjoy collaborative projects and thus don’t compete for key individual resources (Dwyer, 1991). Part o f the argument appears to be that women are simply less competitive by nature. Thus they do not fit the mold o f demands in higher education research. 48 In more recent research Simeone (1987) suggests that inherent differences are often the explanatory framework for wom en’s relative allocation o f time between teaching and scholarship. Simeone suggests that the explanation frequently forwarded about wom en’s preference for teaching over scholarship is a nurturing "people orientation" which is unique to their character as women. W hile Simeone does not totally accept this as the only or best explanatory framework, she believes innate differences by gender are part of the explanation for differences. Aisenberg (1988), is also prone to this line o f argument. She indicates that women have unique pedagogical approaches in the interest o f transforming their students and unique approaches to scholarship in the interest o f transforming their disciplines. Aisenberg does not accept a heavy investment o f women in teaching to the exclusion o f scholarship as natural; yet she does argue that there are naturally unique qualities among women academics which will make both their teaching and scholarship necessarily different from that of males. She suggests that women naturally gravitate toward the study o f human nature which explains the higher proportions o f women in certain disciplines. The school o f explanation which puts such a heavy emphasis on "naturalistic" differences is not the dominant one. The majority o f researchers and writers investigating causes of male/female differences do not seem to accept the primacy o f inherent differences because that belief can be used to justify differences in gender status within the academy 49 and place the responsibility for change upon individuals alone. Yet several researchers believe that at least a portion o f their explanation is based on innate differences. Socialization As The Explanation For Individual Differences A second explanation for the differential role o f women in higher education is that women make difference choices, not because they are natural, but because their individual development is shaped by the certain socialization processes. The result of socialization, which differs between males and females, is that women construct a different developmental life cycle and balance a different set o f societal roles than men. Career choices still follow preferences. Individual preferences are still the dominant factors in choice making, but these preference have been shaped by social, not innate biological or physiological forces. Socialization within the family, school, and community may result in differences by gender in needs and drives, satisfaction sources, self perceptions, ethical frameworks, and societal perceptions (Sheehy, 1976; Gilligan 1982; Belenky 1986). The results of such socialization differences will be differences by gender in the adult life cycle (Sheehy, 1976; Levinson, 1977), Shakeshaft (1989) and life roles (Bernard, 1964; Fox, 1989). 50 F or female academics the result is that their career choices are often different in terms o f both the process o f choice and the actual career choice made. The decisions following that career choice in terms o f preferred activities within the higher education system will thus also vary from those o f males. Finally the balance o f professional and personal roles will be different. Because o f differences in socialization women will choose to develop their total human capital differently than men. Specifically among higher education faculty such socialized differences result in differences by gender in degree status, academic rank, disciplines o f expertise, work allocation in teaching and scholarship, and the productivity o f work. Bernard (1964) first constructed the argument that women who decide on a career in higher education will have key differences in factors which socialize them. At that time women P hD ’s came from higher class homes and were intellectually superior to male Phd’s in terms o f test intelligence. Yet they were not as productive as researchers, were associated with teaching institutions, and saw themselves more as transmitters of knowledge than as social critics. Women focused on more stable subjects with less controversial bodies o f knowledge already in existence. Bernard explained productivity differences by relying heavily on the explanation of socialization. thinking. She argued that women have been socialized into two key patterns of First they are socialized not to put all o f their life fulfillment eggs in one 51 basket-either in the basket o f their paid profession or heavily in one facet o f their paid profession. Women have also been socialized to expect more domestic responsibility. Thus the amount of energy that they perceive they can channel into their w ork is more limited. Bernard suggested that women therefore have a different attitude toward professional advancement and simply invest less o f themselves in the necessary activities to achieve it. The women in her studies demanded less overall satisfaction from their paid professions. M arried women in particular saw marriage as a career hindrance and children as a potential liability to their advancement, particularly because they tended to set very high standards for their children’s behavior and achievement. They also saw less income need for their advancement. Thus, if they chose to marry and bear children, they believed they were choosing against career advancement comparable to that o f males. She also noted that even within the profession, women were socialized into roles whose function was conserving, stabilizing, and appeasing. controversy within their fields. They were socialized to avoid Thus they tended to favor teaching roles o f stable developed material, rather than the controversial role that the development o f theory and pure research can play within a discipline (Bernard, 1964). Aisenberg (1988) piggybacks on Bernard’s socialization theme. In her interviews with women she found that most o f them worked to balance sources o f satisfaction between 52 personal/family and professional lives. She views this struggle between roles as a struggle for balance between the "marriage plot", a privatized anti-intellectual socialized role and the "quest plot", a public intellectual socialized role. The balance that women try to maintain differs from that o f male academics in terms o f the amount o f satisfaction they seek from family life. W omen in her study assumed the burden o f trying to live two full lives, both those o f domestic and professional fulfillment. Often women, because o f beliefs that grow from their socialization, adopt a strategy o f securing one of these two areas o f their lives before diving into the other. Thus, sequencing, instead of the simultaneity o f these roles, results in a differing amount o f career progress when compared to males. Aisenberg (1988) also believes that differential socialization has resulting in different pedagogical styles for men and women. W omen prefer to use active discourse with students, questioning, engaging, and discussing with them. In contrast she suggests that males tend to lecture more, using the idea that they are handing down knowledge to students, rather than nurturing their development. Similarly, Aisenberg suggests that in the area of scholarship it is difficult for women to speak as voices o f authority after being socialized to ingratiate, self censor, self efface and avoid contention. Thus they have difficulty knowing the acceptable boundaries of their public voice in scholarship. They have been socialized toward tentativeness in 53 expressing their findings and viewpoint and either distrust or dislike the prevailing models for professional discourse which often require extensive debate. She also notes that, "Overall, women scholars are heavily engaged in integrating knowledge. Their work combines disciplines, combines theory and reality, combines a commitment to change with a commitment to humane study. Its approach to knowledge is inclusive to the point o f ambiguity, rather than exclusive to the point o f certainty. Its social vision is o f an integrated whole with the characteristics and interests o f diverse groups honored and supported" (Aisenberg, 1988, p. 105). Aisenberg believes that these characteristics o f women are among the primary reasons for their exclusion from the world o f accepted scholarship. Their work in her mind does not fit the existing male paradigms for scholarship within the profession. Thus she sees exclusion and devaluing o f wom en’s contribution in higher education as related to w om en’s abilities and interests. Aisenberg writes, "What we see generally in the range o f work reviewed above is a strong thread o f resistance by women to academic conventions establishing the boundaries o f knowledge-from the rules o f scientific certainty, the logic o f abstract theory, and the division o f subject matter into discrete disciplines, to the rules that include or exclude material as relevant to a particular discipline" (Aisenberg, 1988, pp. 100-101) 54 Finkelstein (1984) too argues that prior socialization is one o f the key reasons for differentiation o f men and women academics. He documents the fact that choosing to enter an academic profession is based on intellectual abilities and family statuses that are shared by both males and females; but then individual choices resulting from sex role socialization have a differential impact on men and women academics. Once choosing an academic profession, males and females develop differently within it. He explains w om en’s stronger orientation toward teaching and at schools ranked lower in the institutional strata as the result o f prior socialization. He notes that the perception o f one’s academic role differ for men and women with women seeing it as a role which competes with other familial roles much more frequently than men. Finkelstein documents that academic women spend twice as much time on family chores as academic men and that academic women frequently revert to traditional sex roles in their family lives. In his mind this may account in part for lesser degree attainment or less geographic mobility related to career development. His explanation does not necessarily question the appropriateness o f the socialization which women experience. Typically it accepts that socialization as a given and then proceeds to examine the differences in human capital development which result from it. Chamberlain (1988) seems as well to attribute some significant gender-related differences to socialization. She suggests that women have been socialized to view the resources needed for research productivity differently than men. She concludes that women attribute research productivity to "personal" variables like hard work, motivation, interest 55 in the topic and skills while men attribute it to organizational factors, such as assistants, funds, and institutional resources. She writes that "...w om en are less likely than men to endorse the importance o f organizational and/or structural variables in enhancing their research or scholarly productivity..."(C ham berlain, 1988. p. 267). In summary this second position emphasizes that women academics, while not by birth different than male academics, become different from male academics because o f their entire socialization. That socialization results in differences in educational background, experience in graduate school, degree attainment, and choice o f institutional type of employment. This school o f thought emphasizes that women must be socialized differently so that they make different choices in the development o f their human capital, choices which will guarantee them more rewards, secure placement, and prestigious positioning within the academy. Structures As The Explanation For Individual Preferences A third set o f theorists argues that the role and functioning o f women in academia is differentiated prim arily because of institutional and societal structures and their effect on women’s roles and effectiveness in academia. All researchers grant that overt discrimination against women has sometimes been a problem; however this group of theorists believes that the cumulative disadvantage o f women faculty results from more than individual discrimination. It is the result o f systemic forces. This tradition 56 emphasizes that external constraints on women, often outside their individual control, have significantly hampered the advancement o f women faculty in academia (Clark and Corcoran, 1985; Smart, 1991). Those who argue from within the liberal feminist tradition typically attached themselves to this stream o f argument. They see the state as the agent o f redress because it is the one entity which can alter the institutional and societal structures which hamper the progress o f women academics. Such groups favor strong affirmative action programs within colleges and universities, laws defining sexual harassment and penalizing its perpetrators, and publicly financed support for programs encouraging women in academia. (Stromquist, 1989). Within this third set o f theorists three sets of foci emerge, those emphasizing organizational behavior theory, those emphasizing bias in the broader culture o f academia, and those examining gender inequality from the standpoint o f labor economics. The organizational behavior theorists emphasize the effects o f the specific organizations in which women participate. Often women do not fit the current configuration o f the organization, feel uncomfortable with the lack o f fit, and may choose to leave the organization as a result. Both Kan ter (1977) and M orrison et al (1987) suggest that organizational structures, cultures, and climate result in traditional roles for women as nurturers, tokenism for the few women who function as equals, and harassment for women who threaten the male-dominant cultures. Such theorists argue that organizational cultures, such as those of colleges and universities, will only change in their reward structures and organizational values when a sizeable number o f women have joined them, 57 becoming not tokens but a viable force for change. They also stress that institutional policies must change to allow acceptance o f women. Such theorists stress the importance o f mentors for women, w om en’s support networks, and a careful definition o f job responsibilities not based on gender. Such theorists also recognize that organizational behavior does not always follow a rational linear path. Yoder, Crumpton, and Zipp (1989) noted that particularly in academic hiring, women candidates were favored in departments that had moderate numbers o f females already in their ranks, represented between 16% and 65% o f their total academic faculty. In contrast, departments with more than 65% females were not too willing to hire additional women; and departments with few or no women were also less likely to hire women. A linear logic might suggest that the departments with few or no women would be the most likely to work at redressing the balance. Yet organizational theorists contend that group dynamics and organizational culture in that environment may prevent serious consideration o f women applicants. A different group o f theorists focuses on another level o f structural influence. They emphasize bias in the broader culture o f academia, arguing that the whole structure of academia in the United States is male dominated. For some this male dominance has resulted in broadly shared institutional morays which thwart the success of women faculty. The hiring and evaluation processes typical to many higher education institutions frustrate access to and promotion within the system for many women. Fox (1989) notes 58 that women were often less likely to be chosen for a teaching and research assistantship in graduate school. This often led to the exclusion o f women from the informal networks o f their discipline and resulted in a marginality that continued to afflict their futures in finding employment and research funding. Kaufman and Perry (1989) cite traditional and broadly shared university policies which prohibit the hiring o f spouses, the hiring of one’s own graduates, and the shift to tenure track positions by previously employed lecturers as examples o f policies which have discriminatory effects. Cole (1979), Theodore (1986), and Simeone (1987) note that the criteria for promotion in rank often included those which were not related directly to perform ance in teaching and scholarship but also considered issues of marital status and undocumented presumptions about future productivity. For other researchers within this school o f thought, the prim ary structural barriers for women desiring a wider range o f academic career options focus on reasons typically given for lower rates o f productivity among women scholars. They see this as a key issue in the development and promotion o f women faculty for positions in major research universities. The whole issue o f how scholarly productivity is counted is an open question. Bernard (1964) first pointed out the flaw o f counting quantity o f publications without evaluating quality. Dwyer(1991) argues that quality o f scholarship is consistent with that o f males even though quantity is not equal. However she also points out that quantity varies 59 across the disciplines and that blanket statements about gender-related productivity should be suspect. Chamberlain (1988) similarly questions the idea that scholarly productivity can be easily counted. She notes that married women and single women publish different kinds o f results. W hile married women are often counted as more productive than single women because o f the number o f publications, single women have the higher book publication rate over a career span. In fact single women publish books at a more prolific rate than married men. Access to the inner circles o f scholarship within a discipline is also hampered by the lack o f women who function within that inner circle (Dwyer, 1991). Cole argued that scholarly productivity differences did not correlate significantly with differences in either occupational location or marriage and family status(except for 3 or more children). Thus he rejected the argument that differential productivity in scholarship occurs because of institutional location or family situation. Instead he ascribes more o f the difference to the invisibility and differing perceptions o f work quality which arise because women are not part o f the informal inner circles within a discipline.(Cole, 1979, 1987). M oore and Sagaria (1990) note the absence o f women who function as editors o f leading journals in many fields and suggest that the result is a lack o f sponsorship o f w om en’s scholarship and an inability for many women to penetrate the editorial networks required to secure publication in the field. Though Cole(1987) argues that the turn down rate by journals is not higher for women scholars, he still agrees that there are substantial 60 differences in visibility and in the perceived quality o f work produced by men and women. He attributes this invisibility and perception o f lesser quality to the lack o f entry into these inform al inner circles. M any o f these writers also recognize that women faculty often have not had a conducive environment for research. Traditional locations have hampered productivity and access to research networks. Bernard (1964), Cole (1979), Finkelstein (1984), Simeone (1987), and Dwyer (1991) noted that the prim ary location o f women in smaller colleges and institutions further from the mainstreams o f their disciplines has placed women at a considerable disadvantage in finding time and disciplinary circulation for their research. Professional rank also influences scholarly productivity. Productivity increases for both sexes with promotions in rank (Astin and Davis 1980 study in Cham berlain, 1988). Since w om en’s rates of promotion are slower than those o f men, the access to resources and the potential shift in responsibilities that accompanies a promotion in rank accrue to them m ore slowly as well (Dwyer, 1991). Overall, the conventions for assessing scholarly productivity, limited access to the inner circle o f editors, traditional geographic locations, and slower promotions in rank have thwarted the ability o f women academics to both enter and thrive as scholars and thus as faculty for research universities. 61 In sum, writers in this subset believe that the primary reasons for the differential progress o f women academics in higher education are related to policies, procedures, and norms for scholarship within the higher education culture that have fit reasonably well with male needs and experience, but thwart the needs, access, and development of women faculty. A final subset among those who lean most heavily toward institutional and structural causes are those who examine the results from the standpoint o f labor economics. Their analyses concentrates heavily on the supply o f and demand for faculty in higher education, the relative effect o f that balance on gender desegregation, the income levels o f academics, and mobility within the academic market. They lean most heavily on the presumption that a faculty m em ber’s goals and values, work satisfactions, and allocation o f work effort are the result, not the cause, o f placement in a given position. Bernard (1964) first commented on the placement o f women faculty. She noted the relatively low numbers o f women faculty, but also that they were located prim arily in institutions ranked lower in the higher education hierarchy. In her analysis o f this situation Bernard argued that between 1930 and 1960 the supply of qualified women faculty had declined substantially. Thus the low numbers were the results o f an insufficient number of women choosing to pursue academic credentials at the doctoral level. She also cited the lack of doctorates as one substantial reason for the differential contribution o f women academics. Without the doctorates which provided some 62 important research skills and credibility among peers, wom en’s prim ary focus was teaching with much less energy committed to scholarship. Part o f her solution was for women to improve their degree qualifications so that they became a qualified labor pool for the developing demand in major universities. Improving the supply o f qualified women would help their placement and their ability to function in diverse academic roles. Rossi (1973) analyzed the situation somewhat differently. She argues that the supply o f women doctorates in raw numbers has increased steadily since the 1920’s. She writes, " ...fo r the past fifty years there has been a steady increase in the number o f women who earned the doctoral degree. It is striking that the number o f degrees granted to women continued to climb during the 1950’s and 1960’s despite the great pressure on women to live in conventional domesticity. W hat is especially important is that there was a dramatic increase in wom en’s numbers in the late 1960’s, that is, before academic women became actively concerned about their position in higher education."(p. 516). However, the great difficulty occurred because the rates o f increase differed proportionately between female and male academics. While the supply o f women was consistently growing, the supply o f white male academics was growing even faster in the 1950’s and 1960’s. The result was that women’s proportion o f the aggregate supply was smaller. Rossi argues that this contributed greatly to their loss in relative status (Rossi, 1973). 63 Chamberlain (1988) and D w yer (1991) both indicate that the supply and demand for faculty have changed substantially since Bernard did her research. Chamberlain notes the significant rise in the number o f women doctorates since the m id-1960’s. In fact "the number o f doctorates awarded to men peaked in 1972 and has declined steadily since then, while the number awarded to women has continued to grow" (Chamberlain, 1988, p. 256). The relative proportion o f women in the academic supply pool has increased. However, that supply has not resulted in massive increases in the number o f women faculty in major universities. As Dwyer notes, during the 1970’s and 1980’s major universities were not expanding their academic staffs substantially. Many had made a substantial number of new appointments during the 1960’s and had no need to conduct massive hiring campaigns into tenure-track positions after that time. W hile there were some growth fields such as business and engineering, these were not fields that had traditionally attracted significant numbers o f women. Thus, the limited areas in which university demand did grow still did not match the areas in which women doctorates were in greatest supply. Instead the growth in demand occurred much more heavily among the community colleges as they steadily expanded during the past two decades. Thus, Dwyer and others (Freeman in Becker, 1979) suggest that a share o f the differential placement o f women among types of higher education institutions can be attributed to strong demand for 64 faculty in the community colleges and weaker and highly selective demand for faculty in four-year institutions and research universities. Other researchers criticize the presumption that a faculty member initially hired at the community college level will necessarily stay in the community college circuit. They suggest that faculty can move into major four-year institutions from that initial placement. However, given weak demand in general among major universities, the path into them frequently requires geographic mobility. Kaufman (1989) documents the "moving ethic" for faculty as they try to improve their relative positioning in the mainstream o f teaching and research. She suggests that this ethic has hampered opportunities for women academics who frequently perceived their geographic mobility to be more constrained than that of male academics. Yet Breneman and Youn issue a caution. From their research on the academic labor markets and prestige, it appears much less likely that such a transition from two-year to four-year institutions will occur. They indicate that one’s career path is determined prim arily by the prestige of one’s graduate institution and the prestige associated with a faculty mem ber’s first appointment. Furtherm ore they also note that it is more likely for a faculty member to move from research into teaching than vice versa. Strong teaching evaluations and high scholarly productivity at the two-year college level may be insufficient to overcome the prestige bias. (Breneman and Youn, 1988). 65 Even for women who have found some marginal employment with universities, typical university policies related to hiring often promote this "moving ethic" as the route to tenure track employment, advancement, or a better fit o f a faculty mem ber’s interests and needed resources with what an institution will readily provide. Rossi (1973) and later Kaufman (1989) document several typical university policies which thwart the promotion o f women in oversubscribed fields when a tenure-track opportunity does arise. Often policy stipulates that tenure track appointments may not be filled by spouses o f faculty, those who have previously lectured in the department, or the university’s own graduates. Some schools even have policies that require non-tenure track lecturers to shift to part time status after a certain num ber o f full time years o f teaching. There appear to be few good routes from a marginal position into a tenure-track one without moving to another institution. Pfeffer and Ross (1990) hypothesize that the strength o f the moving ethic and the geographical constraints on women, whether external or self-imposed, have contributed to substantially less bargaining power related to salaries. In their studies o f higher education salaries, after controlling for variations in education and relevant experience, they attribute much o f the remaining difference in salaries between academic men and women to differences in negotiating strength. Because o f geographical limitations, women seem to have a much small set o f viable alternatives to their present situation. Thus they have much less power in arguing for market adjustments to their salaries in institutions that have discretionary funds available for salary adjustments. 66 Pfeffer and Ross extend the impact o f mobility further. They suggest that those who are geographically bound within the higher education system frequently have less discretionary power within an organization. They then relate it to gender-based wage discrimination. Their research indicates that greater compensation discrimination related to gender occurs in private, larger colleges and universities with slack resources. (Pfeffer and Ross, 1990). In other words those institutions with more resources balance their discretionary spending decisions related to compensation against the negotiating power o f individual faculty members. That power of faculty members is frequently tied to issues o f relative geographic mobility. Thus, women with less geographic mobility wield less negotiating power and are frequently paid less. During the 1970’s several studies noted that women academics were still more likely to regard their spouse as the more significant determinant o f where they would live. (Centra, 1974 and Wallston, 1978 in Simeone, 1987). These researchers would agree with C ole’s reasoning about geographic mobility when he wrote: "Nonetheless, on the whole, women scientists are not as mobile as men, more often feeling tied to a particular geographic location because o f the work requirements o f their husbands. To what extent do women scientists refrain from applying for or accepting positions in outstanding departments located away from their husbands’ place o f work? How does this limit their bargaining position, as compared to men scientists, in the use o f offers from competing colleges and universities to improve their salaries and other perquisites at their own institutions? The combined results o f accommodative self-selection resulting in restrictions on the actual 67 mobility o f married women scientists and the immobility imputed to women scientists by their colleagues obviously contribute to a process o f accumulating disadvantage." (Cole, 1979, p. 12). The later 1990’s could open new opportunities for replacement faculty as those hired during the expansive 1960’s retire. However, institutions will be looking nationally for those replacements. If women academics are still less mobile or at least perceive themselves to be less mobile than men, their ability to seize such opportunities is severely hampered. (Bowen and Shuster, 1986). The above findings fit a more general pattern o f occupational sex desegregation related to supply and demand in professional fields as researched by Jolly, Grimm, and W ozniak(1990). These researchers found that sex desegregation occurred most rapidly in fields with high levels o f occupational growth, a trait not characteristic o f academic employment in the past two decades. Furthermore their research indicated that women tend to enter professional fields that are male dominated as demand in those fields begins to decline. Typically, although these fields are paying men relatively less, they are paying women relatively more than other professions in which their gender is dominant. These general findings about patterns o f occupational sex desegregation seem to fit the situation o f higher education during the past two decades. With the exceptions engineering, business, and computer science, the 1970’s and 1980’s were not decades of strong demand for faculty. The purchasing power o f faculty declined because o f severe 68 inflation. (Bowen and Shuster, 1986). Greater numbers o f academic women entered the profession as demand was declining and real income levels were decreasing. Typically their specialties were not in the few growth fields. Yet for many o f them it still represented a better income opportunity than many traditional avenues o f career employment. The income of the academic profession related to labor demand when compared to other professions has a significant impact on women faculty. Because peak incomes in higher education as an occupation are substantially lower than those in some other professional fields such as law or medicine, the occupation itself begins to shape the choices which women make. Cooney and Uhlenberg (1989) in their research on family building patterns o f professional women note that the decisions o f women regarding marriage, divorce, and child birth differ among these three occupations significantly. Women faculty are less likely to marry, divorce more frequently, and have fewer children than women doctors. They are also more likely to work part time than women in either of the two other professions. W hile these differences might be explained in part by personality, values, and the structure o f career paths, Cooney and Uhlenberg suggest that the opportunity costs for women vary among these professions. Because academia pays less than the medical or legal fields, women in the other two professions can more easily afford quality child care and household help. Thus marriage and children are more o f a detriment to career progress and related income for women within the academic profession than they are in professions with higher levels o f income. 