' - ‘ ’ ~ ‘1 . I v .R._¢..-n.;.z‘..~_‘_.~. “M- —<—..-._-.._ mm. \ « f-Y‘C- V ThEo...) j ‘ __ H i 1.5, .f W ..-" ’ w a; l "T Oran . R‘s”. w”. t? ! #kistg. ,{J‘gg-w :.. 1:4 7,;- ratifi- w ~ — -- 4'! .1. {g ‘r ‘1')! ’; fla'l U ., = v. _ ~._ :*-- g- M. 3 $54“ ‘4 {3% -:-_‘-a‘ ‘1‘” ‘g‘ 7' x V . 7 a .. I This is to certify that the thesis entitled THE PREDICTION OF SUCCESSFUL JOB PLACEMENT FOR UNEMPLOYED OLDER WORKERS RECEIVING ASSISTANCE FROM A SENIOR EMPLOYMENT SERVICE presented by JOSEPH M. BORNSTEIN has been accepted towards fulfillment of the requirements for M.A. Psychology degree in Major professorK—fi 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution MSU ’" lululllllllllm:lululummmmul'" 10481 4094 RETURNING MATERIALS: ' Place in book drop to remove this checkout from LIBRARIES .—:—. your record. FINES will be charged if book is returned after the date stamped below. i: 'C ' “‘~““‘Qk .l .371 IN; . ', THE PREDICTION OF SUCCESSFUL JOB PLACEMENT FOR UNEMPLOYED OLDER WORKERS RECEIVING ASSISTANCE FROM A SENIOR EMPLOYMENT SERVICE BY Joseph M. Bernstein A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1983 in O h )4 ABSTRACT THE PREDICTION OF SUCCESSFUL JOB PLACEMENT FOR OLDER WORKERS RECEIVING ASSISTANCE FROM A SENIOR EMPLOYMENT SERVICE BY Joseph M. Bornstein This study was designed to investigate the effects of unemployment on the occupational status and earnings of dis— placed workers 55 and older. Also examined were the varia- bles that predict successful job placement and job search behaviours. A sample of 26S CETA and non-CETA jobseekers was drawn from the clientekaof a Senior Employment Service (SES) located in southern Michigan. Archival records and. tele- phone interviews were the principle data collection methods. The findings indicated that occupational status did not vary after the client became reemployed. However there was some doubt regarding the validity of the status measure. When clients were reemployed they experienced a significant decrease in earnings (p<.05). Variables that predicted placement outcomes were either motivational or reflected the Human Capital value of the jobseeker. The results indicate that senior employment services should address their client's labour market expectations. To my Mother and Father, who laid the groundwork of my education ii ACKNOWLEDGEMENTS During the course of this study many people gave freely of their time and energy so that this thesis could reach fruitation. I would like to express my sincere appreciation to all of those people. Although space does not allow me to acknowledge everyone personally, I would like to mention some specific people. I owe a great deal of thanks to my advisor, Charlie Johnson. Charlie always made himself available when I needed advice and always helped me to exceed my own expectations. His confidence in my abilities to conduct this project aided me in overcoming my own self-doubt. Special thanks also goes to my other committee members, Neal Schmitt and Mary Zalesny. Each has made a significant contribution to this study and my professional growth. This study could never have been realized without the constant cooperation of the directors and staff of the Senior Employment Service. Their devotion to helping older workers enabled them to ask questions and look for answers beyond the boundaries of their own agency. I owe a great deal of thanks to them for their assistance, advice and insight. I also owe an unestimatable debt to the older workers who willingly gave their time and experience as participants. iii I would like to acknowledge the hard work and persistance of the over twenty undergraduate students who assisted during the course of this study. On a personal and emotional level, I would like to acknowledge the support I received from Denis Gray, Isidore Flores, and Dave Roitman. Each in his own way helped me through the rough spots. Finally, this thesis could never have been completed without the constant encouragement, support and love I received from my wife, Linda. iv LIST OF CHAPTER I TABLES. . . . INTRODUCTION. . . . . TABLE OF CONTENTS Human Capital Theory . . Labour Market Value . . . . . The Current Research Research Questions and CHAPTER II METHOD. Special Factors Motivation. Two Factor Model. Prior Research . Labour Force Status Propensity to Work. Labour Market Education . Decision to Withdraw. . . . . Early Retirement Myth . . . . Value of Work Health. . . Unemployment. Structural Unemployment . . . Work History. Long- Term Unemployment. . . . Age . . . Job Search. Summary . . Sample . . . . . Measurement. . . Archival Data Training and Data Reliability . Interview Data. Training and Data Reliability . Scale Construction. . . . . . Validity. . Hypotheses. Collection. Collection. Page ix 29 29 30 30 36 37 37 38 4O 40 41 CHAPTER II RESULTS. . Chara Stabi Effec Inter Varia Discr Regre I cteristics of the Older Jobseeker. Sex. . . . . . . . . . . . . . . Age. . . . . . . . . . . . . . . Ethnic Status. . . . . . . . . . Education. . . . . . . . . . . . Income . . . . . . . . . . . . . . Work History . . . . . . . . . . . Occupational Status. . . . . . . lity of Occupational Status. . . . t of Unemployment on Wages . . . . Summary. . . . . . . . . . . . . . correlations Between Independent bles . . . . . . . . . . . . . . . iminant Analyses . . . . . . . . . Employed or Unemployed . . . . . Employed or Retired. . . . . . . . Continued Job Search - Reemployed. Continued Job Search - Unemployed. Full-Part Time Job Status. . . . . ssion Analyses . . . . . . . . . . Duration of the Job Search . . . . Wages on New Job . . . . . . . . . Reemployed Job Satisfaction. . . . Job Search Activity During First Month at SES . . . . . . . . . . . Additional Findings. . . . . . . . CHAPTER IV DISCUSSION Major Metho Theor Findings. 0 O O O O O O O O O O dological Issues . . . . . . . . . etical Implications. . . . . . . . Economic Value of Workers in Labour Market . . . . . . . . . . . . . . Job Placement. . . . . . . . . . Duration of Job Search . . . . . . Job Search Activity. . . . . . . . Value of Work. . . . . . . . . . . Work Reasons . . . . . . . . . . Prolonged Job Search . . . . . Summary. . . . . . . . . . . vi Page 44 44 44 46 48 50 52 55 56 58 60 65 66 7O 71 77 81 84 88 90 9O 93 96 99 102 104 105 106 109 111 113 115 115 116 117 119 120 Page Policy Implications. . . . . . . . . . . . . . . 121 Outreach. . . . . . . . . . . . . . . . . . 121 Eligibility Criteria. . . . . . . . . . . . 122 Underemployment . . . . . . . . . . . . . . 123 Employability Review. . . . . . . . . . . . 125 Retraining. . . . . . . . . . . . . . . . . 127 Future Research. . . . . . . . . . . . . . . . . 127 Conclusions. . . . . . . . . . . . . . . . . . . 130 APPENDICES A. CETA Income Levels for Determining Eligibility. . . . . . . . . . . . . . . . . 131 B. DOL Application Form . . . . . . . . . . . . 132 C. DOL Intake Form. . . . . . . . . . . . . . . 133 D. Archival Data Codebook . . . . . . . . . . . 134 E. Worksheet Instructions . . . . . . . . . . . 139 F. Interview Instructions . . . . . . . . . . . 142 G. Classified Index of Industries and Occupations 1980 . . . . . . . . . . . . . . 145 H. Socioeconomic and Prestige Scores for Major Occupational Groups, 1970 Census Classification. . . . . . . . . . . . 146 I. Interview Schedule . . . . . . . . . . . . . 147 J. Administrative Agreement . . . . . . . . . . 160 K. Intercorrelations of Job Satisfaction Scale Items (Before Unemployment). . . . . . 161 L. Intercorrelations of Job Satisfaction Scale Items (Reemployed) . . . . . . . . . . 162 M. Intercorrelations of Extrinsic 'Work Value' Scale Items . . . . . . . . . . . . . 163 N. Intercorrelations of Intrinsic 'Work Value' Scale Items . . . . . . . . . . . . . 164 O. Intercorrelations of Health Scale Items. . . 165 vii Page P. Percent Agreement — Archival Items. . . . . 166 Q. Percent Agreement - Interview Items #1. . . 167 R. Percent Agreement - Interview Items #2. . . 168 S. Percent Agreement - Occupation Classification. . . . . . . . . . . . . . . 169 REFERENCES. . . . . . . . . . . . . . . . . . . . . . 170 viii LIST OF TABLES Table Page 1 Empirical Support for Predictors. . . . . . . . . 25 2 Major Predictors and Criteria . . . . . . . . . . 28 3 Major Variables with Descriptions . . . . . . . . 31 4 Frequencies and Chi-Square - Sex and Marital Status. 0 O O O O C O O O O O O O O O O O 45 5 Frequencies and Chi-Square - Sex and CETA Status. 0 O O O C O O O O O O O O O O O O O O O O 47 6 Frequencies and Chi-Square - Ethnic and Age 0 O O O O C C O O O O O I O O O C O O O O O O 49 7 Frequencies and Chi-Square Education and Age 0 O O O O O O O O O O O O O O O O O O O O 5 1 8 Frequencies and Chi-Square - Age and CETA status 0 O O O O O O O O O O I O O O O O O O 53 9 Frequencies and Chi-Square CETA and Social Security Status. . . . . . . . . . . . . . 54 10 Occupational Group and Status . . . . . . . . . . 57 ll Correlation Matrix - Major Variables . . . . . . 69 12 Discriminant Analysis on Employed- Unemployed Status . . . . . . . . . . . . . . . . 72 13 Discriminant Analysis on Employed- Retired Status. . . . . . . . . . . . . . . . . . 78 14 Discriminant Analysis on Continued Job Search for Employed Group . . . . . . . . . . . . 82 15 Discriminant Analysis on Continued Job Search for Unemployed Group . . . . . . . . . . . 85 16 Discriminant Analysis on Full-Time Job Status. . . . . . . . . . . . . . . . . . . . . . 89 ix Table Page 17 Regression Analysis on Duration of Job Search. . . . . . . . . . . . . . . . . . . . 91 18 Regression Analysis on Re-Employed wage O O O O O O O I O O O O O O O O O O O O O O O 94 19 Regression Analysis on Re-Employed Job Satisfaction. . . . . . . . . . . . . . . . . 97 20 Regression Analysis on Job Search Activity During First Month. . . . . . . . . . . . . . . . 100 CHAPTER I INTRODUCTION Unemployment among older workers has been a longstand- ing problem. Government publications addressing the ques— tion of 'what to do with the older workers' date back to the 1920's. More recently Gordon (1959) wrote, "Throughout the last decade there has been wide— spread concern over the employment problems of older workers, but at no time has the concern been more evi- dent than in 1958-1959" (p.1198). Currently the unemployment of older citizens remains a con- cern. ”The most significant labour market problems of middle-aged and older workers arise from the unemploy- ment experience... Unemployment among senior workers may mean a pronounced reduction in income during the retirement years, especially if the unemployment is pro- longed and/or intermittent." (DOL, 1979, p. 13). While unemployment has profound effects on the older workers there has been scant literature on the factors which influence the re-employment of older job seekers. Much of the industrial gerontology literature which does address these issues is of a general nature, not lending itself to applied purposes. The lit— erature is marked by a noted lack of empirical research. An exception is the increasing interest and development of employment services specifically for job seekers 55 years and older. These agencies can be found scattered through— out the United States (Note 1); and locally in Michigan the Department of Labor funds more than eight programs. Gray (1980) using a randomized design showed that a senior em- ployment service (SES) utilizing the Job Club concept (Azrin, 1975 ) could significantly increase the likelihood of older jobseekers securing employment. Gray's study is important in that it is the only empirical research which has examined the effectiveness of these employment services. Other literature has only presented descriptions of model programs (Brodsky & Robinson, 1981; Greymountain, 1981). Since Gray's (1980) research the proliferation of SESs noted above has occurred. Little is known about the type of client that uses these services. In addition little atten- tion has been directed at determining what variables are important in predicting the likelihood of an older job seek- er finding employment. Despite the success of the Job Club (Gray, 1980) there is still a significant proportion of older jobseekers who remain unemployed (Hadden, 1981). This research attempted to examine and document the personal characteristics of the clients of the Tri-County Senior Employment Service in order to determine the type and range of clientele that used the service. The impact of unemployment on the older worker's financial and occupation- al status was also of interest. The major focus of the in— vestigation was the identification of the type of informa— tion available during the intake process that predicted the probability of a client finding a job. A modified version of Human Capital theory was used to determine the variables likely to be associated with labour market behaviour and placement success. Human Capital Theory "Except within the framework of concepts and theories there are no scientific facts but only chaos" (Myrdal, 1954, p.vii). Even research which is directed at primarily applied problems requires the guidance of theoretical sys- tems in order to understand how real world forces impinge on the problem under study. Human capital theory augmented by job search and moti- vation theory provided the necessary perspective from which to select variables which should predict the probability of older workers finding employment. The concept of human cap- ital dates as far back as 1691 to Sir William Petty who con- sidered that labour was the father of wealth (Kiker, 1966). The contemporary form of the theory is attributed to Becker (1962). Sobel (1972) presents a modified version of human capital intended to encompass the correlates of aging. It is this latter interpretion of human capital which will be chiefly discussed. Human capital pertains to the attributes of people which are quantifiable or measurable as an input to growth. The value of human capital represents a form of investment in human beings (Davis, 1973). This model is an extension of capital theory which explains how objects obtain some value and what forces determine whether that value increases or decreases. Labour Market Value. Human capital is usually consider- ed to be developed as a function of the acquisition of edu- cation, general and specific job skills and other nonspec- ific life skills. The value of human capital is also affected by forces in the labour market. For instance "in- creasing levels of education and training mean that the or- iginal stock of human capital possessed by each of the older age groupings is substantially below that possessed by each succeeding younger one" (Sobel, 1972, p. 7) In addition changes in the structure of the economy affect the value of human capital. As capital intensive industries give way to newer high technology organizations, the demand for higher level job skills increases and consequently reduces the val- ue of older workers who on average have less education (Collins, 1978). "Human capital, like its physical counterpart is sub— ject to obsolescence and depreciation -- both of which oper- ate cumulatively over time to reduce human capital value ... In short, reduction of human capital value due to obsoles— cence and depreciation is cumulative with age" (Sobel, 1972, p. 7). Although it is theoretically possible for older workers to increase their human capital value by upgrading their education, it is rare for this to actually occur. Sobel (1972) in a study involving BUEB workers over the age of 45 found a general reluctance to use training programs. Only 10 percent of the sample had been involved in any upgrading program. For the majority of the working life an individual can generate additional human capital through work experiences and informal on-the-job learning. This increase in value is recognized in most organizations by institutional practices such as seniority. However "in most jobs, the human capital acquired through experience is highly specialized to the firm, industry or even the process in which the experience was acquired. This experience is largely non—transferrable to other employers, industries or fields of endeavors" (Sobel, 1972, p. 8). Therefore the value of the human cap- ital obtained from on-the-job experience is maintained only as long as the individual remains with the same firm. Thus the older worker displaced from a long tenure job experi- ences a substantial loss of human capital. This puts older job seekers at a severe disadvantage. Special Factors. There are two specialized forms of human capital: 1) job market information and 2) motivation or drive. Job market information involves a knowledge of job search methods and the availability of jobs combined with an awareness of one's own skill levels, and how to sell one's self to prospective employers. Given that the majori- ty of older job seekers are usually separated from long ten- ure jobs it is probable that their job market information is depreciated. In order to find employment the job seeker must be able to match his/her own skills with those required in the labour market. An imperfect knowledge of either his/her own job skills or the labour market reduces the like- lihood of an older worker finding employment. Motivation. The value that an individual's human capi- tal has in the labour market depends somewhat on how much they need a job. If the jobseeker possesses the job market information without the motivation to make the best use of it, then the probability of him/her securing employment is reduced. In other words there must be some reason for the older worker to want to stay attached to the labour force. Maslow's (1954) need theory, which proposes a need hierarchy, supports the inclusion of motivation as a part of human capital theory. Motivation to work can come from either basic financial need or from the higher need to be active and productive. Job search theory (Lippman and McCall, 1976) also sup- ports the position of financial considerations as affecting an individual's motivation to find work. "The wealth posi- tion of the job searcher also may influence his search be— havior. As his assets decline the searcher may become more willing to accept employment" (p. 176). In other words as the need to work increases, the job seeker seeks to maximize the use of his/her job market knowledge. Another factor affecting the motivation to work is the value that the jobseeker attributes to work. In other words if work is valued for the money it provides then a jobseeker might not accept a job that paid less than his/her previous employment, but if the jobseeker wanted to work to keep busy or feel useful s/he might be less discerning about the type of jobs s/he would accept. Essentially this is an extension of the Expectancy-Valence model of motivation (Vroom,l964; Porter & Lawler,l968;Campbell & Pritchard,1976). The possible outcomes are full employment, partial employment, unemployment, or retirement. The valence attributed to each of these outcomes, combined with the expectation that a given amount of effort can lead to the desired outcome, should influnce the choices made by the individual. Two Factor Model. Sobel (1972) does not directly ad- dress the relationship between the components of Human Capi- tal, but he implies that labour market outcomes are the res- ult of these multiple forces working in combination. The role of motivation is deemphasized in his paper primarily because the research presented to support the theory was at a macro-economic level of analysis. However without includ- ing motivational factors the theory does not provide an ad- equate model from which to predict the probability of a suc— cessful placement. For instance without considering motiva— tion the theory cannot explain why a worker with high capi— tal value continually refuses offers of employment. By in- cluding motivation it is possible to posit that the fit be— tween the jobseeker's 'work value' and the type of jobs of- fered will influence his/her decision regarding accepting employment. Therefore as a placement model the theory needs to consider both the economic value and the motivation of the jobseeker. The contribution of each of these factors was examined during the course of this research. The main objective of this study was to determine the variables that influenced and therefore predicted the prob- ability of an older jobseeker securing employment. The two factor-Human Capital theory discussed above posits that those factors which effected the value of an individual's human capital would influence the probability of them find- ing a job. The theory suggests that although changes in the value of an older person's human capital were possible, the human capital possessed by the older jobseeker was likely to be set by the time s/he sought aid from an employment agency. Therefore it should be possible to estimate the value of human capital when a client first visits an employ- ment agency and to use that estimate as a predictor of the likelihood of a successful job search. From an economist's point of view, human capital theory is limited in that it does not lend itself to precise esti- mates of the actual monetary value of specific components of human capital. However, by using multiple regression it is possible to use the theory to identify predictor variables. The following literature review provides empirical support for human capital theory, documents the need for the study and identifies appropriate predictor variables. Prior Research Labor Force Status of Older Workers. The activity of workers 55 and older in the labour force has been steadily declining since at least 1890 (Marshall and Cottam, 1981). The general consensus among population demographers is that only 20 percent of males aged 65 and older were in the labour force in 1979 as compared to 46 percent in 1950. Men 60 to 64 have reduced their participation 25 percent from 1950 to 1979 (Rosenfield and Brown, 1979; Chan and Fowles, 1980; Marshall and Cottam, 1981; OECD, 1979; Gillapsy, 1980, Morse, 1979). "Change in the