|lll\._.. ‘ «I I I I 4.117)!- . ' I , Vt. o S. I. a. ll \l ‘01. 00 ‘v c Av o , .al: « I n rig“; ain‘t- a. f. ‘1‘ is} I‘ \ll Mr“! «Ad-{.611 ‘ \l. snub-Rte. I‘M! D “I o I l. l I I‘SI III ll‘l 011.1,. till! Univ.“ xiv/‘HJ‘I Inl‘ll ill 1|"\‘. l- bnlir .al.||nc Ill! 0 IIIII'HI 'l. "Illa?!“ 1 I‘. 4| V tl'l I. v...»|3- ‘-\ --Iia}x..v\ 31.1 « Edam!» . n‘il‘f‘tq‘i I! ‘ 3‘ . . l‘itpllllvv‘ «. II Un|| I It . I . MH‘. u... E ufil. .GHVHKIc‘I , ‘.i§l‘s\. .hfl‘dll‘mnuhuh HUN” - J. x - i.fifl1 «1%flEE 1. - . ‘ . - \J \fi‘lfid tflnui 44‘1{\3-1Iofl1l.\tpll 0‘3”... u \ufltflznahhfiaxwkzumw‘mwbfiumilfi 15.x B L . . ti“ In“ ‘ ' "" " "git” '9 TH my: .. THEE???- ‘ LEM!” Michigan flute _ ; ynlveuity . a? , , A i! This is to certify that the dissertation entitled AN INTERACTIONIST PERSPECTIVE 0N PERSONNEL SELECTION presented by Michael P. Fitzgerald has been accepted towards fulfillment of the requirements for Ph. D. degree in Psychology Major professor Date 5/ l 2/83 MSU is an Affirmative Action/Equal Opportunity Institution 0— 12771 IV1ESI_J RETURNING MATERIALS: Place in book drop to LJBRARJES remove this checkout from .Jll-(ynnnm. your record. FINES will be charged if book is returned after the date stamped below. acsou use Oil-EH AN INTERACTIONIST PERSPECTIVE ON PERSONNEL SELECTION By Michael Patrick Fitzgerald A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Psychology 1983 ABSTRACT AN INTERACTIONIST PERSPECTIVE 0N PERSONNEL SELECTION By Michael Patrick Fitzgerald The traditional selection research model is discussed and critiqued. The phenomena of the validity ceiling and the vanishing validity coefficient are described as evidence of the inaccuracy of two assumptions of human behavior inherent in traditional selection research. This research is described as "trait" oriented, because it denies the existence of person/work environment interaction pro- cesses. An interactionist approach to selection research is developed. This perspective assumes behavior to be a function of individual characteristics, the environment and their interaction. The present study addressed the question of how an environ- ment might affect the expression of performance behavior. Based on Mischel's (l977) work in the area of personality research, it was hypothesized that different work environments provide various levels of information concerning appropriate behavior, that ultimately influence the performance effectiveness of individuals in that environment. This leVel of information inherent in an environment was tapped via the construct of climate. In particular, the concept of "climates for the expression of specific criterion behavior" was developed. Michael Patrick Fitzgerald Empirical evidence of the effect of highly developed criterion- related climates on the performance behavior of ninety-nine school administrators was sought. Individual characteristics of these administrators were measured in an assessment center. Their perfor- mance and climates at their schools were subsequently evaluated. Results of hierarchical regression analyses indicated an increase in prediction when climate scores were entered for two of six criterion behavior dimensions examined. The findings are described as evidence that environmental factors can influence the behavior of individuals and, ultimately, prediction levels. More importantly, the investigation of these environmental influences provides a means for understanding prediction-criterion links. By incorporating criterion-relevant situational characteristics in selection research, low or vanishing levels of prediction may be alleviated. Limitations and suggestions for future research are discussed. ACKNOWLEDGMENTS It has been a great privilege to work with the members of my dissertation committee. I would like to thank Neal Schmitt, Ben Schneider, Mary Zalesny and Ray Frankmann for their guidance and support. Now that my graduate training is completed, I hope to be able to repay a small portion of my great debt to them by striving for excellence in my professional work. I would also like to express my gratitude to my parents (who, after years of training can identify what an Industrial] Organizational psychologist does), my family and my confidant and wife to be, Rose. ii TABLE OF CONTENTS LIST OF TABLES . LIST OF FIGURES. INTRODUCTION. THE TRADITIONAL SELECTION MODEL--A CRITIQUE AN The Validity Ceiling. . Explanations for the Validity Ceiling Summary . . The Vanishing Validity Coefficient . The Traditional Selection Model and Implicit Assumptions of Human Behavior . . . . . . . . . . Summary INTERACTIONIST PERSPECTIVE ON WORK BEHAVIOR Trait vs. Interactionist Perspective on Human Behavior and the Traditional Selection Model . . . The Interactionist Perspective. . Predictor- Criterion Links and the Prediction of Work Behavior . Summary THE SITUATION AND WORK BEHAVIOR Work Behavior and the Situation as Viewed by Industrial/ Organizational Psychology . . A Review of Ability/Situation Interaction Research. Summary . . A Social Learning Conceptualization of the Person/Situa- tion Interaction in the Work Setting. . . A Climate Approach to Defining Criterion— Relevant Situa- tional Characteristics . . Summary . . What is an Interaction? . Page vii Tl I] I4 IS 17 T7 18 20 23 25 25 33 34 36 4O SUMMARY AND OVERVIEW OF RESEARCH . METHODOLOGY . Sample. Measures . . Data Analysis RESULTS AND DISCUSSION Interrater Agreement of Climates fer the Expression of Behavioral Criteria and Behavioral Performance Ratings. Regression Analyses CONCLUSIONS . Limitations and Suggestions for Future Research . APPENDICES APPENDIX A - BEHAVIORAL PERFORMANCE EVALUATION SCALES. APPENDIX B - CLIMATE SURVEY. . . REFERENCES REFERENCE NOTES. REFERENCES iv Page 46 49 49 59 61 61 75 78 83 98 103 104 TABLE know (I! 10 11 12 LIST OF TABLES Abilities/Individual Characteristics Assessed . Criteria Measurements . Outline of Prediction Equations. Within- -group Reliability (rwg(x1)) for Schools with Complete Climate Composite Scores . Interrater Reliabilities of Performance Ratings Means, Standard Deviations and Intercorrelations Among Individual Characteristics (I. C. ), Climate, and Performance Behavior Dimensions . Curriculum and Instructional Leadership Dimensions: Correlation Among Tenure, Individual Characteris- tics (I.C.), Climate, I.C. x Climate and Perfor- mance Behavior . . . . . . . . . . . Student Activities Dimension: Correlation Among Tenure, Individual Characteristics (I.C.), Climate, I.C. x Climate and Performance Behavior . . School Support Services Dimension: Correlation Among Tenure, Individual Characteristics (I.C.), Climate, I.C. x Climate and Performance Behavior . . Staff Development and Evaluation Dimension: Correla- tion Among Tenure, Individual Characteristics (1. C. ), Climate, I. C. x Climate and Performance Behavior . . . . . . . . . . Community Relations Dimension: Correlations Among Tenure, Individual Characteristics (I.C.), Climate, I.C. x Climate and Performance Behavior . . School Plant Maintenance Dimension: Correlation Among Tenure, Individual Characteristics (I.C.), Climate, I.C. x Climate, and Performance Behavior. Page 51 54 6O 63 64 66 67 67 68 68 69 69 TABLE Page 13 Curriculum and Instructional Leadership Dimension: Summary of Regression Analysis. . . . . . . 7T 14 Student Activities Dimension: Summary of Regression Analysis . . . . . . . . . . . . . . 7l 15 School Support Services Dimension: Summary of Regression Analysis . . . . . . . . . . 72 l6 Staff Development and Evaluation Dimension: Summary of Regression Analysis . . . . . . . . . 72 l7 Community Relations Dimension: Summary of Regres- sion Analysis . . . . . . . . . . . . 73 18 School Plant Maintenance Dimension: Summary of Regression Analysis . . . . . . . . . . 73 vi LIST OF FIGURES FIGURE Page 1 Relation of Predictors to Criterion as a Function of Time. . . . . . . . . . . . . . . . l3 2 Dunnette' 5 Modified Model for Test Validation and Selection Research. . . . . . 2] vii INTRODUCTION If one were to consider the ultimate goals of Industrial] Organizational psychology as a field of scientific inquiry, a major objective could be described as the prediction and understanding of human behavior in the work setting. Generally speaking, traditional industrial psychologists and organizational psychologists have approached this objective from two different perspectives. Indus- trial psychologists have eschewed organizational factors and concen- trated on the assessment of individual differences for the predic- tion of work behavior. Conversely.organizational psychologists have concentrated on the properties and processes of organizations, not the personal attributes of people, as the important data in predict- ing and understanding behavior in the work setting (Schneider, 1978). Many psychologists have pointed to the tenuous marriage of industrial psychology and organizational psychology (c.f. Guion, 1976). Although "married" in the sense of ultimate goals, research- ers in the two subdisciplines have often ignored one another's theories, research paradigms and variables of interest (TenOpyr, 1981). For example, in a recent review of organizational behavior research in the Annual Review of Psychology, the concept of ability is never mentioned (Mitchell, 1979). The purpose of the present research is to attempt to inte- grate industrial and organizational concepts in an appropriate 1 theoretical framework for predicting and understanding work behavior. The need for this integration will be demonstrated by critiquing the "traditional" selection model and its implicit assumptions concerning human behavior. The traditional model's individual difference or "trait" orientation, which assumes individual characteristics as the prime determinant of behavior, will be contrasted with an “inter- actionist perspective," which assumes that the individual, environ- mental stimuli and their interaction are the prime determinants of behavior. In particular, the model of behavioral consistency and the assessment center method, one of the most sophisticated personnel research measurement techniques based on this model, will be critiqued. The interactionist perspective is presented as a more appro- priate model of human behavior, one that will allow personnel researchers to understand why long-term prediction of work behavior based on individual difference (person) variables alone is not possible. A climate approach to indexing the situation will be presented as a viable method to understand how person and situational variables interact to influence human behavior in the work setting. The present research effort investigates the role of individual dif- ference data as measured in an assessment center and criterion- relevant climate data in realizing the ultimate goals of both pre- dicting and understanding work behavior. THE TRADITIONAL SELECTION MODEL-- A CRITIQUE The first recorded piece of research in the use of tests for the prediction of job performance was Muensterburg's (1913) experi- ment investigating the selection of motormen. Shortly thereafter, during World War I, large-scale testing of soldiers provided a basic methodology and respectability to the use of preemployment tests as predictors of performance. This prototypical research led to a post- war surge of systematic research in personnel testing. During the twenties a great deal of data became available concerning the validity of various sorts of tests for many different jobs. Accept- able practices in scientific personnel selection were recorded in Freyd's (1923) Measurement in Vocational Selection. Freyd's review of the steps to conduct selection research is surprisingly up to date. He described ten steps, outlined below: Step One. Job analysis--the purpose of which is to determine individual characteristics leading to success or failure on a particular job. Step Two. Criterion development. Step Three. Sample selection. Step Four. Development of a list of abilities, skills, knowledge and traits required for success. Step Five. Find or develop appropriate measuring instru- ments to determine level of individual charac- teristics deemed necessary for successful per- formance. Step Six. Administration of "predictor tests" under care- fully controlled conditions. Step Seven. Statistical comparison of predictor and crite- rion scores. Step Eight. Combination of predictor scores for optimal correlation. Step Nine. Utility analysis. Freyd noted that tests must be justified by comparing their predictive accuracy with that of the methods of selection already in use. Step Ten. Installation of new predictors in selection procedure and continual evaluation of predic- tive accuracy. Freyd's basic steps have stood the test of time. Indeed, these basic ten steps describe quite adequately the major components of what will be referred to as the traditional selection model. The majority of personnel selection research conducted today uses Freyd's model. Guion (1976) expressed dismay that "little of substance has been added or changed in the same century since [Freyd's time]" (p. 782). It is unfortunate that Guion has taken such a negative view of this time-worn model. Rather than point to its age as a symp- tom of obsolescence, the fact that it has undergone so little revision in sixty years reflects the traditional selection model's success as a research paradigm for the prediction of work performance. Additionally, research conducted within the traditional selection model framework has provided major contributions to the personnel selection process. Reflecting the functionalist tradition of the field, industrial psychologists, for the most part, chose not to question that which seemed to work so well. Much of the work done in Industrial Psychology since Freyd's time has focused on raising the level of sophistication of various elements of the traditional selection model. As a result, great progress has been made in many areas, such as the selection of pre- dictors, more extensive criterion assessments, the development of taxonomies of human abilities and personalities, procedures with which we can combine information into predictive equations, the global assessment of individuals (i.e., assessment centers) and quantification of the utilities of selection procedures. Despite the fact that there have been many developments in the field, some have pointed to this plethora of research dealing with the components of the traditional selection model as an obfuscation of questioning any real problems of the model. In fact, Roberts, Hulin and Rousseau (1981) in their critical evaluation of selection research, proclaim most of selection research as simply "tinkering with the same shoddy paradigm" (p. 11). These authors question the ability of the traditional model to adequately predict performance variance within the confines of the traditional model. They point to the inability of personnel researchers to predict high levels of performance variance asha failure of the traditional selection model. However, the ceiling on levels of prediction was hypothesized many years ago. The Validity Ceiling, The first attempt to review personnel selection research was made by Hull (1928) in his classic book, Aptitude Testing. Hull concluded that upper limits for the validity of tests were about .50, corresponding to a forecasting efficiency of only 13%. He felt that with continuing research efforts, the typical predictor-perfor— mance correlations of that time (the 20's) would improve. However, he felt that validity coefficients of over .70 were inaccessible because of performance variance due to "chance or accident." Over thirty years later, results of Ghiselli's (1966) com- prehensive review proved consistent with Hull's earlier prophecy. Examining both published and unpublished studies, he showed average validities ranging in the .305 and low .405; validity of over .50 was a distinct rarity. In an update of this review, Ghiselli (1973) found little improvement in validity coefficient levels. More recently, Boehm (1982) reviewed criterion-related validity studies published in the Journal of Applied Psychology_and Personnel Psychology between 1960 and 1979. She found that although there has been an increase in the average sample size, there was no significant change in the absolute magnitude of reported validity coefficients. The mean validity over the 176 studies that the author reviewed was .22, with a standard deviation of only .056. Low validities have apparently led some psychologists to become disenchanted with selection research. Roberts, Hulin, and Rousseau (1978) claim, "This lack of progress was noted in all areas related to selection and placement: levels of empirically estab- lished validities, major theoretical development, innovations in paradigms, and development of more appropriate modes of task and worker characteristics" (p. 112). Although levels of validities have not increased dramatically over 50 years of selection research, such a denouncement of all aspects of this research is unwarranted. Rather than simply attacking all aspects of the traditional selec- tion model and relevant research, it is more appropriate to deter- mine why the traditional selection model research paradigm prohibits more accurate prediction. This question of why the validity ceiling exists has been addressed by several personnel researchers from various perspectives. Explanations for the Validity CeiTing Currently, the most popular explanation for the validity ceiling can be found in Schmidt and Hunter's (1976, 1979) validity generalization hypothesis. In their work, unreliable measures, biased criteria, restrictions of range and small sample sizes are listed as sources of low validities and variable validity. However, the corrected population validities