69 As a school o f thought socialist feminists align themselves most closely with the explanations outlined above. They would suggest that the white male majority, operating from positions o f capitalistic and patriarchal power in society have both the economic and political resources to oppress the females in society. They would indeed agree that the individual’s role is predominantly and perhaps essentially shaped by the broad structures o f society (Stromquist, 1989). There is little belief in the possibilities o f human agency, resistance, and change in the struggle against currently dominant forces. For socialist feminists the struggle pits white, well-educated, financially endowed males as a class against women, black and white, who are both less economically and politically powerful. Class struggle ideology assigns a clearer perspective on reality to those who are oppressed; but short o f an uprising among the oppressed there would be little hope that the dominantly male structure o f influence and power in academia would change. They would agree that past legislation against sex discrimination and protests against unfair university personnel practices have brought little or no success to women victims o f discrimination personally; instead such efforts have only negative consequences for their future careers (Theodore, 1986). In summary various explanations for the differences in status, work allocation and preference, and career strategies and mobility are given. Both theorists with varied disciplinary backgrounds and feminists with differing strategies and ideologies for change fall at different points along the explanatory continuum. Some theorists along with radical feminists place more emphasis on individual choices which result from natural 70 differences. In such cases the burden is on individuals. They, must develop their human capital to fit a seemingly gender-neutral structure o f activity, evaluation, and reward. The argument is that women more likely fit different functional niches than men academics, with more emphasis on teaching, a different type o f scholarship agenda, and different professional needs. With more developed individual capital via excellent channels o f education and experience, natural differences will be allowed to flourish. However, this strategy may not lead to balanced faculty staffing in all types o f higher education institutions. socialized ones. Therefore other theorists argue that the root differences are Many o f them along with liberal feminists agree that changing the socialization of females will be necessary to expand the range o f individual preferences that find their fulfillment in a broader set o f institutions. Yet in both cases placement and function follow individual preference. Neither o f these two groups argues otherwise; the only question is the origin o f individual preferences. A different set o f theorists places more emphasis on male-dominant structures, cultures, and policies which determine and constrain individual activity and choices. In their eyes placement and function have effected individual preference. In such cases organizational policies, the whole academic culture, and the larger labor market structures must be changed. Only then will more diverse options in preferred work roles be allowed to flourish and further change the climate o f higher education institutions. 71 COM M ENTARY ON THE LITERATURE ON W OMEN FACULTY IN HIGHER EDUCATION How should the above literature be critiqued? There are several difficulties with the descriptive and explanatory literature as it has been developed to date. They involve either the nature o f the analytical sample and the resulting content o f the analysis, the assumptions o f the analysis, or the disciplinary source o f the analysis. The Content o f the Analysis The first difficulty is a lack o f consistent categorization o f the women academics who have been sampled. While one study will differentiate female academics by disciplinary interest, another will not. Thus some studies offer general explanations without examining the variety o f responses that may be tied to different academic interests. W hile very little research has focused on differences in activities among the disciplines, the small amount o f research that has been completed indicates that the preferred allocation o f time varies by discipline. It appears that faculty in education and the fine arts have a stronger preference for teaching than those in natural science. (Austin and Gamson, 1983). Astin and Bayer (in Becker and Lewis, 1979) also found that scholarly productivity is greatly affected by one’s discipline. Natural science faculty produced more books and articles than social scientists and those who taught in the arts or humanities. 72 Similarly, when analyzing differences between male and female academics, the type of employing institution from which responses are drawn is often not a controlled variable. Such studies do not frequently differentiate whether their responses related to measures o f work satisfaction and preferred allocation o f work time vary by type o f institution. Such a lack of differentiation is counter to evidence that the allocation o f faculty members’ time is related to the type o f institution (Austin, 1983). Furtherm ore, scholarly productivity is high even in universities o f lower quality than in schools classed as colleges (Fulton and Trow, 1974 as cited in Becker and Lewis, 1979). In addition the studies that have been done do not differentiate female academics by personal factors which may be key sources o f differences. Only Rossi (1973) substantially addresses the perspective o f black women academics differentiating it from that o f white women academics. Only Bemard(1964) and Chamberlain(1988) differentiate the choices o f women by looking at their marital status, though they did find substantial differences. There appear to be no substantial analyses which evaluate the cross-section o f women faculty by either class origin or religion, two factors which could make substantial differences in their perspectives and motivation. Some studies such as those o f Aisenberg and Simeone do not control for sampling variations because o f geography. Both o f their studies drew their sample from women academics in northeastern urban areas o f the United States. Their findings may not represent a broad cross section o f women academics since the northeast United States has 73 certain factors unique to its historical involvement in the enterprise o f higher education. The strong population o f well endowed private universities and o f traditional wom en’s colleges in that geographic area may have a significant effect on the sampled population. These studies placed limited control on and discussion o f factors other than gender which could significantly affect the findings. Conclusions from research findings are also hampered by the shifting time frames in which the research was conducted. Research has been gathered over approximately thirty years. During the 1960’s only Bernard’s book (1964) could be counted as a major work. The same is true of Rossi’s w ork (1973)in the 1970’s. Not until the mid to late 1980’s are there more researchers in this field. Over the decades the women academics who have responded to these studies have certainly changed in character. The proportion who have married is higher. The range o f disciplines in which they are involved is substantially broader. The range o f institutions in which they are employed and in what capacities has diversified somewhat. Thus generalizations about current women faculty based on past research must be regarded with caution. W hile there may be some enduring characteristics, it is much more likely that by the 1990’s the character of women as faculty members has changed from what it was in the 1960’s. Historical findings about women academics must be treated historically, not as generalizations about current women faculty. 74 The content o f the research findings also seems to focus heavily on analyzing women’s scholarly productivity and documenting salary differentials by gender. W hile both of these are worthy topics for research and analysis, by comparison, less has been done with other worthwhile topics. Almost no researchers have substantial comments (Olsen, 1991 excepted) about service opportunities both within and outside the academy. There is very little known about whether male and female academics shape their careers in distinctly different ways related to service opportunities. Finally, although geographic mobility or lack thereof is often the source o f musings, it is not always a factor that is carefully documented in terms o f real differences between male and female academics. Differences between one’s perceived mobility and one’s actual ability to move are not clearly distinguished. Nor have many links been drawn between perceived mobility and one’s choices in career development. Could it be that a perceived lack of mobility is a significant factor in one’s decision to invest energies in teaching, service, or administration? As geographic mobility becomes a more important factor in the national academic market possibly affecting rank, salary, and academic recognition, it should be a factor for more attention. The Assumptions o f the Analysis Professional rewards come from personal, institutional, and disciplinary sources, and the importance o f each o f these reward sources has shifted over time (Finkelstein, 1984; 75 Baldwin, 1985). The stakeholders in higher education have not embodied a constant culture o f values and the rewards associated with them. Historical research documents significant shifts from early 19th century colleges emphasizing teaching to growing institutions that balanced teaching and public service in the mid 19th century to the development o f large universities and professional disciplinary societies that have been increasingly focused on research since the 1950’s (Jencks and Reisman, 1968). Differing assumptions about institutional cultures, faculty life cycles within them, and faculty reactions to rewards offered influence the resulting analyses that have been done. Furthermore, women may have different sensitivities to rewards than men. Male and female faculty members may react differently to the institutional cultures represented in their systems o f reward, evaluation, and development. In some ways male and female faculty are becoming more similar in their sources o f rewards. For males, the power of the institutional reward system over their behavior has been tempered by their strong ties to national societies in their disciplines and the related research visibility. Women, though less tied to such societies in the past, now regard them important like the men. Yet in other ways male and female faculty are still dissimilar in their valued sources of reward. W hile women and men have equal regard for their disciplines, women faculty are still less loyal to the concept o f faculty tenure than their male counterparts (Boyer, 1989). Given such shifting valuation o f reward sources, male and female faculty may not react to rewards in the same way. 76 In addition, incongruence with one’s institutional reward system may or may not result in significant dissatisfaction with the institution and a decision to change institutions. Finkelstein (1984) citing research by Borland(1970) and DeVries(1975) is not convinced that incongruities between personal work load preferences and institutional evaluation/ reward systems result in changed individual choices. Responsiveness to the institutional reward system is conditioned by how much weight an individual attaches to it and how much power it has over them. Research by McKeachie (in Becker, 1979) suggests that the job itself, its intellectual stimulation, collegial environment, and social significance may provide personal rewards that are much more highly valued than rewards provided by any outside group. Not all faculty members, particularly in larger institutions, believe they must conform closely to the prevailing institutional model. Thus, even when institutional rewards are not as congruent for females as for males, they may not proclaim serious public dissatisfaction or a desire to change institutions. Furtherm ore, most o f the research does not link findings about faculty to the emerging literature about faculty life cycles. Research to date has been based on the presumption that the norm for a faculty career is the straight path from undergraduate to graduate school, from graduate school to full-time employment, and from an entry level position to a consistent climb through the faculty ranks, typically moving from more emphasis on teaching to more emphasis on scholarship. 77 During the 1970’s and 1980’s new research began to link adult life cycle theories (Sheehy, 1976, Levinson, 1978, Aslanian and Brickell, 1981) with theories o f faculty evaluation and career development. Both Fumiss(1981) and Baldwin(1982, 1985) described faculty development as a series o f stages, each o f which may have differing needs and therefore differing emphases on teaching, service, or scholarship. Austin(1983) cites research indicating a saddle-shaped curve for scholarly productivity over the course o f a faculty career may be more normative. It also appears that faculty service increases as faculty members progress in their careers. Baldwin(1985) along with others began to argue for more flexible models o f faculty work. He emphasized the need for structuring faculty reward systems to foster faculty vitality and to create greater emphasis on systematic faculty development. Based on a broader approach to faculty life cycles and development Eble and McKeachie(1985) recommended changes in faculty development approaches to improve the quality o f undergraduate instruction. Most recently Schuster (Bowen and Schuster 1986; Schuster, 1990) has argued that because the quality of the national professoriate is endangered, special attention must be given to systems which enhance faculty motivation, productivity, and effectiveness throughout the course o f faculty life cycles. Life cycle theory may not be the perfect basis for faculty development; yet at least such strategies require that institutional systems o f faculty evaluation, reward, and development recognize differences in the process o f faculty development and respond appropriately. 78 Researchers in the area o f faculty development provide im portant background for understanding differences in positions, behaviors, and concerns that are found among male and female academics. Their findings result in institutional strategies for cultivating an institution’s faculty, namely the structures for rewarding, evaluating, and developing faculty, strategies that embody the operational culture o f a college or university. Yet these recent findings about faculty life cycles and arguments for more flexible faculty evaluation and reward systems have not yet been tied to the literature on the work efforts o f academic women. M ajor researchers, with the exception o f Jesse Bernard (1964, 1981) have not carefully linked distinct differences between the work efforts o f male and female academics to differences in the male and female life cycle. Even faculty life cycle theory to date seems to presume a reasonable homogeneous stage theory that does not differentiate by gender. It presumes a similar set o f stages at a similar pace for both sexes. Overlooking possible gender differences in faculty life cycle leaves a tremendous gap in the foundation o f most scholarship about male and female faculty differences. The Disciplinary Source o f the Analysis Finally, it is clear that the research is fragmented by discipline. With the exception of Bernard and Dwyer, other researchers seem to search for explanations about differences between male and female academics only from the perspective o f their home disciplines without full recognition o f the multitude o f factors involved. The explanation for the current state of affairs is a complex combination o f personal, organizational, economic, 79 and cultural variables. Researchers who tend to generalize from one discipline miss the rich contributions o f other disciplines to the explanation. Explanations from the field o f organizational behavior seem particularly underdeveloped. The primary w ork that has been done is either by psychologists, sociologists, or economists. Psychologists may attribute too much influence to personal factors. Sociologists and economists focus on societal factors. There has been little recognition o f the role that discrete organizations play in shaping the behavior patterns and preferences of individuals. W hile some o f these organizations may reflect more dominant patterns in society, it is also quite possible that their organizational cultures run contrary to the dominant culture for a variety o f reasons. Such reasons may have to do with powerful leaders, unique histories, unique missions, unusual faculty cohorts, or peculiar resource configurations which congeal into unique organizational cultures. With the exception o f Kanter (1977) and Morrison (1987) few theorists have evaluated the status o f women academics from the perspective o f organizational theory. Kanter and M orrison’s works are focused on organizational analysis within business corporations, not within higher education. There appear to be no major theorists who have developed analytical frameworks from the school of organizational behavior and culture, carefully using an institutional typology to evaluate the status o f women in higher education. 80 TH E RESEARCH UNIVERSITY AS A CONTEXT FOR THIS STUDY The women attracted to a research university climate may be far different from those who choose employment in liberal arts colleges. Both Jencks and Reisman (1968) and Finkelstein (1984) document the development o f the unique character o f the research university. Since W orld W ar II American society has gradually developed and supported a limited set o f higher education institutions in which research is a key justification for existence. The development o f such institutions was spurred by the growth in federally funded research in the past four decades, the spreading affirmation o f PhD specialization among faculty, the rise o f the graduate research model o f professional development, and the broadening power o f faculty and their disciplinary societies regarding the range and configuration o f the curricula. It is important to note that the difference between the research university and other types o f higher education institutions is not in whether faculty have an allegiance to their disciplines. Boyer documents that self-reported allegiance to one’s discipline is about the same for faculty across all institutional types. While 77% o f faculty in research universities regard such allegiance as "very important", so do 76% o f those in liberal arts colleges and 81% o f those in two year schools (Boyer, 1989). The difference is how that allegiance is manifested. It is confounded by institutional type. Institutions, identified as Research I and II schools in the Carnegie classification 81 scheme, are distinctly different from other types o f higher education institutions in the operation o f disciplinary allegiance. Clark (1987) compares the unique patterns o f time allocation in the research university with those o f other schools. The general pattern for Research I and II universities is a teaching load o f 4-6 hours per week. This compares with a normal load o f 9-12 hours per week at other four year institutions. Thus much more time is available for scholarship in the research universities. In Research I schools 58% of faculty spend over ten hours per week in research compared to 23% in the Liberal Arts I college (Clark, 1987). The 1988 NSOPF Faculty Survey (SRI International, 1990) confirms these results. From the table below it is apparent that faculty spend a higher percentage o f their time doing research in the research and doctoral universities than in other types o f institutions. While percentages o f time spent in administration, community service, and professional development do not differ appreciably by type o f school, the shift from teaching to research activities is a major one for those in research institutions. 82 TABLE 5 - PERCENTAGE OF TIM E SPENT ON VARIOUS ACTIVITIES BY FU LL-TIM E REGULAR FACULTY, BY TYPE AND CONTROL OF INSTITUTION - FALL 1987 Percentage o f time spent: Type and control o f institution All institutions Public research Private research Public doctoral Private doctoral Public comprehensive Private comprehensive Liberal arts Public two-year Other Teaching Research Admin. Community Service Other work Prof. Devel. 56 43 40 47 39 62 62 65 71 59 16 29 30 22 27 11 9 8 3 9 13 14 14 14 13 13 14 14 10 15 4 3 2 3 2 4 5 5 5 5 7 7 11 9 14 5 6 4 5 7 5 4 4 5 4 4 4 4 5 6 Source: SRI International for the National Center for Education Statistics, 1990. Furthermore, in Research I universities 66% o f the faculty teach undergraduate students less than four hours per week; Thirty-one percent o f them teach no undergraduates. Seventy-nine percent o f these faculty teach some graduate students; twenty-three percent o f them teach only graduate students. By comparison in the Liberal Arts I college only 12% o f the faculty teach undergraduate students less than four hours per week; only 1 % o f faculty in such colleges teach only graduate students. (Clark, 1987). It is apparent that when teaching does occur, the research universities have a very strong orientation toward graduate students. 83 Both the orientation toward graduate students and the time available for investigation shape research university character. The research interests o f the faculty are the primary force in the organizational culture. These faculty do not believe that the primary criterion for promotion should be quality teaching (Boyer, 1989). As a replacement for that belief, 94% o f them see the number o f publications as important for tenure, a measure o f merit in which research faculty place great faith. (Boyer, 1989). power is dominant, particularly related commitments. Faculty to their disciplinary involvements and In general there is little faith in central academic administration; the structure o f it is typically lean and control over faculty is loose. Clark does note that the structure o f control varies by discipline with the humanities having the loosest structure o f control within the research university and the professional schools having the tightest structure. Yet, overall, the research university is characterized by great professional flexibility, power and commitments at the departmental level, and a sense o f calling first to one’s disciplinary research, not to institutional missions in education. These are the highest sources o f career satisfaction for faculty in research institutions. The resulting pattern o f organization represents a loose coupling between the technical core and the administrative core o f the university. In some typologies o f higher education institution, such patterns are referred to as "organized anarchy" (Bimbaum, 1988). Yet there is an underlying logic to the resulting power o f departments and individual scholars if one accepts the premise that individual research and scholarship are the driving forces of such institutions. 84 As has been noted throughout this literature review, research and writing about women academics has not treated differences in institutional typology carefully. Thus there is little scholarship specifically about women faculty in research universities. W hat is known is that they are few in number and typically spread in the lower ranks. They comprised only 8.0% o f the full professors in 20 major research universities in 1989 (AAUP, 1989). Yet very little is known about their job satisfaction, allocation o f time, or their geographic mobility. Often the presumption is that they fit more general patterns among female faculty who seem to prefer teaching, gravitate toward a few selected disciplines, have complex career development paths, and prefer scholarship which is applied instead o f basic. Only O lsen’s 1990 study at Indiana University, a Research I institution, gives some indication that their career paths and preferences may be more similar to the male faculty in that institution than to female faculty who populate other types o f higher education institutions. Further investigation o f the research university context is needed to provide a fuller picture of women as academic professionals. Thus, the exploratory research of this dissertation on the faculty within Michigan State University, a Research I university, was undertaken. The purpose o f the study was to begin to fill this gap in knowledge about possible gender differences among faculty within such research institutions. By looking intensively at the experiences o f faculty in one institution, an in-depth investigation can provide a nuanced analysis o f professional satisfactions, the allocation o f work effort, and geographic mobility. By studying one institution, these variables can be closely 85 examined. The results should contribute to knowledge about any differences between male and female faculty and possible reasons for these differences in the research university context. CHAPTER 3 RESEARCH METHODOLOGY Researching whether faculty gender differences are significant in a research university context requires a substantial data base which can be analyzed from several vantage points to address key questions. This chapter will describe the source and configuration o f the original data base on Michigan State University’s faculty used for this study. It will then explain how that data base was manipulated for the purposes o f this study. Finally the chapter will focus on the research methodologies used on this data base to explore the following specific questions: 1. Are male and female faculty in Michigan State University, a Research I institution, different from each other on measures o f professional satisfaction, allocation of work effort, and geographic mobility? 2. If significant differences between male and female faculty are found, can they be prim arily and/or directly attributed to the effects o f gender? 