labour force depends on population change and on change in the inclination of people to be in the labour force (either by working or actively looking for work)” (Collin, 1978, p. l). The number of males 55 to 64 in the labour force between 1950 and 1979 actually rose, but not as rapidly as the total number of males in that age cat— egory (Chan & Fowles, 1989). Labor force participation has responded to economic 10 conditions. "The decline in labour force rates of men 62 to 64 years generally has slackened during period of low un- employment and has intensified when unemployment rose substantially. For example, between 1966 and 1969, a period of very low unemployment, the labour force partic- ipation rate of these men decreased by 2.8 percentage points or 4 percent. On the other hand, between 1973 and 1976, when unemployment rates reached their highest level in three decades, the labour force rate for these men dropped by 6.4 percentage points or 10 percent. Obviously some men may have retired rather than try to compete for scarce jobs with younger, better educated workers” (Rosenfield & Brown, 1979, p. 15). These findings suggest that older workers during the current economic crisis and rising unemployment will face a diffi— cult task in securing employment. Older workers can leave the labour force either as the discouraged unemployed, i.e., they still wish to work but have given up any hope of finding a job, or as retirees. As long as they continue to search actively for work they are considered members of the labour force. "The hypothesis which has been put forth to explain this phenomena [reduction in labour force participation] has been that of the economic man making a rational and volun- tary choice between work and leisure, and opting for the latter" (Bould, 1980, p. 123). An alternative hypothesis is that older workers are being forced out of the work force by a variety of factors such as age discrimination and tech- nological obsolescence (Bould, 1980). Propensity to work. Sheppard and Rix (1977) point out that older people may choose retirement not because they 11 literally want to retire completely, but because they do not wish to remain in the same dissatisfying jobs. Retirement may be their best option given the difficulties encountered by older jobseekers (Rones, 1980). The proposal that older workers are reducing their labour force participation as a consequence of external forces rather than as a head long rush to embrace a leisure status is further substantiated by Harris et a1 (1981). In a survey of the general population, with an oversample of people aged 65 and older, Harris found that 79 percent of 55 to 64 year olds and 63 percent of those 65 and older expressed a desire to work at least part-time after the normal retirement age. Thus it would appear that a significant number of people in the older age brackets leave the labour force because alternatives to trad- itional retirement are not widely available. Labour Market. Another explanation of the decline in participation is found by examining the industrial and occu- pational structures of the labour market. Analysis of these structures shows that older workers are essentially concen- trated in capital-intensive industrialized sectors as opposed to those sectors which depend on recent technologic- al developments (OECD, 1979). In other words, older workers are found predominantly in older, less profitable indus- tries. The growing dominance of industries that provide few opportunities for older people is contributing to their de- creased participation (Marshall & Cottam, 1981). 12 Education. The bottom line is that most older workers do not have the requisite education needed to compete for jobs in the new industries. The human capital of these workers has depreciated. "The decrease in labour force participation of men has been relatively greater for the least educated than for those who attended college. Among men 55 to 64 years, for example, the labour force participation rate from 1962 to 1978 fell by over 20 percentage points for those who had not graduated from high school but only approximately 7 points for those with at least one year of college. Among men 45 to 54 years old too, the larg- est decrease in rates was also among the least educated" (Rosenfield and Brown, 1979, p. 1?). Chan and Fowles (1980) report that labour force par- ticipation rates for 55 to 64 year olds with elementary edu- cation was 64 percent, with high school was 77 percent and with 5 or more years of college was 88 percent. The prob— ability of an older person being in the work force seems, therefore, to be related to his/her educational level and its associated value in the labour market. Discerning the reasons for labour force withdrawal is a complex task con- sidering that virtually all the research in the area is cor- relational. Identifying relevant factors is a fairly clear process. Determining causal links between the factors and labour force withdrawal is more difficult. So far we have considered the influence of unemployment levels, lack of alternatives to retirement, job dissatisfaction, industrial and occupational structures, and education on labour force participation. Rosenfield and Brown (1979) identify ill 13 health, absence of dependents, desire for leisure, increases in monthly social security benefits payments, other pensions increases and job conditions as factors influencing labour force withdrawal. All of these factors either affect the worker's human capital or his/her motivation to work. Decision to Withdraw. The main controversy in this literature, referred to earlier in this paper, is whether the decision to withdraw from the labour force is rational and voluntary or whether withdrawal is involuntary as a result of external forces. The notion that the majority of older persons withdraw or retire voluntarily is associated with societal expectations that reinforce the idea that in- dividuals are entitled to, and therefore should enjoy a per— iod of leisure when they are old. This is derived from the typical linear life plan, adhered to by most of society, which begins with a prolonged period of education followed by a period of work, then leisure (McConnell, 1980). This life plan however is a relatively recent phenomena created by the introduction of retirement during the late 1800's. In order to ensure that young workers had jobs, older workers were encouraged through social security and private pension plans to withdraw from the labour force. Within a short period of less than one hundred years, retirement be- came the expected behavior of older workers (Sheppard and Rix, 1977). Prior to the creation of retirement, older in— dividuals normally worked until their health restricted them 14 (Sheppard & Rix, 1977). Early Retirement Myth. There is a belief prevalent in society that, given the opportunity, most people would choose to retire early. This belief has been applied to the interpretation of the increasing decline in labour force participation, particularly among the age cohorts 55 to 62; the early retirees. Kingson (1981) shows that this belief is actually a myth; that early retirees usually withdraw from the labour force involuntarily. The decision to volun- tarily choose between labour and leisure is an option avail- able only to individuals who are in good health and well educated with access to a secure pension. The majority of early retirees have a long history of chronic unemployment, health limitations, and limited retirement incomes. Kingson's (1981) study, based on data collected over a nine year period as part of the National Longitudinal Study of Men Aged 45-59 in 1966, found that 80 percent of the black and 66 percent of the white very early retirees retired in- voluntarily. Also men in blue collar occupations were more likely to withdraw involuntarily than men in white collar jobs. These findings strengthen the position that labour force participation is not simply a voluntary choice, but rather in most cases the result of a complex interaction of factors. The factors mentioned above can influence whether an individual remains or leaves the labour force. Several of these factors, as will be shown later, should also be pre- 15 dictors of job search success or failure. The retirement literature addresses the role of several factors discussed above as well as a few additional ones. Walker and Price (1976) in a review of the literature found that the following affected the retirement decision: econo- mic necessity, fear of inflation, work ethic, job satisfac- tion, fear of death and the belief that retirement signals death, and perception that retirement reduces overall life satisfaction. Value of Work. Rones (1980) proposes that the desire to work stems from either economic necessity or a need to feel useful and productive. Morrison (1979) states that the "propensity to leave the labour force is also affected by the availability of early retirement benefits" (p. 224). Afflu— ence or poverty affects the person's attachment to the labour force. An individual with an adequate income who valued work for financial reasons would have less motivation to accept a low paying job than the jobseeker who has no money for the next mortgage payment. Health. Health status is also a major predictor of labour force status in most of the literature (Kingson 1981, OECD, 1979, Arden & Johnson, 1980). The poorer an individ— ual's health, the less likely they will remain employed. An exception to the general trend of the literature is a study by Schmitt and McCune (1981). In a group of civil service employees, job attitudes and financial variables were found 16 to be significant predictors of retirement status, but health was not found to be a statistically significant fac- tor. Since the sample were civil servants who were assured of above average pension benefits, the decision to retire early may have been overly influenced by financial incent- ives, thus minimizing the effect of health. Unemployment. Another variable which may be pushing older workers out of the labour force is unemployment (Bould, 1980). "Regression analysis using the National Longitudinal Survey shows that weeks of previous unemployment is sig- nificantly related to early retirement for both black and white males. This relationship holds when controlling for social security pension eligibility, assets, health, family responsibilities, occupation, changes in unemployment rate and urban residence" (Bould, 1980 p. 123). The choice be- tween chronic unemployment and retirement is mediated by the legitimizing effect that retirement can have for the indi— vidual. The role of a retiree is more socially acceptable than that of being unemployed. The early retiree is often an individual with a history of chronic unemployment and low occupational status (Kingson, 1981). Occupational status and tenure on longest job held differentiates between volun- tary and involuntary withdrawal (Kingson, 1981). Structural Unemployment. Kingson (1981) has constructed a path model which supports the idea that events occurring early in the lifecycle affect later labour force status. 17 "The model does suggest that events occurring early in the life of a very early withdrawee (VEW) are influential in terms of the control experienced over withdrawal. The back- ground variables of race, occupational status of household head, and education of household head operate through the education of the VEW to influence the variables representa— tive of health, occupational status and second pension coverage" (p. 85). The findings support the notion that inequalities experienced early in the life cycle establish a lifetime pattern of employment disruption. Further, the early experiences and education of most persons now in the 55 and older cohort dictated that they would stay in what are now declining industries (Nelson, 1980). The net effect is that many older individuals experience structural unem- ployment (Marshall and Cottam, 1981; Sobel, 1972). Work History. The experience of unemployment for the older worker is likely to be difficult. The duration of unemployment increases rapidly with age (DOL, 1979; Rosenfield & Brown, 1979; Collins, 1978). Unemployed males over 55 in 1979 experienced an average of 19 weeks without work while males 16 to 19 were only unemployed for an average of 8 weeks (Chan & Fowles,1980). While the older worker is less likely to be unemployed, once they lose a job they have more difficulty finding another (Pampel, 1979). Several explanations for this phenomena have been of- fered. The length of tenure at the most recent job held is 18 associated with the probability of being unemployed for a longer period (Hill, 1977). Vandergroot et al (1979) notes that specific training and experience obtained from one job is often of no value to another employer. Employers are reluctant to provide general training which would increase the mobility of employees. Older workers who have been with one firm for an extended period become a devalued commodity in the labour market (Sobel, 1972). Displaced workers are often unaware of their devalued status. "The older job seeker may prolong the job search in hopes of finding a position that matches the previous one in terms of skill requirements, salary, status, and per- quisites" (DOL, 1979, p. 13). As discussed earlier the av- erage educational level of workers over 55 years is lower than those in younger cohorts (Chan & Fowles, 1980) making it more difficult for them to compete in the labour market. Although this gap may be closing present generations still encounter this problem. The very organizational structures which are meant to protect the employed older worker disadvantage him/her when s/he is seeking employment (Dunn, 1981). Seniority, pension and promotional plans within organizations dictate that most jobs, with the exception of entry level positions, are fill- ed from within (Morse, 1979). The consequence is that the older job seeker must move downward to less attractive, low- er paying jobs and while lowering their status must also 19 compete with younger, better educated workers (Sobel, 1972). In a study involving 3000 workers over age 45 registered with six state employment services Sobel (1972) found that 66 percent of the sample that had previously been employed in long tenured jobs were forced downward in occupational status and pay scales. Some older workers may refuse to accept downward mobility and choose instead to remain un- employed or if the option is available, retire (OECD, 1966). This decision will be mediated by the extent to which the jobseeker's human capital has been devalued and by his/her motivation to work. Long—Term Unemployment. The general picture which is emerging is that the unemployed older worker fits into the category of the long-term unemployed. "In April 1979, 64 percent of the unemployed out of work for less than one year were under 35, while 65 percent of the long-term unemployed were aged 35 or older” (Colledge & Bartholomew, l980,p.9). In fact with the exception of the teaching professions, at least half of unemployed men in all occupation groups were 55 and older (Colledge & Bartholomew, 1980). While these statistics represent British labour trends, a recent publi- cation (OECD, 1979) shows that European and North American labour force participation rates are very similar. In order to understand the problems of unemployed older workers it is necessary to consider the effects of long term unemployment. "There comes a point when people can no longer sustain their motivation in the face of continued rejec- 20 tion, heightened awareness of their own shortcomings, disillusionment with job finding services, belief that all available options have been covered, and a knowledge that jobs are scarce anyway" (Colledge & Bartholomew, 1980, p.10.). The long-term unemployed become discouraged workers who for all practical purposes have given up hope of finding a job. The consequences of long-term unemployment are not only psychological, but are also, of course, financial. The older unemployed person is in a more severe position than the younger worker in that his/her period of unemploy- ment is likely to have long-term effects on his/her retire- ment income, as well as immediate consequences for his/her mortgage payments, and the expenses associated with having dependents (Dunn, 1981). Research reported by Colledge and Bartholomew (1980) indicates that there are multiple variables which interact to explain why some unemployed workers are more likely than others to be unemployed for long periods. Some of these variables are: 1) age, 2) level of qualifications and skills, 3) type of occupation and industry, 4) health, 5) regional location, and 6) employment history. These factors were identified by Colledge and Bartholomew in a survey of 1,698 long—term unemployed people randomly selected from British manpower offices and from 50 indepth interviews with long-term unemployed people. The application of these re- sults is limited in that only frequency data was generated. A multiple regression analysis which attempted to identify 21 which factor best predicted success at finding employment would better help the development of specific manpower pro- grams for the long—term unemployed. Age. Another study by the U.S. Department of Labor (1956) also found support for age as a variable associated with the probability of obtaining work. In a review of the characteristics of 7,361 applicants aged 45 and older who had been successful in securing employment they found that of those in the 45 to 59 age cohort, 56 percent found jobs, as opposed to 23.5 percent of those 65 and older. The sig- nificance of this finding is that age may be a predictor of job search success even within the relatively narrow range that constitutes the older age brackets. So far we have addressed the external forces which work to keep older workers unemployed. The structures of the labour market, the devaluation of the workers experience, and the occupational status of both the worker and his/her father. We have also discussed the implications that age, education, psycho—social stress of unemployment and work history have for the probability of an unemployed older worker finding a job. All of these variables are similar in that the older person has little or no impact on how they will affect him/her. For instance it is not possible at age 60 to change your father's occupational status. The older jobseeker can, however, have an active role in how job search variables effect him/her. 22 Job Search. Reid (1972) describes two categories of job search methods: Informal and formal methods. The infor- mal methods include asking friends and relatives about pos- sible job opportunities, unsolicited applications and check- ing notices at factories and offices. The formal methods are using state services, advertisements, trade unions and private employment agencies. Reid found that informal methods were as effective as formal methods. Examining the job search and job finding behaviors of 876 redundant work- ers in England he found that the higher an individual's job skill, the more likely s/he was to begin the job search be— fore being laid off. When age of a worker was considered with job skills it was the lower skilled and older workers who waited until they were laid off before looking for new work. In order to determine whether imperfect knowledge of the labour market explained this finding, workers were asked if they expected that it would be easy or difficult to find new work. Reid found that the lower skilled older workers accurately per- ceived themselves as having a more difficult task ahead of them as compared with the perceptions of the high skilled younger workers who expected to have an easier time of it. Since considerable differences exist between the groups, the imperfect knowledge hypothesis was not supported. Further, when those who expected a difficult time finding work were compared to those who thought their job search would be rel- 23 atively easy, the distributions of length of time out of work were virtually identical. "60 percent of those expect- ing that it would be easy to find a job had found one within 8 weeks" (p.483). 