computed using Schmidt and Hunter's procedures tend not to exceed .50. Recently, Schmitt and Schneider (1983) have questioned the appropriateness of corrections for attenuation and for restriction of range used in the Schmidt/Hunter procedures. The two authors claim that if the rate-rerate performance interval includes changes in the importance of performance factors, an inappropriately low estimate of reliability will result. If this is true, the correction for unreliability in the criterion would overestimate the "true" population test validity. Providing a convincing argument that job experience serves to enhance individual differences, Schmitt and Schneider claim that the Schmidt/Hunter correction for range restric- tion procedures seriously overcorrected "true" validities. Schmidt and Hunter (1976) have typically assumed range restriction effects of about 40% of the original range; consequently, their corrections of observed validities for restriction of range are substantial. An earlier attempt to deal with the question of the vali- dity ceiling was concerned with the use of tests as predictors. In 1968, Wernimont and Campbell (1968) developed the "behavioral consis- tency model" in response to concern for the low validity coefficients found in personnel research. They argued that the traditional selection model's focus on the use of tests as signs of predisposi- tions to behave in a certain manner was the reason for depressed validity coefficients. In contrast, the behavioral consistency model emphasized samples of work behavior as predictors of future work behavior. Wernimont and Campbell (1968) identified two basic methods of trait measurement in personnel selection. The more traditional psychometric measurement method, the authors asserted, was unsatis- factory because it focused on the use of tests as "signs" of predis- positions to behave in certain ways on the job. They proposed an emphasis on samples of actual work behavior as predictors of future work behavior. The behavioral consistency model seeks to establish consistencies between relevant dimensions of job behavior and pre- employment behavior samples obtained from real or simulated situa- tions. Assessment center research has often been discussed as demonstrating the advantages of the consistency approach in select- ing managers (Campion, 1972). The assessment center method is based on the theory of behavioral consistency (Thornton and Byham, 1982). The exercises are designed so that the job applicant undergoing assessment displays actual behavior relevant to the behavioral domain of a particular job. In contrast, typical paper and pencil type selection instruments are usually only signs of that behavioral domain. A review of assessment center validation studies (e.g., Hucky, 1973) could lead one to believe that the ultimate selection device for management personnel has been developed. Comparisons of the assessment method to more traditional psychometric methods for predicting managerial success reveal that the average validity of 10 the assessment center is about as high as the maximum validity attained by the use of these traditional methods (Norton, 1976). However, the vast majority of these validation efforts have failed to conform to the behavioral consistency model. All too often, the criteria used in these studies have been some form of promotion/progress or general ratings of performance. Although management progress is an important criteria, it does not necessarily reflect a behaviorally based performance criteria. The consistency model, to be consistent, requires that the measures to be predicted must also be measures of behavior. Wernimont and Campbell (1968) specifically warned against the use of a behavioral sample to predict such criteria as salary progression or organizational level achieved. The promotion criteria are obviously several steps removed from actual job behavior. These authors claimed that the use of such non-performance behavior criteria will place a ceiling on the maximum predictive efficiency to be expected. However, when this statement is compared with validation results, the opposite is found. After surveying hundreds of studies, Thornton and Byham (1982) found a mean correlation of .63 between assessment center ratings and ratings of management promotion, but found a mean correlation of only .33 between assessment center pre- dictions and ratings of actual job performance. Although the correlation of .33 is by no means trivial, it is significantly lower than the correlation of .63. In fact, the assessment center method, when judged against performance ratings, seems less than the 11 ultimate selection device and must be reviewed as another falli- ble measurement tool in the prediction of job behavior. Summary Both the validity generalization procedures and the behavioral consistency models have attempted to deal with the problem of the validity ceiling. However, empirical results have indicated that the predictiveness of work behavior remains low. Claims for the utility of the procedures have led personnel researchers to be relatively comfortable with the validity ceiling. Indeed, when costs and savings are computed, the average validity of .30 does have substantial utility. However, the validity ceiling does hint at the inadequacy of the traditional selection model as a paradigm for predicting and understanding the complexities of work behavior. When one considers job performance over time, the traditional model's deficiencies become even more apparent. A phenomenon labelled as the "vanishing validity coefficient", is discussed in the context of the prediction of work behavior in the following section. The Vanishing Validity Coefficient An even more discouraging aspect of the results of selection research can be referred to as the "vanishing validity coefficient." This phenomenon is evidenced by levels of validities that become depressed over time. As an employee accumulates experience at tasks, initial levels of validity drop. Even if one uses initial 12 levels of performance on the task or job, rather than some ability measure, the capacity to predict later performance drops (Alvares and Hulin, 1972, 1973; Dunham, 1974; Humphreys, 1962). The case of the vanishing validity coefficient is best described in Figure l. The curve is the expected relation between predictor and criterion in selection research if one were to sum- marize predictor-criterion relationships as a function of time at which the criterion is collected. The curve indicates that the more experience employees have in a work environment, the lower the validity. It should be noted that this depression of vali- dity over time is also evidenced in areas of research which attempt to predict other types of performance. For example, Humphreys (1968) found that American College Testing (ACT) scores obtained during a student's junior year of high school used to predict first college semester G.P.A. have validities in the thirties. However, by the eighth college semester these validities drop to almost zero. One type of personnel selection predictor that appears to defy the law of the vanishing validity coefficient is the assessment center. Mitchell (1975) found that when 24 assessment center pre- dictors were correlated with a criterion of salary growth measured one, three, and five years after assessment, the validity tended to increase over time. In the Hinrich's (1978) research, predictive validity of overall assessment ratings against a criterion of manage- ment promotion after one year on the job was .26; eight years later it was .46. However, these studies did not predict management per; formance, but rather management progress. Klimoski and Strickland 13 .mEFH mo cowuucsu m we co_cmpweo op meouowumca we cowpmpmm--.. mgzm_m cowpmscoecm cowemu_go do pcmsmmmmm< mo wave xgpcm me$< coom agacm emum< mcmm> x )3 D J U. 0 . e 4 III}. .. I = w J // O a ”W -- ..I no a. 1;. :1 L. _/, 0 II. I u x// /. a //I. m I 3 //, 8 /. U / /. d J )N a J O ._ w m. I m. an o 03 J ( by U p 3 u 1 3 J Lo 0 )d u J a J .. I: a 1.3 . .4 0 0 ( 14 (1977) questioned these studies as real tests of validity. Their criticism assumes that promotion/survival is not an adequate measure of management performance. On an empirical basis, one would also expect validities to increase over time with a criterion of manage- ment promotion, simply because of the increase in range expected in the criterion over a longer period of time. The Traditional Selection Model and Implicit Assumptions of Human Behavior The traditional selection model used over the past sixty- five years has provided personnel researchers with a valuable paradigm for the prediction of employee performance. However, the validity ceiling forecast by Hull (1929) remains. In addition, evidence of vanishing validity coefficients over time has not been addressed by many personnel researchers. Empirical evidence of lowered predic- tiveness over time implicitly suggests the short-term perspective of personnel research. This concern for short-term results in the absence of long-range prediction reflects the functionalist tradi- tion of industrial psychology. The question of "how much" as opposed to a real understanding of why and how predictors work has been the primary focus of personnel research. The question of why validities drop over time would require an understanding of processes within the work environment that affect worker attributes and, consequently, performance. The traditional selection model is based on an implicit theory of human behavior which, if followed, usurps the researchers ability 15 to understand these processes. This implicit model of behavior is evidenced by two assumptions personnel researchers must make when working within the confines of the traditional selection model. First, it is assumed that experience in an organization (a situation) has no systematic effects on the members of that organization in terms of the characteristics on which selection is based. This is evidenced in the frequent use of range restriction corrections which assume that the work environment, over time, has no effect on the individual characteristics measured as predictors. However, Alvares and Hulin (1973), Brousseau (1978) and the sociologists Kohn and Schooler (1973, 1978) all provide evidence that experience on a job can have a marked effect on individual characteristics. Second, there is an implicit denial that the work environment (a situation) interacts with individual characteristics to produce a certain level of performance. Summar The traditional selection model has provided personnel researchers with a successful research paradigm which allows for highly utilitarian prediction of work behavior. However, the phenomena of the validity ceiling and the vanishing validity coeffi- cient provide evidence of problems with the model in terms of level of prediction and the short-sightedness of that prediction. Attempts to deal with the validity ceiling, such as validity generalization procedures and selection-based on a behavioral consistency model, 16 have failed to account for greater levels of work performance vari- ance. It is suggested that research conducted within the boundaries of the traditional model is hampered by an inaccurate theory of human behavior. This implicit theory of behavior precludes an examination of work processes that can influence individual character- istics and determine performance. In the following section, this implicit model of behavior will be contrasted with a model which suggests an investigation of these work processes in relation to criterion-based behaviors. This alternative model of behavior will permit personnel researchers to investigate predictor-criterion links which, hopefully, provide for an understanding of the vanish- ing validity coefficient and how the validity ceiling can be alleviated. AN INTERACTIONIST PERSPECTIVE ON WORK BEHAVIOR Trait vs Interactionist Perspec- tive on Human Behavior and the Traditional Selection ModeT The traditional selection paradigm is cast in a psychological tradition of individual differences research. In the traditional model, the focus of the research is on individual characteristics that are considered the prime determinants of behavior. During the job analysis phase of selection research, all job analysis methods seek to determine what level of individual skill, ability, or other personal characteristics (traits) are needed to perform successfully in a particular job. These individual characteristics are assumed to be stable dispositions which serve as the major determinant of later success or failure. As suggested earlier, prior research has addressed the stability of these individual characteristics in varied situations over time. When considering skills, research by Alvares and Hulin (1973) and Dunham (1974) provides empirical evidence of the development of abilities and skills when the individual gains experience at a particular task. In the realm of personal character- istics (traits), Brousseau (1978) has shown how individual traits might be affected by exposure to qualitatively different work environ- ments. 17 18 This individual difference perspective assumes a model of human behavior which is parallel to the trait model of human behav- ior developed by personality psychologists. The trait theory of behavior assumes that the sources for the initiation and direction of human behavior come primarily from within the organism. The individual can be categorized by traits which are tendencies or predispositions to respond that are not linked to specific stimuli or responses (Allport, 1937). Thus, it is assumed that the rank order of individuals with respect to a particular behavior is con- sistent across a wide variety of situations (Endler and Magnusson, 1976). The Interactionist Perspective The assumptions of the trait model of behavior described above are in contrast to the theoretical assumptions of inter- actional psychology. Interactional psychology (Endler and Magnusson, 1976; Pervin and Lewis, 1977) is an approach to the study of behavior that emphasizes a continuous and multidirectional interaction between person characteristics and situation characteristics. Unlike the trait perspective, the interactional model assumes that the sources for the initiation and direction of behavior come primarily from the continuous interaction between the person and the situation. The individual's behavior is influenced by meaningful aspects of the situation, and in addition, the individual selects and interprets the situation in which behavior occurs. Subsequently, the person's behavior affects these situations and their meanings. Therefore, 19 accurate measurement of both individual traits and relevant aspects of the situation are necessary to account for behavior. Endler and Magnusson (1976) presented the basic propositions of the interactionist approach to the study of human behavior. They are: 1. Actual behavior is a function of a continuous process or multidirectional interaction (feedback) between the individual and the situation. 2. The individual is an intentional active agent in this interaction process. 3. On the person side of the interaction, cognitive factors (including ability, motivation, and emotional components) are the essential determinants of behavior. 4. On the situation side, the psychological meaning of the situation and the behavior potential of the situation are the essential determinants of behavior. Despite the fact that the trait and interactionist models of human behavior were developed in personality research, they delineate some key theoretical issues when one considers the models of human behavior that are implicit in the traditional selection model. In personnel selection research, it is assumed that the abilities, skills and traits (referred to hereafter as individual characteristics) measured as predictors are stable individual char- acteristics which serve as the prime determinants of work behavior. However, the phenomena of the validity ceiling and the vanishing validity coefficient suggest that work behavior is determined by 20 more than simply individual characteristics. The model of human behavior that is implicit in selection research could be charac- terized as a "trait perspective." This trait perspective of the traditional selection model ignores the fact that behavior in a work setting is best understood as a function of both individual characteristics pad the environment within which that behavior occurs. Indeed, some have claimed that the slowly accumulating literature in organizational research sug- gests that characteristics of intraorganizational environment account for a greater portion of responses by people than individual differ- ences do (Roberts, Hulin, & Rousseau, 1978). The denial of situa- tional influence in the traditional selection paradigm treats the effects of the environment on performance as error variance. This trait perspective inhibits personnel researchers' ability to fully understand predictor-criterion links that would provide for an under- standing and possibly an alleviation of the validity ceiling and the vanishing validity coefficient. Predictor-Criterion Links and the Prediction of Work Behavior Concern for the processes by which individual characteristics measured as predictors and environmental influences interact to influence work behavior in the personnel selection process is not new. Dunnette (1963) characterizes the traditional selection model's 21 attempt to index the relationship between predictors and criterion by a simple correlation as simplistic. Guetzkow and Forehand (1961) first suggested a modification of the classic validation model. This richer schematization for selection research allows for an accounting of the mediating events between predictor and criterion. Figure 2 presents this revised model as modified by Dunnette (1963). . CONSEQUENCES "microns means no moons muons (Inland o. 0.0013306011016001.) I s.