3. If these differences cannot be directly attributed to gender, what other factors are significant contributors to differences between male and female faculty on these three 86 87 measures? Specifically do age, marital status, children in the family, or the culture of one’s college within the university, contribute to noticeable male/female differences? These foci, professional satisfaction, allocation o f work effort, and geographic mobility, were chosen for investigation because little is known about these factors and their relation to gender in the research university context. The literature search revealed that much more investigation had occurred into differences in professional degree, rank, salary, and some other factors of career mobility in the research university. However, it appears that differences between male and female faculty regarding the nature o f the work itself, namely effort exerted and satisfactions gained, and the issue of geographic mobility related to career mobility have been less carefully explored. The intent o f this exploratory dissertation is to begin to fill some o f the gaps in the knowledge base about gender and faculty in the research university context. TH E SURVEY DATA BASE The research data base was constructed from a survey which was sent to all tenured and tenure track faculty at Michigan State University during the spring o f 1991. Several researchers contributed to the development o f the survey instrument. The chief contributors were Dr. Kathryn Moore, chairperson o f the Department o f Educational Administration at Michigan State University; Dr. Philip Gardner, Director of the 88 Collegiate Employment Research Institute housed on the campus o f Michigan State; and Dr. Linda Forrest, Professor o f Counseling and Educational Psychology at Michigan State University. They were ably assisted by Bob Nienhuis, then graduate assistant to Dr. Kathryn Moore. The primary objective o f the survey was to explore faculty career choices and challenges in a changing academic environment. Particular emphasis was placed on relative job satisfaction, the allocation o f work effort between a variety o f academic tasks, the likelihood and rationale for leaving or staying at Michigan State University, challenges presented by dual career employment among academic partners, and university issues of concern to faculty. Some sections of the survey were constructed by the researchers. Other sections were borrowed from related national surveys o f faculty with the hope that some o f the data gathered could be compared with national norms. administered is shown in Appendix A. The survey as All respondents to the survey were promised confidentiality regarding their responses. The survey was mailed to all Michigan State University faculty who were tenured or in tenure track appointments. Subsequent to the initial mailing, a second copy o f the survey with a cover letter was sent to all non-respondents to increase the response rate. Phone calls were made to several colleges within the university to encourage higher response rates in schools with low response rates. A decision was made not to further pursue non­ respondents beyond this point. The end o f the spring academic term was approaching; 89 the researchers did not want to unnecessarily irritate busy faculty; and the response rate was judged to be sufficient for the researchers’ purposes. A comparative summary o f the survey response rates to total university faculty by unit o f university appointment and gender is shown in Table 6. The table indicates that a slightly higher percentage o f women than men faculty responded. Yet in both cases a usable sample o f those surveyed did respond. TABLE 6-COLLEGE DISTRIBUTION OF UNIVERSITY FACULTY AND THEIR SURVEY RESPONSE RATES BY GENDER College Total College Faculty F M 246 178 119 47 96 124 11 94 16 271 2 106 152 1 88 24 13 1588 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 All co lleg es Total survey respondents M T 42 70 24 15 56 9 33 21 3 43 24 30 43 - 12 11 4 440 288 248 143 62 152 133 44 115 19 314 26 136 195 1 100 35 17 2028 F 156 81 49 29 64 48 5 49 5 113 2 35 82 3 37 8 7 773 Survey respondents as % o f college population T 15 44 11 10 31 7 27 14 3 20 10 7 27 - 5 1 5 237 M 171 125 60 39 95 55 32 63 8 133 12 42 109 3 42 9 12 1010 63.4 45.5 41.2 61.7 66.7 38.7 45.5 52.1 31.3 41.7 100.0 33.0 54.0 100.0 42.0 33.3 53.8 48.7 plus missing cases = _14 survey respondents = 1024 C o lle g e K ey: 35.7 62.9 45.8 66.7 55.4 77.8 81.8 66.7 100.0 46.5 41.7 23.3 62.8 - 41.7 9.7 100.0 53.9 59.4 50.4 42.0 62.9 62.5 41.4 72.7 54.8 42.1 42.4 46.2 30.9 55.9 100.0 42.0 25.7 70.6 49.8 50.5% response rate 1 = A g ric u ltu re an d N a tu ra l R eso u rc e s 7 = H u m a n E co lo g y 13 = S o cial S c ie n c e 2 = A rts an d L etters 8 = H u m an M ed ic in e 14 = U rb a n A ffa irs 3 = B usiness 9 = la m e s M ad iso n 15 = V e terin a ry M ed icin e 4 = C o m m u n ic atio n s A rts 10 = N atural Science 16 = N o n -C o lle g e F acu lty 5 = E d u c a tio n 11 = N u rsin g 12 = O steo p ath ic M edicine 17 = O th e r 6 = E n g in ee rin g Total F 91 The survey data base can also be compared in the distribution o f responses by academic rank to the total university rank distribution as shown in Table 7 below. TABLE 7-COM PARATIVE RESPONSE RATE BY ACADEM IC RANK Academic Rank M SU faculty distribution Survey response distribution Professor Associate Assistant Specialist and Other* 54% 27% 15% 4% 55% 24% 17% 4% TOTAL 100% 100% *Note: Specialist and other includes all members o f the academic staff who have teaching/advising responsibilities but do not have faculty rank. From Table 7 it is clear that the survey elicited responses by rank which approximate the rank distribution in the university. Survey responses can also be described by examining the number o f male and female respondents by academic rank in each unit o f appointment as shown in Table 8. Most colleges had a substantial number o f respondents, although the number was very small in a few colleges. TABLE 8-SURVEY RESPONDENTS BY UNIT OF APPOINTMENT, RANK, AND GENDER Unit of Appointment Specialist M 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 Totals Assistant F 2 1 1 3 2 9 M 2 1 3 1 4 2 1 5 3 22 18 8 6 8 3 13 1 5 2 18 2 2 13 7 106 Associate F M 5 14 6 4 7 2 4 2 2 4 5 4 9 1 1 70 30 14 14 9 18 7 12 15 13 18 2 10 1 163 Full F 3 14 4 3 12 1 12 7 5 2 2 6 4 75 M 105 59 28 12 42 27 4 32 3 77 20 49 1 20 7 486 Total Other F M 5 16 11 9 4 1 6 3 1 9 65 F 1 1 7 9 _ 5 5 Plus missing cases Key: 1 2 3 4 5 6 = = = = = = Agriculture and Natural Resources Arts and Letters Business Communications Arts Education Engineering 7 = Human Ecology 8 = Human Medicine 9 = James Madison 10 = Natural Science 11 = Nursing 12 = Osteopathic Medicine 13 14 15 16 17 = = = = = Social Science Urban Affairs Veterinary Medicine Non-College Faculty Other 171 125 60 39 95 55 32 63 8 133 12 42 109 3 42 9 12 1010 14 1024 93 Finally, the relationship o f survey respondents’ age to their rank and sex should be noted. Table 9 displays the mean age o f the survey respondents by both rank and sex. TABLE 9-M EAN AGE OF SURVEY RESPONDENTS BY RANK AND SEX Rank Full Professor Associate Professor Assistant Professor All Ranks Female 53 44 39 45 Male 54 45 37 50 Sex Note: Average age o f all survey respondents is 49. There is little difference between the male and female respondents in terms o f the relationship between their mean ages and related academic ranks. At most they differ by an average o f two years. Yet the total mean across all ranks varies a bit more with females averaging five years younger than males. To a large degree this simply reflects the larger percentage o f women respondents from the assistant and associate ranks. As Table 8 shows, female respondents are more evenly spread by rank, while the average o f the male respondents is much more heavily weighted by those in the full professor rank. In general the survey respondents’ group corresponds well to the rank distribution o f the university. The survey results also indicate that age o f respondents, whether male or 94 female, is very similar by rank. In these regards the survey is a very adequate sample o f the population. Yet the response rates by college show more variation. The range by college is from a 25.7% to 72.7% response rate. Thus it is clear that some colleges are more strongly represented in the survey results than are others. W hile the respondents’ rank distribution may be representative, their distribution by college is less representative. There are also significant variations in the male versus female response rates. The average female respondent was younger than the average male by five years. The female response rate to the survey was also 5.2% higher than the male response rate. These variations do color the aggregate results. The variation in response rate was also greater when it was reviewed in particular colleges. Only three o f the colleges have response rates for males and females within five percentage points o f each other. In all o f the other colleges the spread in response rates is larger. In the remaining fourteen colleges men have higher response rates in six o f them and women have higher response rates in the other eight. In examining individual colleges with differences between male and female response rates greater than 25 %, it is clear that in most cases it is the result of a very small group o f males or females from whom to draw responses. For example, in the College o f Nursing, males have a 100% response rate. However, this simply indicates that the two total male faculty in that college did respond. When there are only two faculty o f a given sex who could possibly respond, the differences in response rates 95 are understandable. Only in the College o f Agriculture and Natural Resources and the College o f Human Ecology are the pools of both males and females large enough to conclude that the differences in response rates are highly significant. Given that the survey results show such differences in response rates by colleges and by the male and female faculty within them, it is more logical to w ork with the survey data in the aggregate than by specific college. Specific differences by college will be smoothed out in a larger pool, although it is still important to note the overall higher response rate o f female faculty. USE OF TH E SURVEY Given the broad range o f questions asked in the original survey it could be used to investigate a wide range o f faculty characteristics and attitudes. However, given this study’s focus on selected areas o f analysis related to possible gender differences, only certain survey questions were used for further statistical analysis. Each o f these questions appears below with the assigned variable names appearing on the right side of the page with their respective questions. The following questions were used from the Faculty Mobility Survey (see Appendix A for original survey): 96 FACULTY MOBILITY SURVEY - SELECTED QUESTIONS Part I. 1. Questions in this section concern your academic appointment and the general level of job satisfaction you experience in your current position. W hat is your current academic rank at Michigan State University? (PLEASE CIRCLE ONE NUMBER) Original Variable name: Rank Professor Associate Professor Assistant Professor Instructor Specialist Other: 5. 1 2 3 4 5 6 In which college or unit is your primary appointment? (PLEASE CHECK ONE) O riginal Variable Name: College ___ a. Agriculture and Natural Resources b. Arts and Letters c. Business d.Communication Arts e. Education f. Engineering g. Human Ecology ___ h. Human Medicine ___ i. James Madison ___ j. Natural Science ___ k. Nursing ___ 1. Osteopathic Medicine m. Social Science n. Urban Affairs o. Veterinary Medicine p. Non-College Faculty q. Other: 97 8. How satisfied or dissatisfied do you personally feel about each o f the following aspects o f your job at Michigan State University? (PLEASE CIRCLE ONE NUM BER FOR EACH ITEM) Very Som ew tai DiaM iitfied D uM U ifkd Neutral Some whet S u iifie d Very Satisfied Not Applicable Original Variable Name My w ork load 1 2 3 4 5 8 Sail My job security 1 2 3 4 5 8 Sat2 T he authority I have to make decisions about w hat courses I teach 1 2 3 4 5 8 Sat3 The authority I have to make decisions about content and methods in the courses I teach 1 2 3 4 5 8 Sat4 The authority I have to make decisions about other (noninstructional) aspects o f my job 1 2 3 4 5 8 Sat5 Time available to w ork on scholarship and research 1 2 3 4 5 8 Sat6 The mix o f teaching, research, administration, and service (as applicable) that I am required to do 1 2 3 4 5 8 Sat7 Opportunity for my advancement in rank at M ichigan State University 1 2 3 4 5 8 Sat8 Time available for working with students as an advisor, mentor, etc. 1 2 3 4 5 8 Sat9 98 Very D iuatiafied Somewhat D iualiiiied Somewhat Saliificd Very Satiified Not Applicable Original Variable Name Availability o f support services (including clerical support) Availability o f equipment (personal computers, etc.) Freedom to do outside consulting My salary My benefits, generally SatlO S a tll S atl2 Satl3 S a tl4 Overall reputation o f Michigan State University SatlS Institutional mission to c a n y out teaching, research, and public service S a tl6 Quality o f leadership in my department/program S atl7 Quality o f chief administrative officers at M ichigan State University SatlS Quality o f my colleagues in my department/program S a tl9 Quality o f graduate students whom 1 have taught here Sat20 Quality o f undergraduate students whom I have taught here Sat21 Teaching assistance that I receive Sat22 Research assistance that I receive Opportunities for professional growth and development offered by my academic unit Sat23 Cooperation offered by support staff at Michigan State University Sat24 Quality o f faculty leadership (e .g ., Academic Senate) at Michigan State University Sat25 Relationship between administration and faculty at Michigan State University Sat26 Interdepartmental cooperation at Michigan State University Sat27 Spirit o f cooperation among faculty at Michigan State University Sat28 Quality o f my research facilities and support Sat29 My job here, overall Sat30 99 9. Please estimate the percentage o f your total working hours that you spent on each o f the following activities during the 1990 Fall Term . (PLEASE GIVE YOUR BEST ESTIMATES IF NOT SURE: IF NONE, ENTER "0") Note: The percentages you provide should sum to 100 % o f the total time you spent on professional activities. Percent Original Variable Name Teaching (preparing courses: developing new curricula: teaching; grading papers.) Pteach Research and Scholarship (planning for and conducting research; preparing for and giving performances and exhibitions in the fine arts; preparing or reviewing articles or books; preparing for and attending professional meetings or conferences; seeking outside funding, including proposal writing.) Prsp Advising Students (advising undergraduate and graduate students; working with student organizations.) Pads Professional Development (talcing courses: pursuing an advanced degree or participating in other practices to remain current in your discipline.) Ppdt Service and Extension (preparing and giving speeches that build upon your professional expertise; providing o f technical assistance, policy analysis, program evaluation, medical or veterinary services, psychological counseling and therapy; consulting outside with or without remuneration.) Psex Administration and Governance (participating in facultv governance; participating in departmental or institutional committees and task forces; managing and coordinating programs or personnel.) Padgv Other (PLEASE SPECIFY!: Please be sure that you r percentages total: Poth 100% 100 Part II. In this section, we ask you to consider the likelihood of leaving your current position to do something else. la . 2. If you had the opportunity to restructure your current position, would you want to do more, less, or about the same amount o f each o f the following? (PLEASE CIRCLE ONE NUM BER FOR EACH ITEM) M uch Less Somewhat Less Same Am ount as I Now Do Somewhat M ore Much More Original Variable Name Teaching l 2 3 4 5 Rscpl Research and Scholarship l 2 3 4 5 Rscp2 Advising Students l 2 3 4 5 Rscp3 Professional Development l 2 3 4 5 Rscp4 Service/Extension l 2 3 4 5 Rscp5 Administration & Governance l 2 3 4 5 Rscp6 Given your situation at Michigan State University and the job market in your field, how likely are you to take these actions within the next two years: Very unlikely Somewhat Unlikely Neutral Somewhat Likely Very Likely Original Variable Name a. Seek a new position at Michigan State University 1 2 3 4 5 Tkactl b. Look for a position at another institution 1 2 3 4 5 Tkact2 101 Faculty consider many factors when weighing an opportunity to leave an institution like Michigan State University. Listed below are factors that you may contemplate in deciding to leave the university. Indicate the relative degree of importance each factor could have in making your decision. Not An Important Reason A t All To Leave Somewhat Important Reason to Leave Fairly Important Reason to Leave Very Important Reason to Leave Extremely Important Reason to Leave Original Variable Name Reputation o f institution D eglvl Service Load Deglv2 Availability o f internal research funds Dcglv3 Congeniality o f colleagues Deglv4 Job Security/tenure DeglvS Rapport with departmental leadership Deglv6 Promotion in rank Deglv7 Career advancement Dcglv8 Reputation o f associates Deglv9 Base salary DeglvlO Research load D eglvl 1 Benefit package D eglvl 2 Administrative load Deglvl3 Research opportunities D cglvl4 Teaching load D eglvl5 Teaching assignments and/or opportunities D eglvl 6 R apport with university leadership D eglvl 7 Availability o f internal research funds D eglvl8 Reputation o f department D eglvl 9 Institutional mission/philosophy Deglv20 Influence in department Deglv21 Competence o f colleagues Deglv22 Secretarial support Deglv23 Receipt o f m erit pay Deglv24 Influence in college Deglv25 Library facilities Deglv26 Laboratory/research facilities Deglv27 Office facilities Deglv28 102 Not An Important Reason A t All To Leave Somewhat Important Reason to Leave Fairly Important Reason to Leave Very Important Reason to Leave Extremely Important Reason to Leave Original Variable Name Reduced tuition for family Deglv29 Rapport with college leadership Deglv30 Emphasis on publishing Deglv31 Sabbatical, leave, travel, and study policies Deglv32 Consulting opportunities Deglv33 Spouse's career opportunities Deglv34 Geographic considerations Deglv35 Cultural, recreational, and social opportunities Deglv36 Climate o f region Deglv37 Housing costs Deglv38 Proximity o f extended family Deglv39 Extensive and/or close network o f friends living locally Deglv40 Loyalty to institution D eglv4l Loyalty to departm ent/program Deglv42 Appreciation fo r my work Deglv43 Influence in institution Deglv44 III-7. Considering all the factors that can influence your employment, how interested are you in leaving or remaining at Michigan State University? (PLEASE CIRCLE ONE NUMBER) Original Variable Name: LVORM Very Interested in Leaving for Another Position 1 Somewhat Interested in Leaving for Another Position About Equally Interested in Leaving and Staying Somewhat Interested in Remaining in Present Position Very Interested in Remaining in Present Position 103 IV -16. In making a final decision about leaving or staying, how free (based on your individual desires) or constrained (based on job, family or relationship factors that you may not be able to control) do you believe your decision would be? O riginal Variable Name: FREE C. Totally Free Fairly Free 1 2 Fairly Constrained Totally Constrained 4 5 3 How important do you think the following should be in determining faculty rewards: N ot V ery Som ew hat Fairly Very Extrem ely O riginal Im portant im portant Im portant Im portant Im portant Variable N am e Tenure Teaching Research/Scholarship Advising Service/Extension Admin. /Governance 1 1 1 1 1 2 2 2 2 2 3 3 3 3 3 4 4 4 4 4 5 5 5 5 5 Tenul Tenu2 Tenu3 Tenu4 Tenu5 Promotion in Rank Teaching Research/Scholarship Advising Service/Extension Admin./Governance 1 1 1 1 1 2 2 2 2 2 3 3 3 3 3 4 4 4 4 4 5 5 5 5 5 Promrl Promr2 Promr3 Promr4 Promr5 3. Merit Increases Teaching Research/Scholarship Advising Service/Extension Admin./Governance 1 1 1 1 1 2 2 2 2 2 3 3 3 3 3 4 4 4 4 4 5 5 5 5 5 Mtinl Mtin2 Mtin3 Mtin4 Mtin5 PART VI. Demographic Information 1. In what year were you born? 19____ 2. What is your sex? M ale 4. Female Original Variable Name: Born (PLEASE CHECK ONE) Original Variable Name: Sex What is your current marital status? (CIRCLE ONE NUMBER) Original Variable Name: Marry Single, never married 1 Divorced 4 Married, Cohabitating 2 Widowed 5 Separated 3 104 5. If you have children, how many do you have?________ What are their ages? Original Variable Name: NChild These questions were selected by the researcher as those which most clearly addressed the exploratory scope o f this dissertation without introducing unnecessary overlap into the study. The questions listed above relate to the chosen areas o f investigation in the following way: TABLE 10 - USE OF SURVEY QUESTIONS AREAS OF INVESTIGATION SURVEY QUESTIONS USED Dependent variables W ork Satisfaction Allocation o f W ork Effort Geographic Mobility 1-8; II-3 1-9; 11-11, V-C, 1-3 III-7, II-2a,b; IV -16 Independent variables Demographic characteristics 1-1; 1-5; V I-1,2,4,5 There are several reasons these questions were selected from the survey and others were not used. Several sections o f unused questions did not focus on the topics o f this dissertation. A significant number o f questions asked respondents about their salaries, careers in relationship to those o f partners, and alternative plans, if they believed it likely they’d leave Michigan State University. In general these questions focused either on negotiating a change in current status or hypothetical outcomes o f that change rather than on work itself at Michigan State. While these are worthy topics o f research they were outside the scope o f this dissertation. 105 However, one set o f responses which might be perceived as focused on leaving the university was used, namely section II-3 o f the survey. This section was used to evaluate the importance o f these factors to the respondent as sources o f satisfaction. It was used in relationship to section 1-8 o f the survey which asked about respondent satisfaction in several different categories. The intent o f using section II-3, therefore, was only to measure whether possible sources o f satisfaction in 1-8 were important to the respondents. Section II-3 in the survey was not used to specifically investigate leaving Michigan State University since the focus o f this research was work effort and satisfaction and the likelihood o f geographic mobility, not the reasons for mobility. Some o f the demographic questions were also eliminated as variables for analysis. The focus o f this dissertation is current employment and gender. Thus information gathered about previous employment and minority status was also outside the focus o f this dissertation. Information related to tenure also was not used as a central part o f this analysis. While tenure status would seem to be a logical variable for analysis the high levels o f tenure at Michigan State University (80%) and the strong relationship between rank and tenure make data about tenure less useful. A crosstabulation o f rank and tenure reveal a very strong relationship between these variables within the faculty. Table 11 illustrates that relationship among survey respondents. 106 TABLE 11 - CROSSTABULATION OF RANK AND TENURE STATUS OF SURVEY RESPONDENTS Rank Status Full Prof. Assoc. Prof. Asst. Prof. Not tenure track 4 4 2 Tenure track 3 13 162 553 229 14 Tenured Given this strong relationship between rank and tenure, rank was chosen as the variable on which to conduct further analysis. Other questions appeared to overlap unnecessarily with chosen questions noted in Table 10. Relevant responses in sections V-A and B were also covered in section V-C. Thus, sections V-A and B were not used in this analysis. CONSTRUCTING THE SET OF RESEARCH VARIABLES Several o f the chosen questions had multiple items to which survey participants were asked to respond. Examining the survey in Appendix A, using Question 1-8 as an example, it is apparent that responses on the issue of professional satisfaction were further segmented into 31 areas o f response. Thus question 1-8 in its initial configuration yielded 31 different variables to be measured and manipulated. Among the other 107 questions chosen, several o f them were similarly constructed. In total there were 112 original variables which this researcher handled. Focusing on an analysis with this large number o f variables is difficult. Since many o f these variables overlapped in their foci, it seemed reasonable to consolidate at least some o f these variables. From reviewing the survey questions some consolidations seemed logical to this researcher. However they were then tested for the reliability o f these consolidations by investigating the correlations among the items and using Cronbach’s Alpha as a test o f reliability for the consolidations that were done. The net result was that o f the 112 variables, 82 o f them were collapsed into 14 new variables. If items could not be consolidated with a Cronbach’s Alpha o f at least .7, the anticipated consolidations were not carried out. Table 12 documents which variables were consolidated and the names o f the new variables: 108 TABLE 12-CONSOLIDATION OF SELECTED RESEARCH VARIABLES N ew variable nam e N ew variable m eaning W O R K SA T C L IM SA T Satisfaction with the w ork itself Satisfaction with the departmental and university clim ate for work SU PPSA T Satisfaction support for Importance satisfaction IW O R K SA T IC O SE SA T IC LIM SA T w ith institutional o n e ’s w ork to respondent o f with w ork Importance to respondent o f satisfaction w ith com pensation and job security Importance to respondent o f departmental and university clim ate IO PPSAT Importance to respondent o f career opportunities ISU P P SA T Importance to respondent o f institutional support IN O IN SA T Importance to respondent o f noninstitutional satisfactions RETTTRM Importance o f teaching for tenure, prom otion, and m erit increases Im portance o f research for tenure, prom otion, and m erit increases Im portance o f ad vising for tenure, prom otion, and m erit increases Importance o f serv ice for tenure, prom otion, and m erit increases Im portance o f adm inistration/ governance for tenure, prom otion, and merit increases RERTRM REATRM RESTRM REGTRM Old variables consolidated S A T 1, S A T 6, S A T 7 , S A T 9, SA T 3, S A T 4, S A T 5, S A T 12, S A T 15, S A T 16, S A T 17, S A T 18, S A T 19, S A T 26, S A T 2 7 , S A T 28, SA T 29 S A T 10, S A T 1 1 , S A T 22, S A T 23, S A T 2 5 , S A T 30 D E G L V 2, D E G L V 11, D E G L V 13, D E G L V 15, D E G L V 16, D E G L V 21, D E G L V 25, D E G L V 31, D E G L V 44 D E G L V 5, D E G L V 10, D E G L V 12, D E G L V 24, D E G L V 29 D E G L V 1, D E G L V 4, D E G L V 6, D E G L V 8, D E G L V 17, D E G L V 19, D E G L V 20, D E G L V 22, D E G L V 30, D E G L V 41, D E G L V 42, D E G L V 43 D E G L V 7, D E G L V 8, D E G L V 14, D E G L V 32, D E G L V 33 D E G L V 3, D E G L V 18, D E G L V 23, D E G L V 26, D E G L V 27, D E G L V 28 D E G L V 34, D E G L V 35, D E G L V 36, D E G L V 37, D E G L V 38, D E G L V 39, D E G L V 40 T E N U 1 ,P R 0 M R 1 ,M T IN 1 C ronbach’a Alpha .8330 .8301 .7946 .8283 .7754 .8809 .7634 .7846 .8371 .9547 T E N U 2 , PR O M R 2, M TIN2 .9486 T E N U 3 , PR O M R 3, M TIN3 .9361 T E N U 4 , PR O M R 4, M TIN 4 .9564 T E N U 5 , PR O M R 5, M TIN 5 .9 1 1 0 109 It is important to note that the collapse of a few variables was not carried out as planned. Because numerous combinations o f these variables resulted in Cronbach’s Alphas below .7, much closer to .5, it was deemed more reasonable to allow these variables to remain in their original form. These variables were Sat2, Sat8, S atl3, S atl4 , Sat20, Sat21, and Sat24. In collapsing variables it is important to note the treatment of missing values. A missing values code was entered when survey respondents did not provide a response to an item or provided responses- that violated the survey’s instructions. In SPSS, the statistical software used for this analysis, each time a missing value occurs, that respondent case is normally dropped from the process o f consolidating values for a new variable. Thus, if some numerical response is not inserted when old variables have missing values, the number o f cases on which a new variable was based might drop below 50% o f the survey respondents. This occurred particularly when several old variables were being collapsed into one new variable. To resolve this problem missing responses were replaced with the mean response for each o f the old variables. W hile this kept almost all cases in the pool upon which new variables were created, such a strategy does introduce a bit more conservatism into the statistical results. To some degree it counteracts sensitivity to the range o f the responses by using this central measure with missing value cases. However, this researcher deemed this strategy preferable to eliminating up to half o f the possible cases in the respondent pool in the construction o f many o f the new variables. 110 In addition to these new. variables created through the collapse o f several related variables, several other variables in the survey were used in their original form. All variables used in the original form are shown in Table 13. TABLE 13-OTHER SURVEY VARIABLES USED Variables examining work satisfaction: Sat2 Sat20 Sat8 Sat21 Sat 13 Sat24 Sat 14 Variables examining the allocation o f work time: Pteach RSCP1 RSCP2 Prsp Pads RSCP3 RSCP4 Ppdt Psex RSCP5 Padgv RSCP6 Poth Variables examining geographic mobility: TKACT1 TKACT2 LVORM FREE Variables examining demographic characteristics: BORN SEX MARRY NCHILD RANK Together the variables in tables 12 and 13 comprise the entire set o f 43 research variables that was used to focus on issues o f professional satisfaction, allocation o f work Ill time, and geographic mobility in relation to the demographic characteristics. The 5 demographic characteristic variables functioned as independent variables and the other 38 as dependent variables. STATISTICAL M ETHODOLOGIES Findings are analyzed at an aggregate all-university level to investigate whether there are significant differences between the male and female faculty at Michigan State University using both one-way and multiple analysis o f variance. Initially a simple methodology is used to examine the effects o f a single independent variable, namely SEX, on the 38 dependent variables noted above. This is an initial examination o f the effect o f gender on work satisfaction, allocation o f work time, and geographic mobility. In this first step o f analysis, given gender as a nominal category and Likert scaled responses which can be treated as interval data, the appropriate statistical test is the one­ way anova and its corresponding F-test. The null hypothesis, Ho, is that there is no difference between male and female faculty members in the population on mean responses to variables about work satisfaction, allocation o f work time, and geographic mobility. Alpha is set at .05. The larger the F statistic, the smaller the F significance, and the more likely that this researcher can reject the null hypothesis o f no difference in the mean scores of males and females within the population. The one-way ANOVA test assumes normally distributed sample means and homoscedasticity. 112 Following some initial investigation, the ANOVA methodology is then more extensively used to evaluate the effects of multiple classification variables at the same time, testing not only for the main effects o f individual variables, but also for the combined effects between variables, namely the interaction effects. This methodology provides an easily accessible evaluation o f effects. It also provides a foundation for developing appropriate regression equations. If there are significant interaction effects, interactions terms must be introduced into later regression equations. W ith this ANOVA two null hypotheses are being tested, Ho-1, that each o f the independent variables individually has no significant effect on the dependent variables being evaluated, and Ho-2, that there is no significant interaction effect between the independent variables. ANOVAS at an all-university level are run combining variables SEX, RANK, MARRY, N CHILD, with BORN as a covariate to investigate both the individual and interactions effects among these variables. An alpha level of .05 will also be used in this portion of the analysis. Again if an F statistic generated by the two way anova is larger, the F significance is smaller. If the F significance is less than .05, the null hypothesis can be rejected. If an F significance is greater than .05, the null hypothesis about the relationship between the variables cannot be rejected. 