61 percent of the other group found work within the same period. The significance of these findings is that expectations regarding the degree of diffi— culty of the job search are not valid predictors of job search success. A noticeable omission in Reid's (1972) study was that the effect of motivation on the strength of the job search was not measured. It seems logical that the more a person wants or needs a job the harder they will search. The ef- fectiveness of the job search should also be a factor. Summary. To this point, labour force participation, early retirement and unemployment among older workers has been examined. One purpose of this discussion has been to document that older workers, particularly those who are un— employed, are a population at risk. The literature indi— cates that indeed unemployed older workers are unemployed longer than younger workers (Rosenfield and Brown, 1979) and when re-employed usually drift toward lower paying, lower status jobs (Sobel, 1972). Other reasons for reviewing this literature were to document those variables which have been shown to influence the labour force participation of older workers, and to pro- vide empirical support for Human Capital Theory. The re- 24 search indicated that factors such as age, education, tenure on last job, occupational status, health, length of time unemployed, and financial status all influence the probabil— ity that a worker will find a job. They do this by determ- ining the labour market value of the jobseeker's human cap- ital and by effecting his/her motivation. For instance as the duration of unemployment increases the market value of the worker decreases, while the risk of him/her becoming dicouraged rises. In other words the jobseeker encounters difficulty in finding employment because the skills and ex— perience that s/he has to offer to a prospective employer are obsolete. A summary of the variables that were identi- fied are listed in Table l. The Current Research The research presented in the following section attempted to empirically determine what type of people use a senior employment service and to what extent it is feasible to predict the probability of them finding employment using information available during the intake process. The ef- fects of unemployment on the occupational and financial status of the older worker were also examined. Most traditional research is directed toward answering theoretical questions. Ecological or community psychology, a relatively new field within psychology, promotes an active 25 Table 1 Empirical Support for Predictors Predictor Variables Occupation Education Health Financial obligations and need Chronic or long term unemployment Tenure on last job held Motivation Age Employment history (number of jobs/type) Source OECD (1979), Marshall and Cottam (1981), Kingson (1981), Nelson (1980), Colledge and Bartholomew (1980) Rosenfield and Brown (1979), Chan and Fowles (1980), Kingson (1981), Colledge and Bartholomew (1980) Rosenfield and Brown (1979), Kingson (1981), OECD (1966, 1979), Arden and Johnson (1980), Colledge and Bartholomew (1980) Rosenfield and Brown (1979), Kingson (1981), Walker and Price (1976), Rones (1980), Morrison (1979), Dunn (1981) Kingson (1981), Bould (1980), Marshall and Cattam (1981), Sobel (1972), D01 (1979), Colledge and Bartholomew (1980), Chan and Fowles (1980), Rosenfield and Brown (1979), Collins (1978), Pampel (1979) Vandergroot (1979), Sobel (1972), Kingson (1981). Hill (1977) Colledge and Bartholomew (1980) Colledge and Bartholomew (1980), D01 (1956) Colledge and Bartholomew (1980), Reid (1972) 26 problem-solving role for social scientists (Fairweather and Tornatzky, 1977). The implication is that research should not only address academic questions, but should also have an application. The variables identified by the two-factor Human Capital theory as contributing to successful job placement should make it possible to modify the services so that they deal directly with specific problem areas. The findings of the research should also enable the agency to identify difficult to place jobseekers during the intake process and consequently deliver more individualized serv- ices. Research Qgestions & Hypotheses The literature on the labour market behaviour of the older worker has relied primarily on surveys of the gen- eral population as its source of information. The data produced by these methods has been of limited utility for understanding how services can best be delivered to the un- employed older workers who want assistance from a senior employment agency. The characteristics of the latter popula- tion are not yet known. Therefore the first objective of this study is to describe the characteristics of a sample of older jobseekers drawn from a senior employment service. The research by Sobel (1972), Bould (1980), DOL (1979), and OECD (1966), indicated that when unemployed 27 older workers became re-employed they experienced a drop in the occupational status of their jobs. In order to determine whether this was also true for this sample the following hypothesis was tested: The occupational status of the workers's job held before s/he became unemployed will be higher than the status of the job obtained at reemployment. The research cited above also found that wages decreased when an unemployed older worker became re-employed. The following hypothesis was also tested: The wage/hr received at the job obtained when the worker becomes re-employed will be less than the wage/hr received on the job held immediately before the last period of unemployment. The main objective of this study was to determine what information available to a service provider at intake can help differentiate between several success related criteria. The identification of these variables will further clarify the knowledge base regarding what factors are most important in determining specific labour market behaviours of older workers. In order to address this objective the variables listed in Table 2 will be entered into multivariate discrim- inant and regression analyses. The type of possible pre- dictors include demographics, work histories, and motiva- tional factors. The criteria include employment status, job search activity, job characteristics and duration of the job search. A complete list of criteria is presented in Table 2. 28 Table 2 Major Predictors and Criteria Predictor Variables Occupation Education Health Financial obligations and need Chronic or long term unemployment Tenure on last job held Motivation Age Employment history (number of jobs/type) Criteria Employed-unemployed Employed-retired Continued job search (re-employed) Continued job search (unemployed) Full-part time status Duration of job search Wage/hour (re-employed) Job satisfaction (re-employed) Job search activity (first month) CHAPTER III METHOD Sample The participants in this study were selected from the client files of the Tri-County Senior Employment Service located in Lansing, Michigan. This was the site of the first senior employment service in Michigan and was unique in that the agency had been allowed to serve non-CETA clients in addition to the regular CETA eligible clients. This provided an opportunity to compare the affect of unem- ployment on a wide range of older workers. The original sample consisted of 483 registered clients who had used the service between October 1980 and March 1982. The agency's follow-up interviews were conducted every four months; clients registered after March 15th, 1981 were not included in the sample because information on their employment status was still incomplete. After the telephone interviews the final sample was reduced to 265 participants. The sample was evenly divided into males and females. Approximately 43% of the clients were CETA eligible. This meant that those individuals had incomes below the poverty level (see Appendix A). Clients were required to be 55 years or older and and residing in the Tri-County area. 29 30 Measurement In order to assess the variables listed in Table 1, two types of data collection were employed: archival data ob- tained from the existing client files and telephone inter- view data. The archival data set consisted of information already collected during the agency's routine intake and follow-up interviews. The telephone interviews were conduct- ed during the summer of 1982 in order to obtain additional information not available in the agency's files. The two data collection procedures are discussed separately below. Information on the coding of the variables is presented in Table 3. Archival data. The client files consisted of informa- tion obtained during intake and follow-up interviews. The intake clerks conducted a personal interview with each pro- spective client in order to obtain the information required for the DOL application (Appendix 2). The client read the completed intake form and signed and verified that the in- formation was accurate. The follow-up interviews were con- ducted over the telephone and were intended to determine the employment status of the clients. Active clients were fol- lowed up every four months. The follow-up interviews were conducted by SES staff and should not be confused with the research interviews that were conducted for this study. The data collected during intake and follow-up was 31 wouOprum coHumuHHO COHuwuHMO coHHmuHuU :oHHmuHHO coHumuHuO coHumuHuO :oHumuHuU msoscHucou >Hmuom8mu u N ucwcmfiuwm u H puma u N HHDM u H mConoH uoc u mconoH u H N mcHxOOH uo: n N mConoH u H wooscHucoo mmumMMHp >Hocouum mmuom >Hmcouum u onom ucHom H U)H m mooscHucou pouonEoo OCHHOOS no ucmcmsuwm no no uumm Hm xuoz HHHum mucmHHu xuo3 Honuo HHHum mucmH CO U®>H000H H9 muomH>umm Ioo .coHuoeoum "mEOUH m mo mum .non 3mg nuH3 a mo dump mcHuumum coozuwn common m>m om mo mummy aumuomEou nun MH now many HHsn scam ma non MOM mconoH cmonmEmca How GCHHOOH Ho cmonmEm nofl 30: on mom moms 5m .mumxHOB .xuoz .amm Hmcoo mHmom oHuUMMMHumm now 3mc can oxmucH o no uwnfisz :oHumospm >Hmuomsme lucocmEHmm HHsmluumm Loummm co>onEmcD noumom coonmEm N mom: m coHuommmHumm xuosocam pmonmEmc: u N :oHuwuHuO pmonmEo u H msumum ucmfionmEm msuwum mm>e mchou coHumHuommo A>mm¢v oEmz wHQMHum> mgnmaum> mcoHumHuomwQ nuH3 mmHanum> Hohmz m OHQMB 32 MOHOHCOHQ HouoHpmum uouoprum uOOOprum HOuOprum uouoHcmum uOuoHpmum wouoprum uouoprum “Ouoprum mmw III H 0C H O mw> H H 0C H O mm> H H 0C H O mmummme >Hocouum u H mmumm >chouum n Ln onom OCHOQ m mmummme hchouum mmumm mecouum H II an-t mHmom ucHom m ll r-Il wmummme >Hmcouum venom mecouum u If) onom ucHom m ll r—l moummch >Hocouum moumo >Hmcouum n in onom UCHOQ m on N N m0> N H 0C H N mm» H H msoscHucoo pwuHuou I ucmE>onEo msoH>oum ocH>mmH How Cowmmm uch I ucoeonQEm mcoH>wum mcH>mmH How Cowmwm wmo chH n ucweaonEw msoH>wum mCH>me Mom cemmom nuHmmz mo uuommu MHmm I mHoom EwuH 0H vCquoz How acumen ucmuuomEH umos mH >0co: Socoe MOM cozy wonumu MHomuH HON pmsHo> xuoz .muomH>ummsm .mumxuozioo .coHuoeouQ .xuo3 .>mm umEmuH m mo mumHmcoo mHmom .non msoH>mHm nuH3 coHuoommHuom mcHume nsHo new umHsmou wco umme um popcouud conmwm coHumucmHuo nsHo nofllmum mo wocmpcmuufl Emumoum pwumucm ucmHHo xmp mnu “0m muop cmHHsh 1.»:oov m mHnae “may m commmm Ammv N commmm AHMV H Gommmm nuHmm: xuoz OHmCHHuxm xuoz OHmCHuucH H COHuomMMHuom uwnEmz :oHumucmHuo mxmucH 33 Houoprum HOUOprHm HOUOHpmum HODOprum uouoHpmum coHumuHuU \uouoHUmum nouoprum Mouoprum uouoHcmum HouoHowum mSODCHucou mm> on H o Amxmov mDOSCHucou Amnucoev msoccHucoO dBmUICOZ msn mmmx u N HMHUCMCHM H H mm» H H 0C N O mm> H H o: N O oxMHCH um mod mEoocH HmcoHUprm mm: 053 mmoomm nuH3 mCH>HH dump oxmucH ppm non ummH mo wumo mcH£MHcHw comSuwn pommme mwmp mo Honscz uuommu MHom "mxmucH cucumn mCmemmnon ucwmm mnuGOE mo Hmnfisz mHQHmHHm demulco: mm: H0\mm3 HOCHHU Hmnumcz Emumoum :H nucoe umun mcHusp pmuuommu MOHHH>Huom noumwm now no umnEsz non m50H>mum co uson mom moms condom non HocHoHuo you :0mmmm nuHmo: u ucmfionmEm msoH>mum mcH>mmH How Cowmom Umanch mm3 now xuouomeou I ucmEonme mcoH>mum mcH>MOH you :0mmmm 1.6cooc m magma Odd wmsomm comm cm>onEmcs mconom QBWU NuH>Huo< H moms AHHV H xooq “may m Cowmwm Avmv v Cowmmm 34 uouoHpmum uouoHpmum Houochum wouOHpmum wouOHpmum HouoHpmum uouoHpoum nouoHpmum mDODCHUCOU msoscHucou AHmmH .cmEHonummmv mooscHucoo AHmmH .cmEuwcumwmv mDODCHucou mxmucH ou Honm mummw 0N mcHusp prn mnofl mo Hones: Hmuoe non ummH um mamp mo Hmnfisz coHummcooo nomme mo msumum excuCH mo mEHu um mocmumw (mum non ooumum mo msumum oxmuCH ou uoHnm msucos m mnu MSOSCHOCOU 0C" N mw> u H >uHHOCHE u N wuHc3 n H OHMEOM u N mHmE n H A.ucoov :H pw>Hmomu mEOUCH OCHMD pmusmeoo wEOOCH Hmscc¢ >uHusomm HmHUOm mCH>Hoowm acoum OHCSHM xom m mgnme ON HmuOB ounces coHummsooo mocmummoum mEoocH HmHoom ochum xwm 35 fairly objective, requiring little subjective assessment on the part of the intake or follow-up clerk. Primarily demo- graphic, economic and employment related information was obtained. The information which was collected from the agency's files is described below. This data was used in the analyses on the characteristics of this sample and in the discriminant and regression analyses. The reader should refer to the intake form (Appendix 3) and the codebook (Appendix 4) to determine the specific source of each item. Demographics. This included age, sex, race, citizenship, migrant wofker, veteran, residence, offender, and displaced homemaker status. Job Preferences. Included the type of occupations sought and whether part-time or full time work was desired. Mobility. The ability to use public transportation and/or a car. Health. Whether the client had any health problems that would have limited their job search; whether they were hand- icapped. Job Club Status. Divided into those who just attended a job club orientation and those who were full members of the job club. This division was important because clients who attended an orientation also obtained some job search infor— mation. Family Status. Included the number of people in the family; marital status, and whether the client was the primary wage earner. Educational Status. Included the highest grade completed, whether the clients had completed high school or dropped out, whether they were currently enrolled in school, and whether their english language skills were limited. Economic Status. Included the individual client's annual income, the annual combined family income, the degree of poverty as determined by Department of Labor guidelines, and the intake worker's assessment of whether the client was 36 economically disadvantaged. Non—Work Related Sources of Income. Public Assistance, AFDC, SSI and Social Security. CETA Status. Eligibility for enrollment as CETA or non-CETA client, and whether s/he had ever been enrolled in another CETA program (see Appendix 1). Training & Data Collection. Traditionally archival data is subject to unreliability and missing data. The data in SES's files appeared to be fairly complete, however the organization of the information in the files was somewhat confusing. Several changes in the Department of Labour pro- cedures during the previous two years had affected the loca- tion of where specific items were recorded. These changes only affected the form the information was recorded on, not the method in which the information was obtained. Therefore the reliabilty of the data remained unaffected by the modi- fications made to the procedures. However considerable file searching was required to obtain the data. Nine undergraduates were used to code the archival data. Five of them took independent study credit. Of these, three were also used as interviewers during the next stage of the research. The other four undergraduates were from a senior level methods course and helped verify the transfer of information from the worksheets to the opscan coding sheets. Data collection was conducted at the offices of SES. Training was completed in one evening, during which an ex- 37 planation of SES's filing system and the coding procedures used in this research were presented (Appendix 5). The group was then given sample cases to complete. Coders were re- quired to complete the coding of at least two sample files without any errors before they were allowed to code the actual data. The data was first recorded on precoded worksheets and from there it was transfered to opscan sheets. Verification procedures are described below. The Classified Index Of In- dustries and Occupations (U.S. Census,l980) was used to code occupation data (Appendix 7). Reliability. Inter-rater reliabilities were obtained between all coders. This was accomplished by having each coder re-code files that were coded by each of the other coders. A11 paired comparisons were examined. There was 86% agreement. Opscan sheets were verified by a person other than the coder who completed them. (see Appendices) Interview Data. A telephone interview was conducted to obtain more extensive information on previous and current employment status, perceived 'work value' and reasons for seeking employment. The interviewers were able to contact 265 of the 468 clients. All of the analyses conducted in this study were done on this final sample of 265. The vari- ables that were collected during the telephone interviews are described below. Refer to the interview schedule (Appendix 9) for the actual wording of the questions. 38 Work History. This was for the last twenty years of the respondentrs work life. It included the number of jobs held during the last twenty years, the occupation and tenure of each position, tenure, the length of the most recent period of unemployment, the occupation held during the majority of the respondent's life, the hourly wage of the last job, and the reason the worker left the last job. Primary Reasons for Working. This assessed the individual's reason for working. Employment Status. This was the primary dependent measure. Respondents reported whether they were employed, unemployed or retired. Those that were employed or unemployed were asked if they were still actively searching for work. If the client was employed the occupation, starting date and part—time, full—time status of the job were reported as well as the hourly wage of the job. Education. The number of years of formal schooling. Health. This was assessed by self report on 10 health items. The items were intended to assess the respondent's health as it would effect his/her ability to look for, or perform work. The scaling procedures are described below. Job Search Activity. This was assessed retrospectively using items that required the respondents to report the number of job search activities that they participated in during their first month as clients at SES. Extrinsic & Intrinsic 'Work Value'. These scales consisted of items that assessed the respondents' perceptions of the value of work. The items and scaling procedures are des- cribed below. Training and Data Collection. Ten undergraduate stu~ dents conducted and coded the interviews. Training included a discussion of each of the following topics followed by extensive role playing. 1) Overview of the research 2) Establishing rapport and obtaining informed consent 3) Using the questionnaire 4) Using probes 5) Keeping the respondent on topic 6) Closing the interview 0 39 7) Recording responses The role playing was structured so that first a demonstra- tion of the appropriate interview methods was presented, after which the interviewers were broken up into groups of three to role play interviews. Two of the students role played the interview while the third observed and noted how the interviewer performed. The final training step involved each student conduct- ing at least two interviews with SES clients that were not included in the sample. These were conducted under the su- pervision of the researcher who assessed whether or not the interviewers were ready to conduct actual research inter— views. Inter-rater reliability, as described below, was also assessed before actual interviewing began. After the initial training was completed the students conducted the interviews from their own homes during after— noons and evenings. In order to ensure that consistency was maintained weekly supervision meetings were held, during which the interviewers' performance was monitored by the researcher. These meetings included discussion of the re— search interviews and additional role playing. Data collect- ion was completed within a seven week period. The actual interviews took an average of 20 minutes to conduct. For the most part the respondents were cooperative and did not demonstrate any difficulty understanding the questions. Attrition from the original sample of 468 was 4O chiefly the result of disconnected telephones, wrong numbers and changed addresses. Approximately 5% of the clients that were contacted refused to participate. All respondents were read a brief explanation of the research and given the op— portunity to decline. Reliability. Inter-rater reliability was assessed be- fore actual research interviews were conducted. Inter- viewers listened to and coded a taped interview. This was conducted before the interviews began and again at the mid way point in the data collection. The respective inter-rater percent agreements were 93.2% and 98%. (see Appendices) Inter-rater reliability for the occupational classific— ations was computed separately. The interviewers coded a sheet containing ten occupations. There was 93% aggreement. Scale Construction. Several scales were constructed in order to measure job satisfaction, 'work value', and health. The scale construction procedures were as follows: A ration- al process generated the items that were included in the interview. Next, factor analysis with varimax rotation (Hunter, 1980) was used to examine the factor structure within each sub-scale. The item total correlations and the internal consistency were also examined. The conceptual meaning of the scales and items was weighed strongly when final decisions were made on which items to include. The inter-item correlation matrices for each scale can be found in Appendices K through 0. Descriptions of the 41 scales are included. The alpha for each scale and the inter—scale correlations are presented in Table 11. The con- struct validity of the scales is discussed below. .yglidigy. The issue of validity is concerned with the veracity of measurement. In this study there were two major types of data: The objective items such as the wage or starting date of a job; and the scales measuring constructs such as 'work value'. There was no practical method to ver— ify the accuracy of the objective information provided by the client. However very few clients demonstrated any diffi- culty remembering their work histories. Any errors resulting from memory lapses were probably randomly distributed. The retrospective nature of this study did not provide optimal conditions for assessing the construct validity of the scales. In order to accurately determine construct val- idity it will be necessary to replicate this study so that responses are obtained within appropriate time frames. Also different data collection procedures should be used to en- able assessment of method variance. The job satisfaction and health scales were adapted from existing measures that have already been validated (Smith et a1, 1968; Kane & Kane, 1980) The extrinsic and intrinsic 'work value' scales were contructed for this study. The 'work value' construct represents the value that the jobseeker attributes to working. For instance one person might value work because it provides an income while 42 another might place a higher value on the challenges associ— ated with a job. It was hypothesized that 'work value' should influence job placement. Generally the correlations between the scales and other variables were consistent with the construct. The extrinsic scale was negatively correlated with