‘ I l 'i r“ / P' s- / c' ? l I *1 E I I 9 pa 5 u-—————+ | 3| r o l 4 l ‘\\\’ I &| ‘\\\\\\‘\\*c. v.1 l l- l- | "| \ Source: A Modified Model for Test Validation and Selection Research. Marvin D. Dunnette, Journal of Applied Psychology, Vol. 47 October 1963, pp. 317-23. Figure 2.--Dunnette's Modified Model for Test Validation and Selec- tion Research. 22 This modified model takes into account the complex inter- actions which may occur between predictors and various predictor combinations, different groups of individuals, different behaviors on the job, and the consequences of these behaviors relative to the goals of the organization. More importantly, the model recognizes the fact that job behaviors and situation characteristics are both responsible for ultimate performance consequences. Dunnette claimed the model implies nothing concerning the form of relationships to be expected among the different components of the model. One of the unfortunate consequences of utilizing the classic validation model is its overemphasis on the correlation as the sole statistics of validation research. Dunnette's more complex prediction model focuses on complex linkages between predictors and performance and necessitates more complex and SOphisticated tools of analysis in studying these linkages. Dunnette saw the basic problem with the traditional model as its lack of concern for an understanding of the process of how applicants who perform well on a selection test 90 on to, presumably, perform well on the job. He claimed that most previous selection research had been rather fixedly concerned with predicting organiza- tional consequences directly without first seeking to learn the nature of possible linkages between such consequences and all that is used to predict those consequences--most notably, the patterns of situa- tional circumstances and the possible differences in job behavior that result. Dunnette's model is no less concerned in predicting organizational outcomes than the traditional model, but it does 23 direct research efforts toward a more careful analysis of behavioral and situational attributes with the hope of understanding these organizational outcomes better and of predicting them more accurately. Dunnette's revised selection model was an attempt to develop a more interactional model for personnel research. Unlike the tradi- tional model, it realizes the role of situational characteristics in accounting for performance variance. However, Dunnette's model does not specifically address two questions that must be answered in research conducted within an interactionist framework. First, Dunnette does not outline what situational characteristics are rele- vant or how one can determine such relevant characteristics. Secondly, his revised model does not explicitly state how or why situational and individual characteristics might interact to result in the response of interest. Parenthetically, the model does not include a temporal dimension which addresses the question of indi- vidual characteristic/situation interactions over time on a job. Summary The traditional selection model has been characterized as a trait-oriented model of human behavior. The trait orientation of the model is evidenced by a lack of attention to predictor-criterion processes by which environmental influences are accepted as deter- minants of work behavior. It is suggested that an interactional model would be more appropriate for understanding the complexities of behavior in the work environment. Although Dunnette's revised model for selection provided a recognition of environmental 24 influences, it failed to delineate how situations and individual characteristics might interact to produce a certain level of work behavior over time and it did not address what form this interaction might take. If, indeed, personnel researchers are to accept an inter- actionist perspective, these two questions must be examined. In the next section, past research by Industrial[Organizational psychologists which invokes this question will be critiqued. A cognitive social learning conceptualization of situation influence is suggested as an alternative to past theory and a climate approach to situation categorization is developed. Finally, the question of what form an interaction can take is reviewed. THE SITUATION AND WORK BEHAVIOR Work Behavior and the Situation as Viewed by Industrial/Organi- zational Psychology The interactionist perspective stresses the hypothesis that behavior is a function of both person and situation. Industrial] Organizational psychology has accomplished much in developing stan- dardized measures of many dimensions of individual characteristics (abilities and traits) emphasized as predictors of job performance. However, little work has been done in developing criterion-relevant dimensions of environments and relating these to individual charac- teristics. Guion (1976) recognized this problem and commented, "The problem is that environmental factors influencing performance have not been considered very often in attempting predictions during the hiring process" (p. 798). This problem is most obvious when one considers the behav- ioral consistency model. Behaviors exhibited in a standardized job simulation are assumed to match the behaviors to be exhibited in the complex work environment. The interactionist perspective ques- tions the appropriateness of this cross-situational behavioral con- sistency assumption. Indeed, even in the standardized environment of an assessment center, Turnage and Muchinsky (1982) found evidence of behavioral inconsistencies within individuals across different exercises. 25 26 Evidence of behavioral inconsistency is even more apparent when one considers the results of the Management Progress Study, (Bray, Campbell and Grant, 1974). These authors indexed the degree of "job challenge” various employees experienced over the course of eight years. Job challenge was defined as a composite of four dif— ferent constructs, (l) the achievement models provided by super- visors, (2) job stimulation and challenge, (3) supervisory respon- sibilities and (4) the degree of unstructured assignments. When fre- quency data reported by the authors is converted to correlations, individual characteristics measured via behavior observed during the assessment center were found to have substantially lower validity (.15) for employees in situtaions characterized by high challenge than the Validity (.45) for those employees in low challenge positions. The discrepancy between these validity coefficiency suggests that situational variables can affect performance and hence, the predictiVeness of individual characteristics. The authors concluded that the situation had a direct effect on individual motivation, which in turn translated into higher job performance, as indicated by the statistically greater proportion of promoted employees in the high challenge environments. Had this actually been the case, higher validity would have been expected in those situations in which employees were motivated to display individual abilities. An alter- native explanation is that these high challenge environments, charac- terized by fewer constraints on performance, allowed employees the opportunity to display successful behavior that might have been inhibited in the assessment center situation. 27 Unfortunately, the results of this research are confounded by the composition of the job challenge variable. Although concepts such as the degree to which a supervisor provided an achievement model and the degree of unstructured assignments are highly corre— lated situational influences on performance, they may produce differ- ent psychological effects. For example, the degree to which super- visors provide an achievement model suggests employees might exhibit more effective behavior on the job by shaping their behavior to cor- respond to that of an effective role model. On the other hand, the degree of structure in job assignments suggests that the situation simply inhibits or constrains the expression of effective behaviors. This question of how the situation psychologically affects the individual is further clarified when empirical research and conceptual models of performance are reviewed in the next section. A Review of Ability/Situation Interaction Research Empirical research in Industrial/Organizational psychology that has addressed the person/situation interaction in the prediction of work behavior is characterized by the assumption that generic situational properties primarily affect behavior through their direct impact on motivation, as concluded in the Management Progress Study by Bray, Campbell and Grant (1974). In a review of ability/situation research, Schneider (1978a) noted that all of the research reviewed tended to view the situation as an inference of an individual's internal motivational state. In particular, the author found 28 that only three sets of variables, incentive systems, job character- istics and leadership style/management philosophy, have been investi- gated as situational variables. Following is a summary of some of the major laboratory and field studies investigating ability/situa- tion interactions in the prediction of work performance. Weinstein and Holzbach (1973) conducted a laboratory study in which they investigated task-flow interdependence and type of reward (differential vs. equal) as situational characteristics in the prediction of productivity by individual ability. The investigators conducted a differential validity analysis and found that both of the situational characteristics moderated the validity of the Minnesota Clerical Test in the prediction of task performance. More importantly, an ANOVA analysis uncovered main effects for both reward conditions and task interdependence. The authors pointed to this finding as evidence of the limitation of the traditional personnel- differential approach to selection because of its lack of account- ability for mean differences in productivity across various job conditions. Another set of laboratory experiments that indexed type of reward condition as a situational characteristics was conducted by Dunnette (Note 1). Dunnette found that, in general, ability is reflected in performance more often under incentive reward conditions than under hourly reward conditions, but that these differences are not large. Dunnette also found that changes in reward systems may depress the ability-performance relationships. He found that this effect is especially critical when moving from an incentive reward 29 system to an hourly pay condition, regardless of the equity condition. As in Weinstein and Holzbach's research, Dunnette found differences in performance levels in different reward conditions. In essentially every case in this group of studies, average performance was superior under incentive conditions as compared to performance under hourly pay conditions. Lawler (1966) also investigated the type of reward system as a situational characteristic in a study of the ability/performance relationship in the field. Although his research was hampered by methodological problems, he did find an ability/type of reward contingency algebraic interaction. He claimed that ". . . (T)he performance scores supported the hypothesis that a multiplicative relationship exists between the [type of reward] contingency and ability" (pp. 159-160). However, this clearly significant inter- action term must be questioned in light of Lawler's dichotomization scheme and subsequent analysis of variance procedures. It is important to note, however, that a main effect was also found for type of reward contingency. A laboratory study which examined reward contingencies as a situational variable that could interact with individual abilities in the prediction of performance was conducted by Terborg, Richardson, and Pritchard (1980). In this laboratory study, the researchers acknowledged the fact that different types of interaction could be examined. However, the researchers sought out evidence of differen- tial validity and a significant multiplicative interaction term in a regression analysis. Although main effects were found for both 30 ability and reward contingency, there was no evidence of differen- tial validity or significant increases in validity when cross pro- ducts were entered into the regression equation. However, the main effects for both abilities and reward contingency did provide support for the interactionist position that behavior is a function of both the person and the situation. Howard (Note 2) investigated task characteristics as measured by the Job Diagnostic Survey (JDS) as a situational factor. She regressed supervisory performance ratings on ability ratings, total JDS score, and various multiplicative ability x JDS subdimension score interactions. She found main effects for both the ability and JDS situation measures, but the interaction terms were found to be nonsignificant. Forehand (1968), using an inferred climate approach, divided 120 government executives into Group-Centered or Rules-Centered work climates. His results revealed a strong effect of climate on both levels of performance and patterns of correlations between ability and performance. Frederickson, Jensen, and Beaton (1968) inves- tigated the effects of different types of organizational climate on the performance of an In-Basket Test. Inferred climates, as opposed to self-perceived climates, were created by experimentally manipulat- ing administrative procedures ("to be innovative" or “to follow the rules") and supervisory style ("global" or "detailed"). They then defined "consistent" climate (innovative/global or rules/detailed) and "inconsistent“ climate (innovative/detailed or rules/global). In consistent climates,production on an in-basket test was higher 31 than in the inconsistent climates. Additionally, the researchers found that relationships between the dimensions of performance on the in-basket varied as a function of the type of climate under which people worked. Thus, not only were effects on the lgygl of perfor- mance found, but patterns of performance were also shown to be attributable to the climate under which the people worked. Hoffman (1979) conceptualized instructor behavior as a situational variable in the same manner that Hellriegal and Slocum (1974) interpreted leadership or managerial style as a climate vari- able. His research sought a moderated multiple regression inter- action of student ability and instructor behavior on student perfor- mance. Although Hoffman found no interactions in his results which utilized aggregated instruction behavior ratings for each classroom, he found a larger proportion of the performance variance is accounted for by an additive combination of ability and climate measures than by any one kind of variable by itself. Additionally, mean levels of performance were higher for those students in more favorable instruc- tional environments. Schneider (Note 6) conducted a study in which the results suggest that ability predicts success in an environment in which an employer has almost no contact with the person whose behavior is being predicted. He concluded that two kinds of environments exist in which ability is most likely to predict behavior. The first type of environment he refers to is one which supports, encourages, and rewards the display of individual differences. The second is one in which the person is left alone. This second condition can 32 be thought of as a natural selection environment in which the indi- vidual's ability is the best predictor of performance. In a final recommendation, Schneider (1974) urges that research focus on the criterion behavior of interest and ask, "What kinds of situational variables might constrain or facilitate the display of abilities necessary to perform on the criterion (or criteria) of concern?" Forehand (1968) alludes to constraints and facilitators of the display of individual attributes and interests and the rele- vance of the situational issue for the performance of concern. He claims, "An environment (a) may or may not demand manifestation of a trait (for example, intelligence or aggressiveness) and (b) may or may not sustain such manifestations“ (p. 67). Dunnette (1968) focused on constraints and facilitators of individual performance: however, his research dealt only with different incentive systems and only inferred that they constrain or facilitate individual per- formance. In later work summarizing this research, Dunnette con- cluded that an employer's major goal, quite simply, should be to do everything he can to assure each employee gives full expression of his abilities, skills and aptitudes (1973). Peters and O'Connor (1980) presented a conceptual framework specifying the hypothesized influences of situational constraints on work outcomes and individual differences to work outcome associa- tions. Their conceptual framework borrows much from Schneider's (1975, 1978) discussions of performance and situational constraints. A proposed taxonomy of "Situational Resource Variables" stems from a piece of research conducted by Peters, O'Connor, and Rudolf (1980). 