113 As a final step in this study selected variables are analyzed in more detail using regression analysis to investigate some o f the key findings regarding allocation of work effort. The central mission o f a research university involves the balance o f commitments to research and teaching. Preliminary analyses have indicated that this balance differs by sex. Regression analysis is used to investigate the degree o f direct gender effects as well as the effects o f other variables which may contribute to differences in the allocation o f w ork time among males and females. Since the core culture o f a research university is to the greatest degree determined by its dual commitments to both the creation and dissemination o f knowledge, regression analysis focuses on four dependent variables which analyze these two activities. They are PTEACH, PRSP, RSCP1, and RSCP2. These variables measure both the current allocation o f time to teaching (PTEACH) and research (PRSP) and the preferred allocation o f time to teaching (RSCP1) and research (RSCP2). The independent variables in these four regressions are SEX, RANK, BORN, MARRY, NCHILD, and two newly created variables named GC and ROC. SEX, RANK, BORN, MARRY, and NCHILD are independent variables which represent a faculty member’s personal characteristics or situation. GC and ROC as defined below try to capture the culture o f the college structures in which faculty members operate. The intent is, therefore, to allow both personal and structural explanations for differences to surface. 114 GC, one o f the two new independent variables, reflects the gender composition o f faculty in each o f the colleges. The underlying assumption is that the balance o f male and female faculty members may effect the culture of the college and resulting attitudes about teaching and research. GC was constructed by calculating the percentage o f female faculty members in each o f the colleges. This calculation, therefore, presumes that the gender composition o f the entire college, not just those who responded to the survey, may influence the allocation o f work effort. GC for each o f the colleges is shown below: TABLE 14 - FEMALES AS A PROPORTION OF MSU COLLEGE FACULTIES College GC Agriculture and Natural Resources Arts and Letters Business Communication Arts Education .146 .282 .168 .242 .368 Engineering Human Ecology Human Medicine James Madison Natural Science .068 .750 .183 .158 .137 Nursing Osteopathic Medicine Social Science Urban Affairs Veterinary Medicine .923 .221 .221 .000 .120 (Note: Non-College Faculty who retain faculty status were excluded from the analysis) 115 ROC, the second o f two new independent variables, is used as a measure o f the relative research orientation of each of the colleges. The measure o f ROC in each college is the mean survey response on the variable PRSP o f those who represent the majority sex among faculty in each college. The PRSP variable measures the current percentage o f time allocated to research among survey respondents. This presumes that numerical dominance equates to power in shaping the relative nature o f the research/teaching culture within the school. The use o f this dominant faculty group mean prevents a blurring o f distinctions between males and females within each college. Such a blurring may occur if an average combining male and female scores on research orientation within each college is used instead. The ROC for each o f the colleges is shown below: 116 TABLE 15 - PERCENTAGE OF TIM E SPENT ON RESEARCH (ROC) FOR MAJORITY SEX AMONG FACULTY BY COLLEGE College ROC Majority Sex Agriculture and Natural Resources Arts and Letters Business Communication Arts Education 27.89 20.47 32.02 21.72 21.59 M ale M ale Male Male Male Engineering Human Ecology Human Medicine James Madison Natural Science 31.06 17.84 26.78 17.00 34.90 Male Female Male Male Male Nursing Osteopathic Medicine Social Science Urban Affairs Veterinary Medicine 13.32 20.31 25.58 26.67 25.13 Female Male Male Male Male (Note: Non-College Faculty who retain faculty status were excluded from the analysis) The assumption underlying the use o f both GC and ROC is that one’s college of appointment will provide an appropriate basis on which to differentiate. It likely does differentiate the organizational culture in which faculty members work to some degree. However, it can be argued that a better differentiator would be a faculty m em ber’s field of study. Unfortunately, survey respondents were not asked to indicate their academic disciplines. In addition no department codes, indicating department within a college, were included in the data base. Instead respondents were only asked in which college 117 they held appointment. Thus, differentiations at the all-college level must suffice to simulate some o f the effects of chosen field o f study and specific college culture. To run regression analysis four nominal variables were converted to dummy variables. SEX, being a dichotomous variable, was already effectively in dummy form. RANK was changed by differentiating only between assistant professors (primarily non-tenured but tenure track) and a grouping o f associate and full professors (tenured with very few exceptions - see Table 11). Thus RANK now reflected prim arily differences in experience as well as tenure status. MARRY was converted into a dummy variable by collapsing the response categories o f single, separated, divorced, and widowed into one category reflecting current non-partnered status. This could be contrasted with "m arried/cohabitating" which reflected partnered status. Thus, in effect, the variable MARRY now compares partnered and non-partnered status. converted into a dichotomous variable. Similarly NCHILD was All distinctions in numbers o f children were blended together to form a dichotomy between no children and any children. Thus NCHILD now represents that dichotomy. Tests for linearity, normality and homoscedasticity have been run on PTEACH, PRSP, RSCP1, and RSCP2 to determine whether the assumptions o f regression analysis methodology, namely linearity, normality, and homoscedasticity in the population, can be met. Scatterplots were examined for evidence o f non-linearity. Skews and their standard errors were reviewed to detect whether the residuals were normally distributed 118 in the population. Tests for homoscedasticity were run to determine whether the variance in the dependent variables in the population was the same at all levels of the independent variables, using C ochran’s C a s a measure o f equality o f variance. The results o f these tests are reported in Chapter four. LIM ITATIONS OF THIS STUDY W hile this study may prove helpful in unraveling the complex differences between male and female faculty, it is important to recognize its limitations because o f its sample population, the survey itself, and the methodologies employed in the analysis. This dissertation presents only a snapshot o f one university at one point in its development by surveying the faculty population o f Michigan State University. It is important to recognize this study as an intrauniversity analysis. Given that, the results at other institutions may vary. W hile classified as a Research I institution, Michigan State University is also one o f 68 land grant colleges and universities in the United States. The land grant nature o f the institution along with other historical and political factors affect the configuration o f colleges and universities. Many non-land grant schools would not have a College o f Human Ecology in which female faculty are predominant; nor would many o f these universities have Colleges o f Agriculture and Natural Resources. Many other research I institutions would have law schools while Michigan State University does not. Not all research I universities would include human medical schools, while Michigan State University has two such entities. With a different 119 configuration o f colleges, it is possible that another institution may find somewhat different results if this survey were administered. Besides the range o f colleges in the university, differences might also arise because of Michigan State University’s status as a Midwestern public university located in Michigan. It is possible that Midwestern culture may affect professional satisfactions and allocation of time among faculty differently than either o f the coastal U .S. cultures. Furthermore, the changing economic fortunes o f Michigan over the past fifteen years may have influenced the range o f faculty who have chosen to affiliate with its state universities in ways which differ from those in other states or in private institutions. Thus, it would be unwise to claim too much from the results o f this study by attempting to generalize to all Research I institutions. The survey itself could have been improved to provide better results. The large number o f response items could have lulled some survey participants into unthinking patterns of response, resulting in less clear distinctions than may have been possible with a somewhat shorter survey. Because these many response items were consolidated into a smaller number o f workable variables, it is also likely that some nuances in responses have been lost. One crucial piece o f information not asked in the survey was each respondent’s disciplinary field. While such information (i.e. department o f respondent) was known 120 by those to whom the survey was returned, it was not entered into the survey data base. With such information a better analysis could have been conducted on the effects o f one’s chosen field o f study on satisfactions, allocation o f time, and even mobility considerations. After the survey was already administered, the literature search, later conducted, indicated that field of study is a very influential factor, particularly in the allocation o f work time. Certain disciplines are more prone to activities traditionally classified as research than others. W ithout this information in the data base, it was not possible to adequately examine disciplinary field as a factor o f influence. GC, gender composition o f the college, and ROC, research orientation o f the college, are rougher approximations of the effects o f faculty members’ disciplines. However, they do not allow the finer distinctions that could have been drawn with knowledge o f disciplinary field. It is quite possible that gender affects chosen field o f study which in turn determines relative teaching or research orientation. Analyzing the effects o f disciplinary field would open up another branch o f research. While an aggregate survey response rate o f 50.5% is within an acceptable range for survey methodologies, it is important to note that variations among the response rates of the colleges and between the male and female faculty within them are limitations. Some colleges are more strongly represented in the survey data than others. Some colleges have substantial differences in the male and female response rates. These variations also limit the value o f any college level analyses. Thus, the picture o f the university which emerges from this data must be balanced against these limitations. 121 The statistical methods used also have some limitations. When significance is reported, this simply indicates that, based on the sam ple’s results, at a 95% confidence level the null hypothesis can be rejected for the population. Such a result will frequently mean that a null hypothesis o f no difference between the groups is equated to the conclusion that the two groups are indeed different in the population. It is possible that such a conclusion is incorrect. Thus, conclusions about the population from which the sample is drawn should be regarded somewhat tentatively. One can conclude that there is every appearance that groups differ in the population. Yet there is always a chance that such a conclusion is wrong. Inferences about differences should be handled carefully. In addition, there may be some lingering difficulties in the regression analyses caused by heteroscedasticity and multicolinearity which will be noted in chapter four. With almost all variables, the assumptions necessary for regression were met. However, the low Plevel o f C ochrane’s C for the variable PTEACH and RSCP1 should add a bit o f caution to use o f those regression results. Furthermore, the higher levels o f correlation between some o f the variables could create a small colinear effect in the regression results. Neither o f these problems were deemed material in light o f other factors which supported the appropriateness of regression analysis. However, it is important to recognize that the assumptions for regression analysis may not have been perfectly met. Finally, this study does not purport to respond to all the reasons for gender differences posited by different groups o f scholars and described in chapter two. W hile it appears 122 that gender does have a significant effect on differences among faculty in the university, the reasons for those differences have not been completely resolved. This dissertation does not draw clear distinctions between biological nature and sociological nurture. Nor can final conclusions be reached about the relative strength o f personal versus structural explanations for differences. The sensitivity o f the structural variables in the regressions is simply too limited to draw such conclusions. In summary, the research methodology proposed in this chapter is based on the Michigan State University Faculty Mobility Survey from which selected questions were used to investigate possible faculty differences by gender on measures o f professional satisfaction, allocation o f work effort, and geographic mobility. In this investigation several variables were collapsed into more condensed categorical variables when the consolidations met the Cronbach’s Alpha standard o f .7. ANOVAS were then run on the new dependent variable set, checking first and singly the effects o f SEX and then the effects o f SEX in interaction with RANK, MARRY and NCHILD with BORN as a covariate. Regression analysis was then used to explore the relative strength o f gender in relation to other personal and structural variables which might explain both current and preferred allocations o f effort to teaching and research. Within the limitations o f the university, the survey, and the statistical methods proposed, the intention is to look closely at selected areas o f possible gender differences. Chapter four o f this dissertation presents the results o f these research methodologies. CHAPTER 4 STATISTICAL RESULTS Chapter four presents the statistical results o f the research methodologies proposed in Chapter three. First, the one way ANOVA with sex as the independent variable and 38 dependent variables will be examined. Second, the ANOVA using SEX, RANK, MARRY, and NCHILD as independent variables with BORN as a covariate will be reported for this same set of dependent variables. Third, the testing o f the proposed regression models will be explained. Fourth, the regression analyses themselves will be reported. ONE-WAY ANOVA RESULTS Table 16 presents the results o f the one way anova for the 38 dependent variables. 124 TABLE 16 RESPONDENT RESULTS-ONE-W AY ANOVA D e p e n d e n t v a ria b le s P ro fe ssio n a l S atisfaction ♦ W o rk sa t C lim sat ♦ S u p p sat *Sat2 *Sat8 S at 13 S a tl4 Sat2 0 Sat21 ♦ S at24 Im p o rta n c e o f P ro fe ssio n a l Satisfaction Iw o rk sat Ic o sesat Iclim sat ♦ Io p p sat Isu p p sat In o in sat C u rre n t a llo c atio n o f w o rk tim e ♦ P teach ♦ P rsp P ads P pdt P sex P adgv Poth F e m ale M ean 2.71 3 .4 6 2 .9 7 4 .1 7 3 .4 5 3 .3 9 M ale M ean 3 .3 8 3 .5 2 3 .2 0 4 .5 5 3 .8 5 3.41 F Statistic 7 0 .2 5 5 1.407 11.1 7 7 2 7 .0 6 8 2 3 .0 4 0 .030 S ig o f F .000 .236 .001 .000 .000 .863 4 .0 2 3 .7 5 3 .3 9 3 .0 0 4 .0 8 3 .83 3 .2 9 3 .3 8 .607 1.3 4 7 1.9 3 4 13.535 .4 3 6 .2 4 6 .165 .000 2 .8 2 2 .8 5 2 .9 0 3 .0 9 2 .8 0 2 .4 7 2.71 2 .9 2 2 .8 6 2 .9 2 2 .8 2 2 .4 8 2.9 7 1 .744 .517 5.5 0 3 .042 .046 .085 .389 .472 .019 .837 .831 3 8 .0 7 2 1 .8 4 8 .2 4 2 .9 4 11.31 16.96 .73 3 3 .6 6 2 7 .2 0 7 .9 6 2 .5 2 11.64 15.76 1.19 5 .9 0 6 12.488 12.888 1.339 17.943 .609 .535 .015 .000 .621 .248 .807 2 .4 8 4 .1 4 2 .8 3 3 .8 3 2 .8 8 2 .3 2 2 .8 4 3 .8 0 3 .03 3 .5 9 2 .9 0 2 .3 4 3 2 .8 4 9 2 9 .4 2 3 14.2 3 2 19.7 6 0 .110 .112 .000 .000 .000 .000 .741 .737 4 .0 6 4 .2 6 2 .7 8 .313 .024 1.404 .576 .876 2 .9 4 2 .4 5 4 .0 2 4 .2 7 2 .6 8 2 .9 8 2.21 .265 10.391 .236 .607 .001 1.72 2.81 3.41 3 .1 8 1.76 2 .5 7 3.63 2 .8 6 .189 4 .3 9 9 4 .9 1 4 15.479 .664 .036 .027 .000 .435 .465 P re fe rre d A llo catio n o f W o rk T im e ♦ R sc p l *R scp 2 *R scp3 ♦ R sc p 4 R scp5 R scp6 Im port, o f task s fo r te n u re , ra n k , and m e rit in c re a se R ettrm R e rtrm R eatrm R e strm ♦ R eg trm G e o g rap h ic m obility T k a c tl * T k act2 ♦ L v o rm ♦ F re e N = 9 5 5 , 7 4 7 m ale s, 2 0 9 fem ale ♦ V aria b le s fo r w h ic h null h y pothesis can ae rejected 125 As the table above indicates, on 16 of the 38 variables the null hypothesis (no difference between the male and female faculty in the population) can be rejected. For these 16 variables the significance o f F falls below an alpha o f .05, the basis for rejecting the null hypothesis. The null hypothesis can be rejected for some variables in each o f the three research foci, professional satisfaction, allocation o f work time, and geographic mobility. Results related to professional satisfaction Specifically the significance of F is below the alpha o f .05 for the variables WORKSAT, SUPPSAT, SAT2, SAT8, SAT24, and IOPPSAT within the ategory o f professional satisfaction. WORKSAT as a variable reflects satisfactions with the work itself and SUPPSAT measures satisfactions with institutional support for one’s work. In both cases female faculty respondents to the survey were less satisfied. The pattern o f survey results with other satisfaction variables should also be noted. SAT2 measured satisfaction with job security. SAT8 measured satisfaction with opportunities for advancement at Michigan State University. SAT24 measured satisfaction with opportunities for growth and development within one’s current academic unit. In all three cases females were less satisfied than males. W hile female faculty did not differ significantly from males on ICOSESAT, the importance o f satisfaction with job security, they did differ on their satisfaction with job security, SAT2. Furthermore, female respondents regarded IOPPSAT, the importance o f their satisfaction with opportunities, as more important than did the males, yet deemed actual satisfaction, with 126 opportunities for advancement and professional development, SAT8 and SAT24, to be at lower levels than did the male faculty. The gap between male and female respondents on each o f these variables measuring work satisfaction is great enough to allow rejection o f the hypotheses that their levels o f satisfaction in these selected areas and their sense of the importance o f satisfaction with opportunities are the same as males in the Michigan State University faculty population. On several other measures of professional satisfaction there were insufficient differences to reject the null hypothesis. These included CLIMSAT, SAT13, SAT14, SAT20, SAT21, IWORKSAT, ICOSESAT, ICLIMSAT, ISUPPSAT, INOINSAT. These results indicate that it is possible that male and female faculty in the population have similar reactions to the departmental and university climate for their work(CLIM SAT), similar satisfactions with levels o f salary and benefits (SAT13, SAT14), and similar reactions to the quality o f students at Michigan State University (SAT20, SAT21). There is also no basis for rejecting the null hypothesis about the importance o f many aspects o f their job to them (IWORKSAT, ICOSESAT, ICLIM SAT, ISUPPSAT, INOINSAT). It is quite possible that males and females in the faculty population weigh the importance o f satisfactions with the work itself, compensation and job security, the departmental and university climate, noninstitutional satisfactions similarly. institutional support for their work, and m Allocation of W ork Time Among the variables measuring current, preferred and ideal allocations o f work time there were some notable differences. Specifically PTEACH, PRSP, RSCP1, RSCP2, RSCP3, RSCP4, and REGTRM had an F significance allowing the rejection o f the null hypothesis in the population. Fem ale respondents spent a significantly larger percentage o f their time in teaching(PTEACH) and a significantly smaller percentage o f their time in research (PRSP) when compared to male respondents. On other current work effort variables related to advising(PADS), professional development(PPDT), service and extension(PSEX), administration and governance (PADGV), and other activities(POTH), there were insufficient differences between the respondent groups to allow rejection of the null hypothesis. When asked what one would change with the opportunity to restructure one’s position, female respondents wanted somewhat less teaching tim e(RSCPl) and somewhat more research time(RSCP2) than did their male counterparts. Female respondents also preferred significantly less time in advising students(RSCP3) and more time in professional development(RSCP4). The null hypothesis can be rejected for the population on these four variables. However, it cannot be rejected concerning desired time for service and extension(RSCP5) or administration and govemance(RSCP6). Despite the possibility o f differences between male and female faculty in current and preferred allocations o f time, their sense o f relative importance o f teaching(RETTRM ), 128 research(RERTRM ), advising(REATRM ), and service/extension(RESTRM ) for tenure, promotion in rank, and merit pay increases may be the same in the population. None of these four variables had a F significance below .05. Interestingly, only on the importance of administration/governance activities (REGTRM) for tenure, promotion, and merit increases can the null hypothesis be rejected. On this variable female faculty in the population seemed to think that such activity was more important than did male faculty. Geographic Mobility W hile males and females in the population may have similar responses regarding their likelihood o f seeking a new position at Michigan State University (TKACT1), the null hypothesis can be rejected for the other three variables measuring perceived geographic mobility. The likelihood o f seeking a position at another institution(TKACT2) and the level o f interest in leaving Michigan State University(LVORM ) are not likely the same for male and female faculty in the population. Female respondents are somewhat more likely to seek a new position outside Michigan State and somewhat less likely to remain at Michigan State. However, females perceive that they are more constrained than males in their freedom to make a final decision about staying or leaving(FREE). M ULTIPLE ANOVA The one-way ANOVA by sex begins to provide some insight into possible differences between male and female faculty in the population. However, SEX as a variable may 129 be masking the effects o f interactions with other variables with which it may be associated such as age(BORN), academic rank(RANK), partner(M ARRY), or commitments to children(NCHILD). commitments to a Therefore an ANOVA including SEX, RANK, M ARRY, and NCHILD as independent variables with BORN as a covariate was run to determine main and interaction effects which are significant in the population. Tables 17 through 23 present the main effects o f this ANOVA sorted by seven groupings o f dependent variables. These tables are presented below. 130 TABLE 17 ANOVA ON PROFESSIONAL SATISFACTION VARIABLES Dependent Variables Independent Variables Sat2 Sat8 W orksat Climsat Suppsat 51.330 .000 5.590 .018 7.264 .007 96.870 .000 .822 .365 3.25 3.51 3.16 4.48 3.78 2.74 3.38 36.042 .000 3.48 3.52 .122 .727 2.98 3.20 4.879 .027 4.21 4.55 1.304 .254 3.45 3.86 14.093 .000 2.91 3.30 5.530 .019 3.40 3.53 3.298 .070 2.96 3.19 5.206 .023 4.14 4.54 5.341 .021 3.39 3.84 11.422 .001 2.99 3.27 .354 .552 3.42 3.52 .289 .591 3.00 3.17 .257 .612 4.16 4.51 .061 .805 3.75 3.78 1.858 .173 Rank Full/Assoc. mean Asst. - mean F-Stat Sig. o f F 3.30 3.00 1.604 .206 3.52 3.44 .063 .802 3.15 3.19 8.565 .004 4.69 3.54 151.525 .000 3.80 3.65 .299 .584 Two-W ay Interactions Sex/M arrv F-Stat Sig. o f F .041 .840 7.375 .007 .175 .676 1.138 .286 1.408 .236 2.440 .119 .518 .472 .043 .836 .524 .469 .010 .921 Covariate Born F-Stat Sig. of F Grand Mean Main Effects Sex Female Mean M ale Mean F-Stat Sig. o f F M arrv Not partnered - mean Partnered - mean F-Stat Sig. o f F NChild No child - mean Child - mean F-Stat Sig. o f F Sex/NChild F-Stat Sig. o f F 131 TABLE 17 (Continued) W orksat Climsat Suppsat .008 .927 .006 .941 .735 .391 1.088 .297 .692 .406 1.187 .276 6.147 .013 1.980 .160 1.008 .316 .743 .389 Rank/M arrv F-Stat Sig. o f F .262 .609 .227 .634 .088 .767 .014 .904 .152 .697 Rank/NChild F-Stat Sig. o f F 1.525 .217 .075 .784 .800 .371 6.269 .012 .658 .418 .953 .329 .145 .703 .030 .862 3.860 .050 .942 .005 Sex/NChild/Rank F-Stat Sig. o f F .739 .390 1.239 .266 .620 .431 2.631 .105 1.120 .290 Sex/M arrv/Rank F-Stat Sig. o f F 1.681 .195 .059 .809 .399 .528 .422 .516 3.068 .080 NChild/M arrv/Rank F-Stat Sig. o f F 3.665 .056 .031 .861 .915 .339 .195 .659 .848 .357 1.516 .219 .276 .600 2.420 .120 .286 .593 .020 .887 7.218 .000 1.618 .058 2.199 .004 18.278 .000 2.801 .000 M arrv/NChild F-Stat Sig. o f F Rank/Sex F-Stat Sig. o f F Three-W ay Interactions Sex/M arrv/NChild F-Stat Sig. o f F Four-W ay Interaction Sex/NChild/M arrv/Rank F-Stat Sig. o f F Explained F-Stat Sig. o f F N-916 Sat2 Sat8 132 TABLE 18 ANOVA ON PROFESSIONAL SATISFACTION VARIABLES - PART II Dependent Variables Independent Variables Satl3 Sat 14 Sat20 Sat21 Sat24 Covariate Bom F-Stat Sig. o f F 2.130 .145 .756 .385 19.101 .000 15.043 .000 6.537 Grand Mean 3.42 4.07 3.80 3.30 3.27 Main Effects Sex Female Mean M ale Mean F-Stat Sig. o f F 3.41 3.43 .181 .670 4.02 4.08 .001 .978 3.72 3.82 .017 .897 3.31 3.30 1.053 .305 2.99 3.35 5.158 .023 3.20 3.45 4.227 .040 3.89 4.10 4.029 .045 3.66 3.83 1.366 .243 3.29 3.30 .004 .948 2.91 3.33 7.422 .007 NChild No child - mean Child - mean F-Stat Sig. of F 3.45 3.41 .098 .754 3.81 4.09 3.884 .049 3.52 3.83 2.346 .126 3.24 3.30 .028 .867 3.03 3.29 .127 .722 Rank Full/Assoc. - mean Asst. - mean F-Stat Sig. o f F 3.41 3.44 .150 .698 4.05 4.16 6.399 .012 3.84 3.61 .112 .738 3.36 3.12 .447 .504 3.29 3.20 1.692 .194 Two-W ay Interactions Sex/M arrv F-Stat Sig. o f F .000 .991 .087 .768 1.048 .306 .155 .694 1.529 .217 .019 .891 .020 .889 4.119 .043 1.480 .224 .074 .785 M arrv Not partnered - mean Partnered - mean F-Stat Sig. of F Sex/NChild F-Stat Sig. o f F .011 133 TABLE 18 (Continued) M arrv/N Child F-Stat Sig. o f F .137 .712 1.957 .162 1.504 .220 3.026 .082 .573 .449 Rank/Sex F-Stat Sig. o f F .205 .651 .445 .505 .001 .976 .047 .828 4.278 .039 Rank/M arrv F-Stat Sig. o f F .338 .561 3.308 .069 .793 .373 .197 .658 1.517 .218 5.255 .022 9.586 .002 2.038 .154 1.382 .240 1.246 .265 .261 .610 .900 .343 .235 .628 .338 .561 .127 .721 Sex/NChild/Rank F-Stat Sig. o f F 1.477 .225 .403 .526 .250 .617 .162 .687 .025 .875 Sex/M arrv/Rank F-Stat Sig. o f F .018 .893 1.413 .235 .035 .852 .950 .330 .153 .696 NChild/M arrv/Rank F-Stat Sig. o f F .199 .656 2.111 .147 .017 .895 .336 .562 .841 .359 .001 .978 .053 .818 .033 .857 3.396 .066 1.670 .197 1.019 .433 2.242 .003 2.426 .001 1.734 .036 2.295 .003 Rank/NChild F-Stat Sig. o f F Three-W ay Interactions Sex/M arrv/N Child F-Stat Sig. o f F Four-W ay Interaction Sex/N Child/M arrv/Rank F-Stat Sig. o f F Explained F-Stat Sig. o f F N-948 134 TABLE 19 ANOVA ON IM PORTANCE OF PROFESSIONAL SATISFACTION VARIABLES Dependent Variables Independent Variables rW orksat IC osesat IC lim sat Ioppsat Isuppsat Inoinsat Covariate Born F-Stat Sig. of F 7.739 .006 5.595 .018 8.663 .003 52.425 .000 7.710 .006 25.054 .000 Grand Mean 2.73 2.90 2.87 2.96 2.81 2.48 2.83 2.71 1.303 .254 2.85 2.91 .821 .365 2.91 2.85 .064 .801 3.11 2.92 .806 .370 2.82 2.81 .568 .451 2.50 2.48 .040 .841 Marry Not partnered - mean Partnered - mean F-Stat Sig. of F 2.86 2.71 3.112 .078 2.87 2.91 .025 .874 2.93 2.85 1.096 .295 3.08 2.94 1.081 .299 2.87 2.80 .798 .372 2.30 2.51 6.821 .009 NChild No child - mean Child - mean F-Stat Sig. of F 2.67 2.74 3.541 .060 2.70 2.92 5.449 .020 2.79 2.87 3.008 .083 2.94 2.96 5.005 .026 2.85 2.81 .109 .742 2.42 2.49 1.648 .199 Rank Full/Assoc. - mean Asst. - mean F-Stat Sig. of F 2.72 2.81 .149 .700 2.88 2.98 .027 .870 2.84 2.96 .004 .953 2.87 3.36 7.790 .005 2.78 2.94 .611 .435 2.44 2.65 .355 .551 Two-Way Interactions Sex/Marrv F-Stat Sig. of F .282 .596 .966 .326 .166 .684 .273 .601 .163 .687 .935 .334 .958 .328 1.873 .172 .158 .691 .961 .327 .082 .774 .054 .817 Main Effects Sex Female Mean Male Mean F-Stat Sig. of F Sex/NChild F-Stat Sig. of F 135 TABLE 19 (Continued) IW o r k s a t Ic o s e s a t I c lim s a t Io p p s a t Isu n n sa t I n o in s a t Marrv/NChild F-Stat Sig. of F .600 .439 3.091 0.79 1.460 .227 1.813 .178 1.894 .169 1.852 .174 Rank/Sex F-Stat Sig. of F .180 .672 .377 .539 .617 .432 1.077 .300 3.284 .070 .038 .846 .005 .944 2.846 .092 .015 .901 1.552 .213 .274 .601 .378 .539 .312 .577 .111 .739 .001 .974 .971 .325 2.305 .129 .004 .949 1.730 .189 1.956 .162 2.194 .139 1.456 .228 .970 .325 .087 .768 Sex/NChild/Rank F-Stat Sig. of F .033 .855 .387 .534 .080 .777 .933 .334 .205 .650 .464 .496 Sex/Marrv/Rank F-Stat Sig. of F .456 .500 3.710 .054 .240 .625 2.139 .144 .020 .887 2.964 .085 NChild/Marrv/Rank F-Stat Sig. of F .166 .683 .143 .705 .810 .368 .853 .356 .019 .891 .091 .763 .071 .720 .624 .430 .270 .603 2.058 .152 .790 .374 .008 .931 Rank/Marrv F-Stat Sig. of F Rank/NChild F-Stat Sig. of F Three-Way Interactions Sex/Marrv/NChild F-Stat Sig. of F Four-Way Interaction Sex/NChild/Marrv/Rank F-Stat Sig. of F 136 TABLE 20 ANOVA ON ALLOCATION OF WORK TIME VARIABLES D e p e n d e n t V a ria b le s In d e p e n d e n t V a ria b le s P te a c h E rsfi P ad s Ppdt P sex P adsv P o th Covariate Born F-Stat Sig. o f F 4.059 .044 17.293 .000 10.192 .001 1.041 .308 .005 .944 4.242 .040 2.090 .149 Grand Mean 34.32 26.05 8.06 2.63 11.69 16.08 1.12 Main Effects Sex Female Mean Male Mean F-Stat Sig. o f F 37.01 33.60 2.408 .121 22.37 27.04 14.896 .000 8.43 7.96 .116 .734 3.06 2.52 1.719 .190 11.45 11.76 .051 .822 16.99 15.84 1.847 .174 .78 1.20 .001 .975 Marry Not partnered-mean Partnered - mean F-Stat Sig. o f F 38.65 33.60 2.743 .098 23.00 26.56 2.859 .091 7.93 8.08 .139 .709 3.19 2.54 1.303 .254 10.61 11.87 .189 .663 16.81 15.96 .494 .482 .37 1.24 .826 .364 NChild N o child - mean Child - mean F-Stat Sig. o f F 35.01 34.25 .004 .951 29.78 25.72 3.115 .078 7.99 8.07 .608 .436 3.57 2.55 3.233 .072 8.36 11.99 2.741 .098 15.05 16.17 .025 .874 .43 1.18 .056 .813 Rank Full/Assoc-mean Asst. - mean F-Stat Sig. o f F 33.99 35.78 3.599 .058 25.02 30.75 4.293 .039 7.84 9.09 .193 .661 2.82 1.82 7.745 .005 11.79 11.25 .047 .828 17.28 10.64 13.102 .000 1.26 .44 .110 .740 Two-Way Interacts Sex/Marrv F-Stat Sig. o f F 4.761 .029 2.313 .129 .152 .697 1.678 .195 1.077 .300 .043 .836 .048 .826 1.015 .414 .168 .682 .935 .334 1.429 .232 .673 .412 .840 .360 .003 .960 Sex/NChild F-Stat Sig. o f F 137 TABLE 20 (Continued) Pteach Prsp Pads Pgdt Psex Padgv Poth Marrv/NChild F-Stat Sig. o f F .633 .426 .242 .623 2.588 .108 3.316 .069 .724 .395 .221 .639 .130 .719 Rank/Sex F-Stat Sig. o f F .375 .540 .541 .462 .262 .609 1.315 .252 1.101 .294 .340 .560 .019 .889 Rank/Marry F-Stat Sig. o f F .533 .466 .537 .464 2.662 .103 .091 .763 .043 .836 .247 .619 .116 .733 Rank/NChild F-Stat Sig. o f F .002 .966 1.859 .173 .109 .741 1.595 .207 .077 .781 1.433 .232 .067 .796 .662 .416 .160 .689 .665 .415 1.147 .284 .209 .647 .367 .545 .119 .730 Sex/NChild/Rank F-Stat Sig. o f F 5.710 .017 .187 .665 .658 .417 6.006 .014 5.529 .019 .023 .879 .038 .846 Sex/Marrv/Rank F-Stat Sig. o f F 2.934 .087 .472 .492 .011 .916 .006 .936 .087 .768 2.393 .122 .041 .840 .009 .924 2.482 .116 .343 .558 .013 .908 7.534 .006 .339 .561 .018 .892 .620 .431 .435 .510 .586 .444 .042 .837 .521 .470 .213 .645 .002 .963 1.980 .012 3.216 .000 1.498 .093 2.036 .009 1.157 .298 1.559 .073 .262 .998 Three-Way Interacts Sex/Marrv/NChild F-Stat Sig. o f F NChild/Marrv/Rank F-Stat Sig. o f F Four-Way Interacts Sex/NChild/Marrv/ Rank F-Stat Sig. o f F Explained F-Stat Sig. o f F N-980 138 TABLE 21 ANOVA ON PREFERRED ALLOCATION OF WORK TIM E VARIABLES Dependent Variables Independent Variables R scpl Rscp2 Covariate Born F-Stat Sig. of F 16.743 .000 37.254 .000 2.77 Rscp4 Rscp5 1.179 .278 13.339 .000 14.905 .000 10.944 .001 3.88 2.99 3.64 2.90 2.33 2.49 2.84 16.755 .000 4.12 3.81 11.833 .001 2.82 3.03 17.863 .000 3.83 3.59 16.091 .000 2.89 2.91 .190 .663 2.30 2.34 .101 .750 2.55 2.80 4.056 .044 4.00 3.86 .429 .513 2.98 2.99 .324 .569 3.64 3.64 1.129 .288 2.99 2.89 1.283 .258 2.38 2.33 .521 .470 NChild No child - mean Child - mean F-Stat Sig. of F 2.63 2.78 .245 .621 4.07 3.86 .032 .857 3.03 2.98 1.736 .188 3.70 3.64 .241 .623 2.97 2.90 1.810 .179 2.31 2.34 .068 .795 Rank Full/Assoc. - mean Asst. - mean F-Stat Sig. of F 2.80 2.60 .102 .750 3.84 4.06 .174 .677 2.99 2.99 1.373 .242 3.62 3.74 .015 .902 2.90 2.92 6.967 .008 2.35 2.28 .998 .318 Two-Way Interactions Sex/Marry F-Stat Sig. of F .925 .336 .051 .821 2.921 .088 .086 .770 .100 .752 .226 .635 2.763 .097 1.086 .298 .184 .668 .142 .707 .361 .548 1.405 .236 Grand Mean Main Effects Sex Female Mean Male Mean F-Stat Sig. of F Marry Not partnered - mean Partnered - mean F-Stat Sig. of F Sex/NChild F-Stat Sig. of F Rscp3 Rscp6 139 TABLE 21 (Continued) R scn 1 Marrv/NChild F-Stat Sig. of F Rscr>2 R scd 3 Rsct>4 R sc d 5 R sc d 6 1.317 .251 3.952 .047 2.035 .154 .091 .763 4.929 .027 .140 .709 Rank/Sex F-Stat Sig. of F .857 .355 .957 .328 .677 .411 .044 .834 2.013 .156 4.221 .040 Rank/Marrv F-Stat Sig. of F .055 .814 .088 .767 1.236 .266 2.144 .143 .451 .502 .008 .928 Rank/NChild F-Stat Sig. of F .438 .508 .141 .708 3.689 .055 .001 .975 .281 .596 .004 .951 .281 .596 2.895 .089 .196 .658 5.257 .022 .774 .379 .003 .956 Sex/NChild/Rank F-Stat Sig. of F 2.352 .125 1.008 .316 .384 .536 1.491 .222 .036 .850 .368 .544 Sex/Marrv/Rank F-Stat Sig. of F .032 .858 .251 .616 .437 .509 .034 .853 .003 .953 .289 .591 NChild/Marrv/Rank F-Stat Sig. of F .003 .954 .004 .948 .833 .362 1.179 .278 .377 .539 .747 .388 Four-Way Interaction Sex/NChild/Marrv/Rank F-Stat Sig. of F 1.241 .265 .103 .748 1.255 .263 5.104 .024 .070 .791 .780 .377 Explained F-Stat Sig. of F 3.334 .000 3.895 .000 2.105 .007 2.824 .000 2.500 .001 1.287 .198 Three-Way Interactions Sex/Marrv/NChild F-Stat Sig. of F N-980 140 TABLE 22 A N O V A O N IM P O R T A N C E O F T A SK S FO R T E N U R E , R A N K , A N D M E R IT IN C R E A SE S Dependent Variables Independent Variables Covariate Born F-Stat Sig. of F Rettrm Rertrm Reatrm Restrm Reetrm 30.753 .000 19.558 .000 13.811 .000 12.109 .001 3.547 .060 4.03 4.27 2.71 2.98 2.26 4.04 4.03 2.754 .097 4.26 4.27 1.176 .279 2.77 2.69 2.751 .098 2.95 2.99 .516 .473 2.44 2.21 10.820 .001 4.00 4.03 .039 .843 4.28 4.27 .028 .867 2.71 2.71 .246 .620 2.86 3.00 1.090 .297 2.33 2.25 .010 .921 3.76 4.05 4.470 .035 4.41 4.26 1.264 .261 2.55 2.72 .974 .324 2.70 3.01 2.214 .137 2.26 2.26 .071 .790 Rank Full/Assoc. - mean Asst. - mean F-Stat Sig. of F 4.06 3.88 .113 .737 4.26 4.33 1.970 .161 2.69 2.76 10.225 .001 3.02 2.83 .007 .931 2.26 2.24 .223 .637 Two-Way Interactions Sex/Marrv F-Stat Sig. of F 4.235 .040 .830 .362 1.226 .268 2.524 .113 2.999 .084 .414 .520 .503 .478 .663 .416 .549 .459 1.260 .262 Grand Mean Main Effects Sex Female Mean Male Mean F-Stat Sig. of F Marry Not partnered - mean Partnered - mean F-Stat Sig. of F NChild No child - mean Child - mean F-Stat Sig. of F Sex/NChild F-Stat Sig. of F 141 TABLE 22 (Continued) Rettrm Marry/NChild F-Stat Sig. of F Rertrm Reatrm Restrm Reatrm 3.213 .073 .036 .850 4.763 .029 .165 .685 .000 .995 Rank/Sex F-Stat Sig. of F .006 .938 1.891 .169 5.699 .017 2.061 .151 .078 .780 Rank/Marrv F-Stat Sig. of F .000 .991 .238 .625 .196 .658 .663 .416 .410 .522 Rank/NChild F-Stat Sig. of F 1.746 .187 1.990 .159 4.789 .029 2.101 .148 .038 .846 1.122 .290 .012 .912 3.291 .070 2.245 .134 .375 .540 6.148 .013 .038 .846 .040 .842 2.705 .100 .068 .794 Sex/Marrv/Rank F-Stat Sig. of F .038 .846 13.258 .000 1.062 .303 .008 .929 .007 .931 NChild/Marrv/Rank F-Stat Sig. of F .444 .506 9.229 .002 .782 .377 .463 .497 .167 .683 .218 .641 .114 .736 .006 .940 .184 .668 .266 .606 3.450 .000 3.031 .000 3.082 .000 1.779 .030 1.306 .186 Three-Way Interactions Sex/Marrv/NChild F-Stat Sig. of F Sex/NChild/Rank F-Stat Sig. of F Four-Way Interaction Sex/NChild/Marrv/Rank F-Stat Sig. of F Explained F-Stat Sig. of F N-980 142 TABLE 23 ANOVA ON GEOGRAPHIC MOBILITY VARIABLES Dependent Variables Independent Variables Tkactl Tkact2 Lvorm Free Covariate Born F-Stat Sig. of F .520 .471 165.440 .000 38.456 .000 21.377 .000 Grand Mean 1.75 2.61 3.59 2.93 Main Effects Sex Female Mean Male Mean F-Stat Sig. of F 1.70 1.76 .004 .948 2.79 2.57 .789 .375 3.45 3.63 .001 .980 3.17 2.87 6.807 .009 Marry Not partnered - mean Partnered - mean F-Stat Sig. of F 1.63 1.77 .348 .555 2.81 2.58 1.660 .198 3.31 3.64 5.861 .016 3.09 2.91 1.199 .274 NChild No child - mean Child - mean F-Stat Sig. of F 1.65 1.76 .031 .861 3.07 2.57 .079 .778 3.26 3.62 .503 .479 3.19 2.91 .763 .383 Rank Full/Assoc. - mean Asst. - mean F-Stat Sig. of F 1.82 1.43 16.588 .000 2.48 3.23 .266 .606 3.65 3.34 .649 .421 2.92 2.98 6.850 .009 Two-Way Interactions Sex/Marrv F-Stat Sig. of F .167 .683 .225 .635 .104 .747 2.981 .085 .023 .880 1.300 .254 2.354 .125 .071 .790 Sex/NChild F-Stat Sig. of F 143 TABLE 23 (Continued) Tkactl Marrv/NChild F-Stat Sig. of F Tkact2 Lvorm Free 2.273 .132 .180 .672 .022 .882 1.044 .307 Rank/Sex F-Stat Sig. of F .163 .686 .319 .572 .786 .375 4.306 .038 Rank/Marrv F-Stat Sig. of F .009 .926 .028 .867 .768 .381 1.525 .217 Rank/NChild F-Stat Sig. of F 1.656 .198 3.881 .049 4.088 .043 .046 .830 .004 .950 .019 .891 .114 .736 .428 .513 Sex/NChild/Rank F-Stat Sig. o fF .644 .423 1.038 .309 .005 .945 .896 .344 Sex/Marrv/Rank F-Stat Sig. of F .031 .859 .342 .559 .179 .673 1.925 .166 2.349 .126 3.517 .061 3.502 .062 1.254 .263 .020 .889 .190 .663 .140 .708 1.907 .168 1.610 .060 11.167 .000 3.685 .000 3.412 .000 Three-Way Interactions Sex/Marrv/NChild F-Stat Sig. of F NChild/Marrv/Rank F-Stat Sig. of F Four-Way Interaction Sex/NChild/Marrv/Rank F-Stat Sig. of F Explained F-Stat Sig. of F N-980 144 ANOVA RESULTS By drawing from Tables 17 through 23 above it is possible to isolate the dependent variables for which the null hypotheses can be rejected when the effects o f age(BORN) are removed. Table 24 below summarizes the dependent variables for which main and interactions effects had an F significance below alpha. TABLE 24 D EPEN D EN T VARIABLES FOR W HICH NULL HYPOTHESIS CAN BE REJECTED Independent Variables Main Effects Sex Professional Satisfaction W ork Time Allocation Geographic Mobility W orksat Suppsat Sat8 Sat24 Prsp Rscp4 Rscpl Regtrm Rscp2 Rscp3 Free Rank Suppsat Sat2 Sat 14 lopsat Prsp Ppdt Padgv Reatrm Rscp5 Tkactl Free M arry W orksat Suppsat Sat2 Sat8 Sat 13 Sat 14 Sat24 Inoinsat Rscpl Lvorm Nchild Sat 14 Icosesat Ioppsat Rettrm Climsat Pteach Rettrm Interactions Sex/marry 145 TABLE 24 (Continued) Sex/Nchild Sat20 M arry/Nchild Rscp2 Rscp5 Reatrm Rank/Sex Climsat Sat24 Reatrm Rscp6 Free Rank/Nchild Sat2 Sat 13 Satl4 Reatrm Tkact2 Lvorm Sex/M arry/Nchild Rscp4 Rank/M arry/Sex Rertrm Rank/M arry/Nchild Rertrm Psex Rank/Nchild/Sex Pteach Ppdt Psex Rettrm Rank/Sex/M arry/Nchild Rscp4 Main effects o f SEX Among the three independent variables, SEX is one o f the two variables for which the null hypothesis can be most frequently rejected. For 11 o f the 38 dependent variables the null hypothesis can be rejected even when age is no longer a factor. It appears that males and females may differ in certain areas o f professional satisfaction, work allocation, and geographic mobility. 146 Specifically, female survey respondents seem less satisfied than their male counterparts regarding the work itself (WORKSAT), support for their work(SUPPSAT), opportunities for advancement(S A T8), developm entsA T 24). and opportunities for professional growth and In each case the mean satisfaction response o f the female respondents was less than that o f the males. Sex also appears to influence the current and preferred allocation o f work time as well as perceptions o f the importance of governance in evaluation and promotion. While differences in the current percentage o f time spent in teaching(PTEACH) were not significant once the effects o f age were controlled, the female mean for percentage of time spent in research(PRSP) was 22.37% , while the male mean was 27.04% . That represents a significant difference in the population. In terms o f preferred allocation o f work time it appears that there may be significant differences among male and female faculty in the areas o f teaching, research, advising, and professional development. On several measures o f preferred work allocation, the null hypothesis can be rejected for the population. The female means for restructuring their time indicate that they would like to do somewhat less teaching(RSCPl = 2 .4 9 ) and advising (R SC P3=2.82) and spend somewhat more time on research(R S C P 2= 4.12) and professional developm ent(RSCP4=3.83). This compares with male faculty mean responses that are looking for somewhat less restructuring. Males desire to teach somewhat less(R S C P l= 2.84) but not as much as the females. Males, overall, appear to be content with their advising loads(RSCP3=3.03). While they would like to spent 147 more time in research(R SC P2=3.81) and professional development (R S C P 4= 3.59), it does not appear that their desires to restructure their work time are as strong as those of female faculty. Thus, while many o f the preferences for change are the same for male and female faculty, the strength o f the preferences is not the same. Despite these desires for a similar shifts in the direction o f their responsibilities, in the population the importance o f governance activities for determining tenure, rank, and merit increases were not the same for males and females when age was controlled. The female mean(REGTRM) for importance was high, 2.44, compared with the male mean o f 2.21. In terms o f geographic mobility females in the population may be similar to the males on several measures when age is controlled. There is no basis on which to reject the null hypothesis for their seeking a new position at Michigan State (T kactl), or at another institution (Tkact2). N or can it be concluded that males and females are different in their relative desires to leave or stay at Michigan State University as faculty members. However, the male and female faculty may be different in their relative sense of constraint about a decision to leave or stay. Females report a mean sense of constraint(Free) o f 3.17, which is significantly higher than that o f the males at 2.87. This remains a significant difference in the population when age is removed from consideration. 148 In summary, when age is held constant, there may be real significant differences between male and female faculty at Michigan State University. It appears that females allocate less time to research than males. Yet female faculty have stronger preferences for shifting more o f their professional time to research and professional development than do their male counterparts. However, curiously, females place more importance on administration/governance activities for tenure, promotion, and merit increases than do male faculty. There appear to be some conflicts for female faculty. They seem to value teaching and advising as do their male colleagues but have stronger desires to redirect their energies into research and professional development than do their male colleagues. The effects of age differences have been neutralized. Thus such differences must arise from other sources. Despite these apparent conflicts few significant differences were found in the desire or likelihood o f leaving Michigan State University. Though female faculty perceive themselves as more constrained in such decisions, the effect o f such perceptions on real decisions may be quite limited. Main Effects o f RANK Rank is a second independent variable which has a significant effect on several measures o f professional satisfaction, work time allocation, and geographic mobility. Those in the rank o f assistant professor are slightly more satisfied with support for their work than 149 those o f superior rank (SU P PS A T = 3.19 vs. 3.15) as well as compensation-related benefits (S A T 14= 4.16 vs. 4.05). Yet they express much less satisfaction with job security (S A T 2= 3.54 vs. 4.69) when compared with those o f associate and full professor rank. The higher score given to the importance o f satisfaction with opportunity (IO P PSA T = 3.36 vs. 2.87) is significant. They appear to value satisfaction with opportunity more highly than those o f higher rank. Apart from these four measures it appears that those o f lower rank are as professionally satisfied as those o f higher rank. Several measures o f w ork time allocation were affected by rank. The allocation of current work time to research is significantly higher for those in the assistant professor rank (P R S P = 30.75 vs. 25.02) than those o f associate and full rank. Yet the allocation of current work time to professional development is lower (PPDT = 1.82 vs. 2.82) among assistant professors as is the allocation o f time to administration and governance (PADGV = 10.64 vs. 17.28). Only two other small differences surfaced in this part o f the analysis. Assistant professors appear to desire less restructuring o f time related to service and extension than their counterparts of higher rank (R SC P5= 2.92 vs. 2.0). They also appear to value advising somewhat more highly for tenure, rank, and merit increases (REA TRM = 2 .7 6 vs. 2.69). Rank also appears to effect geographic mobility. Assistant professors were somewhat more likely to seek a new position at Michigan State University (TKACT1 = 1.43 vs. 1.82) than their superiors in rank. That may match with their perception that they are 150 more geographically constrained (F R E E = 2 .9 8 vs. 2.92) than the associate and full professors. Overall, RANK is a significant independent variable which appears to affect many o f the dependent variables in several areas. In some instances, such as job security satisfaction, the nature o f the effect seems very logical and reasonable. Yet in other areas the effect of rank on faculty satisfactions, w ork time allocation, and mobility focus on variables and directions of findings that seem somewhat surprising. Main effects o f MARRY When MARRY is used a an independent variable to distinguish the effects o f a faculty member currently being in an adult partnership, there are ten dependent variables for which the null hypothesis can be rejected. Most o f these variables are in the area of professional satisfaction. The non-partnered faculty members seem less satisfied with the work itself(W QRKSAT), having a mean o f 2.91 compared to 3.30 for those who were partnered. They are also less satisfied with university support for their work (SUPPSAT), with a mean o f 2.96 compared to 3.19 for partnered faculty. Similarly the non-partnered faculty register less satisfaction with job security(Sat2, Mean = 4 .1 4 ), their current salaries and benefits (S A T 13= 3.20; S A T 14= 3.89), opportunity for advancement (Sat8, M ean= 3 .3 9 ), and opportunities for professional development (Sat24, m ean= 2.91), when compared with their partnered colleagues who register respective mean satisfactions o f 4.54, 3.45, 4.10 and 3.33 on these five measures. 151 Interestingly, the importance o f different sources o f professional satisfaction does not appear to vary significantly in the population by one’s marital status (Only INOINSAT, the importance o f satisfactions with noninstitutional factors differs). Thus while it is likely that whether one is partnered does have an effect on actual professional satisfaction, the importance o f different sources o f satisfaction may be unaffected by marital status. M ARRY, as an independent variable appears to influence only two other dependent variables significantly. The null hypothesis can be rejected for the preferred amount of time allocated to teaching (R SC Pl). The non-partnered group indicated a stronger preference for somewhat less teaching (2.55) than did the partnered group (2.80). Aside from this variable there were no clear differences in the two groups in terms o f their preferred allocations o f work time. The null hypothesis can also be rejected in reviewing whether these two groups have similar levels o f desire to stay at Michigan State University (LVORM). W hile both express a desire on balance to stay, the nonpartnered group’s mean desire is weaker (3.31) than that o f the partnered group (3.64). In summary, the main effects o f marital status in the population are somewhat more limited in range than the main effects o f sex. The null hypothesis can be rejected prim arily in the area o f professional satisfaction, where the nonpartnered group registers significantly less satisfaction on eight different measures. 152 Main effects o f NCHILD When NCHILD is analyzed as the independent variable it appears that it have main effects on only four o f the dependent variables. The null hypothesis can be rejected for two variables measuring the importance o f satisfaction with compensation and security (ICOSESAT) and with professional opportunities(IOPPSAT). In each case those with children reported a significantly higher mean importance o f these as factors in their decisions than did those without children. Yet despite these differences in the importance o f these sources o f satisfaction, only one significant differences in actual professional satisfaction within the population could be detected for those with or without children. That was satisfaction with compensation-related benefits(SAT14), with those with children expressing greater satisfaction. W hether one has children does not appear to affect the allocation o f w ork time. Those without children reported a similar mean percentage o f time spent in teaching(PTEACH), research(PRSP), and advising(PADS) and other variables measuring w ork time allocation to those with children. On only one other dependent variable does having children have an apparent effect. Those with children appear to place a higher mean importance on teaching (Rettrm) for tenure, promotion, and merit increases than do those without children. In summary, the effects o f faculty members having children appear to be limited to a very few areas. W hile those with children may deem a couple areas o f satisfaction as 153 more im portant, there is no indication that they are any different in real satisfactions from faculty without children. Not only was there no indication that the presence or absence o f children made any significant difference in how faculty viewed the likelihood or freedom to move; but also the allocation o f work time appeared to remain similar for faculty members with or without children. Overall, the main effects show a pattern in which SEX and RANK may have more significant effects than marital status or having children in the family. The most consistent areas of effect may be in professional satisfaction and allocation o f work time with less effect on issues o f geographic mobility. Significant interaction effects The analysis o f possible interactions between and among the four independent variables yielded only a few significant results for the population. W hile it would be hypothetically possible to find 228 significant two-way interactions, only 18 such interactions were found, and while 152 three-way interactions were possible, only 8 such interactions were found to be significant. O f 38 possible variables with four-way interactions, only one variable showed a significant four-way interaction. Thus most of the significant effects were main effects. Far fewer involved interactions among the independent variables. It is, however, instructive to note the dependent variables and the interactions terms for which there was significance. In cases where significance is found, the null hypothesis 154 o f no interaction effect can be rejected. Thus for a few selected dependent variables the interaction o f two or more independent variables may be significant in the population. It appears that SEX and MARRY have a significant interaction for one’s satisfaction with the climate of one’s work(CLIM SAT), the percentage of time spent in teaching(PTEACH) and the importance attached to teaching(RETTRM ). Each o f these variables and their means are presented in Table 25. 155 TABLE 25 DEPEN D EN T VARIABLE MEANS FOR SIGNIFICANT TW O-W AY INTERACTIONS BETWEEN SEX AND MARRY Variable/ M arital Status Sex Male Female Climsat Partnered Non-partnered 3.54 3.28 3.46 3.54 Pteach Partnered Non-partnered 33.60 33.60 33.65 44.11 Rettrm Partnered Non-partnered 4.04 3.89 4.00 4.12 For CLIM SAT, the satisfaction for non-partnered males is the lowest with partnered males and all females appearing to be closer to each other in their satisfaction with the climate for their work. For PTEACH, the greatest difference occurs for non-partnered females. Their mean percentage o f time spent in teaching is far above that o f both married females and all males. For RETTRM, the lowest importance attached to teaching related to tenure, promotion and merit increases occurs for non-partnered males. The non-partnered females attachment of importance is consistent with the fact that they spend a higher percentage o f their time in actual teaching. Significant interaction between SEX and NCHILD is limited to one dependent variable, SAT20, the means of which are shown in Table 26. 156 TABLE 26 DEPENDENT VARIABLE MEANS FOR SIGNIFICANT TWO-WAY INTERACTIONS BETWEEN SEX AND NCHILD Variable/ Nchild Sat20 No children Children Sex Male Female 3.23 3.86 3.79 3.70 SAT20 findings, namely satisfaction with the quality o f graduate students, indicate that the most satisfied group is males with children, closely followed by all females. The least satisfied group is males without children. W hile the statistical results indicate that the null hypothesis o f no interaction effect can be rejected, it is difficult to determine a connections between satisfaction with students, one’s sex, and whether one has children. Significant interactions between MARRY and NCHILD are limited to three dependent variables, RSCP2, RSCP5 and REATRM . Their means are shown in Table 27. 157 TABLE 27 D EPENDENT VARIABLE MEANS FOR SIGNIFICANT TW O-W AY INTERACTIONS BETW EEN MARRY AND NCHILD Partnered Non-Partnered Rscp2 No children Children 4.15 3.84 3.93 4.02 Rscp5 No children Children 3.08 2.88 2.79 3.05 Reatrm No children Children 2.72 2.71 2.28 2.82 Variable/ Nchild M arry RSCP2, the desire to restructure time allotted to research, resulted in a pattern of similarity between partnered faculty with children and non-partnered faculty without children. Yet partnered faculty without children and non-partnered faculty with children expressed a stronger desire to restructure their time toward more research. RSCP5, restructuring time allotted to service/extension work, showed the same pattern, grouping partnered faculty with children and non-partnered faculty without children as desiring a little less time spent in such activities. Partnered faculty without children and non-partnered faculty with children appear to desire a little more time spent in such work. 158 REATRM , the importance o f advising for tenure, promotion, and merit increases, indicates that the non-partnered faculty without children are quite different from all partnered faculty as well as from non-partnered faculty who have children. W hile the number of two-way interactions generated by the combinations discussed above are quite limited, the interactions o f RANK with other independent variables produced many more significant sources o f interaction. While there were no significant interactions between RANK and MARRY, there were several significant interactions between RANK and SEX and between RANK and NCHILD. Table 28 documents the interactions between RANK and SEX below. TABLE 28 DEPEN DEN T VARIABLE MEANS FOR SIGNIFICANT TWO-WAY INTERACTIONS BETWEEN RANK AND SEX Rank Sex Male Female Climsat A ssoc./Full Prof. Asst. Prof. 3.52 3.50 3.56 3.32 Sat24 A ssoc./Full Prof. Asst. Prof. 3.34 3.42 3.06 2.84 Rscp6 Assoc./Full Prof. Asst. Prof. 2.34 2.37 2.37 2.14 Reatrm A ssoc./Full Prof. Asst. Prof. 2.67 2.83 2.83 2.66 Free Assoc./Full Prof. Asst. Prof. 2.88 2.80 3.12 3.27 159 For CLIM SAT, satisfaction with the university climate for work, female assistant professors reported lower levels o f satisfaction than those in all other categories including female professors o f higher rank. Female assistant professors were also the least satisfied with opportunities for professional growth and development, SAT24. However, in this case female professors o f higher rank were also less satisfied than males o f all ranks who were similarly satisfied. Only two differences in the allocation of work time were noted for this two-way interaction. Again, female assistant professors were more eager to be less involved in governance, RSCP6, than were any o f the other three groups which were similar. Yet in the area o f the value o f advising for tenure, promotion, or merit increases, REATRM, a similar grouping did not exist. In this case the female assistant professors were similar to the male associate/full professor group, both o f which assessed advising’s import at a lower level than the male assistant professors and the female full professors. The interaction o f RANK and SEX did generate differences in the perceptions of constraints on geographic mobility. Interestingly, those feeling the most constrained were female assistant professors while male assistant professors felt the least constrained o f the four groups. Yet all females experienced a higher sense o f constraint regardless o f rank than did the males. 160 The two-way interaction between RANK and NCHILD also generated some significant results as noted in Table 29 below. TABLE 29 D EPEN D EN T VARIABLE MEANS FOR SIGNIFICANT TW O-W AY INTERACTIONS BETW EEN RANK AND NCHILD Sex Rank No Children Children Sat2 A ssoc./Full Prof. Asst. Prof. 4.84 3.19 4.68 3.62 Sat 13 A ssoc./Full Prof. Asst. Prof. 3.12 3.90 3.43 3.33 S atl4 A ssoc./Full Prof. Asst. Prof. 3.43 4.32 4.09 4.12 Reatrm A ssoc./Full Prof. Asst. Prof. 2.38 2.81 2.71 2.75 Lvorm A ssoc./Full Prof. Asst. Prof 3.13 3.45 3.68 3.32 Tkact2 A ssoc./Full Prof. Asst. Prof 3.26 3.08 2.44 3.06 The three satisfaction variables, SAT2, satisfaction with job security, SAT13 and SAT14, satisfaction with salary and with benefits respectively, do not have similar results. The group least satisfied with job security is assistant professors with no children. While assistant professors with children are the second least satisfied group, there is a 161 reasonably large gap between others in their rank without children. Yet all assistant professors are less satisfied than those with higher rank. Satisfaction with salary and with benefits show the same pattern. In both cases professors o f higher rank without children are the least satisfied with compensation. Faculty o f all ranks who have children are similarly satisfied at a somewhat higher level. Yet the most satisfied are those without children at the assistant professor level. In terms o f work time allocation there was only one significant two-way interaction related to restructuring o f advising time, REATRM. Professors o f higher rank without children clearly desire to do less advising than their counterparts with children or anyone in the assistant professor rank. Finally, the interaction o f RANK and NCHILD showed significant effects on the likelihood o f leaving or remaining at Michigan State University, LVORM. Faculty of all ranks with children and assistant professors without children all had a stronger propensity for remaining at the university than did professors o f higher rank without children. A similar pattern was found regarding the likelihood o f seeking a new position at another institution, TKACT2, with associate and full professors without children expressing the greatest likelihood o f doing so. 162 This statistical analysis yielded only eight three-way interaction which showed any significance. O f those only one o f eight was unrelated to RANK in some way. In that case it appears that the interaction o f one’s sex, marital status, and children does affect a faculty mem ber’s desire to restructure time for professional development(RSCP4). The results are shown in Table 30. TABLE 30 D EPENDENT VARIABLE, RSCP4, MEANS FOR SIGNIFICANT THREE-W AY INTERACTIONS AMONG SEX, M ARRY, AND NCHILD Children No Children Non-partnered Partnered Non-partnered Partnered Male 3.22 3.72 3.57 3.59 Female 3.90 3.67 3.71 3.89 Table 30 indicates that those who least desire a restructuring o f time to allow for more professional development as non-partnered males with no children. Their mean is most distant from that o f non-partnered females with no children and partnered females with children, both o f who have a stronger desire for restructuring than do these males. The other seven three-way interactions are related to RANK in combination with other independent variables. The three-way interaction o f RANK with MARRY and SEX 163 generated some significant differences on the variable, RERTRM, the value placed on research for tenure, promotion, or merit increases. These results are shown in Table 31. TABLE 31 D EPEN D EN T VARIABLE, RERTRM, MEANS FOR SIGNIFICANT THREE-W AY INTERACTIONS AMONG RANK, MARRY, AND SEX A ssoc./Full Prof. Asst. Prof. Male Female Male Female Non-partnered 4.08 4.53 4.30 4.15 Partnered 4.45 4.01 4.25 4.33 Those groups lending the least and most importance to research were both within the assistant professor ranks. However within that rank non-partnered females and partnered males were most similar in their stronger belief in the importance o f research. The other end o f the spectrum joined non-partnered males and partnered females together in similar views. Interestingly the males and females of higher rank were more similar and represented views between the highs and lows generated by the assistant professor cells. Two significant three-way interactions were generated by the combination o f RANK with MARRY and NCHILD. Tables 32 and 33 show the results for the variable RERTRM , the importance placed on research for tenure, promotion and merit increases, and the variable PSEX, the percentage o f time spent in service/ex tension activities. 