the wage earned before unemployment (r=-.22) and with the reem- ployed wage (r=-.18). The primary value attributed to work by a person earning low wages was the the income provided by the job. A similar association was found with occupation (r=-.22). As the status of the respondents' occupation in- creased their extrinsic 'work value' score decreased. Exam- ination of the correlation matrix showed that all financally based variables exhibited the same relationship with this scale. The intrinsic 'work value' scale was constructed to assess the importance of intrinsically motivating factors. This scale was correlated with unemployed clients that were still looking for work (r=-.43). A negative correlation was found between retiring from the last job and this scale (r=—.13). An individual that chose retirement was demon- strating low intrinsic work motivation. Similarly the relationships between the scale and job search activity (r=.22) and part/full-time job preference (r=.18) were con- sistent with the construct. A high score was associated with high job search activity and with a preference for full-time 43 employment. These relationships suggest that the extrinsic and intrinsic scales did measure 'work value'. However given the problems discussed above the construct validity was not suf- ficiently demonstrated. Future research should address these issues. CHAPTER III RESULTS In this section, results relevant to the research ques- tions and hypotheses will be examined. The characteristics of the client population at SES will be presented using des- criptive statistics. The stability of wages and occupation- al status will then be examined using T-tests. The final portion of this section will focus on the major research questions, regarding the prediction of employment status, continued job search, full or part-time job status, duration of the original job search, reemployed wage, job satisfac— tion and job search activity. Stepwise discriminant and regression procedures were used to conduct these analyses (Nie et al,l975). Characteristics of the Older Jobseeker S35. The client population was evenly split; with 50.2% male and 49.8% female. The chi-square for sex by marital status was statistically significant (Hi44.44,DF=2,p<.001), (see Table 4). Considering the males; 81.1% were married, 4.9% were single, and 13.9% were widowed, divorced, or separated. This is in contrast to the females of which 40.6% were married, 44 Frequencies and Chi-Square - Sex and Marital Status Married Single Widowed, Divorced or Separated X = 44.44 P 45 Table 4 Male Female 99 52 151 6 10 l6 17 66 83 122 128 48.8% 51.2% < .001 60.4% 6.4% 33.2% 46 7.8% were single and 51.6% were widowed,divorced, or separ- ated. The male clients also differed significantly from the females in regard to their average wages and incomes. The mean wage per hour for males on their last job was $7.37 (SD=$3.53), while the mean for the females was only $4.47 (SD=$2.31), (T=7.42,df=208.79,p<.001). The mean income for the males in the year preceding their intake into the pro- gram was $5430 (SD=$5775, range=$0-$32,000) while the mean income for the females was $3856 (SD=$5021, range=$0-$33,600), (T=2.26,df=239,p<.03). Since CETA eligibility is dependent on income, it is not surprising, given the above wage differential, to find that significantly more women (50.4%) than men (36.7%) in the program were registered as CETA clients. These results suggest that the female clients, as a group, were in less desirable life situations. Their financial need was greater, and at least to the extent that a spouse can provide vital support to an unemployed person, the women appeared to have fewer family supports. These results are consistent with the general employment literature (Sobel, 1972) that shows women of all age categories receiving less pay then men. Ag_. The mean age was 61 years with a standard devia— tion of 5.8 years. The sample broke down into the following divisions, which were predetermined by social security eli— giblity criteria: 66% of the clients were between the ages 47 Table 5 Frequencies and Chi-Square - Sex and CETA Status CETA Non-CETA Male 47 81 128 50.2% Female 64 63 127 49.8% 111 144 45.5% 56.5% x = 4.31 p < .04 48 of 55 and 61; 8.7% were between 62 and 64 years old; and 25.3% were between 65 and 92 years old. The 55 to 61 year olds were too young to receive social security, which means that the majority of the clients did not have a choice be- tween working and retiring. Since social security cannot be collected until the minimum age of 62 this 'younger' group of jobseekers must rely on resources other than social security benefits to moderate their periods of unemployment. This in contrast to the 77.3% of those between 62 and 64, and the 88.9% of the clients 65 and older who were receiving payments from social security. These 'older' clients had additional income, thus reducing the urgency for them of finding a job. Minority clients belonged to the latter category. Ethnic Status. Minorities accounted for only a small portion of the sample (10.9%), but they were significantly older than the rest of the clients. Over 64% of the minor- ities were 62 and older as compared to less than 29% of the whites (gil4.72,df=2,p<.00l). This means that they belong to the 'older' group of jobseekers who were collecting social security. It is unclear why so few minorities are using the services of this agency. Blacks and Hispanics comprise 20% of the area's population (U.S. Census,l980). The literature indicates that older members of minority groups have a higher risk of experiencing sustained unem- ployment (Bould,l980). Therefore, one would expect to find 49 Table 6 Frequencies and Chi-Square - Ethnic and Age Age Ethnic 55-61 62-64 65+ White 164 18 48 230 Minority 10 4 14 28 174 22 62 67.4% 8.5% 24% x = 14.72 p < .0006 89.1% 10.9% 50 a larger percentage of minorities using the services than was found here. The implication of this finding is that in some way the services provided by SES are not known to the minority population or do not meet the needs of this group. (see Table 6) Education. The mean education level was 11.86 years of schooling (SD=2.77 years) with 28.9% completing high school or higher. At the low end only 3.8% reported having com- pleted less than 7 years, while at the upper end only 3.8% had graduate degrees. 18.6% of the clients had completed some college with 6.5% having received undergraduate degrees. 42.6% of the sample had completed high school. The level of education does not differ significantly with the age of the jobseeker (£26.84,df=4,p>.10). Sobel(l972) found that older workers were less educated than their younger counterparts. He posited that this dif- ference would put the older worker at a disadvantage when competing in the job market. The average educational level of this sample is relatively high when compared to the median reported for this age cohort (12.0 years) in the 1980 Census. This group also fares well when compared to the av- erage education of the population younger than 55. The 1980 Census reports that the median years of schooling for the 25-54 cohort ranges from 12.5 to 12.9. The older workers in this sample do not appear to be at a disadvantage because of their education. In fact the results of this study and the 51 Table 7 Frequencies and Chi-Square - Education and Age Age Education 55-61 62-64 65+ Less than high school 41 9 25 75 28.5% High school 79 10 23 112 42.6% graduate Partial college 53 4 19 75 28.9% or higher 173 23 67 65.8% 8.7% 25.5% x = 6.84 p N.S. 52 Census both indicate that the education gap is closing. As was expected education was found to be significantly correlated with income (r=.21,p<.00l). Income. The mean annual income for individuals was $4634 (SD=$5467). The combined income for the 43.8% of the sample with income earning spouses was $8308 (SD=7288). The mean for those with combined incomes was significantly higher than for those with only one income (T=-8.8, df=237,p<.0001). In the majority of cases CETA eligiblity was calculated using only the income of the individual client. In this sam- ple 43.5% of the clients were registered as CETA eligible with the remaining 56.5% considered ineligible. This is equivalent to saying that 43.5% of the clients had incomes below the poverty line. The mean income for CETA eligi- ble clients was $2470 as compared to $6141 for the ineligi- lDle clients (T=5.78,df=228,p<.00l). CETA status also differed significantly with the age of tlae client. As age increased the number of CETA eligible crlients decreased. Between the ages of 55 and 61 50.9% were CETA eligible. From 62 to 6.4 34.8% were eligible and for tlflose 65 and older the percentage dropped to 27%. The c=l'li—square for age by CETA status was Xgll.45,df=2,p<.003. Ir! order to determine whether the social security income of tliea older clients was responsible for them being ineligible ft>r' CETA status a chi-square analysis was conducted on CETA 53 Table 8 Frequencies and Chi-Square - Age and CETA Status Age 55-61 62-64 65+ CETA 86 8 17 111 43.5% Non-CETA 83 15 46 144 56.5% 169 23 63 66.3% 9% 24.7% x = 11.45 p < .003 54 Table 9 Frequencies and Chi-Square — CETA and Social Security Status Social Security Recipient CETA Status Yes No Yes 37 70 107 43.9% No 59 79 138 56.3% 96 149 39.2% 60.8% 2 x = 1.36 p = N.S. 55 by social security status. The result was not significant (£21.36,df=1,p>.05) indicating that factors other than social security benefits were responsible for the higher incomes of the older clients. CETA status also varied significantly as a function of the client's sex; 36.7% of the males were CETA eligible ver- sus 50.4% of the females (£24.31,df=1,p<.04). This is con- sistent with the results for sex by income. Women were in worse financial situations in all respects. The majority of clients using SES had relatively low incomes; the source of which might be linked to poor work histories or the relatively low occupational status of many of the older workers in this sample. work History. The mean number of jobs held by an indi- vidual during the last 20 working years was 2.39 (SD=1.3). The mean length of time spent at the last place of employ- ment was 4128 days (SD=4117.27) or approximately 11.31 years. The mean length of time spent unemployed before intake into SES was 621.99 days (SD=949.35) or approximately 1.7 years. It appears that the average client did not change jobs often but instead stayed at one job for a considerable length of time. These workers did not fare well when they became unemployed. Chan & Fowles (1980) found that older workers remained unemployed for a longer period than younger workers. The older jobseekers in their sample were unemploy— ed for an average of 19 weeks. This is considerably less 56 than the unemployment experienced by the jobseekers in this sample; even when different economic and geographic condi- tions are considered. The low occupational status of the sample also does not completely account for the long periods of unemployment. Occupational Status. Occupational status was measured with an instrument designed by Stevens & Featherman (1981). This was a socioeconomic measure using a combination of education, income and prestige factors. The mean status for the client's major occupation was 28.91 (SD=14.09), with a range spanning from 14.04 to 68.63. As shown in Table 10; 15% of the sample was engaged in private household services (includes housewives), 1% were labourers, 19% were classi— fied as operatives, 5% were involved in transportation, 11% were in general service areas, 1% were farm owners, 3% were craftspersons, 19% were in clerical positions, 13% were in sales, 9% were in management related occupations and 4% were professionals. The majority of the jobseekers (62%) there- fore, had been employed as factory workers, general service personnel, or sales and clerical staff. The mean occupation- al status falls between the status of craftpersons (25.63) and clerical workers (31.99). The data portrayed a relatively homogeneous population. Given the generally low status of the clients the possibil- ity of any further downward mobility is questioned. Occupational Group and Status Occupational Group Professional Managerial Sales Clerical Crafts Operatives Transport Labourers Farm Owners Service (executive private household) Service (private household) 57 Tablelfl Percentage 4 9 l3 19 19 11 15 Mean Status 68.63 51.07 42.30 31.99 25.63 18.24 20.37 15.99 22.29 20.81 14.04 58 Stability of Occupational Status The hypothesis that occupational status would decrease after a period of unemployment was not supported by these results. The status of the jobs obtained by the re-employed clients did not differ significantly from the status of their previous jobs. There were 124 subjects who had both pre and post status scores. The mean status for the last job was 28.92 (SD=14.1) and the mean for the new job was 28.06 (SD=12.2). The difference between the status of the last job and the new job was not statistically significant (T=.60,df=123,p>.05). This finding is contrary to the liter- ature which has shown sharp decreases in the occupational status of older workers who have been unemployed (Sobel,1972). There are several reasons which might account for the lack of status shift in this sample. The research reported by OECD (1979) and Sobel (1972) included a large percentage of workers who had previously been employed in labour inten- sive industries that were on the decline. When those workers became unemployed the dwindling number of jobs in their pre- vious occupations, combined with the effects of seniority systems on hiring policy, forced them to take lower status jobs. This was not the case with this sample. Most of the participants were already in relatively low status occupa- tions that were not associated with declining industries. 59 Only 19% were previously employed as operatives (factory work). It is possible that no shift in status was found because: a) an extreme shortage of jobs at the jobseeker's previous status was not evidenced; or b) the occupational status of SES clients was too low to decrease substantially. The coding format used to determine occupational status lacked some precision. The codebook used by Stevens et a1 (1981) to classify jobs was different than the 1980 Census codebook used in this study. This required the aggregation of specific job types into larger categories. The subsequent reduction in variance might explain why no differences were found between the status of the two jobs (see Appendix H). A third alternative is that the occupational status measure used in this study was not sensitive to variants, such as full or part time employment, that may affect the actual prestige of a job. For the most part the last job reported by most respondents was considered full-time. This is in contrast to the new jobs acquired; of which over 60% were part-time. Since Stevens's measure does not differ- entiate status On part and full-time jobs it is possible that a shift in status was not detected. Still, a fourth alternative, is that individuals seek- ing assistance from a senior employment agency might not lose occupational status because of the aid they received from SES. This explanation is appealing since it suggests the effectiveness of this mode of intervention. Regretably 60 the design of this study does not permit a valid evaluation of this alternative. Sobel (1972) found that decreases in occupational status corresponded with declining wages. This is consistent with the concept of occupational status which includes the earning power of an occupation as well as educuation and prestige. If occupational status declines then the wages received on a job should also decline. Given the low sens- itivity of the status measure it is possible that wages could decline even though occupational status did not vary. Effect of Unemployment on Wages Downward shifts in occupational status were of concern because they are usually associated with a loss of prestige and income. The hypothesis that wages would decrease from the last job to the new job was supported, despite the sta- bility of occupational status. Wages received on the job after a period of unemployment were an average of 94 cents/hr lower than on the job held immediately prior to unemployment. The mean wage for the first job was $5.73 (SD=$3.34) and the mean wage for the new job was $4.79 (SD=$2.14). The mean difference of .94 cents was significant (T=3.34,df=102,p<.001). These results were consistent with the literature on older jobseekers. Sobel (1972) reports that 66% of older workers are required to take a drop in pay 61 after a period of unemployment. Sobel(l972) posits that the drop in pay experienced by many older jobseekers is partly due to a decrease in the market value of their human capital. In other words, the jobseeker's skills are not highly valued in the current labour market. Another possibility is that jobs available in the occupations held by older worker's are limited, by hiring policy, to entry level positions. If the latter is true it explains why no shift in occupational status was found even though wages changed significantly. An entry level position would not be differentiated from a higher level position by the occupation status measure used in this study, but the entry level job might have a lower compensa- tion rate. A third alternative is that the difference between the two rates of pay is a consequence of a shift from full to part time employment. The majority of all jobs held before unemployment were full-time. After re-employment the respon— dents reported that 67% of the new jobs were part-time. When full and part-time wages/hr on the new jobs were compared, full time workers earned $1.38 more an hour than part-time workers (T=2.95,df=52.9,p<.005). This difference in pay rates is consistent with Morse's (1969) findings that part-time employees usually receive less renumeration than full-time employees. Further examination of these results revealed that 62 those workers who were re—employed at full time jobs exper- ienced a significant drop in pay from their previous jobs (d=$1.39,SD=$2.43, T=3.43,df=35,p<.05). Part—time employees however, encountered a much smaller decrease in their wages (d=$.7l,SD=$3.07, T=l.88,df=66,p<.05). A T-test comparing the drop in wages for full and part-time workers was not statistically significant (T=l.15,df=l0l,p>.05). This in- dicates that full or part-time job status was not interact- ing with the time elapsed between jobs. There was, however, a significant difference in the wages received by workers now working full or part—time on the last job that they held. Jobseekers now working full-time received $6.96/hr (SD=$3.04) on their last full-time job, while respondents who were re-employed at part-time jobs only received $5.04/hr (SD=$3.01) on their last full-time job (d=$1.92,T= 3.20,df=l08,p<.002). There are at least two explanations for the discrepancy between the earnings of these two groups. The first, which is suggested by Human Capital theory, is that individuals re-employed at part-time jobs had a limited choice of jobs as a consequence of poor employment histories. Therefore, they were forced to accept jobs at low wages. Their poor employment backgrounds would also explain the significantly lower wages of this group on their last job. The workers re-employed at full-time jobs were able to get these jobs at higher wages, because they had held better jobs before they 63 became unemployed. (It should be noted that even though the workers re—employed at full-time jobs maintained higher wages than their counterparts who received part—time jobs, they still experienced a significant drop in pay as a con- sequence of being unemployed.) A second alternative is that respondents did not fully understand the instructions given during the telephone in- terviews. The questions pertaining to the most recent job specified that the interviewer was interested in the participant's last full-time job. It is possible that des- pite these instructions the last job reported by some inter- viewees was actually only part-time. If this was the case, the lower wages received by individuals now employed part-time could have been because these clients also held part-time jobs before they became unemployed. This explan- ation is doubtful. Pilot interviews and discussions held with the student interviewers suggested that the respondents did understand this question. Furthermore, as a precaution- ary measure, the interviewers periodically probed to ensure that the interviewee understood the nature of the question. Three alternatives have been posited so far to explain the decrease in wages experienced by SES clients; a decrease in the value of a jobseeker's human capital; limited open- ings for individuals seeking employment at higher levels than entry positions traditionally provide; and the affect of full and part-time job status on wages. A fourth possi- 64 bility is that older jobseekers received lower wages, when they were forced to compete on the job market, as a conse- quence of age discrimination. SES clients may take lower paying jobs because of discrimination they encountered look- ing for employment. This study did not directly addresses this issue, however, respondents who were still unemployed at the follow-up interview did report their age as one of the main reasons that they were still unemployed. They felt that employers saw their age as a negative factor. Another issue raised by this data is how the dif— ferences between wages on the last job and the current job varied as a function of sex. As discussed above, females received lower wages than males on both jobs. Breaking this down further, it was found that males experienced a signif- icant drop in pay (d=$1.63, T=3.5,df=54,p<.001), while fe- males only had a small non-signficant decrease (d=$.15, T=.63,df=47,p>.05). These results show that despite a sig- nificant drop in earnings males still maintained higher wages than females when they became re-employed (d=$l.22,T=3.23,df=85.1,p<.002). A similar shift in wages was observed as a function of CETA status. CETA eligible clients received lower wages than their non-CETA counterparts on the job held before register- ing with SES (d=$1.37,T=3.06,df=206,p<.002). However, they did not differ significantly from them when they became re-employed (d=$.37,T=.89,df=l07,p>.05). The CETA eligible 65 clients did not experience a significant drop in pay (d=$.28,T=.62,df=43,p>.05), but the non-CETA clients lost an average of $1.55/hr when they became re-employed (T=4.19,- df=55,p<.0001). The fact that the wages of non-CETA eligi— ble clients dropped to the same level of the CETA eligible clients after both groups became re-employed suggested that the experience of unemployment had an adverse effect on the non-CETA jobseekers. Although the design of this study does not allow assessment of the agency's effect on the clients participa- ting in this research, one possibility is that the program moderated the effects of unemployment. The client's encount- er with SES may have increased or decreased the likelihood of him/her finding a job at a lower wage. The veracity of this hypothesis should be empirically tested. Senior employ- ment agencies in Michigan are no longer allowed to serve non—CETA clients. If the agencies positively moderate the effects of unemployment for this group, then current practices can only lead to worse hardships and possibly in- creased social welfare costs for the older worker who be— comes unemployed. Summary. In this subsection the results pertaining to the hypothesis that wages/hr would decrease when the older jobseeker became re-employed were discussed. The hypothesis was supported, alternative explanations were considered and changes in wages/hr were discussed for specific subgroups. 