33 This taxonomy includes dimensions such as Amount of Job-Related Information, Tools and Equipment, and the Work Environment. Unfortunately, Peters and O'Connor's work practically ignores social interpersonal factors of the situation that might also con- strain or facilitate work performance and affect validities. Hackman, (1976) among others, has already shown quite convincingly that group norms do affect individual work behavior. A social phenomenon, such as group norms, could restrict behavior, even when the Peters and O'Connor physical-technological constraints are accounted for. Summary This discussion of the person/situation interaction litera- ture has raised many issues. First, what aspects of the situation are salient when one considers individual differences as predictors of job performance? The studies reviewed have shown that researchers have been guided in their choice by motivational theory rather than interest in the specific criterion behavior of interest. Secondly, why and how does the situation interact with individual characteris- tics to influence behavior in the work setting? For the most part, the empirical research has viewed the situation as an inference of individual motivation. On the other hand, others have suggested the situation serves as "gatekeeper" for the display of individual abilities and characteristics. However, these theories often treat the situation as an entity that the individual behaves in rather than addressing the issue of how the situation might psychologically affect the individual. These two 34 questions will be addressed next from an interactionaist perspective that will provide further insight. A Social Learning Conceptualiza- tion of the Person]Situation Interaction in the Work Setting The interactionist perspective emphasizes that individual and situational characteristics should be studied as joint deter- miners of individual behavior. However, the interactional approach does not provide a clear guide as to what aspects of the situation are relevant. Terborg (1981) believes this lack of guidance is important to the interactionist perspective because researchers are required to consider person and situation factors prior to research design and data collection. The first step is to clearly identify a dependent variable of interest. Once this dependent variable is identified, the researcher must consider individual and situational characteristics which might have impact. By considering the depen- dent variable as the focus, the research will, hopefully, consider only those individual and situational characteristics that have direct relevance to that criterion behavior of interest. This con- cern for the criterion behavior was suggested by Schneider (1978a). He concluded that motivation theory-based situation indices are inade- quate for realizing the true potential for environmental influence on specific behaviors. Before salient situational characteristics can be extracted from the complex work environment, one must understand how behavior 35 enfolds in that environment. Mischel (1973, 1977) has addressed this issue by conceptualizing the situation as a cognitive social learning environment. This conceptualization provides a basis for understanding how "situations" function psychologically. The environ- ment is described as providing information to the individual as to the most appropraite behavior in a given situation. Situations are viewed as powerful determinents of behavior to the degree that they lead individuals in environment to construe the particular events (stimuli) in the same way, induce uniform expectan- cies regarding the most appropraite response pattern, and provide adequate incentives for the performance of that response pattern. Conversely, situations are weak to the degree that they are not uni- formly encoded, do not generate uniform expectancies concerning the desired behavior, do not offer sufficient incentives for its perfor- mance or fail to provide the learning conditions required for success- ful genesis of the behavior (Mischel, 1977). Situations function psychologically by providing individuals information concerning appropriate behavior. But why and how do individuals perceive infor- mation in the environment? Social psychologists from the Gestalt tradition have stressed that individuals strive to create order in environments and behave on the bases of that apprehended order (cf. Chaplin and Krawiec, 1968). People create perceptions of total realities consisting of inferred as well as actual practices and procedures that exist. While psychologists working in a gestalt tradition claim that 36 individuals have no choice but to perceive this order, those working from a functionalist tradition propose that people apprehend and create order in their environments so that they can function adap- tively (Schneider, 1975). The question of how individuals derive meaning and understand- ing from a work environment has been the source of much debate in organizational psychology. Certainly one of the most popular and controversial approaches to this question is found in the construct of climate. Briefly, climate is a construct which suggests that individuals apprehend order in their work environment based on per- ceived and inferred cues and behave in ways that fit the order they apprehend (Schneider, 1975). Next, the construct of climate will be discussed as a phenomena by which a work environment provides information concerning appropriate behavior in that environment. A ClimateyApproach to Defining Criterion-Relevant SituationaT Characteristics The concept of climate has been the focus of 12 reviews in the past 20 years (cf., Forehand and Gilner, 1964; Taguiri and Litwin, 1968; James and Jones, 1974; Schneider, 1975). These reviewers have all adopted different positions on the conceptualizations of climate and its measurement. For example, Payne and Pugh (1976) view climate as a function of various macro-organizational concerns. In another reView and conceptual piece, James and Jones (1979) approach the climate construct from a social learning perspective. 37 Schneider (Note 7), in a review of the climate literature reviews, however, found several major theoretical advances that seemed to run through many of the conceptually different reviews, two of which are reviewed below. First, he claims there has been a demise of the search for "objective" organizational structural antecedents of climate. Practically all climate researchers feel that climates reside in the perceptions of organizational process, not in "objective" data on organizational structure. Second, there is a general concensus concerning the distinction between psychologi- cal climate and organizational climate. Briefly, psychological climate is the meaning an individual attaches to a work context; organizational climate is the aggregated meaning over individuals. In both forms, climate refers to a psychological construct, through which people are able to make sense out of all the stimuli that surround them (Naylor, Pritchard, and Ilgen, 1980). Schneider (1975) refers to climate as "molar perceptions people have to their work setting" (p. 473). He clarifies this statement in his 1981 review by describing it as an attribute of the setting in a sense that people in a setting are an integral part of it and their perceptions pf it may be important for their behavior in it. In his paper presented at the Third Annual Symposium on Applied Behavior Science in 1980, Schneider (Note 7) described the processes by which a climate might emerge. His major contention was that climates are not just there, such as aspects of the physical] technological environment, or even more concrete concepts, such as 38 organizational structure. Rather, they "emerge out of the goal- oriented activities of people which, in turn, emerge from similar people interacting with each other and their changing environments" (p. 24). In 1973 Dunnette asked "Performance equals ability and what?" Certainly, organizational psychologists have addressed this question time and again as a function of individual motivation. Schneider (1975), however, directly addressed the lack of situational influences when accounting for performance variance. He answered Dunnette's question,"Performance equals ability and a climate which stresses the display of individual differences" (p. 457). Schneider (1973) reviewed a number of studies investigating climate as a moderator or predictor-criterion relationships which suggests the accuracy of this contention. However, the majority of the research which has addressed this issue (cf. Frederiksen, Jensen, & Beaton, 1968) has dealt with inferred climates from objective situational attributes as opposed to perceptions of climate. Perhaps even more importantly, the climate indices used were not specifically criterion-relevant. It is proposed that response-relevant situational characteristics (response-relevant climates) should be developed for specific compo- nents of job performance. Most applied psychologists agree that job performance is multi-dimensional in nature, and that adequate measurement of job performance requires parallel multi-dimensionality in criteria. If it is argued that criteria represent behavioral or psychological constructs that are behaviorally heterogeneous 39 (Dunnette, 1963, Guion, 1965), is it not expected that different aspects of the situation will affect different criterion-behaviors? For this reason, it is proposed that different dimensions of climate will affect performance in different behavioral dimensions of a job. As it was found that predictor measures not tailored to specific behavioral criteria are not valid (Guion and Gottlieb, 1965), it is expected that climate indices not tailored to specific criteria will not be valid or provide for an understanding of how much information the situation provides the individual as to appropriate behaviors. Summar The question of what aspects of a situation are salient to performance behavior has been addressed from a cognitive social learn- ing perspective. Situations are seen as important influences on behavior to the extent that they provide the individual with suffi- cient information concerning appropriate behavior. The climate approach is cited as a method of tapping this concept in the work setting. Indeed, past research suggests that this hypothesis has merit. However, inadequate conceptualization and measurement has not fully addressed the concept of climate in relation to specific criterion behaviors. It is suggested that climates for the expres- sion of specific behavioral criteria may indeed exist and provide for an understanding of work experience processes that inhibit long- term prediction solely from individual characteristics, including abilities or individual characteristics. 40 Past research efforts in search of empirical support of person/situation interactions in the prediction of work performance have often met with failure (cf. Schneider, 1978a). However, this finding must be viewed in the context of the narrow definition of an "interaction" in industrial/organizational research. Next, the traditional view of the interaction is examined and an expanded view is suggested. What is an Interaction? Another important characteristic of interactional research, when compared with either a trait or situational approach, is the expanded view of the concept of an interaction. As noted earlier, the basic propositions of interactional psychology encourage many different meanings of person/situation interaction. All too often in the past, the view of interaction in personnel research has been limited to the notion of the moderator variable. This use of the multiplicative term in regression analysis as the sole determinant of an interaction must be reviewed and an expanded view of how an interaction can be measured must be acknowledged. Schneider (1978b) speculated on why statistically signifi- cant multiplicative interaction terms are so rare in organizational research in field settings. First, in most work settings extreme scores on either predictor or criterion, needed for statistical signi- ficance, are rare. The range of abilities people bring to work settings and the range of competence displayed or rated in work settings is most probably rather restricted. Second, the level of 41 measurement needed to create multiplicative interaction terms is not typically found in behavioral research. Third, and most importantly, the multiplicative interaction might be redundant in research already employing a work condition variable as a predictor of performance. This is most likely in those studies utilizing climate as a situa- tional attribute contributing to the prediction of performance. The natural interaction of people in work settings creates the very work conditions being used as a predictor and the creation of multiplica- tive interaction terms over and above a linear combination of ability and situation would be redundant. This argument is similar to the criticism Mobley and Locke (1970) made of attempts to weight job facet satisfaction responses by importance. If Schneider's (Note 7) presumptions fbr assessing climate as an aggregate of the perceptions of individuals is followed, then the collective data should be less systematically contaminated by individual differences and therefore more likely to exhibit person/situation interactive effects. Interactional psychology has broadened the view of what form an interaction can take. In fact, the basic propositions of inter- actional psycholOgy (Endler and Magnusson, 1976) encourage many dif- ferent meanings for the concept of person/situation interaction. Pervin and Lewis (1978) present five different meanings of inter- actions: 1. Descriptive interaction. This refers to a mere description of the interpersonal relationship(s) 42 between two people. There is no reference to personal and contextual characteristics which might explain the nature of the interaction. Statistical interaction. This is the form of inter- action which most psychologists are familiar with. It includes both the lack of additive effects in an Analysis of Variance (ANOVA) design and the search for a moderator in a multiple regression design, which were discussed earlier. Unlike the regression interaction, the ANOVA interaction does consider mean levels on the criterion, as opposed to the strength of relationship among groups. An ANOVA interaction is present when the effect of one independent variable on the dependent variable varies across different levels of a second independent variable. This type of interaction is typically found only in controlled and contrived laboratory settings. Schneider (1978b) has speculated that this type of interaction is most likely to be found statistically significant in laboratory research. However, the extremes of either persons or situations necessary for such terms to be significant rarely occur in natural settings (c.f., Cronbach and Snow, 1977). Terborg (1981) points to the fact that because researchers have control over how they select levels of person character- istics and levels of situation characteristics in the 43 lab, main effects and interactions can be viewed as artifacts except in truly random ANOVA designs. Additive influences. A third interpretation of the person/situation interaction is one in which two or more person and situation factors may make independent and linear, but not nonlinear, contributions to a depen- dent variable. Behavior is viewed as the jOint function of both the person and the situation. This type of interaction is represented by main effects of both person and situation attributes, rather than an algebraic interaction term. This type of interaction was con- sidered in the work of Howard (Note 2), but is all too often implicitly denied by organization and personnel researchers. Interdependent interactions. This fourth meaning of situation/person interaction refers to the case in which facets of people and setting are potentially independently measurable but, because the people and situation do not exist in isolation from each other, they are understood only by assessing reciprocal effects. This type is more easily understood when one adds a dynamic time element, which is considered as a fifth form of interaction. Reciprocal action-~transaction interaction. People not only react to situations, they also create or enact them (Weick, 1979). The new situation influences future 44 behavior, which, in turn, can change the situation again. Hackman's (1976) task redefinition process is one which captures the idea of reciprocal causation between the situation and the individual. In most Industrial[Organizational research, when an interaction is considered, a moderator effect is often what is sought. A moderator variable is one which affects the bivariate relationship between a predictor and a criterion. Nothing is implied about differences in group means in either the predictor or the criterion. The failure to consider mean differences in performance, in addition to rela- tionship differences, which may be caused by situational differences, is a major concern for those who are interested in using the inter- actionist approach as a source of understanding human behavior. Schneider (1978b) commented on the search for the statis- tically significant algebriac interaction term and furnishes an explanation for why it is inadequate for many types of interactional research. He states, "The problem is that whenever we think of predicting performance by combining the effects of ability and some other variable, we think of a multiplicative combination implying not only effects on intercepts (main effects) but, slopes (inter- actions) as well" (p. 301). Locke, Mento, and Katcher (1978) developed a theoretical rationale as to why multiplicative interactions should not be strongly considered when investigating ability-performance relation- ships under different goal-setting conditions. They developed a homogeneity hypothesis which stated that moderator effects are due 45 to different