164 TABLE 32 DEPENDENT VARIABLE, RERTRM, MEANS FOR SIGNIFICANT THREE-WAY INTERACTIONS AMONG RANK, MARRY, and NCHILD A ssoc./Full Prof. Asst. Prof. Non-partnered Partnered Non-partnered Partnered No Children 4.25 4.76 4.47 4.24 Children 4.41 4.24 4.18 4.26 TABLE 33 D EPENDENT VARIABLE, PSEX, MEANS FOR SIGNIFICANT THREE-W AY INTERACTIONS AMONG RANK, MARRY, AND NCHILD A ssoc./Full Prof. Asst. Prof. Non-Partnered No Children Children Partnered Non-Partnered Partnered 16.46 5.00 5.00 8.66 7.83 12.58 11.90 12.09 For RERTRM the import given to research is highest for partnered assistant professors without children while it is lowest for non-partnered professors o f higher rank with children. The range of responses is also broader at the assistant professor level than at the associate professor level. For PSEX, time spent in service and extension, the range is from 5% for two o f the ranked groups without children to 16.46% for non-partnered assistant professors without children. Those with children did differ but within a narrower range. Yet in general 165 those with children in most ranks and marital situations tended to devote more time to service and extension than did those without children. The interaction o f RANK with NCHILD and SEX results in four three-way interactions with significance for the variables PTEACH, the percentage o f time spent in teaching, PPDT, the percentage o f time spent in professional development, PSEX, the percentage o f time spent in service and extension, and RETTRM, the importance placed on teaching for tenure, promotion, and merit increases. Tables 34-38 below indicate the results. TABLE 34 DEPENDENT VARIABLE, PTEACH, MEANS FOR SIGNIFICANT THREE-W AY INTERACTIONS AMONG RANK, NCHILD, AND SEX Asst. Prof. Male A ssoc./Full Prof. Female Male Female No children 26.17 42.63 36.23 31.32 Children 35.46 36.09 33.35 37.40 From Table 34 it is apparent that male and female assistant professors without children represent the greatest disparity in results on teaching time (PTEACH). These males report the smallest percentage o f time spent in teaching while these females report the largest. Aside from these two lower-rank groups without children, all other groups are reasonably similar. It is interesting to note however that there is some difference among females in the higher ranks on the basis o f NCHILD. Those without children report a smaller percentage o f time spent in teaching. 166 TABLE 35 DEPENDENT VARIABLE, PPDT, MEANS FOR SIGNIFICANT THREE-WAY INTERACTIONS AMONG RANK, NCHILD, AND SEX Asst. Prof. A ssoc./Full Prof. Female Male Male Female No children 3.42 1.11 2.96 6.98 Children 1.60 2.14 2.62 3.10 The percentage o f time spent in professional development, PPDT, fluctuates widely as noted in Table 35. Female assistant professors without children report the least amount o f time spent followed closely by male assistant professors with children. Female associate/full professors without children report the greatest amount o f time spent and as a group appear to be quite different from all other groups. TABLE 36 D EPEN D EN T VARIABLE, PSEX, MEANS FOR SIGNIFICANT THREE-W AY INTERACTIONS AMONG RANK, NCHILD, AND SEX A ssoc./Full Prof. Asst. Prof. Male Female M ale Female N o children 10.83 9.16 3.96 12.00 Children 10.34 14.02 12.33 10.69 Table 36 shows that the percentage o f time spent on service and extension work, PSEX, also varies widely. Fem ale assistant professors with children report the highest percentage o f time spent in this way while male professors at the associate and full rank report the lowest percentage of effort. All other groups fall in a relatively small range. 167 TABLE 37 DEPENDENT VARIABLE, RETTRM, MEANS FOR SIGNIFICANT THREE-WAY INTERACTIONS AMONG RANK, NCHILD, AND SEX Asst. Prof. M ale A ssoc./Full Prof. Female Male Female No children 4.17 3.58 3.60 3.91 Children 3.78 4.13 4.08 4.10 From Table 37 above, it is apparent that the importance placed on teaching for tenure, promotion, and merit increases, RETTRM, also varied by the interaction o f one’s rank with one’s sex and household children. The range was broadest among those without children. Curiously the biggest gap among those without children was between male associate/full professors with the lower rating o f import and male assistant professors with the highest rating o f import. Female assistant professors were quite similar to the male professors of higher rank in their rating. Among those with children there was less disparity. Yet the male assistant professors with children still rated teaching o f less importance than did the other three groups. The statistical analysis yielded one significant four-way interaction among the independent variables o f RANK, SEX, MARRY, and NCHILD shown in Table 38 below. 168 TABLE 38 DEPENDENT VARIABLE, RSCP4, MEANS FOR SIGNIFICANT FOUR-WAY INTERACTIONS AMONG RANK, SEX, MARRY, AND NCHILD A ssoc./Full Prof. Asst. Prof. Female Male Male Female Partnered and No children 4.00 3.44 3.60 3.89 Partnered and Children 3.68 4.07 3.58 3.82 Non-partnered and No children 3.00 4.00 3.33 3.80 Non-partnered and Children 3.54 3.56 3.58 3.80 On the variable RSCP4, the desire to restructure the am ount o f time spent in professional development, there was substantially more diversity among assistant professors than among those o f higher rank. Yet in all categories for analysis except one, the female professors desired more restructuring than did their male counterparts in a like category. Only among partnered females with no children did their desire for additional professional development time lag behind that o f partnered males with no children. In summary, the two, three, and four-way interactions effects are somewhat limited. However, the bulk o f the significant interactions are related to RANK in some way. O f 27 interaction effects in total, only 7 are not related to RANK. Among those, there appears to be no definitive pattern, although the interactive effect on specific dependent variables is of interest. Those results do not lead to obvious conclusions. 169 However, it is clear that RANK as a variable interacts very significantly on many variables related to professional satisfaction, work time allocation, and geographic mobility. W ith this background the regression analyses were undertaken to further explore teaching and research activities. TH E REGRESSION ANALYSES The regression analyses investigated some o f the findings about current and preferred allocation o f work effort in regard to teaching and research efforts, the two activities in which faculty spend the majority o f their time. Legitimate regression analysis may be conducted only when the assumptions o f regression analysis have been met. The sections below first present evidence that the assumptions o f regression analysis have been met and then present the results o f the regressiona analyses that have been conducted. Meeting the Basic Assumptions o f Regression Analysis Regression analyses can only be run when the results o f tests for linearity, normality, and homoscedasticity indicate that regression analysis is an appropriate methodology. These tests were run for the four dependent variables, PTEACH, PRSP, RSCP1, AND RSCP2, which were not constants. Scatterplots were examined for evidence o f nonlinearity and homoscedasticity. Sample Skews were examined to determine whether population skews were likely to be between + 1 and -1, an acceptably range o f skew. Cochrane’s C and its p-level were used to examine homoscedasticity as well, using the criterion that the C- 170 value should not exceed .667 for a two-group analysis. The results o f these tests are shown in Table 39 below. TABLE 39 RESULTS OF TESTS ON REGRESSION ASSUMPTIONS Analysis of Scatterplot Residuals Standard­ ized Sample Residual Skews Standard Error of Skew Cochrane’s C P-level PTEACH Current % o f time spent in teaching No evidence of nonlinearity. Some evidence o f unequal variance. .461 .079 .5603 .010 PRSP Current % o f time spent in research No evidence of nonlinearity. Some evidence o f unequal variance. .964 .079 .5147 .545 RSCP1 No evidence of nonlinearity. Some evidence o f unequal variance. .131 .079 .6463 .000 RSCP2 No evidence of nonlinearity. Some evidence o f unequal variance. -.611 .079 .5191 .457 Tests Dependent Variable Tested 171 Although the scatterplots showed some evidence o f unequal variance for all four cases, the skews o f the standardized residuals fell into an acceptable range in all four o f the cases. In each case the Cochrane’s C was within the range that indicates acceptable levels o f homoscedasticity. Yet on PTEACH and RSCP1 there were some questions because the p-levels o f Cochrane’s C were .010 and .000 respectively. However, because o f the large sample, small differences in the sample variances may show statistical significance but still be substantively trivial. In this particular case, the ratio between the maximum and minimum variances should also be considered. In the case o f PTEACH this ratio is 1.274 and for RSCP1 it is 1.827. Such results would seem to indicate that, despite the p-level, the data is sufficiently homoscedastic to work with it effectively. Thus the results o f these tests indicate that the data is sufficiently linear, that residuals have a reasonably normal distribution in the population, and that the variance in the dependent variables in the population is reasonably the same for all levels o f the independent variable. Given these results, the assumptions o f regression analysis have been met without the need for data transformations. The regression analyses focused on the effects o f seven independent variables, SEX, RANK, BORN, MARRY, NCHILD, ROC and GC on four dependent variables, PTEACH, PRSP, RSCP1, AND RSCP2. The zero order correlation for the independent variables is shown below in Table 40. 172 TABLE 40 ZERO ORDER CORRELATIONS FOR IN DEPENDENT VARIABLES Sex Rank Born M arry NChild Inti Int2 Int3 ROC Sex 1.000 Rank -.2 0 2 1.000 Born .201 -.535 1.000 M arry -.268 .147 -.053 1.000 N C hild -.211 .176 -.212 .206 1.000 In ti .730 -.119 .175 .166 -.080 1.000 Int2 -.2 8 0 .099 -.156 .833 .613 .0661 1.000 Int3 .716 -.286 .199 -.106 .124 .676 -.0397 1.000 ROC -.2 7 0 .022 .068 .072 -.025 -.211 .054 -.254 1.000 .374 -.033 -.046 -.085 .013 .319 -.062 .334 -.691 GC GC 1.000 M ost o f the correlations among the independent variables are below .375, indicating a reasonably low level o f correlation among the independent variables. One o f the exceptions to this relatively low level o f correlation was the correlation o f RANK and BORN at -.535. Although, it appears to be a negative correlation, the appropriate interpretation o f it is that as birth year declines, e.g. faculty are older, rank increases. Although this correlation was relatively high, the importance o f both variables to the analysis precluded the option o f dropping either o f them. Another high correlation, at -.691, was between GC, gender composition o f the college, and ROC, research orientation o f the college. Both GC and ROC are specific numerical 173 calculations for each o f the 15 colleges included in the analysis. The strong correlation indicates that the lower the percentage o f female faculty in the college, the higher the research orientation o f the college. There were also some understandably high correlations related to interaction term variables which were newly created for the regression analyses. Because of the significant two-way interaction in the ANOVA o f SEX and MARRY related to PTEACH, IN T I was created as a new variable for the regression analysis o f PTEACH. Similarly, INT2, the two-way interaction o f MARRY and N CHILD , was created as a new variable for the regression analysis o f RSCP2, since it had shown significance related to this dependent variable in the ANOVA. INT3, the three-way interaction o f RANK, NCHILD and SEX also affected PTEACH significantly in the ANOVA and was entered into the analysis as well. Because each o f these interaction terms includes the independent variables o f SEX, MARRY, RANK, and NCHILD in combination with each other, the high correlations o f IN T I with SEX (.790), INT2 with MARRY (.833), and N CH ILD (.613), and INT3 with SEX (.716) and IN TI (.676) are not surprising. Although these higher correlations may cause some problems related to multicolinearity in the regression analysis, the overall impact on the R2 may not be significant enough to require their removal from the analysis. 174 The results o f the regression analysis on the four dependent variables are shown in Table 41 through 44. TABLE 41 REGRESSION RESULTS ON PTEACH CURRENT PERCENTAGE OF TIM E SPENT IN TEACHING Variable GC NChild Rank Bom M arry Sex ROC In ti Int3 (Constant) Un standardized Equation Standard Error 9.336118 -1.234773 -5.590537 -.267300 .549046 6.783652 .059177 -11.102484 4.994440 49.831615 R2 for PTEACH Standard Error o f Estimate F Statistic F Significance 7.518692 3.125249 2.629547 .097007 2.954116 4.636663 .200660 4.651656 4.120152 9.503022 .02829 = 22.99653 = 2.96983 = .0017 Sig. o f T .2147 .6929 .0338 .0060 .8526 .1438 .7681 .0172 .2257 .0000 175 TABLE 42 REGRESSION RESULTS ON PRSP CURRENT PERCENTAGE OF TIM E SPENT IN RESEARCH Variable GC NChild Rank Bom M arry Sex ROC (Constant) Unstandardized Equation Standard Error 9.819539 -3.787041 -4.659461 .149691 2.875124 -3.950075 1.031019 -3.717136 -1.708542 -8.302730 -6.061154 .019296 .655543 -7.203335 .720297 -16.202265 R2 for PRSP Standard E rror o f Estimate F Statistic F Significance Sig. o f T .0949 .1001 .0000 .0245 .1104 .0174 .0000 .5592 = .12340 == 18.40213 = 19.10469 == .0000 TABLE 43 REGRESSION RESULTS ON RSCP1 DESIRE TO RESTRUCTURE AM OUNT OF TEACHING TIM E Variable GC NChild Rank Bom M arry Sex ROC (Constant) Unstandardized Equation Standard Error -.623375 -.057250 .139850 -.004776 .134066 -.259192 -.022947 3.585283 R2 for RSCP1 Standard Error of Estimate F Statistic F Significance = = = .249171 .097603 .030297 .002818 .076313 .070317 .006716 .269857 .06484 .78057 9.40985 .0000 Sig. o f T .0125 .5576 .0000 .0905 .0793 .0002 .0007 .0000 176 TABLE 44 REGRESSION RESULTS ON RSCP2 DESIRE TO RESTRUCTURE AM OUNT OF RESEARCH TIM E Variable GC NChild Rank Bom M arry Sex ROC Int2 (Constant) Unstandardized Equation Standard Error -.050835 .196571 -.048188 .013212 .236496 .181365 -.008450 -.344389 3.443123 R2 for RSCP2 Standard Error o f Estimate F Statistic F Significance = = = .250422 .164341 .030444 .002836 .184081 .070709.006750 .200457 .289053 Sig. o f T .8392 .2319 .1138 .0000 .1992 .0105 .2109 .0861 .0000 .05565 .78434 6.99001 .0000 In each o f the four regressions the Rho-squared was significant, indicating that the null hypothesis, that Rho-squared is zero in the population, can be rejected. Thus it appears that this set o f independent variables is significant in explaining differences among female and male faculty in the population related to teaching and research. However, since the total proportion o f each regression’s variance explained by this model is small, other variables not included in these regression equations must play a significant role in the explanation. Ill The standard errors o f estimate should be handled carefully. In PTEACH and PRSP the responses were calculated in percentages. In RSCP1 and RSCP2 the responses were on a Likert scale from 1 to 5. Thus standard errors should not be compared across the regressions because they are calculated in two different units. However, within their units o f the same type, the larger standard errors indicate wider variability around o f the slope o f the regression line. A summary o f the slopes between the independent and dependent variables is shown in Table 45. For the cases where "Yes" is noted in the table, the slope o f the regression line was significant using an alpha o f .05. With non-dummied variables this indicates that the null hypothesis o f a zero slope in the population can be rejected. With dummy variable slopes this indicates that no significant difference between the measured groups (female, partnered, with children, associate/full professor rank) and the excluded groups (male, non-partnered, without children, assistant professor rank, respectively), exists in the population. 178 TABLE 45 SUM MARY - REJECTION OF NULL HYPOTHESIS ALLOW ED Dependent Variable Independent Variable Pteach GC NChild Rank Bom M arry Sex ROC In ti Int2 Int3 Prsp Rscpl Rscp2 Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes The results displayed above indicate that some o f the independent variables have little significance in relation to the dependent variables which measure current or preferred allocations o f teaching and research time. W hether a faculty member is partnered (MARRY) or has children (NCHILD) do not appear to be significant related to the amount o f time currently spent in teaching and research or one’s preferences in the future. INT2 (MARRY and NCHILD) and INT3 (RANK, NCHILD and SEX) also generate no significant results. The gender composition o f one’s college(GC) appears to have significance only in relationship to faculty preference regarding the amount o f time spent in teaching (RSCP1). W hile partnership (MARRY) by itself affects no dependent variable, IN TI (SEX and MARRY) appears to be significant for the percentage o f time spent in teaching (PTEACH). 179 The research orientation o f one’s college(ROC), seen as the percentage o f time spent in research by the majority sex o f faculty in that college, may have some significance for the percentage o f time spent in research (PRSP), but also for a faculty m em ber’s preference in terms o f time spent in teaching (RSCP1). The independent variables affecting the most dependent variables were a faculty m em ber’s academic rank (RANK), gender (SEX) and age (BORN). There was significant relationship between RANK and three o f the four dependent variables, namely PTEACH, PRSP, and RSCP1. Age (BORN) significantly affected results for PTEACH, PRSP and RSCP2. SEX significantly affected results for PRSP, RSCP1 and RSCP2. From another viewpoint it appears that o f the independent variables tested, the ones having the most significant effect on the percentage o f time spent in teaching (PTEACH) were RANK, BORN, SEX and IN TI (SEX and MARRY). The other independent variables GC, NCHILD, MARRY, ROC, INT2 and INT3 did not significantly affect the PTEACH results. The percentage o f time spent in research (PRSP) was most affected by RANK, BORN, SEX and ROC. The variables GC, NCHILD, MARRY, IN T I, IN T2, and INT3 did not significantly affect the results. 180 The desire to change the amount o f time spent in teaching (RSCP1) was most significantly affected by GC, RANK, SEX and ROC. It was not influenced significantly by NCHILD, BORN, M ARRY, IN T I, INT2 or INT3. The desire to shift the amount o f time spent in research (RSCP2) was affected only by BORN and SEX. It was not significantly influenced by GC, NCHILD, RANK, M ARRY, ROC, IN T I, INT2 or INT3. In summary, it appears that the regression analyses indicate that while several o f the independent variables have limited effects on teaching and research, a faculty member’s rank, sex and age have significant effects in these decisions. These findings parallel results from the two way anova analysis which were previously discussed. In that analysis, even when age was held constant, a faculty mem ber’s rank and sex still generated some differences in allocation o f both current and preferred work effort. Based on the statistical results generated above through the ANOVAS and regressions that have been run on this sample o f Michigan State University’s faculty, some tentative conclusions can be reached about the population. conclusions and posssible reasons for them. Chapter five will discuss these CHAPTER 5 DISCUSSION OF TH E RESULTS The purpose o f generating the data summarized in chapter four was to investigate three key questions posed at the beginning o f chapter one. These questions were: 1. A re male and female faculty in the research university different from each other in key areas which affect university activity and culture? 2. If differences between male and female faculty are significant, can these differences be attributed prim arily and/or directly to gender? 3. If differences between male and female faculty cannot be attributed to gender, what other factors contribute to differences among these men and women? After answers to these questions are reviewed the implications o f these findings will be discussed. ARE THEY DIFFERENT? First, the data do reveal some areas in which the null hypothesis, that female and male faculty do not differ in the population, can be rejected with a 95% degree o f confidence. 181 182 Specifically the one-way ANOVA indicated that in the area o f professional satisfaction some differences appear to exist between the two faculty groups. Female faculty reported themselves less satisfied than males when asked about the work itself, institutional support for their work, job security, opportunities for advancement, and opportunities for growth and development. Significant differences for male and female faculty were not found to be related to satisfaction with the university climate for academic work, salary and benefits, or the perceived quality o f Michigan State University students. Although there were several areas in which females were less satisfied than males, the importance male and female faculty placed on the various sources o f professional satisfaction did not vary significantly. The only exception to this general observation is in the area o f professional opportunity. Satisfaction with professional opportunity appears to be more important to women in the population than to men. Aside from the importance of opportunity, the overall balance o f satisfactions that female faculty are seeking does not appear much different than those of male faculty. Females seem to value sources o f satisfaction similarly to males. From these results it is clear that female faculty are less content than men regarding their current status. They desire more change and more opportunity than do their male counterparts. There are two possible explanations for this state. Perhaps female faculty currently have less satisfaction and opportunity than male faculty at Michigan State 183 University. Thus, their desire for change is an effort to achieve parity with the current status o f male faculty in the university. It is also possible that female faculty hold a status equal to that o f their male colleagues. However, they have set higher standards for themselves than the current predominately male norm. Thus, they are not as satisfied with the status quo. W hile either explanation is possible, the data on the allocation of w ork time and the greater percentage o f women in lower ranks tend to reinforce the likelihood o f the first explanation. In some key activities the allocation o f work time differs between male and female faculty. In the one-way ANOVA, the percentages o f time spent in both teaching and research are important areas o f difference. Other faculty activities such as advising, professional development, service/extension, and governance do not yield obvious differences. W omen faculty in the population teach significantly more and do significantly less research. Yet the female faculty in this university do not appear completely content with the allocation o f their time. W hen asked about their preferences, they prefer more restructuring o f their positions than their male counterparts. Their desire for shifts to less teaching and advising and more time for research and professional development are significantly greater than those o f male faculty. This greater desire to restructure is likely reflected in the lower level o f satisfactions about their current situations as reported by female faculty. 184 These desires to restructure their time do not imply that female faculty see the ideal relationship between academic tasks and rewards differently than do their male colleagues. The statistics do not reveal such significant differences. Both groups appear similar in their opinions about the importance o f various academic activities for determining tenure, rank, and m erit increases in the research university context. Male and female faculty both attach themselves to the research university ethos that expects and rewards attention to research. Finally, it appears that female faculty’s current restlessness about the gulf between their current situations and their aspirations is complicated by their perceptions about geographic mobility. They believe themselves less likely to stay at Michigan State University and more likely to look for a position at another institution than do males, although the women feel significantly more constrained in their freedom to move. It is possible that female faculty feel caught between hope and despair. They may hope that moving to another institution can help bridge the gulf from current reality to their aspirations. Yet the perception that their freedom to move is more constrained may lead to despair about their ability to change the situation. Overall, it appears that the male and female faculty at Michigan State University are quite similar in their ideals. The importance o f key professional satisfactions and the link between academic tasks and rewards does not appear to differ for these two groups. Yet the women are less convinced that their current situations reflect their ideals. They 185 perceive their actual satisfactions and actual use o f time to be further from the ideal than do the males. Thus they seek a greater amount o f change and more opportunity to accomplish it. W HY DO M ALE AND FEM ALE FACULTY DIFFER? Given that male and female faculty do differ from each other in terms o f their current situations, the key question is why do they differ? possible reasons for these differences. This study investigated several They could result directly from differences in gender. However, they could also result from differences in age, academic rank, family situation (an active adult partnership and/or children), or the climate o f the university college with which each faculty member is affiliated. Each o f these possible explanations was investigated. Age appears to be a significant, though not exclusive, factor in the explanation. In the regression analyses run on the current and preferred allocation o f time to teaching and research, the slope o f the variable representing age (BORN) was significant in three o f the four equations. Academic rank is also a significant factor in the explanation. Like age the slope o f RANK was significant in three o f the four regression equations. In the ANOVA as well, RANK, by itself or interacting with other independent variables, was substantively 186 related to results in several areas. While some o f the differences in professional satisfaction related to job security (SAT2) and compensation (SAT13 and SAT14) are logically related to RANK, it was also clear that RANK significantly affected work time allocation. The percentage o f time currently spent or preferred for teaching, research, advising, professional development, service/extension, and adm inistration/governance w ere all affected in some significant way by RANK. Particularly when RANK was analyzed interacting with some other independent variable, one pattern was quite noticeable. Very frequently the range o f responses was greatest among those in the assistant professor rank regardless o f sex, marital status, and the presence o f children. Although the responses did not always fit expected patterns, the analyses clearly showed that in most such situations faculty o f associate and full professor rank were more similar to each other than were those o f low er rank to colleagues also in their rank. In both the multiple ANOVA and the regressions, when the effects o f age were controlled and RANK effects were differentiated, gender (SEX) still emerged as a significant variable. In the ANOVA, controls for age and rank eliminated differences in satisfactions with job security and all measures o f geographic mobility except perceived freedom to move. Yet the effects o f sex on differences in the satisfaction with work, university support for work, and opportunities for rank advancement, growth and 187 development remained. Gender had a main effect on these differences even when other sources o f possible explanation were factored into the analysis. Controlling age and rank did eliminate gender’s direct effect on the percentage o f time currently spent in teaching. The regression analysis indicated gender did not have a significant effect on time spent in teaching although the interaction between gender and marital status did. The ANOVA also indicated that gender does not have a main effect. Instead there were significant interaction effects o f gender and marital status and also of gender with rank and the presense o f children. Thus, results in terms o f gender’s effect on teaching are complex, complicated by age, marital status, rank, and children. Gender has some effect on the amount of time spent in teaching but only as coupled with other variables. Together these form a configuration which cannot be explained by individual factors. Perhaps A isenberg’s (1988) assertion that the "marriage quest" has a major effect on academic w om en’s teaching activities should not be dismissed too easily. Gender did have a direct effect on the percentage of time spent in research in the ANOVA, and in the regression analysis. The results indicated that female faculty currently spend significantly less time on research. Yet gender also had a main effect on the desire to change that balance, with women having a stronger desire to spend more time on professional development and research and less time in teaching and advising. 188 Thus, while it is possible that the percentage o f time spent in teaching is substantially affected by other variables interacting with gender, the percentage o f time spent in research is directly influenced by gender. Do family situations have significant effects? It is clear that adult partnerships influence a faculty mem ber’s sense o f professional satisfaction with their work, sense o f job security and opportunity, satisfaction with compensation and satisfactions outside the university. However, aside from those measures, the only other effect is on the desire to restructure the amount o f time spent in teaching and the likelihood o f staying at Michigan State University. Thus, adult partnerships appears to have a limited, though significant, effect on professional satisfaction but little effect on the allocation o f work time or even perceptions o f geographic mobility. Children in the family appear to contribute to only a few minor shifts. The importance o f a few sources o f work satisfaction change for those with children. The presence of children also alters to some degree the value faculty place on teaching. However, the presence or absence o f children does not appear to affect levels o f professional satisfaction, faculty members’ preferred allocations o f work time, their sense of appropriate rewards or geographic considerations. Thus, the presence o f children appears to have very limited effects on faculty satisfactions, perceptions, and choices. 