66 One final implication of these results is worth noting. When asked to give their main reason for looking for work, 51.5% of the sample chose 'finances'. Since these results showed that wages generally decreased, the expectations of these older jobseekers were likely to be violated. If the lower wage jobs did not fit the needs of the jobseeker then s/he might have prolonged the job search. Discriminant and re- gression analyses were conducted in order to determine whether such information could predict placement outcomes. However before these analyses are reported the problem of multicollinearity should be addressed. Intercorrelations Between Independent Variables Intercorrelations among the independent variables were generally low (r<.30). There were, however, some exceptions. Whenever the correlation between two variables exceeded .30 tflae pair was examined and if possible, one variable was dzxapped from the analysis, or a new variable was created Ccnnbining the two. In some cases neither option was feasible arufl both variables were retained. When this occurred the FKDtential problem of multicollinearity became a concern. “filen two highly correlated independent variables both enter- eCl an analysis interpretation of the discriminant and re- QIWession coefficients became problematic. This was mainly a CCUficern when both variables helped to define the same func— 67 tion. Discussion of specific multicollinearity problems will be left until the results of the individual analyses are reported. In this section the correlations among predictor variables greater than .30 will be presented and the action taken will be described. The length of time spent at the last job was negatively correlated with the total number of jobs held during the past twenty years (r=-.62). It was decided to include only one of these variables in each analysis. Based on the cor- relations with the dependent variables, tenure was included in all of the analyses with the exception of the two on cur- rent job search status. Attendance at a job club orientation session was cor- related with subsequent job club membership (r=.61). In or- der to maximize the sample size, attendance at an orienta- tion was retained and job club membership was dropped for all the analyses except for the regression analysis on original job search activity. Age was negatively correlated with social security status (r=-.59). The coding of the social security item (1=yes,2=no) should be taken into account when considering this relationship. As age increased the likelihood of a participant receiving social security also increased. Since more than 75% of those 62 and older received social security it was decided to drop this item and retain age. The wage received at the last place of employment was 68 correlated with the length of time spent on that job (r=.46). Wage was also negatively correlated with the sex of the jobseeker (r=-.43) and positively correlated with re- tirement as the reason for leaving the last job (r=.44). The importance of these variables required that all of them be retained. 'Retirement as a reason for leaving the last job' was also correlated with tenure on that job (r=.45), CETA status (r=.36), and being laid off from the last job (r=-.35). CETA status which was correlated with income (r=.33) was dropped since income was the main variable used to determine CETA eligibilty. Retirement and 'laid off from last job' were retained. An individual's score on the health scale was negative- ly correlated with 'health given as a reason for leaving the last job' (r=-.39). If a person left their last job for health reasons they were likely to have a lower score on the health scale. Since the average time span between the term— ination of employment and entry into SES was almost two years it was decided not to combine or drop these two varia- bles. Education was correlated with both the occupational status of the job preference stated at intake (.33) and the status of the major occupation held during the lifetime (r=.37). The latter two variables were correlated with each other (r=.36). Occupational status was computed by Stevens Status F1 naort Satisfactionz la 2 9e Emloyod Search the-ployed Search PartoFull Penn-Taco Education lntaka Ortentatlon lumbar Satisfaction! Intrinsic Hark Extrlnstc Hort Health Reason l Anson 2 Anson 3 Reason 4 Reason 5 Look 1 flags 1 Act1vtty Cata Houlonq unloved Spin Spam Age Sex Etnntc Social lncouo Preference Octopatlon Tenor! Ful l-Part Prof Status find-sort tlonZ (I) ' (I) - -.12 - -.o7 - -.os . . ' -.05 ' .21A .05 ~.10 .02 -.18 .m -.12 .02 -.07 .03 -.IO -.09 -.03 .II .360 .03 -.lZ -.03 .07 .03 .04 .06 -.22A .04 .04 .04 .248 .08 .288 .05 -.300 -.07 IS .04 - 08 - OI .258 170 I4 .01 -.13 .07 .02 .02 .13 .15A - 05 ° 08 .03 .11 -.IO .01 -.ll .00 .06 —.03 -.05 -.04 -.0'5 ' .II - 02 .07 (Alpha) ' Not Cmutabla A-p:.05 813:0] C-D:.001 0-01.“)0] II III AAA‘. IIIIIIII o....... O AAA—LAAAAA C 9 l L L A ALLAL A —‘ I I 7 I I I D I I O . . . . V ass a .4 as IJ_L1._LLA__ILLAL COO—4° _o_n goumNgbwav ) )) N Full- Part Prof (H 70 et a1 (1981) using a combination of education level, prestige and income data for each occupation. It was ex- pected, therefore, that education would be correlated with the occupational status variables. It was decided, however, to retain both types of variables in the analyses since the occupational status scores were originally calculated on a population which may not be equivalent to the population studied here. The status of the preferred occupation was dropped because it's mean was not significantly different from the status of the major occupation (T=l.72, df=262,p>.05) and because it had excessive missing data. Discriminant Analyses Thus far the characteristics of the jobseekers; the effect of job change and unemployment on occupational status and wages; and the intercorrelations among the independent variables have been examined. The results of the major focus of this research will now be reported. The main purpose of this study has been to assess the multivariate predictive power of the independent variables in relation to the depen- dent variables. The six discriminant analyses which are re— ported below share the following features. The predictor variables listed in Table 3 were included in the stepwise discriminant function analyses using the Wilks Lambda method. Because the nature of this research was 71 exploratory the significance of the F—to—enter criteria was set at p<.l0. Listwise deletion was used to handle missing data, therefore, the number of cases varied for each set of analyses. The mean for each independent variable was sub- stituted for missing values during the classification analy— ses, therefore, the number of cases reported for the pre- diction results approximated the total number of valid cases for the dependent variable in question. This is considered a conservative procedure. One table will be presented for each analysis. The table will include the variables which entered the analysis when the criteria was set at p<.l0. Also included are their respective means on the dependent variable, and their stand- ardized coefficients. The information presented for the dis- criminant function includes the canonical correlation coef- ficient, and the chi-square for the function. The predic— tion results are also found in this table. The analyses dealing with each placement criterion will be presented in sequence. Employed or Unemployed. The research question regarding the best set of predictors that differentiate between employed and unemployed jobseekers is considered here. The results of the discriminant analysis on employment status are presented in Table 12. The analysis is based on a sample of 144 cases. Intercorrelations among the predictors entered 72 Na.mm u ooHuHmmmHo >Huomuuoo w um wv poonQEocs vm. ooonmEmc: me on poNoHEEm 44.- posoHasm oo>onEmcD po>onEm msouu oHouucoO pouomooum pwuwwmoom Hmsu04 mustmm oouoHooum Hooo. u oocmonHcmHm h n .m.o m.mN n owumsqmuch mm. mOH n.~ mo.NH wNH no.N Hm.HH :oHuoosom ov. FOH mm. mH.H vNH mN. no.H dsouu oHccum mp. OHH Nv. NN. nNH mm. SH. non umoq E0uw couHuom om.u OHH Nw.mmom mo.OHnm hNH ~.mNHv Nv.vmmm bow umoq co ouscoe mo. oHH om. om. ANH om. we. mcaxoompon 00w mCOmmom HmHococh Nm. voH mm.onHH mn.Hmn nHH m.Hon om.va ucoEmoHdE0c3 Lo coHuouso novv. :oHumHouuou Howmcocou mucmHoauoooo 2 am cap: 2 am can: moabmppm> own no :ou o mwmmamca m ooNoHQEoc: ooNoHQEm ucmcHEHuomHo oH. w a n wouc010ulm mo cocoonHcmHm msuoum pw>oHQEocoupm>oHdEm co mHm>Hoc< ucocHEHHOmHQ NH OHDMB ewHHz 73 at p<.10 were low (r<.30). Six variables entered the analysis as predictors. The 'duration of unemployment before the jobseeker became a client at SES' was the first variable to enter the analysis. It had a standardized coefficient of (.52) with the discrim- inant function. The next five variables to enter were 'finances given as the major reason for the client's job search'(.65); 'tenure on last job' (-.80); 'retired from last job' (.75); ethnic status (.40); and education (.35). The canonical correlation of employment status and these predictors was (.44) which represented a statistically sig- nificant relationship (p<.02). The Wilks Lambda was (.81). The prediction results indicate that by using this dis- criminant function it is possible to correctly classify em- ployment status 60% of the time. This is significantly dif- ferent from chance which would be approximately 51% (£23.92,df=1,p<.05); 237 cases were used in this classific- ation analysis. According to this analysis an employed participant had the following characteristics. The individual had been un— employed for a shorter time before entering the SES and was looking for work to keep busy, feel involved or simply be- cause s/he valued work. The profile would also include the following: The employed person would have been at his/her last place of employment for a longer period of time, left his/her last job for reasons other than retirement, would be 74 caucasian, and would have a low educational level. The 'retirement from last job' variable is correlated with 'tenure on that last job' (r=.45). The high intercor- relation of these two variables could present a potential interpretation problem. Both variables were excluded from further consideration. The profile outlined above can be interpreted as follows: The relationship between the 'duration of the un- employment period' and subsequent employment may be a func- tion of two factors; the time an individual spends unemploy- ed can effect his/her general psychological well-being and consequently their motivation and behaviour (Hill,l977). Hill found that the longer jobseekers were unemployed the more discouraged and pessimistic they became. Therefore, clients who had already been unemployed for a long time when they entered SES may have been more despondent and less optimistic than those who had been unemployed for a shorter period. This is further substantiated by the small, but significant correlation between membership in the job club and the length of time unemployed (r=.12,p<.05). The longer the period of unemployment, the less likely the individual was to have joined the job club. Since the job club was the main service provided by the agency clients who did not use the service were probably less motivated or did not like the services. The unemployment period may have also effected the value of the client's job skills in the labour market. Human 75 Capital theory suggests that the value of job skills de- preciate when they are not used regularly. At the very least, prospective employers might perceive the abilities of a job applicant who had been unemployed for a long period with some skepticism. The positive relationship between 'finances as the main reason for wanting to find a job', and the subsequent status of being unemployed suggested that clients who said 'finances' actually meant that they wanted a job only if they could find one that provided income significantly high- er than they were currently receiving. Since the wage/hr of the average re-employed jobseeker decreased, individuals in this group were less likely to find an acceptable job, and thus, remained unemployed. On the other hand, respondents that said they wanted a job to 'keep busy' or to 'feel useful' were probably less discerning in the qualities they required of a job and were, therefore, more readily re-employed. The term 'expectations' can be substituted for 'the main reason for wanting to work'. The qualities that clients expected from a job seemed to determine the likelihood of them getting a job. This finding suggests the hypothesis 'that a jobseeker's expectations regarding the purpose and type of employment desired are directly related to the prob- ability of achieving a successful placement'. This is con— sistent with an expectancy model of motivation and shall be 76 discussed further in Chapter IV, as a special case of Human Capital theory. Research by Sobel (1972), DOL (1979), and OECD (1966) also found that older workers delay taking a job in the hope of finding one that meets their expectations. The relationship between ethnic and employment statuses found in this study is consistent with the general litera- ture on the association between race and employment. Bould (1980) found that older black males were more likely to ex— perience longer periods of unemployment than white males. The significantly higher age of the minority group in this sample may also explain their lower employment rate. Those participants that were re-employed had less for- mal education than those who were unemployed. One explana- tion for this outcome is that jobseekers with more education were likely to have left higher status jobs with better pay. This might provide more financial security and consequently less urgency to find a job. Another related explanation is that workers with lower educational levels were less picky about the type of jobs they were willing to accept. Their expectations about prospective jobs influenced their prob- ability of successful placement. The higher incomes of the better educated clients may have given them the leeway to indulge their expectations of receiving a 'good' job. The results presented above are consistent with the literature and with Human Capital theory. They also support an expectancy model of motivation as a key element in under- 77 standing the factors that lead to re—employment. The per- centage of variance explained by these variables, however, is very small (19%), suggesting that there are other vari— ables not included in this study that influence job place- ment. The literature suggests that individuals who choose to retire rather than look for work may be different in some ways from other workers (Kingson, 1981). These differences include health and income. Employed or Retired. It was decided to consider 'retired' as a different status from 'unemployed'. Whenever a client reported that s/he was retired the interviewer probed to determine if the client would accept a job. If the client would take a job they were classified as unemployed-not looking. However, if they were not interested in working they were considered 'retired'. The main intent of this analysis was to determine which variables differen- tiate between 'employed' and 'retired' workers. The results of the discriminant analysis are presented in Table 13. There were 92 cases included in the analysis. Intercorrela- tions among the variables that entered at p<.l0 were low (r<.30). The five predictors are presented below in the order in which they entered the analysis with their respective stand- ardized coefficients in parentheses. Retirement from the client's last place of employment (.60); score on the in— 78 nm.m> n ooHMHmmmHo >Huoouuoo w «H mH coHHuom wh.H oouHuom am mOH oosoHaem mm.) poNoHEEm poHHumm pomonEm msouu moHonucmO oouowomnm coubwmmnm Hopped muHcmom coHHOHomum Hooo. v a u mocmoauacoam m u .m.o NH.mm n ponmaomunzo Hm.l hN hv.Hmom wm.mmvm hNH >.mNHv Nv.vmmm non ummq . :0 ounces em. HN MN.v am.w HHH vH.m on.m non umoH co ommz om. mN Hm. mm.H mNH mv. mm.H cOHumucoHuo QDHO now No. SN mm. mH.m mNH an. mh.m oosuHuu¢ xuoz oncHuwcH om. SN Hm. «v. hNH mm. 5H. non umMH Eoum coHHuom mHmm. coHumHonuou HmoHcocmu mucoHonmooo 2 am com: 2 am new: moHanum> mud no so IIIIIII p mmmsgwcaom ponauom ooNoHaem uGMCHEHuomHQ OH. H m u Houcoloulm mo OOCMOHMHcmHm Nmuz msuoum pwHHuomloomonEm co mHm>H0c< ucmcHEHuomHo MH OHQMB 79 trinsic work attitude scale (-.62); attendance at job club orientation (.54); wage on last job (.54); and 'tenure on last job' (-.51). The canonical correlation between the de— pendent variable and these predictors was .58 which repre- sented a statistically significant relationship (p<.001). The value of the Wilks Lambda was (.66). The discriminant function explained 34% of the variance. The prediction results indicated that the function de- fined by the five predictors can correctly classify indiv— iduals as employed or retired 76% of the time. This classif- ication percentage was significantly different from chance, which would be 71% (Hi3.84, df=l,p<.005). The classification analysis included 154 cases. According to this analysis employed jobseekers were different from retired ones in the following manner. The retired individuals had retired from their last job, re- ceived lower scores on the intrinsic work attitude scale, had not attended a job club orientation session, had higher wages on their last job and had been at their last place of employment for a longer period. The wage and tenure variables should be interpreted cautiously because they are correlated with each other (r=.46) and with retirement from the last job (r=.44,r=.45). The fact that retirement from the last job dif- ferentiated between 'employed' and 'retired' clients may be attributed to several factors. It may be because of social 80 security rules limiting the amount of work a recipient is allowed to perform before penalties are imposed. It may be that individual's who retire from their last job are more financially secure and, therefore, can afford to wait for 'right' job to appear. The association of low intrinsic work attitude scores with being retired suggests that persons retired from their last jobs who have a low score were simply entertaining the idea of going back to work, but subsequently decided to re- main retired. This is further substantiated by the rela- tionship between 'attendance at a job club orientation' and being 'retired'. The retired clients were less likely to attend an orientation session, which leads either to the conclusion that they were not strongly motivated to find work, or that the services of the agency did not meet their needs. These findings lend support to motivation as a key fac- tor in predicting employment status. Low intrinsic work scores mean that the participant does not value 'work' very highly. The retired individual who does not value work would need a substantial incentive to return to the labour market. Why then, did these individuals initially come to an agency that specializes in job placement? One answer is that they may have been in a transition phase between the statuses of employed and retired. During this transition people tend to explore alternatives as they adjust to their new retirement 81 status. Thus, it is possible that these individuals consi- dered returning to the work force and therefore registered with SES. After a period of time they decided that they really didn't want to return to work and discontinued their job search activities. While some individual chose retirement over employment others who were able to find jobs still continued looking for better jobs. Continued Job Search-Reemployed. The analysis reported here presents the set of variables that best differentiate between re-employed jobseekers who reported that 'they were still looking for other work' and 'those who were no longer searching'. The results for this discriminant analysis are presented in Table 14. There were 79 cases included in the analysis. All of the independent variables had low intercor- relations (r<.30). There were three predictors: the dichotomous variable 'the main reason the subject left