degrees of homogeneity with respect to a causal vari- able among different subgroups. Thus, the main focus of an investi- gation of goal-setting impact should be main effects, rather than multiplicative interactions. Locke et al., (1978) manipulated goal difficulty and goal specificity in a laboratory study. It was found that ability pre- dicted performance better in groups that were homogeneous with respect to goal setting than in those groups which were hetero- geneous. More interestingly, a moderated regression analysis showed that most of the differential validity was reducible to the main effects of both ability and goal-setting conditions. SUMMARY AND OVERVIEW OF RESEARCH The phenomena of the validity ceiling and the vanishing validity coefficient are outcomes of the short-term perspective of the traditional selection model. Long-range prediction is improbable given the "trait" orientation of the implicit model of behavior assumed in personnel selection research. It has been suggested that researchers must adopt an interactional model of human behavior to study predictor-criterion links in order to understand and increase long-term predictiveness. This interactionist perspective necessitates an understand- ing of situational influences on behavior. In a review of ability] situation interaction research it has been found that the situation is often treated as an inference of an individual's internal moti- vation state. Conceptual models of performance and other empirical studies have viewed the situation as a source of constraints and facilitators of performance without addressing how the situation functions psychologically to influence behavior. Mischel's (1973) cognitive social learning conceptualization of the situation is cited as a theoretical framework by which one can understand how the environment influences an individual's behav- ior. The concept of climate has been described as a method by 46 47 which the individual is able to attach meaning to the environment and understand what appropriate behavior should be displayed. Although past research has addressed ability/climate interaction in the prediction of behavior, conceptual and measurement problems have prohibited any definitive conclusions with regards to psycho- logical effects of climate. Most importantly, a lack of concern for specific behavioral criteria in explaining salient situational climates has characterized this research. It is hypothesized that highly developed "climates for the expression of specific behavioral criteria" might exist and provide rich cognitive social learning environments which allow for considerable situational influence on performance behaviors. The present research effort was designed to determine if ”climates for the performance of specific behavioral criteria" do indeed exist. The existence of such climates is based on the assump- tion that environmental climates represent a "shared assignment of psychological meaning". From an operational standpoint this implies, and is implied by, a reasonably high level of interrater agreement (James, Note 3). Climate measures have been developed which are designed to measure the extent to which individuals perceive prac- tices and procedures in an environment that would allow an employee in that environment a rich source of information guiding appropriate behaviors for successful performance. These individual level percep- tions can presumably be aggregated within an environment to indicate higher order or molar perceptions of shared meaning. It was expected that these highly developed climates would increase predictive 48 accuracy significantly over validities obtained when only individual characteristics are considered. Individual characteristics con- sidered are skills and traits measured in an assessment center designed to elicit behaviors that are necessary for successful per- formance in the work setting. Although evidence of such a individual characteristic] situation interaction has typically entailed a search for a multi- plicative interaction term, this index might be superfluous when one considers a climate approach to index the situation. Thus, it has been suggested that an expanded view of the interaction be adopted. Given the existence of "climates for the performance of specific behavioral criteria" alluded to above, it is hypothesized that an individual characteristic/climate interaction in the pre- diction of performance behavior could be empirically manifested by evidence of two independent main effects for person and climate variables. METHODOLOGY Sample The study focused on the selection of individuals for the jobs of secondary, middle, and elementary school principal and assistant principal positions. These school administrators had been assessed in a two-day assessment center developed under the auspices of members of the Society for Industrial[Organizational Psychology of the American Psychological Association. After assessment, scales measuring climate for expression of specific criteria were adminis- tered within schools in which the applicants were placed. Performance of these principals and assistant principals was evaluated during the second year of data collection. These three sets of measures. initially developed for a validation study, were used to evaluate how both individual and situational characteristics serVe to predict performance behavior. Ninety-nine school administrators participated in the study, including 33 from elementary schools, 16 from middle schools, and 50 from senior high schools. Of the total group, 13% had been pro- moted to principal, 25% had been promoted to assistant principal and the remaining 62% were assistant principals who had been assessed, but not promoted. At the time of the first data collection, all of the principals had tenure of two years or less. However the range of tenure in position of the assistant principals was much 49 (Ian. 50 greater. This group had an average of sixteen years experience with a standard deviation of seven years. Measures Individual Characteristics Abilities and individual characteristics were measured before promotion to an administrative position using behaviors exhibited in the assessment center. A job analysis conducted prior to the development of the center indicated that 12 abilities and personal characteristics were important for successful performance as a school principal. These predictor variables are outlined in Table l with brief explanations. Assessing these abilities and characteristics involved the use of two in-baskets, a semi-structured interview, a fact-finding and decision-making simulation with an oral presentation and an analysis and group discussion of a case study. Six trained assessors rated each applicant on each ability and characteristic using a five-point Likert-type scale. As in most assessment centers, the assessors met after the completion of the exercises to reach consensus for each applicant on each ability dimension and a summary placement recommendation. These consensus ratings served as predictors for the analyses. Behavioral Criteria Job analysis interviews were conducted in 13 school districts throughout the country. These group interviews included the parti- cipation of principals, students, parents, teachers, support staff 51 TABLE 1. Abilities/Individual Characteristics Assessed l. 10. Problem Analysis Judgment Organizational Ability Decisiveness Leadership Sensitivity Range of Interest Personal Motivation Educational Values Stress Tolerance Ability to seek out relevant data and analyze complex information to determine the important elements of a problem situation; searching for information with a purpose. Skill in identifying educational needs and setting priorities; ability to reach logical conclusions and make high-quality decisions based on available information; ability to critically evaluate written communications. Ability to plan, schedule, and control the work of others; skill in using resources in an optimal fashion; ability to deal with a volume of paper work and heavy demands on one's time. Ability to recognize when a decision is required and act quickly (disregarding the quality of the decision). Ability to recognize when a group required direction; ability to get others involved in solving problems; to effectively interact with a group to guide them to the accomplishment of a task. Ability to perceive the needs, concerns, and personal problems of others; tact in dealing with persons from different backgrounds; skill in resolving conflicts; ability to deal effec- tively with people concerning emotional issues; Snowing what information to communicate and to om. Competence to discuss a variety of subjects-— educational, political, economic, etc.; desire to actively participate in events. Showing that work is important in personal satisfaction; a need to achieve in all activities attempted; ability to be self-policing. Possession of a well-reasoned educational philoso— phy; receptiveness to change and new ideas. Ability to perform under pressure and opposition; ability to think on one's feet. TABLE 1.--Continued 52 11. 12. Oral Communication Skill Written Communication Skill Ability to make a clear oral presentation of ideas or facts. Ability to express ideas clearly in writing; to write appropriately for different audiences 53 personnel, and district staff. The outcome of these interviews was the development of 10 behaviorally anchored rating scales to measure performance on the six criterion dimensions of interest (see Appendix A). Those scales used to measure each criterion dimension are out- lined in Table 2. These behavioral scales were the criterion measures of interest to the study. In all districts, supervisors were asked to supply ratings on the 10 behaviorally-anchored performance scales. Two teachers were also asked to supply a similar set of ratings. The district coordinator was instructed to select senior level teachers who would be able to observe and know the administrator's work. Raters who felt that they did not know the principal's work well enough to make a judgment were instructed not to do so in those cases. In addition, self ratings were collected. Performance ratings from each group for each criterion dimen- sion were averaged. The average response for each group served as a performance index. As is evident in Table 2, five of the criterion dimensions have more than one behaviorally anchored rating scale as an index of that dimension. For these five dimensions, the ratings were sumned as a composite behavioral performance rating. The ratings from each of the three groups were expected to serve as three criteria measures for each performance dimension. Comparisons of different types of rater suggest that, in general, one should expect only low to moderate correlations among different groups of raters (Landy and Farr, 1980). This has been attributed to the fact that appraisers see different aspects of an employee's behavior and employees behave differently with their boss, peers, TABLE 2.--Criteria Measurements 54 Criteria Dimension 1. Curriculum and Instruc- tional Leadership Coordination of Student Activities Direction of School Support Services Staff Evaluation and Development Development and Maintenance of Community Relations Maintenance of School Plant Scales - Monitoring Curriculum Objectives - Monitoring Individual Progress - Supervision - Participation - Direction of Support Services - Directing the Behavior of Students - Staff Evaluation - Developmental Activities - Community Relations - Interpersonal Effectiveness - Community Relations: Parents - Maintenance of School Plant 55 and subordinates (Latham and Wexley, 1981). Also, Barman (1974) has suggested that raters have different perspectives on performance that influence their ratings. Although it cannot be stated that one group of raters will be more "valid," one would expect that the teachers, support staff, and supervisors perceive different aspects of the principal's behavior and will rate accordingly. However, this multi- variate consideration of the criterion was deemed inappropriate, given the small size of the full-data sample and the moderate to high degree of interrater agreement among the rater groups. Outlined below are the specific individual characteristics rated as essential by subject matter experts (i.e. incumbent school administrators) in a related study for each of the criterion dimen- sions of interest (see Schmitt, Meritt, Fitzgerald and Noe, Note 5). These individual characteristics were averaged in a composite for each behavioral criterion dimension. 1. Curriculum and Instructional Leadership Behavior-- Problem Analysis, Judgment, Organizational Ability, and Decisiveness. 2. Coordination of Student Activities Behavior--Judgment, Organizational Ability, and Decisiveness. 3. Direction of Support Services of the School Behavior-- Problem Analysis, Judgment, Organizational Ability, and Stress Tolerance. 4. Staff Evaluation and Development Behavior--Judgment, Organizational Ability and Written Communication. 56 5. Development and Maintenance of Community Relations Behavior--Judgment and Written Communication. 6. Maintenance of School Plant Behavior--Judgment and Decisiveness. Climates for Expression of Specific Behavioral Criteria Climate research is one of many areas of psychological research which has been beset by measurement problems. The present effort has taken advantage of the lessons learned from earlier research. Perhaps the most perplexing concern for the operationali- zation of climate is the confusion of climate measurement with job attitude measurement. Schneider (1975) has argued that climate measures must emphasize descriptions of practices and procedures, rather than evaluations of those practices and procedures. Joyce and Slocum (1979), Newman (1975), and others have shown that when operationalized as different facets, climate (belief) and job satis- faction (affect) do not correlate particularly well. Scales assigned to measure the degree to which a climate for the expression of specific behavioral criteria were developed for six behavioral criterion dimensions listed below. 1. Curriculum and Instructional Leadership Climate. Climate for Student Activities and Participation. Climate for School Support Services. Climate for Staff Evaluation and Development. 01-wa Climate for Community Relations 57 6. Climate for Maintenance of the School Building Facilities. The items contained in each climate scale are presented in Appendix B. Participants responded to the climate items using a 5-point scale as shown below. 1. Strongly agree Agree Disagree Strongly disagree 01wa No opinion/no experience A rating of 5 was treated as missing data and the ratings of 1 through 4 were reverse scored in the analyses conducted. Scale scores index the degree to which a school's practices and procedures fostered a positive environment for each behavioral criterion. Higher ratings are indicative of a more highly developed climate. The climate scales were developed from extensive interview- ing of job incumbents and a "Program Audit" survey used by the Fairfax, Virginia Public School System. This latter resource used in the development of climate scales is rather atypical of most measures of school climate in that it focuses on descriptions of school policies and procedures, rather than an evaluation of those policies and procedures. The field of education has quite a burgeoning literature on the assessment of educational "environment" (cf., Walberg, 1979). Unfortunately, in practically all of the research, climate is measured only at the indiVidual level and is really more of an affective 58 measure than a description of the environment. Within each school environment, four students, two teachers and two support staff personnel were asked to respond to the climate measures. School administrators were requested to select teachers and support staff of sufficient tenure and students more actively involved in school affairs to ensure more accurate climate ratings. As defined earlier, the concept of climate as used in this research refers to the shared perceptions of participants in a work setting. Joyce and Slocum (Note 4) showed that, indeed, people can be clustered into groups based on their shared perceptions of organi- zation events, practices, and procedures. Schneider and Synder (1975) showed that position within the organization can impact per- ceptions of organizational climate. For this reason, in the present research effort, different subgroups within the school setting were surveyed for perceptions of the various criterion climates. Responses from the climate raters within each school were collected and examined for interrater reliability to determine if school environments do exhibit "climates for specific behavioral criteria". Agreement among the school climate raters was examined by a within-group procedure for determining interrater agreement described by James (1982). This within-group approach to reliability is distinguished from the typical between-group procedure which assumes differences among group aggregate scores as a basis for determining "aggregate agreement." 59 Data Analysis Separate hierarchical multiple regression analyses were per- formed for each of the six criterion behavior dimensions as outlined in Table 3. In each regression tenure in position was added first, so that experience in the principalship would not confound the rela- tionships of interest. Behaviorally anchored performance ratings, averaged over three rater groups, served as the dependent variable. Within each performance dimension regression, individual characteris- tics were entered as a composite in the second step. These individual characteristics were hypothesized to account for a significant propor- tion of performance variance. In the third step, climate scores were entered. It was expected that this variable would account for a significant increase in R2. Although this increase in R2 when adding climate to the regression on performance behavior would suggest the existence of an additive influence of person and situation vari- ables, a multiplicative interaction term was examined in the final step of the regression. However, this multiplicative interaction term was expected to be nonsignificant. 