189 Adult partnerships and the presence o f children interacting with gender also resulted in only a few significant differences. The interaction effect o f gender and partnership only on variables measuring the percentage o f time spent in teaching and belief in teaching’s value for rewards confirm the limited degree to which gender and family status are intertwined in professional efforts. Aside from personal attributes or situations, what role might a faculty m em ber’s college have in activities that relate to the core teaching and research missions o f the university? It appears that the gender composition(GC) and research orientation o f the college(ROC) or what they may represent, have some influence on the percentage o f time spent in research and the desire to restructure time devoted to teaching. Thus, it appears that one’s college does influence a faculty mem ber’s allocation o f time to some degree. Yet the pattern o f influence is not consistent. Not all regression slopes related to these two factors were significant. The influence o f one’s college appears to be limited and to prim arily affect research time. In summary, it appears that of the variables examined, gender, rank and age have the more significant effects. They generated the largest and most consistent number of significant differences in the regression analyses when compared with variables examining family situations and the culture o f one’s college. In addition when age was held constant, the differences found in the one-way ANOVA shifted somewhat in the multiple ANOVA. Still, controlling for age did not cancel most o f the main effects 190 associated with gender and also with academic rank. Thus, it appears that age, rank and gender work hand in hand as faculty wrestle with their current situations and aspirations for the future. IM PLICATIONS OF THIS STUDY FOR M ICHIGAN STATE UNIVERSITY Despite some limitations on the data and method o f analysis, the results are consistent enough to w arrant some consideration o f policy changes. It appears that a greater influx o f female faculty into Michigan State University will not alter to ideals o f the faculty dramatically. Overall, female and male faculty seem to balance the importance of teaching, research, and other faculty activities similarly. They also seem to be searching for the same balance o f professional satisfactions. Thus, it appears that the gender balance o f the faculty will not have major ramifications for the broad mission or faculty reward systems o f the university. Concerns that more female faculty will dramatically alter the basic goals o f the institution seem unfounded. It appears that female faculty like males are gradually absorbed into the research university ethos. Furtherm ore, there is no indication that the faculty has significant concerns about institutionalized discrimination in compensation or job security. When age and rank are factored in, significant differences between female and male faculty cannot be documented. Thus, while it is important to remain vigilant in maintaining gender equity in terms o f these basic factors o f employment, there does not appear to be a need to systematically overhaul the systems o f the university at this point. 191 The more appropriate avenues o f response for Michigan State University would appear to be in the areas o f faculty work assignments and faculty development, particularly among less experienced faculty in the lower academic ranks. The broader range o f responses from assistant professors quite possibly indicates that those in the lower academic ranks are not yet completely inculcated into the ethos o f the research university. Since lower academic rank correlates strongly with younger age and with less years o f professional experience, it appears that as faculty members age and gain more experience in a particular environment, the organizational values and culture o f the university are absorbed and more fully embraced. Young faculty who believe they do not fit well with the ethos o f the research university may seek alternative employment. Through a combination o f university evaluation and self-selection only those who fit the dominant ethos remain. Those who don’t fit conclude that they should establish their careers in other types of colleges and universities or outside the sector o f higher education. W hile it is advantageous for any institution to shape a coherent and consistent culture, the danger is for those whose identification with the research university may be in a slightly different form than those who control the ethos. Inexperienced assistant professors who are involved in legitimate forms o f research which do not fit the predominant model for research in their discipline may conclude too quickly that they do not belong. This is a particular danger in research universities which maintain a narrow 192 definition of sound scholarship favoring quantitative research above all else. While such research should be encouraged, it is not the only type o f scholarship appropriate to a research university. The same concern exists when an institution evaluates scholarship only by counting the number o f books published or articles in juried journals. Such rigid evaluation mechanisms often send inappropriate signals to young scholars as they are determining their career plans and working to establish professional confidence. Particularly for females entering research institutions as assistant professors, the culture may not be a supportive one. If the definition o f research is narrow and the evaluation mechanisms are rigid, such an environment could cause great doubt for the many bright women whose career paths have not been linear. Aisenberg(1988), Dwyer(1991) and others have documented the circuitous paths many women take into academic careers. They recognize the hidden passages many women find to combine career and family. Bernard (1964), Rossi (1973) and Dwyer (1991) also note the challenge o f finding a path given perceived geographic constraints. All o f these factors complicate the career planning o f academic women. When they are combined with a narrow and restrictive ethos in the research university, belief in their abilities as scholars could be seriously undercut by this rigidity. The data on young female assistant professors at Michigan State University confirms that they are struggling. They are spending substantially higher percentages o f their time in teaching and service/extension work and substantially less time in professional 193 development than professors of higher rank or even male professors o f the same rank. These aspiring professionals report less comfort with the climate o f the university, desire more opportunity, want to spend less time in administration/governance activities, and are struggling to carve a niche. The patterns noted above do not portend a positive future for many o f these young women faculty. The research university ethos values a different balance of activities than many o f them have undertaken to date. It is possible that unless they are supported by the university in finding appropriate time and avenues for research, and the nature o f appropriate scholarship at Michigan State is publicly discussed, they may conclude unnecessarily that they do not fit at this university. W here women faculty teach or advise more than the man, academic administrators and mentors should intervene by actively planning career development with their faculty. Roger Baldwin’s work on faculty life cycles and careers(1985, 1990) would be particularly helpful for those providing guidance to faculty. Career development patterns and pitfalls should be discussed directly with women faculty, though male faculty would likely benefit from similar discussions. Although career preparation begins already in childhood, a faculty/career orientation program may still help female faculty as they begin their work in the university. However, they may also need help along the way to recognize when and how they should 194 change the balance o f their activities. Colleges and departments should consistently track the type o f w ork activities required o f both genders to insure that they are not subconsciously assigning more teaching/advising responsibilities to female faculty on a continuing basis. Furtherm ore, administrators and mentors should also point out research opportunities, and consider creating internal support for worthy research projects when external support is less available. Because cumulative disadvantages may cause some women to lack the disciplinary connections which result in research funding and personal confidence in their research abilities, assistance from colleagues and administrators may be crucial. Colleges should encourage women to carve out professional development time and should support them in those efforts. Given the scarcity o f such time, even small incentives and opportunities provided by colleagues and deans may be gratefully received. Strong encouragement of professional development time sends the right signals regarding the future. It allows women to believe that the university is investing in their capacities as professionals for the long run. For female faculty in colleges where they are very few in number, it is critical that they nurture networks and mentors among both males and females inside and outside their colleges. It would be all too easy for such women to be seen as tokens among their male colleagues and to encounter all of the career advancement barriers that are associated with tokenism. It is also quite possible that such women would be perceived as "queen 195 bees" by women in other colleges. (Kanter, 1977). Only if female faculty actively seek out males who are supportive within their schools and females who can provide friendship and guidance outside their colleges will they prosper. Such networking and support will provide helpful insights into the college and university ethos as well as suggestions for career advancement in an environment which may seem somewhat foreign. The university could aid such network formation by arranging connections particularly outside one’s college. Helping female faculty find each other would not be difficult. Electronic bulletin boards, newsletters, and occasional inter-college events could help provide the social connecting points needed to begin such networks. W hile many women faculty will undertake such projects themselves, those who are less experienced and buried under heavier teaching assignments will need encouragement and assistance from the institution. Perhaps there is little that universities can do to alleviate the perception o f greater geographical constraint that its female faculty sense. However, that reality, should encouraged even more attention to the quality o f the work experience itself. If women are indeed more constrained in mobility, there is every reason for the university to invest deeply in their professional futures. development could be deep and lasting. The contributions which come from such The professional satisfactions bred by such 196 career development planning could also foster commitments to the university which make constraints on geographic mobility much less central to professional opportunity. CONTRIBUTIONS TO TH E LITERATURE ON W OM EN IN HIGHER EDUCATION The discussion in chapter two noted that there are few contributions to knowledge about women faculty in higher education from the discipline o f organizational behavior analyzing the intersection o f individuals, particularly women, with discrete organizations. Although such work has been done with business organizations (Kanter, 1977; M orrison, 1987), to date little research has been done in higher education organizations from a similar perspective. Furthermore, with the exception o f Olsen’s (1991) study, there appears to be little developed literature about women faculty in the research university context regarding their sources o f satisfaction, use o f time, and geographic mobility. The intent o f this study o f Michigan State University was to fill a portion o f that gap. By looking at one institution and considering how faculty relate to its institutional ethos and policies, issues o f gender difference among faculty in a research university can be analyzed at an organizational level. This study documents that some differences among faculty exist because o f gender. The work-related preferences o f female faculty and the balance o f academic activities which they regarded as important for future rewards did not vary substantially from the male faculty. Yet their current real experience did differ substantially from that o f males, especially for those in the lower academic ranks. Currently women faculty do allocate 197 work time differently but are less satisfied with such arrangements and seek greater restructuring o f their time. This analysis, however, also found that the issue o f gender-related differences is often greatly complicated by the interaction o f gender with age and rank. Many differences which the casual observer might attribute to gender, should more accurately be attributed to differences in one’s age and achieved academic rank. Both o f these factors likely reflect differences in years o f professional experience; and those years o f experience do change the professional satisfactions and the uses o f professional work time. Yet these factors do not cancel out the influence o f gender-based differences entirely. These two themes, not only the continuing direct influence o f gender, but also the interplay o f gender with age and rank, are two documented contributions to the literature which could only be made with this degree o f detail and nuance at a single-site location. Within broader theories explaining the roles o f nature, socialization, and structure on outcomes for women, this study also makes a contribution. The argument that nature inherently differentiates men and women in role and function is not strongly supported by this study. much alike. In many ways the male and female faculty o f this university are very The number of areas in which they report differences in professional satisfaction or differences in professional values are limited. aspirations, even though their real experiences are dissimilar. They have similar Thus, there is little to 198 suggest that the male and female faculty in a research university context are radically different because o f nature. W hile there are some differences in chosen field o f study and perhaps in the definition o f appropriate scholarship, overall these men and women have more similarities than differences. If nature necessarily caused deep divides between men and women such differences would be much more obvious in the results. Socialization is very likely a key part o f the explanation. As Bernard (1964) noted almost thirty years ago, there are key socialization factors which differentiate all people who aspire to academic careers. By the time faculty are employed in a research university, they have already self-selected into an environment with a certain type o f value set and ethos. Their ability to attach themselves comfortably to these values and this culture were very likely shaped by many prior years o f socialization in their families, schools, and communities. Thus in the cases o f both male and female faculty earlier socialization is a very important factor in their career choices. W ithout such socialization neither sex would have aspired to faculty status at a major research university. By the time faculty members have been invited into and elected to join a major research university, their orientations to the balance o f teaching, research, and service may already be somewhat similar. In this context, Finkelstein’s (1984) argument that women have an inherently stronger orientation to teaching does not hold up. The survey shows that while women may teach more they are still committed to a balance o f teaching and research similar to their male colleagues. 199 Despite these similarities, socialization does contribute to differences in the balance o f men and women faculty in several fields o f study. At Michigan State University, like many other colleges and even research universities, the women faculty are still concentrated in a limited number of fields. Areas such as education, human ecology and nursing continue to have disproportionate numbers o f female faculty, while other areas such engineering and business lag far behind. Sex segregation in field o f study is intertwined with occupational sex segregation in the work force. W hile there are glimmers o f change, the messages society sends to its students about appropriate fields o f study and work related to one’s sex are still informed by traditional gender roles. The set o f academic fields chosen by most women has not yet changed substantially. Yet in those fields chosen some major changes have occurred within the context o f the research university. For academic women managing the intersection o f career and personal life is not the same as it was three decades ago. In 1964 Bernard noted that married women academics saw marriage as a hindrance to their careers and children as a liability when related to career advancement. While that may have been the reality o f the 1960’s, the reality o f the 1990’s is different. In general marital status and the presence or absence of children did not have a significant effect on many professional satisfactions outside the realms o f job security and compensation. These factors resulted in almost no significant differences between male and female faculty in the allocation of work time. There was a limited effect on the likelihood o f staying at Michigan State University, but even then the effect was not clear and obvious. Only for those o f either 200 sex who were not in an adult partnership but did have children were there professional struggles. However, males and females in such circumstances were more sim ilar to each other than different. From these results it is clear that the way American society has socialized the women who now populate the faculty o f Michigan State University is quite different from the socialization o f their predecessors. Women seem to have transferred more o f their aspirations to the world o f professional academic work. At the same time, marriage and children are no longer seen as the professional liability that they were thirty to forty years ago. W hile the set o f academic fields considered appropriate for women is just beginning to change, the socialization of women academics regarding the relationship o f marriage and family to career has changed significantly in the last three decades. That difference has altered w om en’s professional goals and their capacity to sustain success in both personal and professional life in the research university context. W hile some changes in socialization have supported women academics in their aspirations within research universities, the organizational structures and cultures o f the research universities have taken only the first steps to open themselves to women academics. At Michigan State University, like many other major research universities, it appears that some progress has been made on achieving gender equity in compensation. W hile some 201 inequities may still exist, this is no longer a source o f major dissatisfaction for most female academics. Yet the issue o f salary equity is only one issue in a much larger field o f concerns. Universities must address the current distribution o f males and females by academic field. W ithout the leadership of the universities in providing a better balance o f males and females in the disciplines, there is less hope that occupational sex segregation will change substantially. University hiring policies and support structures for the few women in atypical fields may need adjustment. Universities should also look closely at their mechanisms for assigning work responsibilities. It would appear that particularly at the assistant professor level something is structurally amiss when female assistant professors differ so strongly from males in the same rank in terms o f the percentage of time spent in teaching. Just because more women are populating research universities does not guarantee that their experience is similar to that o f male faculty. In addition the universities, if they are like Michigan State, must acknowledge that women faculty seek greater opportunities for growth and development and particularly seek time and opportunities to pursue research even more avidly than their male counterparts. Universities must find institutional mechanisms to encourage and support 202 these aspirations. Faculty evaluation strategies should also be reviewed to insure that the criteria used to evaluate appropriate scholarship are not excessively narrow. Research universities must also come to grips with the issue o f geographic mobility. Women academics still do feel more constrained in terms o f geography. With some changes in personnel policy research universities could be open to a broader pool o f qualified women who are working with geographic constraints. This pool could include women currently employed in other college and university settings that are less research oriented. Current placement for women may not be symbolic o f their aspirations. Hiring policies about the hiring of spouses, one’s own graduates, and those serving in non-tenure track positions should be reconsidered to enlarge the pool o f potential female applicants. Compensation strategies based on labor market demand must be sensitive to that fact that women may be less able to use moving to another institution as a bargaining chip in salary negotiations. W hen compensation systems are sensitive to market demand, universities must remain vigilant about the emergence o f new compensation inequities because o f constraints on mobility. On balance simply changing the structural aspects o f major research universities will not resolve all the challenges for women in their faculty ranks. Changes in the socialization of young women regarding appropriate fields o f study and the combinations o f career and personal life are necessary prior steps which prepare women for faculty positions in such institutions. Yet the likelihood that they will be recruited in greater numbers, successful 203 in their professional work, and in concert with the ethos o f the research university can be enhanced by such structural changes. FUTURE RESEARCH OPPORTUNITIES W hile this study adds to the base o f knowledge about female faculty in major research universities, there are many other avenues o f research and analysis that remain fertile ground for work. Since the survey at Michigan State University did not clearly specify academic field, no work could be done to compare levels o f professional satisfactions and the allocation o f work time by academic discipline. Feldm an’s (1974) study o f gender and the academic disciplines provides an excellent foundation for continuing work. Such research could yield significant new knowledge about differences and similarities between male and female faculty. It may be that those who choose a particular academic discipline share more similarities than differences regarding their use o f professional time and their relative orientations toward teaching, research, and service. W ithout further research few documented claims can be made about such issues. It may also be useful to compare the results from this study o f Michigan State University with other major research universities, land grant institutions, and comprehensive schools. Such comparisons could highlight findings from this study that are not more broadly shared with other universities and could make a further contribution to the literature about unique higher education cultures in relationships with their faculty. From the standpoint o f organizational theory, higher education is a set o f organizations about 204 which little is known. research universities. Within that grouping even less is known about the subset of W hile higher education has similarities to other corporate forms o f organization, the particular history, mission in education, special organizational culture, and not-for-profit status o f higher education research institutions warrants the study o f their faculty as a unique subset. CONCLUSION This study o f Michigan State University’s faculty has developed a nuanced understanding o f the relative effect o f gender on the experience, activities, and aspirations o f its faculty. Female and male faculty professionals work together to achieve the university’s mission. Understanding the dynamics o f their work enables this university and other research universities to function more effectively as agents o f knowledge and education in a learning society. APPENDIX 205 APPENDIX A it—Biiaai. FACULTY MOBILITY SURVEY T h an k you fo r taking th e tim e to com plete the Faculty M obility Study. T h is study explores faculty c are er choices in a dram atically changing environm ent. T he inform ation gained from this effort will b e u se d in u nderstanding th e c a re e r challenges facing faculty. Yotfr p articip atio n in this study is voluntary. Y ou may choose n ot to p a rticip a te a t all o r term inate-youx involvem ent a t any tim e. Y ou have th e right to refuse to answ er any question. H ow ever, w e w ould ap p rec ia te it if you co u ld answ er all q uestions in o rd e r to m inim ize th e am ount o f m issing inform ation th a t m akes it difficult to analyze T h e e n tire survey will tak e a b o u t 30 m inutes to com plete. A ll your responses will b e k e p t strictly confidential. If you h ave any q uestions a bout this project, please contact D r. K athryn M o o re a t 355-2395, D r. Philip G a rd n er at 355-2211 o r D r. L in d a F o rre st a t 355-8502. Please re tu rn your survey by M a rch 25,1991 to: Collegiate Employment Research Institute 113 Student Services Building Michigan State University East Lansing, MI 48824 T he re tu rn o f the com pleted survey constitutes your inform ed and voluntary c onsent to p a rticip a te in this research. 206 PART I. Q u e stio n s in th is section c oncern y o u r academ ic a p p o in tm en t a n d th e g eneral level o f jo b satisfactio n you experience In y o u r c u rre n t position. 1. W h at is your c u rre n t academ ic ra n k at M ichigan S tate U niversity? (P L E A S E C IR C L E O N E N U M B E R ) P rofessor 1 A ssociate P rofessor 2 A ssistant P rofessor 3 In stru c to r 4 S pecialist 5 >6 O th er;___________________________________ (Please Specify) 2. In w hat year d id you achieve your c u rre n t rank? 19______ 3. In w hat year d id you begin your em ploym ent as a faculty m em ber at M ichigan S tate U niversity? 19______ 4. W hat is your c u rre n t ten u re status at M ichigan State U niversity? (P L E A S E C IR C L E O N E N U M B E R ) N ot in ten u re system 1 In ten u re system but not te n u re d 2 T e n u re d 3 In w hat y ear d id you achieve ten u re at M ichigan S ta te U niversity? 