their last place of emp- loyment was because the job was only temporary' (.78); 'finances' (.78); and education (.63). The canonical cor— relation coefficient for this variable and the dependent variable was (.41) which represented a statistically signif- icant relationship (p<.0l). The value for the Wilks Lambda was (.83). The discriminant function explained approxim- ately 17% of the variance. The prediction results indicated that the function 82 mm.Nm n ponHmmoHo wHuomuuoo N we 5N mconoq uoz Nm.u oconoq uoz 0N Nm mconoq Hm. mCHHOOH mcHxOOH uoz mcHxOOH QDOHU mpHouucoU couoHpoum pouompoum Hmsuo< muHsmom coHuoHcoum moo. v a u monounmficmam m u .a.o mo.mH u oonmsomuaco mm. mm mm.N SN.HH Hm mm.N mm.HH :oHumosom mconmmn0h mm. mm om. mv. Nm om. om. How mGOmmom HMHocmch me. me om. so. mm am. we. snmnomeoe ma; non ammo He. coHumHouuou HMOHcocmo mucoHonoooo 2 am can: 2 am new: moHanoo> pouHoumocmum mHmmHmcm mcHxOOH uoz mconoq ucmcHEHuomHQ OH. H m n umucoICUIm mo oOCMOHMHcmHm mhuz macho cmonmEm 00w noumom now ooscHucou c0 mHmmHmc< ucmcHEHHomHQ vH magma 83 defined by the predictors can correctly classify employed individuals as looking or not looking for other work 63% of the time. The difference between this classification per- centage and chance (.52) was statistically significant (xi6.39,df=l,p<.02). The following profile can be drawn from these results. An employed individual who is still actively searching for other work would have held a temporary job before coming to SES; 'finances would be the main reason for jobseeking' and the clients still searching would have a higher educational level than the clients no longer searching. If the client's last job was temporary then it is prob- able that s/he either had difficulty finding a permanent job because of inadequate or obsolete job skills or because his/her occupation required temporary placements (e.g. Home care nurses). In the former case the same characteristics could lead to an unsatisfactory job placement and the sub- sequent search for a better one. It is also possible that these individuals originally took temporary jobs because they could not find satisfactory permanent jobs. They may have wanted jobs with higher earning potential or jobs that were more consistent with their prior education and exper- ience. The results suggested that these clients were still looking for better jobs. Their expectations of higher wages and occupational status were not met by the positions they found through the SES. Despite their desire for better 84 quality jobs these individuals wanted or needed to work badly enough to take undesirable jobs. They continued their job search in order to find the type of job which they originally sought. Unemployed clients that were still look- ing for work were motivated by different factors. Continued Job Search-Unemployed. This discriminant analysis examined the set of variables that best discrimin- ated between unemployed clients who were 'still looking for work' and those who were 'no longer actively searching'. The analysis included a sample of 65 clients who were unemployed at the follow-up. Intercorrelations among the independent variables were low (r<.30). These results are presented in Table 15. Five variables met the criteria (p<.l0) to enter. Each variable is presented here in the order in which it entered the analysis, along with the respective standardized coef— ficient in parentheses: Intrinsic work attitude scale (.70), intake date (.66), sex (-.49), health as a reason for term- ination from last job (.44); and job search activity report- ed during the first month at SES (.42). The canonical cor— relation coefficient for this set of predictors and the de- pendent variable was (.73) which represented a statistically significant relationship (p<.0001). The Wilks Lambda was (.46), which indicated that the function accounted for 54% of the variance. The prediction results indicated that the function 85 wH.mn u ooHuHmmoHu >Huoouuou a mm vH oconoq uoz oH Hm ocaxooa oconoq uoz wconoa macho oouoHooum couuHooum Hosuu< muHamom coHuoHooua o~.~- ocaxooq 002 am. oconoq moHouucoU Hoo. v a hm.mv mucouHuHcon .m.o couo90manO No. me vo.o~ wk.m om mh.va Ao.mH >0a>auo< :uuoom non oouuoaom vv. mv vN. co. Ho on. HH. non umoq ucH>moq u0u cowoom cuHoo: mv.u mv 0v. H>.H Ho mv. om.H xow mo. mv 0N.vHH mo.mmmmvH Hm Ho.MNH no.0nomvH ouoo oxoucn oh. mv Hm. mH.n Hm cm. mo.m oosuHuu< nNmb. xuoz uHmcHuucH coHuoHouuou HooHcocou mucoaowucooo 2 am cum: 2 am coo: moflnoaum> nonaouoocoum mHmmHoc< mconoq uoz mconoq ucocHEHuomHa mo.wa a bouzoIOunm uo mucouHuHchm 0:000 oo>oHdEoca new couoow non mH anmF ooscHucou co mHmaHoc< acocHEHuomHo 86 defined by the five predictors could correctly differentiate between unemployed clients who were searching for work and those who were no longer searching 78% of the time. The dif- ference between this classification result and chance (.51) was statistically significant (Ké6.39,df=l,p<.02). These results suggested that an unemployed client who was still looking for work would have a high intrinisic work attitude score, would have been in the program for a shorter amount of time than those who were no longer actively searching, was male, would have left his/her last job because of poor health and would have reported a higher num- ber of job search activities during the first month in the program. The profile presented by these findings is consistent with much of the previous research on unemployed older work- ers. Sobel (1972) observed that the length of time a person was unemployed was associated with the probability that s/he would eventually find work. This has been referred to as the 'discouraged worker syndrome' (Bould,l980;Sobel, 1972; and Colledge & Bartholomew, 1980). The mediating factor seems to have been the motivation of the jobseeker and the reason that s/he is looking for work. In this sample the longer a person was at SES the less likely s/he was to be active- ly looking for work. These jobseekers might have become 'discouraged'. In fact, as reported below, 27% of the res- pondents who were no longer looking for work stated that 87 they had stopped searching because they were discouraged. It appears that the desire to 'work for the sake of working' also effected continued job search behaviour. Individuals with high intrinsic work scores were still looking for work despite the length of time they had been unemployed. The fact that females were less likely to maintain their job search may be related to expectations. The data reported above indicated that females were likely to receive significantly lower wages. The longer a woman encountered unfavourable job prospects the more discouraged she might have become. Another possibility is that males continued their job search longer because they were married and had more dependents. It is not clear from these results why poor health on the last job would account for continued job search be— haviour. One possible reason is that an involuntary term- ination from the last place of employment strengthened the jobseeker's perceived need to re-enter the labour force. It is interesting that the number of job search activ- ities reported during the first month was associated with subsequent job search behaviour. The fact that both of these measures were obtained during the same interview leads to the suspiscion that this finding represents concurrent in- stead of predictive validity. This was also the case with the intrinsic 'work value' scores. Nevertheless the results still support an association between continued job search 88 and motivation or discouragement. In some cases these jobseekers remained unemployed because they were unable to find jobs that met their criter- ia. Sobel (1972) found that many older workers prefered to continue looking for work until their resources ran out be- fore accepting an undesirable job. An alternative strategy was to accept a part-time job until a better one surfaced. Full-Part Time Job Status. The set of variables that differentiates best between re—employed jobseekers receiving full or part-time jobs are presented below. There were 78 cases included in this analysis. The results are presented in Table 16. Interpretation of these results would be difficult because many of the predictors that entered were highly cor- related. The first variable was wage on the last job which had a standardized coefficient of (.86). The second variable was retirement from the last job (-.66). The remaining vari— ables were 'last job was temporary' (.42); spouse with income' (.54); 'satisfaction with last job' (.48); 'laid off from last job' (.42); and 'tenure on last job' (.48). In addition to high intercorrelations among the independent variables there was inconsistency in the relationship of the predictors with the criterion. For example the zero order correlation between 'wage on the last job' and the criterion was (—.29), while the standardized coefficient for this relationship was (.83). Therefore, it wasn't meaningful to 89 NN.Nh u pwHuHmmmHo >Huoouuou w mm mH uuom mm.| upon 5H mN HHsm NH.H HHsm uumm HHSE mocha mpHOHOCOU pwuoHpmum pouoHpmum Hmsuo< muHsmmm coHuoHpoum Hoo. v m u mocmoHuHcmHm h u .m.o NH.mm u boomsqmlHLO mv. vm ov.mHmm ov.hmmm Nv mN.mwwv NH.¢an ouscoe Nv. vm we. om. Nv om. mv. wuo oHoq mv. Hm mm. Nn.m Ne mo. mo.m boo umoq co coHuommeuom 306 cm. vm om. Nv. Nv Hm. om. meoocH nuH3 mmSOQm No. oo oH. oo. me now. moo. spmnanoe mm3 bow umma ww.n vm ov. 0N. Nv mm. NH. non ummq Eouw pouHuom mm. Hn Ho.m vo.m mm vo.m oo.m boo ummq :0 momz mm. coHuonuuou HmoHcocmu . muCOHOwawOU 2 am cow: 2 am coo: moHQchm> ONHpuoocou o mHmNHmc< m puma HHsm ucocHEHuomHo oH. M Q n pouchOuum mo wocmonHcmHm -ilflIMWmm mSuMum non weHe uummlHHsm co mHm>Hoc< ucmcHEHuomHo wH anme 90 continue this analysis. Regpession Analyses In the previous section the placement criteria were dichotomous. In this section the continuous nature of the outcome variables required multivariate regression analyses. The four placement criteria presented in this section include: The duration of the job search; the wage/hr re- ceived on the new job; the worker's job satisfaction; and the number of job search activities reported during the client's first month at SES. Additional findings regarding the unemployed client's perception of why s/he was still unemployed are also reported. Duration of the Job Search. A stepwise multiple regres— sion analysis was used to examine the relationship between the independent variables and the duration of the job search. There were 65 cases used in the analysis. The de- pendent variable was computed by subtracting the client's intake date from the starting date of the new job. The re- sults are presented in Table 17. The first predictor to enter was 'wage at the previous job'. The beta weight for wage was negative (-l8.85). The next variable, 'age', did not contribute significantly to the prediction (p<.08), therefore, the regression equation was defined by only one predictor. The multiple R between 91 mm.m oo oo.m . m mo.m HH.© mm.wH| H coHOMH>oQ oumocmum cows m oHnoHum> moo. v m mv.o No. mH. mv. who. mN.m oo< N ooh woo. v m mN.m om.u vH. om. voo. mN.m umMH :0 woos H oocooHMHcmHm m m m m oocmon noucm QHQMHum> moum HHouo>o onEHm N onHuHsz IHcmHm cu m commom non mo coHuouso co mHm>Hmcd conmoumom NH OHQMB 92 the predictor and the duration of job search (.37) repre— sented a statistically significant relationship (p<.004). The overall R—Square was .14. Since this is an exploratory study the beta weight for age is given in Table 17. According to these results, the higher the wage re- ceived on the previous job the less time needed to find a new job. This makes rational sense since wages are usually associated with the skill level of a job. It should be easier for a highly skilled person to find a job than some- one who has no skills to offer an employer. The amount of time required to find a job would also be dependent on whether there was a demand for the skills of the jobseeker. Someone with lower but more appropriate skills for a specific labour market might find a job quicker. The posi- tive relationship between higher wages and the duration of the job search suggests that more skills lead to quicker placement. Another possibility is that part of the variance assoc- iated with wages could actually be explained by inflation. The length of time a person was unemployed before coming to SES was negatively correlated with the wage/hr received at the last job (-.l4). The longer jobseekers were separated from their last job the lower the wage they received on that job. This suggested that a higher wage might have also rep— resented the recency of the jobseeker's employment exper- ience. Human Capital theory states that the longer somone 93 is out of work the more their skills suffer from obseles- cence. Therefore individuals with lower wages might take longer to find a job because their original skills were not marketable or because their skills had diminished in value. Still a third alternative is that the positive rela- tionship between wages and education (r=.25) somehow re- flected the ability to learn appropriate job search skills. Human Capital theory posits that job search skills effect the efficiency of the job search and the subsequent length of time needed to secure employment. The influence of wages is examined further in the next set of analyses. Wages on New Job. A stepwise analysis was used in order to determine the best set of predictors for the wages re— ceived by the re-employed clients on their new jobs. There were 74 cases included in this analysis. The results are presented in Table 18. The dependent variable, 'wage on the new job', was reported during the follow-up interview. Two variables contributed significantly to the re- gression equation; 'wage on the previous job' (B=.45) and 'retirement as the reason for leaving the last job' (B=-3.55). The multiple R between the two predictors and the criterion (.67) represented a statistically significant relationship (p<.0001). The R-Square for the equation was (.45). According to these results, higher wages on the pre- vious job lead to higher wages on the new job. Individuals 94 on. ma. mm.m- non umMH Eoum pmuHuom eo.~ oa.m co. non ommH co oooz coHuoH>oo oumocoum coo: m mHQMHuo> Hooo. v a m.mm om.- mo. so. ooo. ~.~m non umMH Eoum oouHuom m Hooo. v a o.mm mo. no. mo. ooo. o.mm now ummH :0 0003 H mocnonHcmHm m m m m oUCMOHu umucm oHQMHum> moum HHmoo>o oneHm m oHaHuHsz (Hcon 0» m vouz ommz ooonmEomm co mmeHmc< conmouvmm mH OHDMB 95 retired form their last job were more likely to have re- ceived lower wages on their new jobs. Since wages were associated with the status of the job (r=.25) it is rational that those jobseekers who had jobs with high wages before they became unemployed should be more likely to obtain higher wages on their new jobs. Wages ref- lected experience, skills and occupational status, and were therefore indicative of the value of a worker's human capital. The reason that workers 'retired from their last job' were more likely to receive lower wages when they became re-employed probably had to do with the type of jobs they obtained. The relationship between retiring from the last job and age (r=.18) indicated that individuals who had re- tired were slightly older. As age increased the preference for part-time employment also increased (r=.-15) As dis- cussed above the wages of part-time jobs were usually lower. The association between 'retired from last job' and 'income' (r=.20) provided additional support for this hypothesis. Individuals with greater financial resources should be able to better afford retirement. The correlation between 'retired from' and 'the reason for the job search' (r=-.21) confirmed that the motivation of these jobseekers to work was primarily to keep busy and feel useful. Therefore they readily accepted lower paying jobs. The factors motivating the client to accept lower paying jobs should also influence 96 their reported job satisfaction. Reemployed Job Satisfaction. The next analysis examines the prediction of job satisfaction on the new job. There were 78 cases included in a stepwise multiple regression analysis. The results are presented in Table 19. Five variables contributed significantly to the regres- sion equation: Satisfaction on the previous job (B=.20), retirement from the last job (B=.41), sex of the client (B=.39), intrinsic work attitude scale (B=.21) and spouse with income (B=-.32). The multiple R between these predic- tors and the dependent variable (.54) represented a statis- tically significant relationship (p<.0001). The R-Square was (.29). According to these results, high job satisfaction by workers on the new job was associated with high job satis- faction on the last job; retirement from that job; being female; having a strong intrinsic work attitude and not having a spouse with income. Since satisfaction on both the previous and current jobs and the intrinsic value of work score were all measured concurrently there is a potential confound resulting from possible contamination of the retro- spective scales by current feelings toward work. Terborg et a1 (1980) showed that retrospective questions asked along with questions regarding current attitudes have a stronger association with the same items asked during a previous interview than with the concurrent 97 om. Hm. mm.- m Mb. mm.m HN. v om. mo.H om. m ov. oH. Ho. m mo. oo.m om. H coHumewo pumpcmum coo: m oHanum> Hooo. v a oo.m mm.u om. om. moo. oo.m osooca :uH3 omsomm m Hooo. v a om.o Am. mm. om. moo. mH.v opsuouua xoo3 OHmcHuucH v Hoo. v a oo.o AH. Hm. oo. ooo. mo.m xom m Hoo. v a oo.o Hm. AH. Ho. mmo. oo.m ooh ummH scum ponooom m moo. v a om.o mm. HH. mm. moo. om.o non pound co coHuommeumm H OOCMOHMchHm m m m m OOCMOHw kucm OHDMHHm> mmum HHouo>O onEHm N onHuHsz IHcmHm ou m oouz coHUOMMMHumm non poonmEMIom co mHm>Hoc¢ conmmumwm 0H OHQMB 98 questions about present attitudes. In light of this it was possible to treat the concurrent measures in the present study as if they were asked during different interviews. However, this author prefered to treat these results con— servatively. Therefore, the only conclusion that was drawn was that the relationship between a high intrinsic work attitude score and job satisfaction suggested that the ex- pectations workers had about the type of jobs they wanted could effect their satisfaction with those jobs. If a job- seeker wanted to work just to keep busy then it should take less to satisfy him/her than someone whose prime motivation for working was financial. Workers that retired from their last job probably ex- pected less from their new jobs. This was supported by the fact that they tended to take part-time jobs. Since they wanted less from a job it is logical that they were more satisfied with their jobs. The results reported above showed that females earned less than males, had lower status jobs, and were more likely to be divorced, widowed, or separated. The relationship be— tween the sex of a jobseeker and his/her extrinsic work attitude score (r=.20) indicated that money was a major reason that females in this sample wanted to work. The negative correlation between income and extrinsic work attitude scores (-.22) suggested that the monetary orienta- tion of these women toward work was a function of need. 99 Therefore it is possible that because of their uncomfortable life situations the women were more easily satisfied with their new jobs. It is also possible that women were more likely to be satisfied then men because they did not exper- ience a significant drop in wages when they became re-employed. The reason that workers who did not have an income earning spouse were more satisfied with their jobs may have been related to occupational status. The correlation be- tween the occupational status of the new jobs and the 'spouse' variable (r=.