60 .mcoom mnemogsou m_mc_m m :_ umc_neou mew; :o_m:wewu chow>mcmn some :_;aw3 muwumwcmuomcmsum ozp com» um »:m_n _oo;om Lo mocmcmpc_mz mco_pe_mc pcmcma . mmm=m>_pummwm chomgmacmpcw . meowpummc »u_::seou - 0:» com» um mcowumpmc zuwcseeou co mucmcmacpme new pcmsao_m>ma oz“ Law» an acmerPm>mc use cowuum—mm .cowumz—e>m mempm ozu new» on mmow>cmm aeoaasm mo cowuumc_o cowuoa_owocma . :owmw>cma=m . o3» cam» on mmwa_> -_pum pcmuaum we coppm:_ucooo mmmcmocq _e:c_>_c:w 1 mm>_pumnno s:_=uwcczo . oz“ com» um a_;mgmcemp chowuoagpm:_ use e:_:uwgcsu .m=o cum» on mmvumrwumw new mcwcrwaa Foogum mo mucmcwpcwms com mumswpu mco cam» on mcowp -mFmL Ape:=EEoo so» mamswpo mco cum» on :oFHms—m>m use acmeaopm>mu emmpm Low mumewpu mco Lem» pm mmu_>cmm acoqazm Poonom Lo; mumsmpu mco Law» on :owumavowpcma new mmwpw>wuum pzwnzum Lee mumswpu mco com» um awcmcmcmm_ pmcowpuzcumcw use E:~:Uwccao com mume__u mmmcm>wm_uwo newsmnzc cowmewcstou coupes: su_Fwn< Paeowoauweamco ocmsmvzc cowecuwczseou coupes: »a_ppa< Pacoweauweamco acmEmuaq wocmcmpoh mmmcpm saw_an< _a=owaa~2=aaao pcmsmuzn mwmapmc< Em—nogm mmmcm>wmwumo suep_n< _a=o_ae~weamco pamEmuzw mmmcm>wm_umo sow._a< Facoaoauwcaaco pcmsmnsw mwmxpmc< Ewpnoca .m .< mucmELomcma xaeeH =o_oe=oam auaamwcaouacazu Pasca>meca mcowumzcm :o_uuwuwca co weeppso--.m m4m'mo macaw .w_ ave. use. mo. m_. awn. mo. m_. Fm. mo. ~_. N_. vo.- mo. No. mm. ~_.m mucoeLowLma mmu_>com ucoaaam .m_ amo. mo. aom. amN. o_. mo. mm. o_. 0—. 0—. e_. m_. m_. Km. mm.e oucoeLowLwa mm_uw>_uu< neonaum .8, mm. mN. .NN up. No. «mm. _~. amm. amm. ow. sum. amw. mm. «m.v oucoeLowcoa e:_=u_cc:u .m_ amm. «Ne. «mm. —_. awe. _o. _o. _o. no. co. no. em. P—.m moose—u mocncau:_e: .oogom .N. amo. amm. NN. .NN. no. mo. no. no. 80. so. KN. o_.m muae_—u m=o_u -o_mm xuwcaeeou .__ amm. m_. aha. o_. o_. mo. mp. mp. o_. em. No.m muae__u oceanopo>oo mmoum .QF «mm. amm. NF. mo.- mo.- ~_. o_. NP. mp. cm.~ ouos__u mouv>com utoaaam .m FN. amw. so. No. _~. amm. aom. KN. mm.~ ou~e__u movuw>—uo< ucmvaum .m mo. co. co. .p. we. mp. mp. om.~ ouoev_u es—zupcczu .5 am“. am“. «mm. anm. awm. em. m_.m .u._ mucecou:_uz .oozom .w oo._ am“. aem. com. me. ~_.m .u.~ meow» -23. SEEES .m awn. acw. com. no. P—.n .u.~ u:~qu_o>ma awaum .e «Po. «mm. mm. _o.m .u.~ mouw>c~m utoqqam .m amm. no. ~_.m .u.~ . mapup>auu< acouaum .~ ace. mo.m .u.~ es—aupscau .— m— N— o— m— v— m— NP —— op m w s o m v n N — .o.m m mco_mcme_o Low>acom oucoELouLma nee .uuue__u .A.u.~v mowumwcouuocacu _eavw>wuc~ acoe< weapon—oLLOUqu=~ tea mcovuow>oo uLoUcaum .mcoutii.u u4m

vuc~ .N awum mm.i —m.m mo. MN. «Lunch .— noun 82.5 x .84 32:... .3 3.55; 0.5; 8:25 a a e 832...; 8:38: muco_uwmuoou cowmmocmoa uo~wvcoucoum u amum mm N opnwupaz «3.323 cofimmcmmm mo acme—Em ":22er 3.53.52 «c2531.: 39: mo. v NC .m.— m~.- om.~- mF.- om. Fo. m~.~ mp. mm. macaw—u x .u.~ .e amum mp. ~—. op.- amo.m co. .ms.m FF. cm. mumewru .m amum ON. A... aoo.n co. apm.m mo. mm. A.U.Hv muwumvcmuumcmcu mN. Fmauw>wuc~ .N ampm mp.- p~.~ co. mp. «taco» .p amum muoe_ru x .u.H owner—u mwwH- mcacme o_umm mucosa a a a mpnmwco> cowuorumca muco_uwccoou co_mmmmmwm umNWuccccnum u amum mm N apnea—3: t I. III I I'll Ii‘filhl... a!“ all,” “I‘M.“ E14! I‘ll nan-I'll E mwmx_cc< co_mmmcmom co xcoesam ”cowmcoe_o a_;mcmccm4 poco_uuacumc~ use Eapsowccau--.mp m4m.ec. .~ aaem e..- m~.~ me. e.. etscae ._ amem aaas__e . .e.. aaae._e .e.. acscae a...“ manage a we a apna.ea> =o_ee.amta maea_e..cmoe eopmmacmam ea~.ecaecaem . aaem ~¢ opa.e.=z 39:22 5.53.33. we bee-5m E3233 5.32:3. use 22322.89 teem--.op “SE mo. v m. cm... em. mm.. o~.- .e.. .9. me. eo. «N. meme._e x .e.. .e amem eo. eo.- o~.- me. eo. mo._ mo. Nw. meme..e .m amen mo.- .~.- co. 00. em.. me. .N. ..e... no.3..eoeuaca5e paae.>.e=. .N aaem .~.- em.~ ae. .N. acacee ._ amem meae__e x .e.. meee_.e .e.. acacoe c.3aa manage a a e a.na.ea> eo.ue_eaca £53.:er 5.5333. eefieceeeeum . eoum Na N 323:: flares. 53.3.53. we D253 "eo.m=oe.a mou_>tam “Loans” .oo;um--.m. msmrec. .N amum Pm. acacoh .— emum m open—Le> :e.uuweoca ope.upez 73 mwmxpee< ce_mmocmom we xgeEEem ”ce_mcoe.o euceceucwex peep; Feecem-- m. mem

wec. .N aoum NN. ounceh ._ emum m eFeewce> :e_uu_eeca 32:32 mwmxpec< ce_mmecmem we xLeEEem ”ce_m:oe_o mcewuepem A».c:EEeu--.~_ memqe 74 school plant maintenance dimension. However, these significant increases in R2 in this dimension are impossible to interpret, given the nonsignificant overall F. Although only three of the six regressions showed a signi- ficant level of performance behavior accounted for, two of these 2 dimensions included a significant increase in R when the climate “1., scales were added to the equations. It appears that a climate for curriculum and instructional leadership and a climate for community relations could account for greater performance levels within these criterion dimensions. These results suggest that the degree of information concerning appropriate behaviors for effective community relations and curriculum leadership that exists in an environment can have a marked effect on what types of behavior a school adminis- trator will exhibit. As expected, no evidence of a multiplicative interaction was found in either of these dimensions. Rather than affecting the behavior of various principals in a different manner, climate was found to increase performance levels for all of the administrators. CONCLUSIONS Although combined into the field of Industrial[Organizational Psychology, traditional industrial psychologists and organizational psychologists have approached the problem of predicting work behavior from two different perspectives. Industrial psychologists have dealt almost exclusively with individual difference variables and organi- zational psychologists have focused on organizational properties and process variables. The present research effort has integrated these two different perspectives in an attempt to better understand and, hopefully, better predict work behavior. The phenomena of the validity ceiling and the vanishing validity coefficient have been pointed to as two major areas of concern for industrial psychologists that cannot be understood or alleviated within the research boundaries of the traditional selection model. This model has been described as a trait-oriented research paradigm that denies the existence of individual characteristic/work environment interaction processes. The traditional selection model is hampered by two assumptions of human behavior that do not allow for an examination of individual] environment interaction processes. The experience in a work environ- ment is assumed to have no systematic effect on the characteristics on which selection is based. Additionally, there is an implicit denial that the work environment and the person characteristics interact to produce certain types of behavior that, in turn, lead 75 76 to a certain level of performance. The research reported here has attempted to demonstrate the inaccuracy of this assumption. In particular, the concept of a "behavioral consistency hypothesis" has been questioned. This hypothesis, which provides the basic theoretical framework for the assessment center method, assumes that those behaviors exhibited or inferred during the testing phase of selection research accurately represent those behaviors that will be displayed in the actual work environment. However, the effect of environmental influence on performance behavior has been well documented. Research reviewed has shown that type of reward, contingency, job characteristics and leadership style] management philosophy have marked effects on levels of performance. However, the studies reviewed have generally treated the situation as an inference of an individual's internal motivational state. The present study has addressed the question of how the environment might effect the expression of expected behaviors as measured by employment testing from a social learning perspective. Based on Mischel's (1977) work in the area of personality research, it was hypothesized that different work environments provide various levels of information concerning apprOpriate effective behavior. In particular, certain school environments were thought to directly influence behavior exhibited by school administrators by the level of information concerning appropriate behavior inherent in that environment. This level of information concerning appropriate behavior was tapped via the construct of climate. The concept of "climates for the expression of specific criterion behavior" was 77 developed and high levels of interrater reliability within schools provided evidence that such climates did exist. In particular, climate scales were constructed for each of seven behavioral criteria of interest. It was expected that highly developed climates would provide a principal a rich source of information concerning approp- riate effective behaviors. As a result, these climates would increase the level of prediction above that of individual characteristics inferred from the behaviors exhibited in an assessment center. Additionally, the typical empirical indices of such individual] situation interactions has been challenged. It has been suggested that multiplicative interaction terms are inappropriate for field research and might prove misleading when a climate approach to index-- ing the situation is employed. Thus, it has been suggested that the concept of interaction can also include the presence of “additive influence" of both individual and situational variables on behavior. The results of the hierarchical regression analyses showed that an increase in prediction was evident with the inclusion of climate scores in two of the six criterion behavior dimensions examined. In the curriculum and instructional leadership and com- munity relations dimensions, social learning cues inherent in an environment, as measured by climate, apparently informed all admin- istrators what behaviors should or should not be exhibited. This provides evidence that, indeed, the characteristics of a work environ- ment can have a direct effect on performance behaviors. Evidence of the "missing link" of environmental influences between the measure of individual characteristics as predictors and the measure of 78 criterion behavior suggests that the phenomena of the validity coefficient and the vanishing validity coefficient may, in reality, be pseudo-issues. By incorporating criterion relevant situational characteristics into the framework of the traditional selection model, low or vanishing levels of prediction may be alleviated. Limitations and Suggestions for Future Research Although increased prediction was evident in two of seven criterion dimensions by the inclusion of a climate predictor, this provides scant evidence of a social learning environmental influence on performance behavior. In light of the results, the hypothesis that the concept of climate can be utilized to tap the degree of social learning cues available to the individuals in a particular environment must be questioned. Climate, as operationalized in this study, is premised on the notion of a shared psychological reality. However, this assumption that the climate measures used adequately reflected the degree of social learning information inherent in the school environments must be investigated. Although the climate raters (i.e. teachers, students and support staff) within each school were found to be in agreement on the level of various climates, the administrators might not have shared in these perceptions. In parti- cular, the administrators investigated in the study might not have encoded the policies and procedures (climate) of the school environ- ment as social learning cues. Mischel (1973) has stressed the concept of idiosyncratic stimulus meanings in his discussion of social learning cues. 79 Individuals cognitively transform the meaning and impact of stimuli in any given situation because of idiosyncratic social learning histories. Climate, as employed in this study, assumes that indivi- duals experiencing the same environment over time or choosing to select into a particular environment share common encoding schemes. However, many of the administrators in the sample were relatively new to their schools. It is likely that these principals brought to their new school environments social learning encoding schemes from their previous school experiences. In future research, the assumption that the concept of climate adequately measures social learning cues must be tested. In parti- cular, the degree to which climate induces appropriate performance behaviors for various individuals in an environment must be studied. Parenthetically, the degree to which individuals perceive adequate incentives to perform these behaviors should also be of interest. This suggests a movement away from simply indexing shared perceptions of climates and towards the measurement of individual psychological climate perceptions. Additionally, evidence of a climate effect on performance in the other behavioral dimensions might have been obfuscated by the sample of school administrators examined. This sample included a number of assistant principals whose presence might have confounded the relationships of interest. Although assistant principals are certainly important members of an administrative team within a school, few, if any, have all the responsibilities of the principalship. 80 In many schools the assistant principal is designated a specific function such as student discipline or curriculum development. As a result, these assistant principals' activities and behaviors are limited and an evaluation of climate effects on and the evaluation of certain areas of criterion performance might have been unwarranted. Additionally, these two levels of school administration might have encoded climate perceptions differently in relation to appropriate behavior. Hopefully, future research efforts would include only principals or investigate only those behavioral domains relevant to speCific school administrators. The present research was also hampered by the fact that the reciprocal influence of the situation and the individual character— istics on one another could not be examined. To fully understand predictor-criterion links, research must seek empirical evidence of situational influence on performance behavior, and on the level of individual characteristics measured as predictors. A longitudinal data strategy that provides for a reevaluation of individual charac- teristics after exposure to a particular work environment is a logical next step. Such a strategy would provide for a more thorough test of a social learning hypothesis in that preemployment behavior can be compared with learned apprOpriate behaviors. Although the identification and understanding of predictor- criterion links does not necessarily provide for greater utilitarian prediction than the traditional model, it does provide a key data source when one considers long-term prediction. If personnel researchers understand how behavior "evolves" in a given work 81 environment, long term prediction is possible. This might take the form of two different kinds of prediction in personnel research: (1) prediction of short-term performance and (2) prediction of expected performance after exposure to a given situation over a period of time. Suchauiexpanded view of prediction would not only increase long-term prediction but serve as an aid to management development and affirma- tive action programs. Personnel researchers could become more proactive in terms of placing employees in situations which could allow them to develop needed abilities and personal characteristics. 82 APPENDICES 83 APPENDIX A BEHAVIORAL PERFORMANCE EVALUATION SCALES 84 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. The examples to the right are examples of behavior of in- dividuals who are usually rated "High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Law" on this dimension. Maintenance of School Plant Typical High Behavior -Works with new students to develop school pride 6 reduce vandalism. Older students are given responsi- bility for taking new students "under_their wing". -Initiates a program to clean up graf- fiti in school; provides students with cleaning materials 6 develops a con- test for cleanest area thereby unifying students 6 staff. ~8stablishes a process which allows staff 6 community members to provide input for the orderly improvement of school plant facilities 6 equipment. Typical Average Behavior -Takes an active part in keeping the building clean (e.g., picking up the paper in hallways, repairing bulletin boards). -Obcains necessary materials when things need to be fixed for ongoing daily functions in the building. -Set up a graffiti board in the restrooms to minimise vandalism. Iypicai Low Behavior -Unilatera11y determines which school facilities are to be improved. -Does not provide a safe storage place for expensive merchandise. -Never foresees equipment problems; reacts to peoples' requests 6 emergencies. -Ignores a report of unsafe playground equipment until something happens to a child. 