19_____ 5. In w hich college o r u nit is your prim ary ap pointm ent? (P L E A S E C H E C K O N E ) a. A griculture a n d N atu ral R eso u rces i. Jam es M adison b. A rts a n d L etters j. N atural Science c. B usiness k. N ursing d. C om m unication A rts L O ste o p a th ic M edicine e. E d u ca tio n _____ m. Social Science _f. E ngineering n. U rb a n A ffairs _g. H u m an E cology o. V eterinary M edicine _h. H u m a n M edicine p . N on-C ollege Faculty q. O th er:______________ 6. D o you currently hold a jo in t appointm ent? Yes No If yes, in w hat o th e r departm ent/school/ non-departm entally organized college d o you h old an a p p o in tm e n t(s)? __________________________________ 7- A t how m any o th e r institutions have you held academ ic appointm ent at the level of assistant professor or above? Institutions In w hat y ear did you hold your first academ ic a ppointm ent? year 207 H ow satisfied o r dissa tisfie d d o you personally feel about e ac h o f th e following aspects o f vour job at M ichigan State U niversity? (P L E A S E C IR C L E O N E N U M B E R F O R EACH IT E M ) V ery Dissatisfied Som ew hat D issatisfied N eutral M y w ork load 2 3 My jo b security 2 3 T he authority I have to m ake decisions a bout w hat courses I teach 2* 3 T he authority I have to m ake decisions about content and m eth o d s in the courses I teach 2 3 T he authority I have to m ake decisions about o ther (noninstructional) aspects of my job 2 3 Tim e available to w ork on scholarship and research 2 3 T he m ix o f teaching, research , adm inistration, and service (as applicable) th at I am re q u ire d to do 2 3 O p portunity for m y advancem ent in rank at M ichigan State U niversity 2 3 Som ew hat V ery Not Satisfied Satisfied A pplicable T im e available for w orking w ith students as an advisor, m entor, etc. 3 4 Availability o f su p p o rt services (including clerical sup p o rt) 3 4 3 4 Availability o f equipm ent (personal com puters, etc.) 2 F reedom to d o outside consulting 2 My salary 2 3 4 My benefits, generally 2 3 4 O verall re p u ta tio n o f M ichigan S tate University 2 3 4 In stitu tio n al m ission to carry out teaching, research, and public service 2 3 4 Q uality of lead ersh ip in my d e p artm ent/program 2 3 4 Q uality of chief adm inistrative officers at M ichigan S tate U niversity 2 3 4 Q uality of my colleagues in my d epartm en t/p ro g ram 2 3 4 Q uality of grad u ate students w hom I have taught here 2 3 4 4 208 8. C o n tin u e d Very D issatisfied Som ew hat Dissatisfied N eutral Som ewhat Satisfied Quality of u n d e rg ra d u a te students whom I have taught here 2 3 4 T eaching assistance th a t I receive 2 3 4 R esearch assistance th at I receive 2 3 4 N ot Very Satisfied A pplicable O pportunities for p rofessional grow th and developm ent offered by my academ ic unit C ooperation offered by su p p o rt staff at M ichigan S ta te U niversity Quality of faculty leadership (e.g., A cadem ic Senate) at M ichigan S tate University R elationship b etw een adm inistration and faculty at M ichigan S tate U niversity Interd ep artm en tal c o o p eratio n at M ichigan State University Spirit of co operation am ong faculty at M ichigan State University Quality of my re sea rc h facilities a n d su p p o rt My jo b here, overall 9. Please estim ate the percentage o f your total w orking h ours th at you sp en t on each of the following activities during the 1990 Fall T erm . (P L E A S E G IV E Y O U R B E ST E S T IM A T E S IF N O T S U R E : IF N O N E , E N T E R "0") Mote: The percentages you provide should sum to 100% o f the total time you spent on professional activities. Percent T eaching (p rep a rin g courses; developing new curricula; teaching; grading p apers.) ________ R e se arc h a n d S cholarship (planning for a n d conducting research; prep arin g for and giving perform ances a n d exhibitions in th e fine arts; prep arin g o r reviewing articles or books; preparing for a n d a ttending professional m eetings o r conferences; seeking outside funding, including p ro p o sal writing.) ________ Advising Studentr. (advising underg rad u ate and g rad u ate students; w orking w ith student organizations.) ________ Professional D evelopm ent (taking courses; pursuing an advanced degree o r participating in o th er p ractices to rem ain cu rren t in your discipline.) ________ Service and Extension (preparing a n d giving speeches that build upon your professional expertise; providing o f technical assistance, policy analysis, program evaluation, m edical o r veterinary services, psychological counseling a n d therapy; consulting outside with o r w ithout rem uneration.) _______ A dm inistration and G overance (participating in faculty goverance; participating in dep artm en tal or institutional com m ittees a n d task forces; m anaging a n d coordinating p rogram s o r personnel.) ________ O th e r (P L E A S E SP E C IF Y ):_______________________________________________________ Please be sure that your percentages total: 100% 209 10. Indicate how satisfied you are w ith these facets o f your life a t this tim e. (P L E A S E C IR C L E O N E N U M B E R F O R E A C H IT E M ) V ery D issatisfied Som ew hat D issatisfied N eutral Som ew hat Satisfied V ery Satisfied 5 8 Not A pplicable Life in general, outside o f w ork 1 2 3 4 H ealthful lifestyle 1 2 3 4 5 8 Fam ily life 1 2 3 4 5 8 A m ount o f tim e for leisure activities 1 2 3 4 5 8 G en eral level o f happiness 1 2 3 4 5 8 Level o f physical activity 1 2 3 4 5 8 D egree o f physical fitness 1 2 3 4 5 8 G eographical a re a w here you live 1 2 3 4 5 8 C lim ate w here you live 1 2 3 4 5 8 A bility to cope w ith stress 1 2 3 4 5 8 Social life 1 2 3 4 5 8 O verall health status 1 2 3 4 5 8 PART II. In th is section, we a sk you to c o n sid er the likelihood of leaving y o u r c u rre n t p o sitio n to do som ething else. la . If you h a d the opportunity to re stru ctu re your c u rre n t position, w ould you w ant to d o m ore, less, o r about the sam e am ount of each o f the following? (P L E A S E C IR C L E O N E N U M B E R F O R E A C H IT E M ) M uch Less T eaching 1 Som ew hat Less Sam e A m ount A s I N ow D o 2 3 Som ew hat M o re M uch M ore 4 5 R esearch a n d Scholarship 1 2 3 4 5 A dvising S tudents 1 2 3 4 5 Professional D evelopm ent 1 2 3 4 5 Service/Extension 1 2 3 4 5 A dm inistration a n d G overnance 1 2 3 4 5 If you w ere to leave this jo b to accept a nother position, w ould you w ant to d o m ore, less, o r about the san of each o f the following as you currently do? (P L E A S E C IR C L E O N E N U M B E R F O R E A C H IT E M ) M uch Less Som ew hat Less Sam e A m ount As I Now D o Som ew hat M ore M uch M ore T eaching 1 2 3 4 5 R esearch a n d Scholarship 1 2 3 4 5 Advising Students 1 2 3 4 5 Professional D evelopm ent 1 2 3 4 5 Service/Extension 1 2 3 4 5 A dm inistration and G overnance 1 2 3 4 5 210 2. G iven your situ atio n a t M ichigan S ta te U niversity a n d the jo b m arket in your field, how likely are you to take these actions w ithin th e n e st two years: Very Unlikely a. S eek a new p o sitio n at M ichigan S ta te U niversity Som ew hat Unlikely N eutral Som ew hat Likely 3 b. L ook for a position a t a n o th er institution c. R esign m y c u rre n t position for a sim ilar position a t a n o th e r institution d. R esign m y p o sitio n to retire e. R esign m y p osition to re tu rn to school as a s tu d e n t f. R esign m y position for o th er reasons including c a re e r change, child rearing, providing d e p e n d e n t c are , etc. g. A c ce p t em ploym ent at a(n): * d o c to ra l gran tin g university o r college 4 * o th er 4-year university o r college 4 * 2-year p o stseco n d ary institution 4 * elem entary o r seco n d ary school 4 * hospital o r o th e r health care organization 4 ’ consulting, self-ow ned business, freelancing 4 •p riv ate se c to r for-profit business o r industry 4 * foundation o r o th e r nonprofit organization 4 * federal governm ent (including m ilitary) 4 * state o r local governm ent 4 Very Likely 211 3. Faculty consider m any factors w hen w eighing an o pportunity to leave an institution like M ichigan State University. L isted b elow a re factors th a t you m ay c ontem plate in d eciding to leave th e university. In d icate the relative degree of im p o rtan ce e a c h fa cto r could have in m aking your decision. N ot A n Im p o rtan t R eason A t A ll T o L eave Som ew hat Im p o rta n t R e aso n T o L eave Fairly Im p o rtan t R easo n T o L eave V ery Im p o rtan t R eason T o L eave Extrem ely Im portant R eason T o Leave R e p u ta tio n o f institution 2 4 5 Service L o ad 2 4 5 Availability o f in tern al research funds 2 4 5 Congeniality of colleagues 2 4 5 Jo b Security/tenure 2 4 5 R a p p o rt w ith d ep artm en tal leadership 2 4 5 P rom otion in rank 2 4 5 C a ree r advancem ent opportunities 2 4 5 R e p u ta tio n o f associates 2 4 5 Base salary 2 4 5 R esearch load 2 4 5 B enefit package 2 4 5 A dm inistrative lo ad 2 4 5 R esearch o p p ortunities 2 4 5 T eaching load 2 4 5 T eaching assignm ents a nd/or opportunities 2 4 5 R a p p o rt w ith university leadership 2 4 5 Availability o f in te rn al research funds 2 4 5 R ep u tatio n o f d e p artm e n t 2 4 5 Institutional m ission/philosophy 2 4 5 Influence in d e p artm en t 2 4 5 C om petence of colleagues 2 4 5 S ecretarial su p p o rt 2 4 5 R eceipt of m erit pay 2 4 5 Influence in college 2 4 5 212 3. ( re a s o n s to le a v e c o n tin u e d ) N ot A n Som ew hat Fairly Im p o rtan t Im p o rtan t Im p o rtan t Im p o rta n t Im portant R eason A t A ll R eason T o R eason T o R easo n T o R eason T o L eave L eave T o L eave L eave V ery L ibrary facilities 2 4 L aboratory/research facilities 2 4 Office facilities 1 3 R ed u ced tuition for family 2 4 R a p p o rt w ith college leadership 2 4 E m phasis o n publishing 2 4 Sabbatical, leave, travel, and study policies 2 4 Consulting o pportunities 2 4 Spouse’s c are er o pportunities 2 4 G eographic considerations 2 4 Cultural, recreational, and social opportunities 2 4 Clim ate of region 2 4 H ousing costs 2 4 Proximity o f extended family 2 4 Extensive a nd/or close netw ork of friends living locally 2 4 Loyalty to institution 2 4 Loyalty to departm ent/program 2 4 A ppreciation for my work 2 4 Influence in institution 2 4 Extrem ely Leave 213 4. A re you seriously considering o r actively seeking a jo b change? (P L E A S E C IR C L E O N E ) Y es 5. No M aybe H ave you received a n actu al jo b offer(s) in w riting from a n o th e r institution o r organization in the p eriod betw een S e p tem b e r L 1989 a n d M arch L 1991? (P L E A S E C IR C L E O N E N U M B E R ) 6. 7. 8. Yes, I have received one offer 1 Y es, I have received m ore th an o ne offer 2 N o, I have n o t received any offers 3 (Please go to P a rt III, Q uestion 1) W ho initiated the con tact th at re su lte d in this offer? (P L E A S E C IR C L E O N E N U M B E R ) I m ad e the first contact 1 T he o th er institution m ade the first contact 2 W ith w hom d id you discuss the jo b offer(s) you received? (P L E A S E C IR C L E A L L T H A T A P PL Y ) C olleagues(s) in y our departm ent/unit 1 C olleagues(s) outside of your d ep artm en t/u n it 2 T h e chairperson/adm inistrator of your d e p artm e n t/u n it 3 T h e chairperson/adm inistrator o f a n o th e r dep artm en t/u n it 4 T h e d e a n o r o th e r senior adm inistrator in your college 5 T h e provost o r o th er senior adm inistrators in the university 6 O th ers n o t associated w ith M ichigan S tate U niversity 7 H ow m u ch (fid y our d e p artm e n t chairperson/unit adm inistrator influence the decision on your m ost recent offer? (P L E A S E C IR C L E O N E N U M B E R ) Strongly Influenced M e T o Leave 1 9. Slightly Influenced M e T o L eave H ad No Im pact O n My D ecision Slightly Influenced M e T o Stay 3 4 2 Strongly Influenced M e T o Stay 5 A t this tim e, have you a cc ep te d a job offer from another organization o r in stitution?(P L E A S E C IR C L E O N E N U M B E R ) Yes, a cc ep te d a n o ffer 1 A m still considering offer(s) 2 No, rejected o ffer(s) 3 PART III. T hese q u estio n s deal w ith yo u r p ercep tio n s o f sa lary levels a n d benefits received a t M ichigan S ta te U niversity as c o m p a red to o th e r in stitu tio n s you a re fa m ilia r with. 1. D o you believe your cu rre n t salary, w hen co m pared with the salaries o f p eers in your field at M ichigan State U niversity, to be: (P L E A S E C IR C L E O N E N U M B E R ) M uch low er th an the average 1 Som ew hat low er than average 2 A bout average 3 Som ew hat higher th an average 4 M uch higher than the average 5 214 2. D o you believe y our c u rre n t salary, w hen c o m p a red nationally w ith the salaries o f p e ers in your field, to be: (P L E A S E C IR C L E O N E N U M B E R ) M uch low er th a n th e average 1 Som ew hat low er th an average 2 A b o u t average 3 Som ew hat h igher th an average 4 M uch h igher th an th e average S 3. H ave you h a d y our salary ad ju sted fo r m ark et conditions d u rin g th e m iddle o f th e year? Y es 4.' W hat p e rce n ta g e o f your salary w o u ld a n o th er institution have to offer for you to consider leaving M ichigan State U n iv e rsity ? ________ P ercent 5. Is it a p ractice in your d e p artm en t to solicit a jo b offer from a n o th e r institution for the purpose of: 6. 7. a. enhancing salary Y es______ N o ______ b. receiving a prom otion Y es______ No_____ c. enhancing su p p o rt Y es______ N o______ N o_____ H ow likely is it th a t you could o btain a position at a n o th er institution that is as good o r b e tte r th an your present position a t M ichigan S ta te U niversity? (P L E A S E C IR C L E O N E N U M B E R ) V ery unlikely 1 U nlikely 2 Likely 3 V ery likely 4 C onsidering ail the factors th at can influence your em ploym ent, how interested are you in leaving o r rem aining at M ichigan S tate University? (P L E A S E C IR C L E O N E N U M B E R ) V ery In tere sted In Leaving F o r A n o th e r Position Som ew hat In tere sted In L eaving F o r A n o th e r Position A bout E qually In tere sted in Leaving A n d Staying Som ew hat In tere sted R em aining P resent Position In In Very In terested in Rem aining In Present Position If you are m arried o r in a com m itted relationship, please com plete PART IV; otherw ise please continue with PARTV. PART IV. T h is section exam ines c a re e r com m itm ent a n d em ploym ent o p p o rtu n ities a n d c o n stra in ts faced in d u a l c a re e r situ atio n s. 1. W hat is your p a rtn e r’s last com pleted degree? (P L E A S E C IR C L E O N E ) H igh School M A /M S Ph.D . Professional (M D , RN , T.I.D, D V M ) 2. Is y our p a rtn e r presently em ployed outside the hom e? Yes N o_____ 3. Is y our p a rtn e r em ployed by M ichigan S tate University? Yes N o_____ O th er If your p a rtn e r does n o t w ork at M ichigan S tate University, how m any m iles apart a re your jobs? miles. 215 4. A re yon currently living w ith your partn er? Yes N o _____ If you currently live together, w ould you b e willing to consider living a p a rt to g et th e jobs you b o th d e s ire ? yes no If you a re currently living ap art, how far a p art do you liv e ? 5. W h at is your p a rtn e r’s c u rre n t occupation? m iles ______________________________________ H ow m any years o f c are er experience does your p a rtn e r have?____ years 6. W h at is your p a rtn e r’s c u rre n t jo b title? ______ __ _________ _____________________________ H ow long has your p a rtn e r b e en in his/her cu rren t position?_______years? 7. H ow d o you com pare your stage of c are er developm ent w ith th at o f your p a rtn e r’s career? M y C a re e r is Substantially A h e ad 1 M y C a ree r is Som ew hat A h e ad 2 B oth C areers at Sam e Stages 3 P a rtn er’s C a re e r is Som ew hat A h e ad 4 P a rtn er’s C areer is Substantially A head 5 8. In th e w orld o f w ork th ere are differences in a care er’s status b ased on a n um ber o f variables, such as pow er, social prestige, salary, etc. H ow w ould you evaluate the status of b o th your c are er and your p a rtn e r’s c are er? (P L E A S E C IR C L E O N E .) My Career. V ery L ow Status Partner's Career. Very L ow Status Fairly L ow Status M edium Status Fairly H igh Status V ery H igh Status Fairly L ow Status M edium Status Fairly H igh Status V ery H igh Status 9. C ouples generally m ake decisions about w hose c are e r will take priority. W hich statem en t below b est describes the c are er priority in your relationship? (P L E A S E C IR C L E O N E ) a. P a rtn e r’s c a re e r is th e sole im portant care er in the relationship. b. B oth careers a re im portant but p artn er’s c are er is primary. c. B oth c are ers a re co n sid ered equal. d . B oth care ers a re im portant b ut m y care er is prim ary. e. M y c a re e r is the sole im p o rtan t c are er in th e relationship. 10. O n the following scale, w hat priority do you give to your c are er a n d your relationship/fam ily? (P L E A S E C IR C L E O N E ) Fam ily R elationship T o p P riority 1 2 E qual Priority 3 4 C a ree r T op Priority 5 11. W hich priority d o you believe your p a rtn e r w ould give to his/her c are e r a n d your relationship/ family? (P L E A S E C IR C L E O N E ) Family R elationship T o p Priority 1 2 E qual Priority 3 4 C a ree r T o p Priority 5 216 12. C onsidering all the factors th a t can influence y our p a rtn e r’s em ploym ent, how in terested is your p a rtn e r in leaving o r rem aining in h is/her p re sen t p osition? (P L E A S E C IR C L E O N E ) V ery In tere sted in L eaving for A n o th e r Position 1 Som ew hat In te re ste d in L eaving for A n o th e r P osition 2 A bout E qually In te re ste d in L eaving a n d R em aining 3 Som ew hat In tere sted R em aining in P re se n t Position Very in In tere sted in R em aining in P resent Position 4 5 13. T o w hat extent d o you tak e into account y our c hildren a n d their interests in m aking decisions about your job? L ittle o r no extent 1 M o d erate extent 3 2 G re a t E xtent 5 4 N ot A pplicable 9 14. If you and your p a rtn e r a rc em ployed full-tim e (o r in school) and one of you h a d to stay hom e (for exam ple, to care for a sick child or w ait for a re p air person), who is m ore likely to rem ain at hom e? D efinitely M e U sually M e 1 E qually Likely 2 3 U sually P a rtn e r 4 D efinitely P a rtn er 5 15. D o eith er of your c are ers restrict the locations w here em ploym ent m ay be available? (P L E A S E C IR C L E O N E ) M ine P a rtn e r B oth N either 16. In m aking a final decision ab o u t leaving o r staying, how free (based o n your individual desires) o r rn n strain p .d (b ased o n job, family o r relationship factors th at you m ay n o t be able to control) d o you believe your d ecision w ould be? T otally F re e Fairly F re e 1 2 Fairly C onstrained 3 4 T otally C onstrained 5 17 a. E stim ate the n um ber o f jo b openings in your discipline this year th at w ould be ap p ro p ria te for you based o n your level a n d experience? num ber d o n ’t know b. E stim ate the n u m b er o f jo b openings your p a rtn e r w ould find this y ear that w ould be ap propriate for him o r h e r b a se d o n level a n d experience? num ber don’t know 18. A t this tim e, how m obile d o you consider you a n d your p a rtn e r to be? N either of U s M obile 1 P a rtn e r M o re M obile T han I A m 2 I Am M o re M obile T h an P a rtn e r 3 B oth E qually M obile 4 19. If you a n d your p a rtn e r w ere to begin a jo b search, w hat strategy w ould you likely use? (P L E A S E C IR C L E O N E ) a. I w ould look first a n d receive a jo b offer(s), th en my p a rtn e r w ould look. b. P a rtn er w ould look first a n d receive a jo b offer(s), th en I w ould look. c. W e w ould b o th look independently at the sam e tim e. d. A pply to jobs as a couple C ould you briefly d etail the reasoning b ehind your p re fe rre d strategy._______________________ 217 PART V. A number of issues are of concern to the faculty at Michigan State University. Please indicate the extent to which you agree or disagree with each of the following statements. (P L E A S E C IR C L E O N E N U M B E R F O R E A C H S T A T E M E N T ) Strongly D isagree Som ew hat D isagree N either A gree O r D isag ree A. General Issues: It is im portant for faculty to participate in governing th e ir institutions. 3 Faculty p rom otions should b e b a se d a t least in p a rt o n form al evaluations by students. 3 Teaching effectiveness should be the prim ary c riterio n for prom otion and tenure o f faculty. 3 Service/Extension should be an equivalent criterio n w ith teaching and/or research for prom otion a n d tenure o f faculty. 3 R esearch/publications should b e the prim ary c riterio n for p rom otion and tenure of faculty. 3 Faculty should be free to p re sen t in class any id e a they consider relevant. 3 Private consulting in a reas directly re la te d to a faculty m em b er's field or research o r teaching should b e restricted. B. In stitu tio n al Issues: T he adm inistrative function is taking an increasingly heavy share of available resources. T he university's landgrant m ission is em phasized in m y academ ic unit’s overall objectives. T he university's landgrant m ission receives a p p ro p ria te em phasis in overall university objectives. Service/Extension should carry m ore weight in p rom otion a n d ten u re decisions 3 R esearch should b e rew arded m o re th an teaching. 3 R esearch should be rew arded m ore th an p ublic service 3 Fem ale faculty m em bers a re tre a te d fairly. 3 Faculty who a re m em bers o f racial o r ethnic m inorities a re tre a te d fairly. 3 Som ew hat Strongly A gree A gree 218 C. H ow im p o rtan t d o you think th e following should b e in determ ining faculty rew ards: N ot V ery Im portant Som ewhat Im portant Fairly Im portant V ery Im portant Extrem ely Im portant 1. T en u re T eaching 2 4 5 R esearch/S cholarship 2 4 5 A dvising 2 4 5 Service/E xtension 2 4 5 A dm inistration/G overance 2 4 5 3 4 5 2. P rom otion in R an k T eaching 2 R esearch/S cholarship 2 3 4 5 Advising 2 3 4 5 Service/E xtension 2 3 4 5 A dm inistration/G overance 2 3 4 5 T eaching 4 5 R esearch/S cholarship 4 5 Advising 4 5 Service/E xtension 4 5 A dm inistration/G overance 4 5 3. M erit Increases A s you look to w ard 1995, d o you perceive M ichigan State University will be: (P L E A S E C IR C L E O N E N U M B E R ) M uch W orse O ff Som ew hat W orse O ff A bout T he Sam e A s T oday Som ewhat B etter O ff 1 PART VI. Demographic Information 1. In w hat y ear w ere you b om ? ____1 9 __ 2. W hat is your sex? M ale_ Fem ale 3. W hich best d escribes you? (P L E A S E C IR C L E O N E N U M B E R ) (P L E A S E C H E C K O N E ) A frican A m erican 1 M exican-A m erican/C hicano A m erican Indian 2 Foreign N ational A sian-A m erican o r Pacific Islander 3 Caucasian/W hite H ispanic-A m erican 4 O th e r (Please Specify)______ M uch B etter O ff W h at is y our c u rre n t m arital status? (P L E A S E C IR C L E O N E N U M B E R ) Single, never m arrie d 1 D ivorced 4 M arried /C o h ab itatin g 2 W idow ed 5 S e p a rated 3 If you have children, H ow m any d o you have?_____________ W hat a re th eir ages?____________ A re you a university distinguished p rofessor o r d o you h old a n endow ed chair? No 1 Y es 2 In w hat year did you receive this appointm ent? 19__ W hich o f the follow ing have you received (P L E A S E C IR C L E A L L T H A T A P P L Y ) A university-level excellence aw ard a t M ichigan S tate U niversity 1 A college-level excellence aw ard a t M ichigan S tate U niversity 2 A departm ent-level excellence aw ard a t M ichigan S tate U niversity 3 A sim ilar excellence aw ard at a n o th er institution 4 D o you currently have an adm inistrative assignm ent (p ro g ram c o ordinator, d e p artm e n t chair, assistant or associate chair, e tc ? Yes N o_____ W e have trie d to be com prehensive in addressing em ploym ent issues in this survey. H ow ever, we may not have a d d ressed all the factors p e rtin en t to your d ecision to rem ain a t M ichigan S tate University. W e invite you to use use the following space to elaborate o n those issues that a re m ost p ressing concerning your c are er and the environm ent in w hich you work. Thank you fo r participating in this survey. I f you desire an executive summary o f the survey’s results, please m ail a card separately from the survey to the Collegiate Employment Research Institute. We invite your comments concerning this survey. 220 FACULTY MOBILITY STUDY Thank you for completing the Faculty Mobility Study. There are several issues that we were not able to cover in great detail. We are interested in learning more about the decision process involved in accepting and rejecting job offers extended from other institutions and the job market obstacles faced by dual career couples. If you have received a job offer within the last two years or if you are involved in a dual career relationship, we invite you to participate in these follow-up studies by completing the information below. Yes, I would like to participate in the decision study on job offers: Name: _____ ________________ _________ Address:_____________________________ _______ Campus Phone: _____ ____________________ Yes, I would like to participate in the dual career follow-up study: Name: ________________________________ Address:_______________________________ ______ Campus Phone: ____________ ________________ Please return this form in a separate campus mail envelope. If you include it with your survey, one of the principal investigators will separate it from your survey upon its receipt. Thanks again. Please return to: Collegiate Employment Research Institute 113 Student Services Building Michigan State University East Lansing, MI 48824 BIBLIOGRAPHY BIBLIOGRAPHY AAUP, "The Annual Report on the Economic Status o f the Profession 1988-89," Academ e. W ashington D .C .: American Association o f University Professors, M arch-A pril, 1989. Acker, J. "Hierarchies, Jobs, Bodies: A Theory o f Gender Construction" In J. Lorber and S. Farrell (Eds.) The Social Construction o f G ender. Newbury Park, California: Sage Publications, 1991. Aisenberg, N ., and Harrington, M. Women o f Academe: Outsiders in the Sacred G rove. Amherst, MA: University o f Massachusetts Press, 1988. Astin, H. S. The Woman Doctorate in America. New York: Russell Sage Foundation, 1970. Astin, H. S ., and Leland, C. Women o f Influence. Women o f Vision. A CrossGenerational Study o f Leaders and Social Change. San Francisco, CA: JosseyBass, Inc., 1991. Aslanian, C. Forty Million Americans in Transition. New York: College Board Publications, 1981. Atkinson, P ., and Delamont, S. "Professions and Powerlessness: Female Marginality in the Learned Occupations." The Sociological Review. Vol. 38, #1, Feb., 1990, pp. 90-110. Austin, A ., and Gamson, Z. Academic Workplace: New Demands. Heightened Tensions. ASHE-ERIC Higher Education Research Report #10. Washington, D .C .: George W ashington University, 1983. Baldwin, R. G. "Faculty Vitality beyond the Research University, Extending a Contextual Concept." Journal o f Higher Education. Vol. 61, #2, M arch/April, 1990, pp. 160-180. Baldwin, R. G. Fostering Faculty Vitality: Options for Institutions and A dministrators. W ashington, D .C .: ERIC Clearinghouse on Higher Education, 1982. Baldwin, R. G ., and Krotseng, M. "Incentives in the Academy: Issues and Options" in R. Baldwin (Ed.), Incentives for Faculty Vitality. New Directions for Higher Education #51, Vol. XIII, #3. San Francisco, CA: Jossey-Bass, 1985. 221 222 Becker, W. E., and Lewis, D. R. Academic Rewards in Higher Education. Cambridge, MA: Ballinger Publishing C o., 1979. Belenky, M ., and Clincy, B. W om en’s Ways o f Knowing. New York: Basic Books, 1986. Bernard, J. Academic W om en. University Park, PA: Pennsylvania State University Press, 1964. Bimbaum, R. How Colleges W ork: The Cybernetics o f Academic Organization and Leadership. San Francisco, CA: Jossey-Bass, 1988. Bowen, H. Investment in Learning: The Individual and Social Value o f American Higher Education. San Francisco: Jossey-Bass, 1977. Bowen, H ., and Schuster, J. American Professors: A National Resource Im periled. New York: Oxford University Press, 1986. Boyer, E. The Conditions o f the Professoriate. Attitudes and Trends. Princeton, N. J.: The Carnegie Foundation for the Advancement o f Teaching, 1989. Boyer, E. Scholarship Reconsidered. Princeton, N .J.: The Carnegie Foundation for the Advancement o f Teaching, 1990. Bowles,S., and Gintis, H. Schooling in Capitalist Am erica. New York: Basic Books, 1976. Breneman, D ., and Youn, T. Academic Labor Markets and Careers. Philadelphia, PA: The Falm er Press, 1988. Brown, B. "Gender equity for women professors at research universities". Dissertation Abstracts International. 51, 1925-A. 1990. Caplow, T ., and McGee, R. The Academic M arketplace. New York: Basic Books, 1958. Centra, J.A . Men. Women and the Doctorate. Princeton, N .J.: Educational Testing Service, 1974. Chamberlain, M. W omen in Academe: Progress and Prospects. Sage Foundation, 1988. New York: Russell Clark. B, R. The Academic Life: Small Worlds. Different W orlds. Princeton, N .J.: The Carnegie Foundation for the Advancement o f Teaching, 1987. 223 Clark, S.M ., and Corcoran, M. "Perspectives on the Professional Socialization o f Women Faculty: A Case o f Accumulative Disadvantage?" Journal o f Higher Education. Vol. 57, Jan./F eb., 1986, pp. 20-43. Cole, J. R. Fair Science. Women in the Scientific Community. New York: The Free Press, 1979. Cole, J. R. Fair Science. Women in the Scientific Com munity. M omingside Edition, New York: Columbia University Press, 1987. Cooney, T ., and Uhlenberg, P. "Family Building Patterns o f Professional Women: A Comparison of Lawyers, Physicians, and Postsecondary Teachers" in Journal of Marriage and the Fam ily. 1989, pp. 749-758. Conners, J. V. "Gender differences in perceived advancement problems, stree, interpersonal stress, and satisfaction of university faculty". Dissertation Abstracts International. 51, 2638-A, 1991. Dwyer, M ., Flynn, A ., and Inman, P. "Differential Progress o f Women Faculty: Status 1980-1990." in J. Smart (Ed.) Fligher Education: Handbook o f Theory and Research. New York: Agathon Press, 1991. Eble, K ., and McKeachie, W. Improving Undergraduate Education Through Faculty Development. San Francisco: Jossey-Bass, 1985. Feldman, S.D. Escape from the D oll’s House: Women in Graduate and Professional School Education. Berkeley, CA: The Carnegie Foundation for the Advancement o f Teaching, 1974. Finkelstein, M. J. The American Academic Profession: A Synthesis o f Social Scientific Inquiry since W orld W ar II. Columbus: Ohio State University Press, 1984. Fleetham, D. "Then who will take care o f the home? The experience o f women faculty at Calvin College", Unpublished paper, History 395, Calvin College, 1991. Fox, M. "Women and Higher Education: Gender Differences in the Status o f Students and Scholars." in J. Freeman (Ed.) Women: A Feminist Perspective. Mountain View, CA: Mayfield Publishing, 1989. Fum iss, T. Reshaping Faculty Careers. Education, 1981. Washington, D .C .: American Council on Gilligan, C. In a Different Voice. Cambridge, MA: Harvard University Press, 1982. 224 Howe, F ., ed., W omen and the Power to Change. Berkeley, CA: The Carnegie Foundation for the Advancement o f Teaching, 1975. Jencks, C ., and Riesman, D. Doubleday, 1968. The Academic Revolution. Garden City, New York: Jolly D ., Grimm, J., and W ozniak, P. "Patterns o f Sex Desegregation in Managerial and Professional Specialty Fields, 1950-1980." W ork and Occupation. Vol. 17, # \ , F eb., 1990, pp. 30-54. Kan ter, R. M. Men and Women of the Corporation. New York: Basic Books, 1977. Karabel, J ., and Halsey, A. H. Power and Ideology in Education. New York: Oxford University Press, 1977. Kauffman, D ., and Perry, F. J., "Institutionalized Sexism in Universities: The Case o f Geographically Bound Academic W omen." NWSA Journal. Vol. 1, #4, Summer, 1989, pp. 644-659. Levinson, D .J. The Seasons o f a M an’s Life. New York: Ballantine Books, Random House, Inc., 1978. Levinson, D .J. "A Concept o f Adult Development." American Psychologist. Jan., 1986, pp. 3-13. McKay, N. "Black Woman Professor-W hite University." In R. Klein and D. Steinberg (Eds.) Racial Voices. A Decade o f Feminist Resistance. New York: Pergamon Press, 1989. M oore, K. M ., and Johnson, M. "The Status o f Women and Minorities in the Professoriate: The Role of Affirmative Action and Equity." In G. Lozier and M. Cooris (Eds.) Managing Faculty Resources. New Directions for Institutional Research No. 63. San Francisco, CA: Jossey-Bass, 1989. M oore, K. M ., and Sagaria, M. "The Situation o f Women in Research Universities in the United States: Within the Circles o f Academic Power." In G. P. Kelley and S. Slaughter (Eds.) Women and Higher Education in Comparative Perspective. Boston: Kluwer Academic Publishers, 1991. M oore, K. M ., and W ollitzer, P. Women in Higher Education: A Contemporary Bibliography. W ashington, D .C .: National Association for Women Deans, Administrators, and Counselors, 1979. 225 M orrison, A ., Randoll, P ., and Van Velson, E. Breaking the Glass C eiling. Reading, MA: Addison-Wesley Publishing C o., 1987. National Center for Education Statistics. "Digest o f Education Statistics, 1990." W ashington, D .C .: United States Department o f Education. NSOPF-88. See SRI International listing. Olsen, D. "Report on 1990 Study o f M inority and Women Faculty." Bloomington, IN: Indiana University, 1990, unpublished. Olsen, D. "Women and M inority Faculty Job Satisfaction: A Structural Model, Examining the Effects o f Professional Role Interests, Professional Satisfactions, and Institutional Fit." Presentation at Association for the Study o f Higher Education, Fall, 1991. Pfeffer, J., and Ross, J. "Gender-Based W age Differences-The Effects o f Organizational Context." W ork and Occupations. Vol. 17, #1, Feb., 1990, pp. 55-78. Reynolds, J.A . "Knowing one’s world: gender and world view in a research university culture." Dissertation Abstracts International. 49, 3638-A, 1989. Rossi, A ., and Calderwood, A. Academic Women on the M ove. New York: Russell Sage Foundation, 1973. Rudolph, F. The American College and U niversity. New York: Vintage Books, 1962. Shuster, J., W heeler, D ., and Associates, eds. Enhancing Faculty Careers: Strategies for Development and Renewal. San Francisco: Jossey-Bass, 1990. Shakeshaft, C. Women in Educational A dministration. Publications, 1989. Newbury Park, CA: Sage Sheehy, G. Passages: Predictable Crises o f Adult L ife. New York: Dutton, 1976. Simeone, A. Academic W omen: W orking Towards Equality. Bergin and Garvey Publishers, Inc., 1987. Smart, J. C. "Gender Equity in Academic Rank and Salary." Education. Vol. 14, #4, Summer, 1991, pp. 511-526. South Hadley, MA: Review o f Higher SRI International. "Special Issue, Reports from the 1988 National Survey of Postsecondary Faculty." Menlo Park, CA, 1990, Prepared for the National Center for Education Statistics, W ashington, D.C. 226 Stromquist, N. "Determinants o f Educational Participation and Achievement o f Women in the Third W orld: A Review o f the Evidence and a Theoretical Critique." Review o f Educational Research. Vol. 59, # 2, Summer, 1989, pp. 143-183. Theodore, A. The Campus Troublemakers: Academic Women in Protest. Houston, TX: Cap and Gown Press, 1986. Tum in, M. Readings on Social Stratification. Englewood Cliffs, N .J.: Prentice Hall, Inc., 1970. W eiler, K. Women Teaching for Change. Publishers, Inc., 1988. South Hadley, MA: Bergin and Garvey Yoder, J ., Crumpton, P ., and Zipp, J. "The Pow er o f Numbers in Influencing Hiring Decisions." Gender and Society. Vol. 3, #2, June, 1989, pp. 269-276.