-19) indicated that individuals who did not have a spouse with extra income were employed at higher status jobs. The higher status of these jobs may have been responsible for the job satisfaction reported by these clients. Reid (1972) found that occupational status was also associated with job search activity. The higher the occupa- tional status the more job search activity reported. This may have been partially due to the worker's satisfaction with his/her previous employment. Reid reported that lower skilled workers accurately perceived that their job search would be more difficult. If a jobseeker perceived that the available employment options lead to dissatisfying jobs then his/her job search activity might have been effected. Job Search Activity During First Month At SES. The final analysis pertained to the prediction of job search 100 mm. ho.m hm.H xao3 oncHuucH om.m Ho hm.u 00¢ mm.H vv.N Oh.N ON Hmuoe coHumH>oo oumccmum coo: m OHQmHHm> Hoo. v a Hv.n mH. HH. om. mmo. mm.v xu03 OHmCHHucH m Hoo. v a oo.o o~.u oo. om. oHo. oo.o moo m Hoo. v m mv.oH ON. 00. ON. HOO. m¢.OH ON HMUOB H OOCMOHMHcmHm m m m m OOCMOHM umucm OHQmHuo> moum HHmum>O onEHm N mHmHuHsz IHcmHm ou m nnHuz cucoz umuHm mcHuao NuH>Huo¢ nouoom non ON OHQMB co mHmemc« conmoumom 101 activity during the first month in the program. There were 177 cases included in a stepwise multiple regression. The number of job search activities was reported retrospectively during the follow-up interview (see Table 20). Three variables contributed significantly to the pre— diction equation: The total number of jobs held during the last twenty years (B=2.7);the age of the client (B=-.57);and the intrinsic work attitude (B=l.87). The multiple R between these predictors and the number of job search activities (.34) represented a statistically significant relationship (p<.0001). The overall R-Square was (.11). The profile of a client demonstrating high job search activity would be someone who had many jobs in the last twenty years, was relatively young, and who had a strong intrinsic work attitude. Human Capital theory treats job search skills as a special type of value. The better a person's job search skills the more likely they were to find employment. The data suggested that an individual's job search skills in- creased with the number of jobs s/he had in the last twenty years. The evidence for this is the higher job search activity of workers reported by clients who have had more jobs in the last twenty years. It appears that job search activity decreased as age increased. The relationship between 'age' and 'full/part-time job status' (r=-.15) indicated that 'older 102 jobseekers' required a job for different reasons than 'younger jobseekers'. The reason they were looking for work was to keep busy and supplement pension income. Therefore there was less urgency to engage in job search activities. The amount of effort put into the job search was not necessarily a function of why a person wanted a job; but how much they wanted a job. The association of a high intrinsic work score with job search activity suggested that wanting to work for 'the sake of working' can lead to high frequency of job search activity. The motivation to engage in job search activities, however can wane if the behaviour does not yield results. The evidence for this is reported below. Additional Findings. In order to better understand why some jobseekers remained unemployed and in some cases became discouraged the following questions were asked during the follow-up interviews. Respondents who reported that they were unemployed and no longer looking for work were asked to list their reasons for terminating their job search activities. Participants were allowed to give more than one reason. Poor health (27.3%) and discouragement (27.3%) were the most frequently cited answers. Respondents who reported that they were still unem- ployed were asked what they thought were the reasons that they had not found employment. Participants were encouraged to provide as many answers as possible. Age (22.7%), health 103 (29.5%), and minimal job search effort (50%) were the major reasons cited. CHAPTER IV DISCUSSION The purpose of this study was to examine the con- ditions that contribute or detract from older jobseekers securing employment. The research design was predictive, using data collected from client files during an intake pro- cedure for the Senior Employment Service of Lansing. Addi- tional data was obtained through telephone follow-up inter- views, which asked respondents to provide retrospective, as well as current information on their employment status. This section will focus on the implications of the results for both theory and policy. The major findings will be reviewed along with a discussion of methodological issues. This will be followed by an examination of the fit between the results and the theoretical models presented in Chapter I; a discussion of the policy implications for in- dividual programs as well as state and national policies targeted toward older workers; and finally some concluding remarks along with consideration of future avenues of re- search. 104 105 Major Findings The major findings associated with each of the four questions/hypotheses that were posed in this study are sum— marized below. The research on unemployed older workers has not dealt extensively with the older jobseeker that uses a senior em- ployment service. Since the number of these types of pro— grams is rapidly expanding one objective of this study was to document the characteristics of the clients of one rep- resentative senior employment service. The findings in- dicated that the typical client of SES was between 55 and 61 years old; had completed high school; was caucasian; had previously been employed in low to moderate status oc- cupations like sales or clerical work; was economically dis- advantaged; and had been unemployed an average of 1.7 years before seeking assistance from the agency. Generally females were worse off economically than males. Literature on reemployed older workers indicated that this group generally suffered a decrease in their occupa— tional status. The findings of this study indicated that the jobseekers in this sample did not experience a loss of occupational status. The third major set of findings indicated that the re- employed workers in this study suffered a drop in their wages. This is consistent with the research reported by 106 Sobel (1972). Non-CETA clients experienced the largest decrease in earning power. The findings of the final research question indicated that work history, and motivation were the most important type of variables associated with the prediction of the nine success related criteria. The intent of this study was to examine the effect of unemployment on older jobseekers and to discern predictor variables that enable the prediction of success related criteria. The findings just reviewed indicated that it was possible to address the major research questions. There were some methodological issues that effect interpretation of these findings that should be considered before theory and policy implications are discussed. Methodological Issues Unfortunately because much of the data was collected retrospectively some doubt was raised concerning the valid- ity of some of the predictors. This issue was already dis- cussed in Chapter III. The concern was whether the respon- dents were able to accurately recall their physical or psychological state at intake. Research reported by Terborg, Howard & Maxwell (1980) indicated that answers given by subjects to retrospective questions were more similar to answers given on a pre-test than to concurrent 107 questions regarding their present attitudes or status. The implication of the Terborg et a1 study for the present re— search is that the respondents probably understood the dif- ference between the retrospective and current questions and were able to accurately recall their previous ideas and feelings. Given the above discussion the veracity of the responses to the retrospective questions is less doubtful. Nevertheless further 'predictive' studies should be con- ducted before the retrospective items are included in any screening protocol. Another major methodological issue raised by the re— sults was generalizability. The original intent of sampling from SES was to expand the knowledge base regarding older jobseekers using the services of an agency. In this respect the scope of the findings were limited to this specific pop- ulation. The conclusions that are drawn in this paper are only meant to apply to clients of senior employment services, although it would be possible to argue for gen- eralizability to older workers in general. The following discussion applies to all of the findings reported in this study. In regard to the generalizability of these results to other senior employment services there are two major consid- erations: The program model used by an agency; and the geo- graphic location of the agency. There are various program models used by different senior employment agencies, all of 108 which can be categorized as either individual or group focused. The SES uses a combined approach which should increase the generalizability of the results at least in regard to the treatment model. A more serious consideration has to do with the geographic location of the agency and the accompanying economic situation. The diversified economy of the Tri-County region creates a population that is composed of all facets of the labour market. In this respect the sam— ple should be representative of most major occupation groups. On the other hand the diversified economy has les- sened the impact of the economic depression. This may de- crease generalizability in respect to more depressed regions in the state, but it may improve generalizability with areas outside the state. In regard to either possibility the ef- fect of the economy on this study is probably limited to the actual number of job placements. The characteristics of the sample and job search effort should only be effected minim- ally. In summary the generalizability of these findings should be fairly good. The conclusions drawn below regard- ing theory and policy should be relevant for agencies in other regions using either of the two models outlined above. There is of course a need to replicate this study in order to further improve generalizability. 109 Theoretical Implications In Chapter I Human Capital theory was presented as the most rational model for understanding what happens to an unemployed older worker. The theory as modified by Sobel (1972) accomodates two important aspects of the unemployment/job search phenomena, which combine to in- fluence the 'success' of the jobseeker. The two major parts of the theory can be characterized as: the potential econom- ic value of the jobseeker for the prospective employer; and the potential valence employment represents to the job- seeker. Sobel (1972) does not directly address the rela- tionship or dynamics that function between these aspects of the theory, but his paper implies that placement outcomes are the result of multiple forces working in combination. Despite his inferences regarding the multivariate nature of the phenomena Sobel stresses the role of the jobseeker's 'economic value' to the employer as the key factor moderat- ing job placement. There is little doubt that the value rep- resented by the older worker's skills and experience has important influence on the success of his/her job search. The research reported by Sobel (1972), Bould (1980) and others strongly support this relationship. The deemphasis of the motivational component of the theory however appears to be the result of the type of data available to theorists, rather than a reflection of the important role that the job- 110 seeker's perceptions of the value represented by work has on job placement. Sobel and others have taken a macro-economic approach to the phenomena using large aggregated samples. Their focus has been on tracing the movement of these aggregates in the labour market. Their research has focused on variables such as occupational status, tenure, number of days unemployed and income, but has not assessed 'work value'. In Chapter I it was stated that the role of motivation in Human Capital theory had not been stressed enough. One of the objectives of this study was to empirically determine the extent to which 'work value' influences job placement. Without the motivation component Human Capital theory does not provide an adequate model of job placement. The theory predicts that a jobseeker with low capital value would have a very low probability of successfully finding a job in a labour market that demanded workers with high technology skills. Similarly the theory would predict that a highly skilled worker looking for work in a high technology labour market most certainly would find employment. However in both cases it is possible to demonstrate that these pre— dictions could be wrong unless motivation is considered. If the low skilled worker was highly motivated s/he might put more effort into his/her job search or perhaps be more willing to accept a lower paying job. Either of these mot- ivation related behaviours might result in a successful 111 placement. If the highly skilled jobseeker was dissatisfied with his/her career or if s/he simply wished to obtain a job demanding less responsibility then s/he might decide to turn down job offers from high technology firms. In both of these examples the predictions of the one factor theory would be erroneous. Therefore for Human Capital theory to provide a useful model of job placement it is essential to include motiva- tion. The inclusion of motivation in the theory allows a broader range of predictions. Although the findings of this study support the two factor model it is possible that fu- ture evidence could indicate that other factors should be included in the model. For instance if a highly skilled and motivated jobseeker was unable to get a job in a high tech- nology labour market then the present theory would prove inadequate. The results suggest that both of the components out- lined above have at least equal importance in respect to the utility of Human Capital theory as a model for understanding the search behaviour and ultimate probability of placement for older jobseekers. The evidence supporting each aspect of the theory will be presented separately followed by a dis- cussion of how the two forces are associated. Economic Value of W0rkers in Labour Market. The major tenet of Human Capital theory is that workers represent a commodity in the labour market. Within this framework the 112 worker is thought to have a distinct economic value that is determined by his/her skills and experience and by the de- mands of the labour market. The probability of job place- ment is directly related to the value represented by the jobseeker's human capital. Overall the clients in this study had difficulty find- ing employment. Only 47% had found a job. Of those who were still unemployed 44% could be described as discouraged work- ers. The characteristics of the sample indicated that the human capital of these jobseekers was relatively low. Most of the workers had been employed in low status jobs, had only a high school education and had been unemployed for an average of 1.7 years. These characteristics are consistent with the findings reported by Sobel (1972). The theory posits that another consequence of the dep- reciation of human capital is a drop in the worker's earning power. Depreciation is the result of obsolescence, dis- placement, and demand. All of the clients were displaced from their last place of employment. The economic conditions in Michigan had reduced the demand for labour oriented toward low technology. Given the rising demand for workers in high technology it is likely that the low technology skills of most of the sample were obsolete. The outcome of these factors was a drop in the average wage/hr earned by reemployed workers. Wages received on a job reflected the value of the worker in the labour market. The fact that 113 individuals who received lower wages on their previous jobs also received lower wages when reemployed further supported this aspect of the theory. The findings associated with the final set of research questions also lent some support to the influence of the worker's economic value on job placement and other related outcome criteria. The discussion of these findings will be limited to those variables that either support or contradict the economic aspect of Human Capital theory. The findings that support the motivational aspect of the theory will be discussed later. Job Placement. The length of the unemployment period, the tenure at the last job and the education level of the jobseeker were associated with job placement. The longer a person was unemployed before registering with SES the less likely s/he was to become reemployed. This is consistent with the effects of obsolescence. The skills of an indivi- dual would depreciate from lack of use. The longer the tenure on the last job, the greater the probability of reemployment. Intuitively this seems consis- tent with the theory. A person should acquire more skills and experience with longer tenure. This should result in increased human capital and subsequent reemployment. Sobel (1972) came to a different conclusion. The findings of his research indicated that long tenure lead to longer periods of unemployment. He posited that the skills learned on a 114 job were unique to that place of employment. Thus the value of the employee for the specific employer increased with time. However when the worker became unemployed his/her skills were of little value in the labour market. The con- flicting findings may be due to the nature of the respective populations. Sobel studied a more diverse group of older workers who did not necessarily request assistance from a specialized agency. It might be that the relatively homo- geneous occupational status of the clients at 885 was dif- ferent from a general population sample. The skills ac- quired on a low status job should be less specific than skills learned in a higher status position. If this was the case then long tenure on a low status job would increase general skills and thus would also increase human capital. Therefore both findings may be correct. The association of lower education levels with place- ment seems to contradict the theory. However, given the low status occupations obtained by most of the clients a higher education might actually lower the value of the jobseeker. A prospective employer might see the worker as overqualified and therefore not well matched for a job. The labour market might have demanded workers with skills not requiring much education. In this case the higher educated worker did not fit the needs of the market. This was supported by the analysis on the continued job search activity of reemployed workers. The more highly educated workers were more likely 115 to still be looking for another job despite the fact that they were employed. This suggested that these individuals were underemployed and perhaps only accepted their present jobs because their choices were limited by the labour market. Duration of Job Search. The theory also posits that the value of an individual's human capital will influence the length of time it takes for him/her to find a job. Common sense suggests that the person who was more likely to find a job would also have taken less time to find that job. The analysis on the duration of the job search supported these conclusions. The wage received on the last job significantly predicted the length of the job search. The higher the wage the shorter the job search. As was discussed above the wage received on a job should be indicative of the value of the worker. Job Search Activity. The findings discussed above sup- port the importance of capital in determining the successful placement of older jobseekers. The theory posits that job skills alone are not sufficient to ensure placement. The jobseeker needs to be able to market his/her skills and s/he must be motivated to work. The association between the num- ber of jobs held in the last 20 years and job search activ— ity during the first month in the program supported the notion that job search skills are learned. The more job changes a person experienced the more job search activity 116 s/he demonstrated when unemployed. The motivational aspects of the theory posit that the 'value of work' will also in- fluence successful placement. Value of Work. The second aspect of Human Capital theory deals with the way the jobseeker perceived work. The value held for work should determine both the amount of ef- fort expended in the job search and the probability of the worker accepting a job offer. Essentially this is the same as saying that the motivation of the jobseeker will in- fluence the success of their job search. Motivation in this context is best explained by an expectancy-valence model (Vroom,1964; Porter & Lawler,l968; Campbell & Pritchard, 1976). The possible outcomes are full employment, partial employment, unemployment, or retirement. The valence attrib- uted to each of these outcomes, combined with the expecta- tion that a given amount of effort can lead to the desired outcome, should influence the choices made by the indivi- dual. S/he first must decide how much effort to put into looking for a job and then if offered the opportunity to work s/he must decide whether to accept the job. The prob- ability that the jobseeker will be offered a job is moderated by the value of their skills and experience. Thus both components of the theory work in combination. The actual motivation to work was not directly measured in this study. The 'value of work' was assessed by an open-ended question and by intrinsic and extrinsic 'work 117 value' scales. The association found between these variables and 'success' criteria support that work motivation was an important placement factor. It was shown that reemployed workers experienced a significant drop in pay. One explanation given above was that unemployment decreased the capital value of the job- seeker thus resulting in lower wages. Another alternative is that older workers attributed different values to work. The fact that 62% of the reemployed workers were at part-time jobs, with lower wages, might have been because they wanted to work for reasons other than finances. This would be especially true for previously retired workers who were more likely to report that they were looking for work to keep busy or feel useful. Work Reasons. The reason given for looking for work reflects the value attributed to work. Clients gave two types of reasons. They wanted to work for financial reasons or to keep busy and feel useful. Their responses on this item were significantly related to the success of their job search. Individuals who wanted to work to keep busy were more likely to find work. This is consistent with the theory. If the value of work was seen as providing an opportunity to keep busy or feel useful then the pool of possible jobs was fairly large. These jobseekers were less discerning about the type of jobs they would accept. How- ever if the reason for working was to improve or maintain 118 the individual's financial position, then the pool of avail- able jobs would be much smaller. This pool would be smaller because the jobseeker's depreciated human capital would re- duce the employment options available to him/her. The lat- ter was supported by the fact that most of the reemployed clients suffered a drop in their wages. Therefore financi- ally motivated jobseekers would only accept a job that main- tained or improved their financial situation. Sobel (1972) found that unemployed workers were less willing to accept a job with lower wages if their financial resources were adequate. That was the case in the present study. Re— employed clients who were still searching for work gave fin— ances as their motivation for wanting another job. It ap- pears that these jobseekers could not afford to wait for a job that met their financial