85 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. The examples to the right are examples of behavior of in- dividuals who are usually rated "High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average” on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Law” on this dimension. Fiscal or Monetary Management Typical High Behavior -Involves all staff in establishing priorities for the allocation of resources 6 materials. -Periodically reviews 6 shifts alloca- tions based on current needs; obtains agreement from all department heads. -Maintains strict requirements on the procedures to follow with purchase orders 6 budgets of departments. -Acquaints the staff with the daily operations budget and how it relates to their curriculum. -Calls several other schools 6 sources to get materials when they are not available at the local school. IIQical Average Behavior -Honitors expenditures of students' activity funds. -Submits a yearly budget to the Board of Education or district. -Provides funds for special art pro- jects so that students would not need to bring materials from home. -Keeps staff informed of budget. Typical Low Behavior ~Does not comply with standard accounting procedures. -Hisspends school money in one area 6 compensates by taking money out of budget in another area. -Spends an inordinate amount of school funds on athletics. -Requires staff to fill in requisi- tion forms for small items such as toilet paper 6 light bulbs. 86 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. Coordination with District and Other Schools The examples to the right are examples of behavior of in- dividuals who are usually rated "High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Law" on this dimension. Typical High Behavior -Participates in professional organi- zation problem solving projects aimed at improving the functioning of central administration services which impact directly on the school. -Balances district activities against building priorities. ~Encourages use of district resource personnel to develop needed programs. -lnforms central office of possible complaints or problems at school or in community so they will be aware of it ahead of time. Typical Average Behavior -Arranges meetings wiifi’central of- fice administration to speak to staff at the school on various issues. -Implements decisions at the building which were made at the district level even if not involved in making decisions. -Participates in district-level committees when called upon. Typical Low Behavior -Hakes all decisions without the advice of any immediate superior. -ls so involved in district committees that he/she is always out of the building. -Does not communicate with district personnel unless absolutely necessary. -Hever volunteers for district committees. O 87 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. The examples to the right are examples of behavior of in- dividuals who are usually rated "High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Law" on this dimension. Community Relations: Parents Iypical High Behavior -Starts a procedure whereby parents come to meet teachers individually to discuss child's progress. ~0rganises coffees in parents' homes to interact with parents in a nonschool environment. -Writes a letter to all parents inviting them to school. spends an evening talking to them 6 answering questions. -Distributes a monthly newsletter to parents. as well as an annual report discussing the progress of the school in meeting its goals 6 objectives. Typical Average Behavior -Initiates a program in which parents follow the class schedule of their students for a shortened day. -Organizes a system for buses to pick up parents 6 bring them to school events in order to promote community involvement. -Calls parents to remind them of parent-teacher meetings so as to in- sure attendance. Iypical Low Behavior -Refuses to develop or initiate a parent advisory group. -Controls P.T.A. by dominating of— ficers, promoting participation of only one segment of school population, or failing to follow through on their requests or suggestions. -Does not have a system of reporting students' discipline problems to parents. . -Tells parents they do not really know their chldren when problems arise. Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. The examples to the right are examples of behavior of in- dividuals who are usually rated “High” on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usual ly rated "bow" on this dimension. Interpersonal Effectiveness Typical High Behavior -Shows a sense of humor in times of conflict. pinteracts with students during the lunch hour in the cafeteria. -Encourages parents to go directly to a teacher with a compliment in- stead of relaying the message him/ herself to the teacher. -Displays students' art work in main office as encouragement. -Comes from behind desk to discuss problems. takes notes. 6 refers to this file later when another discussion takes place. Typical Average Behavior -Leaves a note of encouragement to someone who has made a mistake to carry on 6 start over. -Provides staff with knowledge of all other faculty's contributions to school activities 6 communicates appreciation. -Allows staff member to go home during the day to solve a family problem. Typical Low Behavior -Delegates work but then takes work away or interferes because he/ she thinks it is not getting done correctly. -Uses language 6 words which the students do not understand when communicating with them. -£stablishes a policy without giving staff or students a rationale. -Gets into shout-outs with students in the hallways. 89 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. The examples to the right are examples of behavior of in- dividuals who are usually rated ”High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Law" on this dimension. Community Relations Typical High Behavior -Informs community about the basics of operating the school 6 educational objectives; asks for the ideas. opinions, 6 support of these people 6 listens to their suggestion. ~Works with various community 6 local groups to develop cooperation with the school (e.g., coop programs). -Institutes a course which is aimed at developing cultural awareness in an integrated community. Typical Average Behavior -Sands stories to the community news- paper and/or invites newspaper correspondents to school events. -Organizes senior citizen breakfasts to explain school programs. -Coordinates school activities with church 6 community events. -Takes 2 students to the Rotary Club every month as a reward 6 to promote good relations. Typical Low Behavior -Stays away from community organiza- tions such as Jaycees, Lions. Kiwanis. etc. -Defends his/her position to the community rather than listening to their requests for certain programs. -Speaks in educational jargon to explain school programs to community members. -Spesks badly of the school system to the news media. 90 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. The examples to the right are examples of behavior of in- dividuals who are usually rated ”High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "W" on this dimension. Developmental Activities Typical High Behavior -Initiates management by objectives. provides training to write objectives. to determine ways to evaluate. 6 sets a time frame for accomplishment of objectives. -Provides in-service programs for staff which include dealing with student be- havior problems and interactions with parents. ~Attends sesdnars. workshops. etc. for professional growth. -Provides time for teachers who have attended developmental activities to communicate information to the rest of the staff. Typical Average Behavior -Devises class schedule to allow for staff development periods. -Uses staff meetings to encourage and publicise in-service training. -Allows teachers time off to attend meetings and classes. Typical Low Behavior -Tells staff to get involved in staff development problems when helshe does not get involved. -Plans staff development activities with no staff input. -Does not help teachers who have problems but works toward docu- menting problem 6 removing teacher. -Does not allow staff to go to professional meetings even on their own expense. L— 91 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. The examples to the right are examples of behavior of in- dividuals who are usually rated "high" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Low" on this dimension. Staff Evaluation Typical High Behavior -Provides constructive feedback after observations a calls in specialists if additional help is needed; then observes again 8 gives more feedback. -£ncourages evaluation of him/herself by staff. -Consults with individual staff members on a periodic basis to develop in- dividual standards of performance (goals 8 objectives) and reviews subsequent accomplishment of goals. -Hrites a note to a teacher who has led a particular event or done some- thing special to show appreciation and give praise. Typical Average Behavior -Disciplines employee wh6*is not functioning properly. -Asks permission to come to observe classroom. Izzig§l_Low Behavior —Does not question a teacher before taking disciplinary action on the word of 2 of the teacher's students. -Circulates a staff bulletin criticizing teachers when only 3 or 4 teachers were guilty & should be reprimanded individually. -Criticises or belittles staff mem- bers in front of other staff members. students. or parents. 6.—_W__— 5._____. 4.______ 3 ._____. 2 ______ 1 ._____ 92 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. Support Services: Directing the Behavior of Students The examples to the right are examples of behavior of in- dividuals who are usually rated “High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated “Low" on this dimension. Typical High_§ehavior -Haintains up-to-date staff manuals. including statements on discipline policies. which communicate all pro- cedural matters. -worka with teachers I students to handle discipline problems. -Relps counselor develop programs for incomdng students. such as orientation. -Sets guidelines for student behavior which are not threatening a enforces them objectively. -Conducts a cafeteria survey to determine student needs in that area. Typical Average Behavior -Develops a staggered release schedule on rainy days to alleviate traffic jams in front of school caused by parents picking up their children. -Develops special instructional pro- grams for lunch period so that students do not just play in gym or hang around. -Starts a program in which highest attending classroom is recognized with a free period 6 a party. Typical Low Behavior -Discontinues an arts program because a small group of students ruined a kiln. -Suspends a student with whom the counseling staff is working closely-~ does not check with staff first. —Does not discipline students because helshe does not know what to do or does'not have time. -Places a student in a special educa- tion program for reading problems without proper evaluation. 93 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. The examples to the right are examples of behavior of in- dividuals who are usually rated "High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average" on this dimension. The examples to the right are examples of behavior of in— dividuals who are usoally rated "Law" on this dimension. Direction of Support Services Typical High Behavior -Devises a system whereby librarianis able to coordinate library resources (books. magazines) with needs of teachers as indicated by lesson plans. -lncludes clerical/custodial staff in all school parties & meetings -Verbally acknowledges the completion of tasks by school maintenance a food services personnel. Typical ngrage Behavior -Initiates a program to have parents receive first aid training so there can be personnel in nurse's office during lunch hour. -Organizes groups of support services people so they have a more unified voice. -Asks cafeteria & custodial staff for input in school decisions. -Asks teachers to periodically complete a form concerning the quality of support services as an aid in monitoring these activities. Typical Low Behavior —Assigns hall 8 cafeteria duty with- out staff input. -Allows a community group to use a school projector without the permission of the media specialists. -Does not consult with custodians when making requisitions for supplies. -Neglects to inform secretary of school programs 6 events. 6——]—— 94 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. The examples to the right are examples of behavior of in- dividuals who are usually rated "High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated ”Low" on this dimension. Student Activities: Participation Typical High Behavior -lnitlates a program so that handi- capped students are incorporated into various activities, such as plays or sports. -Participates in extracurricular school activity by actually working at the function. such as fun fair. schoo l dinne rs . -Sets up class for student government officers to help them solicit student participation. Typical Average Behavior -Talks to students who are not parti- cipating in extracurricular activities to get them involved. ~Attends extracurricular activities. -Encourages staff to participate in extracurricular activities but does not demand such participation. Typical Low Behavior -Tells the students that certain activities are important a then doesn't support them by his/her presence or allowing time in the schedule for them. -Selectively participates in only certain school activities (e.g. football). 95 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. Coordination of Student Activities: Supervision The examples to the right are examples of behavior of in- dividuals who are usually rated "High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated ”Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Low" on this dimension. Typical High Behavior -Initiates an awards banquet B esta- blishes a letter for superior academic performance. -Evaluates all activities a student needs with input from faculty. student council. S student surveys. then acts on results. -Heets regularly with student leaders to coordinate activities a take suggestions. -Reschedules a homecoming event which had been scheduled for the night before the SAT. Typical Average Behavior -Schedules a sporting event during the academic day for the benefit of bussed and poor students who can not come to night activities. -0rganizes parent-student teams to work on fund raising activities. -Organizes faculty teams to play against varsity teams. Typical Low Behavior -Organizes a student fund raising activity without talking with students. -Cancels a pep meeting because of a drinking party by a few students. -Allows extra-curricular activity 8 school classes to conflict so that students have to be pulled out of class to participate in band or sports. Check to the left if you have insufficient information with which to make a judgment: about this individual's behavior for this dimension-of perfrormance. Currziculum & Instructional Leadership: .Honitoring individual Progress Typical High Behavior The examples to the -lnitiates a program with the help of right are examples of behavior of in- dividuals who are usually rated "High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "bow" on this dimension. student-parent associations to allow students to take advanced classes offered at nearby schools. -Organizes student help sessions which meet after school hours for students who are failing. T gigglnéyerave Behavior -Reviews records of individual—student progress - including standardized scores 6 basic skills program. ~Institutes a classroom where several things are taught (e.g., ceramics. painting) because of a lack of enough interest in any one of them to justify individual classes. -Sponsors a biweekly reading rally in which all students stop what they are doing 8 read for 20 minutes. Typical Low Behavio; -Allows students to hand carry scores of certain tests home; students are able to open these unsealed envelopes and compare scores. -Schedules all advanced courses in the morning so that academic students can take only 2 out of S. 97 Check to the left if you have insufficient information with which to make a judgment about this individual's behavior for this dimension of performance. The examples to the right are examples of behavior of in- dividuals who are usually rated "High" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated “Average" on this dimension. The examples to the right are examples of behavior of in- dividuals who are usually rated "Law" on this dimension. 'Curriculum 5 Instructional Leadership: Honitoring Curriculum Objectives Typical High Behavigg -Takes a stand on issues even when they are controversial; supports 8 stands up for educational values. -Constructs a schedule which maximizes student options, yet miniuutes course conflicts. -Visits classrooms to monitor the cur- riculum actually being taught in school. 112353} Average Behavior -Reads necessary materials to have a thorough knowledge of state or district curriculum standards 8 requirements. -Looks at new texts so that helshe can influence the selection of materials and textbooks. -Reviews 8 makes comments on weekly lesson plans in terms of overall curriculum objectives. Typical Low Behavior -Does not utilize special curriculum programs provided by the district. —Adopts curriculum materials without the input of teachers or in spite of their recommendations. -Does not allow any curricular experi- mentation by faculty members. 98 APPENDIX B CLIMATE SURVEY #00“) 01 10. 11. 12. 13. 14. 15. 16. 99 ASSESSMENT OF SCHOOL CLIMATE Curriculum and Instruction Sometimes students help the teacher plan class activities. Teachers encourage students to think for themselves. Students are assigned too much homework. There are opportunities for small group work and/or independent study. Students feel too much pressure to get good grades. Studentsirnthis school are highly motivated and involved in learning. This school offers the courses students need and would like to take. Teachers challenge students to work to the best of their abilities. Student Activities A sufficient variety of student activities and clubs is available. There is equal opportunity for all to participate in student activities and clubs. Teachers seem interested in sponsoring clubs and activities. The student government is effective. A fair amount of emphasis is given to the athletic programs. Support Services The school rules and regulations are too strict. Students behave well at school (in and outside of classes). Rules on tardiness and class attendance are enforced regularly. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 100 Students are not safe in school. Discipline is administered fairly at this school by teachers and principals. Drugs and alcohol are easily available at this school. Students are treated fairly. The library/media center adequately supports the instructional program and is useful. I am satisfied with the report card used at this school. Counselors and teachers cooperate in giving students extra academic or personal help. It is easy to arrange a conference with guidance counselors. I am satisfied with the quality of the guidance service students receive. Counselors and teachers cooperate in helping students select appropriate courses. Students feel free to talk to counselors about personal problems. Counselors help students make choices about future careers. The transportation of students is provided for in an orderly, safe manner. Lunch period is always a time of chaos; there is little super- vision of students in the cafeteria and the cafeteria staff and school administrators do not seem to cooperate. Secretarial staff seem genuinely helpful to students, teachers, and public. Staff Evaluation and Development Teachers and/or staff receive helpful supervision. There is strong support for professional development activities. Staff development programs meet staff needs. Teachers are well organized and make good use of class time. As a whole, members of the staff respect one another. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 101 Community Relations Businesses/homeowners near the school grounds have relatively few problems with students. Parents feel welcome in the school. Parents are promptly informed about students' academic and behavior problems. Parents received sufficient information about their children's academic achievement. Parents' opinions are requested and respected. Parents received information about curriculum objectives and requirements. The principal and teachers are involved in community activities and affairs. Maintenance of School Building and Facilities School facilities and equipment are adequate. Measures are taken to stop vandalism of the school plant. The school grounds are kept in good condition. Graffiti is at a minimum in the school. 102 REFERENCES REFERENCE NOTES Dunnette, M. D. Performance equals ability_and what? Univer- ity of Minnesota, Department of Psychology,_Technical Report No. 40009, 1973. Howard, A. Intrinsic motivation and its determinants as factors enhancing_theprediction of job performance from ability. Unpublished ONR Technical Report,Department 0? Psychology, University of Maryland, 1976. James, L. R. Organizational climate: Another look at a poten- tially important construct. Unpublished manuscript, 1981. Joyce, N. F., & Slocum, J. Correlates of climate discrepancy. Unpublished manuscript, 1981. Schmitt, N., Meritt, R., Fitzgerald, M., & Noe, R. Results of the validation of the NASSP Assessment Center, Michigan State University, Department of Psychology, Research Report No. 81-4, 1981. Schneider, B. Organizational type, organizational success, and the_prediction of individual performance. University of Maryland, Department of Psychology, Research Report No. 6, 1974. Schneider, B. Nork climates: An interactionist_perspective. Unpublished ONR Technical Report, Department of Psychology, Michigan State University, 1981. 103 REFERENCES Allport, 6. Personality: Aypgychologjcal interpretation. New York: Holt, 1937. Alvares, K. M., & Hulin, C. L. Two explanations of temporal changes in ability-skill relationships: A literature, review and theoretical analysis. Human Factors, 1972, 14, 295-308. Alvares, K. M. & Hulin, C. L. An experimental evaluation of a temporal decay in the prediction of performance. Or ani- zational Behavior and Human Performance, 1973, 9, 69-185. Anastasi, A. Psychological testing (3rd ed.). New York: Macmillan, 1968. Boehm, V. R. Are we validating more but publishing less? (The impact of governmental regulation on published validation research--an exploratory investigation). Personnel Psychol- ggy, 1982, §§, 175-187. Borman, W. C. The rating of individuals in organizations: An alternate approach. Organizational Behavior and Human Performance, 1974, 1g, 105-124. Bowers, K. S. Situationism in psychology: An analysis and critique. ngchological Review, 1973, pp, 307-336. Bray, D. W., Campbell, R. J., & Grant, D. L. Formativegyears in business: A long-term A.T.&T. study of managerial lines. New YBrk: Wiley, 1974. Brousseau, K. R. Personality and job experience. Organizational Behavior and Human Performance, 1978, 22, 235-252. Campbell, J. P., Dunnette, M. D., Lawler, E. E., & Weick, K. E. Managerial behavior, performance and effectiveness. New YOrk: *Macmillan, 1973. Campion, J. E. Work sampling for personnel selection. Journal of Applied ngchology, 1972, 56, 40-44. Chaplin, J. P. & Krawiec, T. S. (Systems and theories of_p§ychology (2nd ed.). New York: Holt,7Rinehart & Winston, 1968. 104 105 Cooper, M. An exact probability test for the use with Likert-type scales. Educational and ngchological Measurement, 1976, §§, 647-655. Cronbach, L. J., & Snow, R. E. Aptitudes and instructional methods; A handbook for research on interactions. New York: Irvington, 1977. Dachler, H. P., & Mobley, W. H. Construct validation of an instrumentality-expectancy-task-goal model to work motiva- tion. Journal of Applied Psychology, 1973, 68, 397-418. Dreher, G. F., & Sackett, P. R. Some problems with applying content validity evidence to assessment center procedures. Academy of Management Review, 6, 551-560. Dunham, R. B. Ability-skill relationships: An empirical explana- tion of change over time. Organizational Behavior and Human Performance, 1974, 3 (12), 372-382. Dunnette, M. D. A modified model for test validation and selection research. Journal of Applied Psychology, 1963, 41, 317-23. Dunnette, M. 0. Personnel selection and placement. Belmont, Calif.: Wadsworth, 1966. Endler, N. S., & Magnusson, D. (Ed.). Interactional psychology and personality. New York: Hemisphere, 1976. Endler, N. S. & Magnusson, D. Toward an interactional psychology of personality. Psychological Bulletin, 1976, 83, 956-974. Finn, R. H. A note on estimating the reliability of categorical data. Educational and ngchological Measurement, 1970, 89, 956-957. Forehand, G. A. On the interaction of persons and organizations. In Tagiuri and G. Litwin (Eds.), Organizational climate: Explorations of a concept. Boston: *Division of Research, HarvardTBusiness School, 1968. Forehand, G. A., & Gilmer, B. V. H. Environmental variation in studies of organizational behavior. Psychological Bulletin, 1964, pg, 361-382. Fiedler, F. E. Engineering the job to fit the manager. Harvard Business Review, 1965, 115-122. Fiedler, F. E. A theory of leadership effectiveness. New York: McGraw-Hill, 1967. 106 Finn, R. H. A note on estimating the reliability of categorical data. Educational and Psychological Measurement, 1970, pg, 71-76. Frederiksen, N., Jensen, 0., & Beaton, A. A. Organizational climate and administrativgiperformance. Princeton, N. J.: Educa- tional Testing Service, 1968. Freyd, M. Measurement in vocational selection: An outline of research procedure. Journal of Personnel Research, 1923, 1, 268-284; 377-385. Ghiselli, E. E. The validiiy of occupational aptitude tests. New York: Wiley, 1966. Ghiselli, E. E. The validity of aptitude tests in personnel selec- tion. Personnel P§ychology, 1973, 26, 461-477. Guetzkow, H., & Forehand, G. A. A research strategy for partial knowledge useful in the selection of executives. In R. Tagiuri (Ed.), Research needs in executive selection. Boston: HarvardTGraduate School of Business Administration, 1961. Guion, R. M. Personnel testing. New York: McGraw-Hill, 1965. Guion, R. M. On trinitarian doctrines of validity. Professional Psychology, 1980, 11, 385-398. Guion, R. M. Recruiting, selection and job placement. In M. D. Dunnette (Ed.), Handbook of Industrial-Organizational Psychology. Chicago: Rand-McNally, 1976. Guion, R. M., & Gottier, R. F. Validity of personality measures in personnel selection. Personnel Psychology, 1966, 18, 135- 164. Hackman, J. R., & Lawler, E. E. Employee reactions to job character- istics. Journal of Applied Psychology, 1971, §§, 259-86. Hackman, J. R. Group influences on individuals. In M. D. Dunnette (Ed.), Handbook of industrial and organizationai_p§ychology. Chicago:TRand‘McNally, 1976. Hinrichs, J. R. An eight-year follow-up of a management assessment center. Journal of Applied Psychology, 1978, pg, 596-601. Hellriegal, D., & Slocum, V. W., Jr. Organizational limits: Measures, research, and contingencies. Academy of Manage; ment Journal, 1974, 17, 255-280. 107 Huck, J. R. Assessment centers: A review of the external and internal validities. Personnel ngchology, 1973, go, 191 212. Hull, C. L. Aptitute testiog. Chicago: World Book, 1928. Humphreys, L. G. The organization of human abilities. American Psychologist, 1962, ii, 475-485. James, L. R. Aggregation bias in estimates of perceptual agreement. Journal of Applied ngchology, 1932s.§Z: 219-229. James, L. R., & Jones, A. P. Organizational climate: A review of theory and research. Psychological Bulletin, 1974, 81, 1096- 1112. James, L. R., & James, A. P. Perceived job characteristics and job satisfaction: An examination of reciprocal causation. Personnel Psychology, 1980,_g§, 97-135. Joyce, W. F., & Slocum, J. R., Jr. Climates in organizations. In S. Kerr (Ed.), Organizational Behavior. Columbus, Ohio: Grid, 1979. Kohn, M. L., & Schooler, C. Class, occupation, and orientation. American Sociological Review, 1973, i8, 659-678. Kohn, M. L., & Schooler, C. The reciprocal effects of the substan- tive complexity of work and intellectual flexibility: A longitudinal assessment. American Sociological Journal, 1978, 85, 24-52. Klimoski, R. J., & Strickland, W. J. Assessment centers--Va1id or merely prescient. Personnel ngchology, 1977, go, 353-361. Landy, F. J., & Farr, J. L. Performance rating. Psychological Bulletin, 1980, 81, 72-107. Latham, G. P., & Wexley, K. N. Increasing productivity through per- formance oppraisal. Reading, Mass.: Addison-Wesley, 1981. Lawler, E. E. Ability as a moderator of the relationship between job attitudes and job performance. Personnel Psychology, 1966, 19, 153-164. Locke, E. A., Mento, A. J., & Ketcher, B. The interaction of ability and motivation in performance. An exploration of the mean- ing of moderators. Personnel Psychology, 1978, ii, 269-80. 108 Magnusson, D. & Endler, N. S. Interactional psychology: Present status and future prospects. In D. Magnusson & N. S. Endler (Eds.), Personality at the crossroads: Current issues in interactionalpsychology. Hillsdale, N. J.: Lawrence Erlbaum Associates, 1977, pp. 3-37. McGregor, D. The human side of enterprise. New York: McGraw-Hill, 1960. Mobley, W. H., & Locke, E. A. The relationship of value importance to satisfaction. Organizational Behavior and Human Perfor- mance, 1970, 5, 463-483. Mischel, W. Personality and assessment. New York: Wiley, 1968. Mischel, W. Toward a cognitive social learning reconceptualization of personality. Psychological Review, 1973, 89, 252-283. Mischel, W. The interaction of person and situation. In D. Magnusson & N. S. Endler (Eds.), Personality at the cross- roads: Current issues in interactional psychology. Hillsdale, N.J.: Lawrence Erlbaum and Associates, 1977, pp. 333-353. Mitchel, J. 0. Assessment center validity: A longitudinal study. Journal of Applied Psychology, 1975, 69, 573-79. Mitchell, T. R. Organizational behavior. In M. R. Rosenzweig & L. W. Porter (Eds.), Annual review of psychology, (Vol. 30). Palo Alto: Annual Reviews, Inc., 1979, 477-525. Muensterburg, H. Psychology and industrial efficienoy. Boston: Houghton-Mifflin,719l3. Naylor, J. C., Pritchard, R. D., & Ilgen, D. R. A theogy of behavior in organization. New York: Academic Press, 1980. Newman, J. E. Understanding the organizational structure--job attitude relationship through perception of the work environ- ment. Organizational Behavior and Human Performance, 1975, 14, 371-3970 Norton, S. D. The empirical and content validity of assessment centers vs. traditional methods for predicting managerial success. Academy of Management Review, 1977, g, 442-453. Payne, R. L., & Pugh, S. S. Organization structure and organization climate. In M. D. Dunnette (Ed.), Handbook of industrial and organizational psychology. Chicago: Rand MCNally, 1976. 109 Pearlman, K., Schmidt, F. L., & Hunter, J. E. Validity generali- zation results for tests used to predict training success and job proficiency in clerical occupations. Journal of Applied Psychology, 1980, 62, 373-406. Peters, L. H., & O'Connor, E. J. Situational constraints and work outcomes: The influences of a frequently overlooked construct. Academy of Manogoment Review, 1980, 6, 391-397. Peters, L. H., O'Connor, E. J., & Rudolf, C. J. The behavioral and affective consequences of situational variables relevant to performance settings. Organizational Behavior and Human Performance, 1980, 66, 79-96. Pervin, L. A., & Lewis, M. Overview of the internal-external issue. In L. A. Pervin & M. Lewis (Eds.), Perspectives in inter- actional psychology, New York: Plenum, 1978. Pugh, D. S., Hickson, D. J., Hinings, G. R., & Turner, C. Dimensions of organizational structure. Administrative Science Quarterly, 1968a,_16, 65-105. Pugh, D. S., Hickson, D. J., & Hinings, G. R. Dimensions of organi- zational structure. Administrative Science Quarterly, l968b,_ii, 91-114. Porter, L. W. Personnel Management. Annual Review of Psychology, 1966, 11, 395-422. Porter, L. W., & Lawler, E. E. Properties of organizational struc- ture in relation to job attitudes and job behavior. Psychological Bulletin, 1965, 63, 23-51. Roberts, K. H., Hulin, C. L., & Rousseau, D. M. Developing an interdisciplinary science of organizations. San Francisco: Jossey-Bass, 1978. Schmidt, F. L., & Hunter, J. E. Moderator research and the law of small numbers. Personnel Psychology, 1980, 66, 41-60. Schmidt, F. L., Hunter, J. E., Pearlman, K. C., & Shane, G. S. Further tests of the Schmidt-Hunter Bayesian validity generalization procedures. Personnel Psychology, l979,_§2, 257-2181. Schmidt, F. L., Hunter, J. E., & Pearlman, K. Task difference as moderator of aptitude test validity in selection: A red herring. Journal of Applied Psychology, 66, 1981, 166-185. 110 Schmidt, F. L., Hunter, J. E., & Urry, V. W. Statistical power in criterion-related validity studies. JoUrnal of Appiied Psychology, 1976, 61, 473-485. Schmitt, N., and Schneider, 8. Current issues in personnel selec- tion. In K. M. Rowland & G. D. Ferris (Eds.), Research in personnel and human resources manogement. Greenwith, Conn.: JAI Press, Inc., 1983. Schneider, 8. Organizational climates: An essay. Personnel Psychology, 1975, 26, 447-479. Schneider, 8. Person-situation selection: A review of some ability- situation interaction research. Personnel Psychology, 1978a, _61, 281-297. Schneider, 8. Implications of the conference: A personal view. Personnel Psychology, 1978b, 61, 299-304. Schneider, 8., & Snyder, R. A. Some relationships between job satisfaction and organizational climate. Journal of Applied Psychology, 1975, 66, 318-328. Shrout, P. E., & Fleiss, J. L. Intraclass correlations: Uses in assessing rater reliability. Psychological Bulletin, 1979, 66, 420-428. Stern, G. G., Stein, M. 1., & Bloom, 8. 5. Methods of personality assessment. New York: Free Press, 1966. Taguiri, R., & Litwin, G. H. (Eds.). Organizational climate: Explorations of a concopt. Boston: Harvard University Press,_1968. Tenooyr, M. Trifling he stands. Personnel Psychology, 1981, 66, 1-18. Terborg, J. R. Validation and extension of an individual difference model of work performance. 0r anizational Behavior and Human Performance, 1977, 16, 88-216. Terborg, J. R., Richardson, P., & Pritchard, R. D. Person-situation effects in prediction of performance: An investigation of ability, self-esteem, and reward contingencies. Journal of (Applied Psychology, 65i, 574-583. Thornton, G. C. & Byham, W. C. Assessment centers and managerial porformance. New York: Academic Press, 1982. 111 Turnage, J. J. & Muchinsky, P. M. Transsituational variability in human performance within assessment centers. Organizational Behavior and Human Performance, 1982, 66, 174-200. Vroom, V. H. Work and motivation. New York: Wiley, 1964. Walberg, M. J. (Ed.). Educational environments and effects. Berkeley, Ca.: TMCCuthan Publishing, 1979. Weick, K. E. The social psychology of organiziog (2nd ed). Reading, Mass.: Addison-Wesley, 1979. Weinstein, A. G., & Holzbach, R. L., Jr. Impact of individual dif- ferences, reward distribution, and task structure on produc- tivity in a stimulated work environment. Journal of_Applied Psychology, 1973,.66, 296-301. Wernimont, P. F. & Campbell, J. P. Signs, samples and criteria. Journal of Applied Psychology, 1968, 66, 372-376. ' .