expectations. Therefore they accepted a less desirable job until they could find approp- riate employment. The clients who chose to retire rather than seek em- ployment had lower intrinsic 'work value' scores and had not attended a job club orientation session. Since most of the clients who retired had pension incomes it is not surprising to find that they had lower intrinsic 'work value' scores. If they were internally motivated to work then they would not have retired. The fact that they didn't attend an orien- tation suggested that their motivation to work was low even when they originally entered SES. It is possible that these 119 individuals were only experimenting with the idea of return- ing to work. Prolonged Job Search. The motivation of a person to work is seen as contributing to the value of their human capital. This was especially true when then the economic value of the capital was not sufficient enough to help the individual find a job quickly. The continued job search activity of unemployed workers was related to the value that they attributed to work. These individuals had high intrin- sic 'work value' scores. In the terms of the theory this meant that the outcome they desired was a job that would enable them to keep busy or feel useful. Since the pool of jobs that met those criteria was relatively large the indiv- idual could expect that a reasonable amount of effort would lead to a job. The fact that these workers were still highly motivated to find work might have been because they had been unemployed for a shorter period than those clients that had given up. The data suggested that the motivation to continue looking for a job decreased as the period of unemployment increased. The association between 'work value' and prolong- ed job search seemed to be related to the sex of the worker. Males were more likely to still be looking for work than females. This may be due to societal norms for this age group that stress that the male identity is associated with work. The job search activity reported during the first month 120 in the program was also related to continued job search be- haviour. This suggested that the motivation level demon- strated by the client when s/he first entered the program persisted over time. The intial job search activity was associated with intrinsic 'work value'. The more intrinsic- ally motivated a client was the more job search activity s/he reported. This is consistent with the way 'work value' was seen as contributing to an overall human capital theory. The value attributed to the placement outcome combined with the expectation that the outcome was achievable, would con- tribute to a successful placement. The intrinsically motiv- ated jobseekers believed that their job search activities would get them a job. Summary. The findings of this study support a Human Capital theory that has two major components. The value of an individual's human capital is a function of the skills and experiences that s/he has accumulated as well as the value s/he attributes to work. Separately neither factor provided an adequate model of job placement. By considering both factors it was possible to significantly predict place- ment related criteria. The contribution of motivation to the overall theory was demonstrated. Nevertheless further re- search is needed to examine the relative contribution of each of these factors in the prediction of placement out- comes. The low percentage of variance explained by some of the analyses suggests that either better measures should be 121 developed to assess the economic and motivation components of the theory or perhaps that factors not presently included in the theory should be considered. The theory has facilitated the identification of spec- ific variables that are relevant to placement success. Con- sidered within the context of the model the economic and motivation factors associated with the theory have impli- cations for policies related to employment interventions for older workers. Policy Implications Rising concern about the employment problems of this nation's older workers has brought with it a call to action. The media and other public forums have pointed to the lack of concrete policy on these issues. There has been little data available to policymakers that pertains directly to the sub-population of older worker's at most risk. The working poor seem to bear the greatest hardship from ageist employ- ment policies. The findings of this study suggest some specific policies that could begin to address the problems of this group. Outreach. The unemployed older workers that used SES could be characterized as the working poor. The low wages and income of this group combined with relatively low status occupations presented a picture of individuals with few 122 financial resources other than the income derived from work. Long periods of unemployment deprived these individuals of their main source of income. The average client did not register with 888 until s/he had been unemployed for 1.7 years. This suggests that these clients waited until their unemployment benefits ran out before seeking assistance. It is not clear to what extent these jobseekers actually looked for work before coming to 888. The findings of this study and of research reported by Sobel (1972), Bould (1980) and Marshall & Cottam (1981) indicated that jobseekers became discouraged and quit searching for work after they had been unemployed for a prolonged period. The implication is that interventions focused on providing individual level assis— tance should somehow reach prospective clients as soon as possible after the end of their employment. Hill's (1977) study of the stages experienced by the unemployed suggests that outreach may be thwarted by the tendency to perceive the first few weeks after termination as a vacation period. It is not clear how these stages are manifested in a de— pressed economy, however efforts should concentrate on facilitating voluntary participation. Eligibility Criteria. The different impact of unemploy- ment on the subsequent wages of CETA and non-CETA clients has important implications for the current eligibility policy of government sponsored employment programs. Present policy restricts eligibility to individuals with annual 123 incomes that are below the poverty level. The income of CETA and non-CETA clients differed by approximately $3600, but the mean income of the non-CETA group was still only $6141. The wage levels of each group corresponded to their income levels. After the period of unemployment the CETA group did not experience a significant drop in wages, but the non-CETA clients suffered a decrease of $1.37 per hour. Research with general population samples (Sobel, 1972) have found the same loss of earning power. Perhaps interventions should target the non—CETA group in order to prevent this decline in wages. Present policy restricting services to this pop— ulation are punitive and counterproductive. The ineligible worker is required to expend all of his/her financial re- sources before receiving assistance. Additionally the delay of service delivery caused by eligibility criteria may ex- tend the length of the unemployment period resulting in a greater risk of discouraging the jobseeker. Underemployment. The structure of the labour market may contribute to the decline in wages. Most job openings are entry level positions. Many older jobseekers found that in order to get work they had to accept jobs that did not re- quire all of their skills and experience. Over 60% of the reemployed workers were only able to get part-time employ- ment. This was especially true for workers who had previ— ously received lower wages. Consequently many of the older workers were underemployed and underpaid. State and federal 124 governments could address this problem by providing sub— sidies and tax incentives to employers that provide appro- priate jobs to unemployed older workers. The findings of the latest Harris (1982) survey on aging in the United States provide a reasonable counter ar- guement to these reccomendations. The survey indicated that most older workers would prefer to work part-time. It could be inferred that because some older workers want to work only part-time that earnings are not a major concern of this group. The data from this study defuses this arguement. More than 51% of the 265 respondents participating in this research indicated that finances were their major motivation for working. The policies discussed thus far range in scope from specific program changes to broad national economic inter- ventions. The common base leading to these reccomendations has been the financial impact of unemployment on older work- ers. Another policy issue raised by this data is the need for specific programming within senior employment services directed towards older women. The findings of this study reveal that women had lower incomes, lower wages and were employed in lower status occupations. Even though women did not experience a significant drop in wages they continued to earn less than their male counterparts. women were also more likely to be widowed, divorced or separated. Given the age of this group it is not clear whether a lifetime of struc- 125 tural underemployment can be rectified by an intervention at this stage in the lifecycle. At the very least agencies should consider programming that addresses the specific sup- port needs of this group and the particular problems assoc- iated with the 'displaced homemaker'. Employability Review. The results of the discriminant and regression analyses indicated that two major categories of variables influenced the probability that an older job- seeker would find work. The motivation of the client and his/her job skills combined to determine the probability of a successful placement. Clients whose job search was motivated by finances were more likely to remain unemployed because they were unable to find jobs at an acceptable pay level. Individuals who retired rather than remain unemploy— ed or work had retired from their last job and were not intrinsically motivated to work. Reemployed workers that were forced to accept lower paying jobs were more likely to continue looking for other work. The implication of these findings is that programs serving older workers could im- prove their services by using an improved employability development review that would enable the agency to get a better picture of the client's strengths and weaknesses and help the clients to clarify their own expectations regarding available employment options. The type of information that would be most useful includes: the length of time the person has been unemployed; 126 their previous wage and occupation; tenure on the last job; reason they left their last job; education; intrinsic and extrinsic 'work value'; and the total number of jobs held during the last 20 years. Unfortunately the accuracy of the classification equa- tions computed in this study was not sufficient to allow agencies to use them to screen clients. However the infor- mation listed above can be used to understand some of the forces that are impinging on particular clients. The opt- imal application of this information is to use time during intake to help the client to clarify his/her expectations regarding the fit between his/her human capital and the la- bour market. This should also involve a discussion of the client's 'work value' and the probability that s/he could find the type of work that would satisfy the attributes that s/he requires from a job. The purpose of such a session is not to discourage the jobseeker, but to help him/her focus job search efforts in directions that would best fit his/her needs. The data suggested that clients with 'work values' that did not match the reality of the labour market were less likely to find employment. It is possible that some clients will re-think their motivation to work. Some may even de- cide to remain unemployed. A large percentage of new clients at SES never return to receive services after their initial intake. Helping clients to clarify their expect- 127 ations during the intial contact may increase the number of clients that actually participate in the program and the number that actually find jobs. Retraining. One final reccomendation derived from this study pertains to the retraining focus of the Job Partner— ship Training Act of 1983. Research studies by Sobel (1972), Chan & Fowles (1980), Vandergroot (1979) and Colledge & Bartholomew (1980) point out that older workers are most likely to be found in declining industries. The present study found that 66% of the sample were in general service or factory jobs. The characteristics of this sample strongly suggested that many of the jobseekers using a senior employ- ment agency could be classified as structurally under- employed. Retraining efforts should keep in perspective the skills and education of older workers. The emphasis on re- training for high technology jobs may raise expectations that cannot be met within the limited number of years that many of these workers want to continue working. Future Research The main objectives of this study have been accomplish- ed. The findings have shed some light on the factors that influence the labour market behaviour of older workers. Sup- port was found for a two factor Human Capital theory. Two major streams of future research are suggested by the 128 conclusions drawn from this study. Future research should improve upon the design of the present study and explore the feasibility of several innovative interventions which deal with the unemployment problems of older workers. This study should be replicated in order to increase the generalizability of the conclusions and to improve the accuracy of the predictive models. Samples should be drawn from programs using different treatment models and from dif- ferent geographic areas. The predictive models should be tested using pre-data that is totally obtained during in- take. Further research should validate the 'work value' scales and examine the relationship between a jobseeker's 'work value' and his/her work motivation. In this respect a better measure of motivation should be used in future studies. More work is also needed to further determine the ver- acity of the two factor Human Capital model of job place- ment. Research should focus on the relative contribution of each factor to the prediction of placement related criteria. Emphasis should also be placed on identifying other factors that would expand the theory and explain the remaining var— iance. Several problem areas were discovered during the course of this research that would benefit from experimentally focused social innovations. The first of these interventions should be concerned with designing and evaluating a method 129 of helping unemployed older workers to clarify the fit between their 'work values' and the 'true' labour market. The potential benefit of this intervention was outlined in the previous section. Programs specifically designed for women should be evaluated. Such programs should consider the family status, income and occupational status of older women. The research should also question whether a program segregated by sex differs significantly from one that is integrated. Almost half of the clients were still unemployed when they were interviewed. Approximately 44% of this number had given up looking for work. These individuals could be class- ified as discouraged workers. Research by Hill (1977) and OECD (1966) indicated that discouragement is a major problem encountered by jobseekers. Specific interventions targeted at the discouraged worker might return some of these workers to the labour market. Research should also focus on pre- venting the 'discouraged worker syndrome'. Job clubs cur— rently encourage clients to use their services for as long as they need. However the repetitive cycle of the program's modules might discourage some clients from using the job club after an extend period of enrollment. A major problem encountered by older displaced workers was the drop in wages they experienced when they become re- employed. Retraining and advocacy interventions may prove to be solutions to this problem. All of these interventions 130 should be empirically evaluated. Conclusions This study has shown that a Human Capital theory com- prised of economic and motivational components provides a good model of the relationship between the characteristics of older jobseekers and their probability of successfully searching and obtaining employment. The findings also indi- cated that placement outcomes can be empirically predicted using information available during routine intake proce- dures. 'w0rk value' was found to be an important factor associated with the clients motivation to find employment. Despite the promising results of the prediction equations the accuracy was disappointingly low. Future research should improve the efficacy of the predictive models and address the problem areas that were identified during this study. One final note should be made. The design of this study lends itself to conclusions at an individual level of analy- sis. Some of the results supported interventions that focus on changing the client. Interventions at this level are needed, but these interventions should not be allowed to obfuscate the changes that are necessary at the systems level. The real problem is the structural unemployment that results from a system that pays more attention to the needs of the labour market than it does to individual workers. APPENDIX A CETA INCOME LEVELS FOR DETERMINING ELIGIBILITY NON-METROPOLITAN AREAS .mam>oa cmuflaomouuwzlcoz mm: ou mum Axucsou cmusmcm> ummoxmv mmmom« "pom xflm um>o HwnEwE Omn.m omm.m oam.a osH.H omm.H Hmcofluaoom 30mm Mom ohm.ma ovo.ma OHN.MH omm.m oH~.HH o ova.oa o~>.ma oom.HH oom.m 0mm.m m omo.mH omo.HH osm.m omH.n omv.m a omo.HH omv.m oms.n omo.o one.» m oso.m oom.o omo.m omm.v omo.m m omm.v m omH.v m cmv.m w omo.m m on.¢ m H ouumzlcoz ouuszcoz ouuwzlcoz Eumm Eummlcoz wNHm haflamh Hmnm-o Hmnmuo Hmumno Hmuaaam wumo m>fluoommm quqq quqq quqq suum>om wooa wmm won mzo «mdmmfl ZdBHAOmomBMZIZOZ VBHAHmHOHQm DZHszmMBMQ mow mqm>mq MEOUZH dBmu d XHszmmd 131 APPENDIX B DOL APPLICATION FORM APPENDLlLIBHfl--.,V -.. .- wouu ov mum no W C AUON [ELIGIBILITY pom 3‘" I. A J —; "AID _ y I! I II. V! gum" u IN lug. "Rial . . I. Info-oily: A. oi: - _ ......umyu...-[I]j CD LIED m-I-MCD gm" II] CD 5,13 um..ovu..(...m) uno- hon-o- hm» n.” .- .. Mil. and (not We) I No.1; mu 0 I. Auk:- [MI-norm”. M noun-o: hum unu- t-Sm-Mtuz 1 I IIITIIIIJIT home] van [1 I u ...... cu: mare-«(Inna mm= rm .. mfiijj) {IDCIID ... m: [mp-q: [I] u-mo-w- ;- mum-u. mm“; 0“- 0'“ _m-" slag W "- 0“ -‘ ~ A- "0:.“— I I gm M A]. IJVMyOIIUmuoSuluaqu? DY. on. flNO.myouamIndau-Iofl DY. Ohio GUAM I4 no you many meadow. and (o: hot-on ”uni/urn)? D No D Fund-n OM4- OW Pad-o I! you on month. new. “Icon I... O! M II Doyoulownflukhoolbtuout or. on. Doyoohnlcm ov- ouo ‘I'Wv QV-DYENB'“ W‘naollcloolw-phutm 7 J.“ ."“Dy. 0'. I6 Do you out a yon-cu o: Inc-Id Map that would Inn out mun-v $Y. DI. .' 'Vmu. 9“". Y :7 run yo‘amnmfli 0.x autumn-«tuneup? DY. No ' "3;. MM'IIISMIOS I75) u ouuamz to ”a. Try-“W 'ospauomuu Crow-r l9.&umn~w¢omcn~dub¢¢ynlmolm~mf DY. 0'“ 9.;MMHW7WMWN :0 KenyoumuauluCfiAho'I-I DY. 0N0 IIYflJoanoI-Ihcfouomfouoau-oyoo mun-I DYQ Du. won I- CETA. USE EXACT DATES. that: w: __ M m Ion w": 0.: ol LII I1 Icon m.°,_~.. a... a... on I“ “MI... I. A. wdmzovm .../mmJ-ontnn. h '- ‘.. W‘— DY. r—‘F W “Fiat: 'm' H M“ Ill-I“: I“ m .0 u. QII for pol out: o tony o! Ind/m Job. 2| Von "mot-y - META: (La ALLyooo fot .1 ho: (to ha ‘1 Doom. LB! EXACT DATES) ‘ Iolovtll m. hunch-.01 hon- I:- u—o E: :3.- o—nu—n— ”WWW: 0"" DY- 2..me ‘ om: W ”.mmmAI DYI- DNo “.11.. h CIA “ IOIINI (m “MIN”: Toul— VIII: Icon VI _ flout CI? icon IS. Funny Fun! In 2 I... I" m Och-r PM I.“ W I“ 22. law Smu 111mm norms SIX (0 ION“. IIDoyou'IOVIOESMamumummIdom-dm DY. 0N0 . - . 1! yrs. w. my? CD turn-£7 u». := “.5.th an: 24 Doyou ucnvuosumuyu-mm-mmmmmmr DY. DNo ”J Wn""""l.;.. ".fu'wm'lzou :5 loco-u won-um; ' . ' .ON Y: gnu. L I. lenmdy'vmcomvloo .- “w‘ ...-.24....“ Inn-am mono-un- “IN-II. U"..- ”my lot CETA. la! «a “may mambo: Imn VII. you. me (but Into-o. Lu ALL family Minoan. m- 1 they do NOT by! m. . b Doyou.otmluuly~-hnmmmwywnath-yoltbolouom? DY. DM 7.... 1.. IIY‘Es.umau: DADC OBI Doc-um . Ow DFooISuwl I...” 11'! 00““ (”CO“) I... " “- ‘m c.DoyouovuyhuuhunhnmuummmMWMMI UV. ONO ‘1“.l “haul. "Jul.” mm; If YES. use“ 0 lam-m 0 8m 0 Duomty mm LB“. “"77““ d Ooyouuwon:Int-(AInbuym-MNMMMmmlmd-dflMu-on)? Enqsgmu 11.355143", Om 0N0 ouaoostuu. u-IMLLSIL 36 a An you euucmly "can“ Unanpbyno-I Conufllbl? DY” D No 0AM I005 LUIL w II YES. nan um: duc‘ 19. loam“ PA wall Hon om >~ "a" you llhcuutd au Sun and chm Uncuployon-I Conn-uno- booofiu-fllunhhyoc? D You O No 10. M5 to . II M: 3" Do you. 0! any memo-v o! your funny do "on! o: anon-I (noun? 0 Yoo No D Y. D N DNIA I! An you now 0: My. 70.; "n bun .- ml. or 00‘ 7'00“”- ov anon? D You No 2!. “4qu In '8‘ Im: 2° 1 Au you "land to any County Comm-mu 0: other the!“ official In the SIM. o! Macho-Y D You O No 8'00“! 0'“ b An you Ruled loony pom mpmnbh lot my pan 0! IN CETA procu- II lIoSIuool WI 0 Y. O No '- mm W— . If YES to cum gunman. ammo mm. ooumy at who and "booming; a. m 1; km - Tm— ' ~07 1 Federal hut" 20 CF! Put 6?! noun um CETA 099mm. to non-nod Io ”on“ (loll Sod-I Soc-Ann Ul- Mon‘ Um Hume-o You my not a: damn: (HNIIIDOIII In a CETA Will H you "In. to dado- you: Soda] Sccunty Number. ”7"“ You! Socul Scum” Numb" an In m noIcIy Io: (In mun-I noudtoom mo! (In uptown" (ITA mm. om 0'” on )o cunncmou uao cuaruuvv 21. My m: E Y- E N- Iamty um I In» nun-vol mu none-non um that the Inlet-um. on. u (m Io tho but of my too-hop. I! V“. W lu- mo unn than "a nalmmuon I Mn now!“ u «Mm Io who no anneal». m I my m to pro- ”. W Mar: 0 You 0 No no- dour-mm w suppon mu o'fluuol. I unannu um Mun-I Io you. a. non-cud am my ro- 34. Minute-doom“: D You O No mu m my mmuuuon from th- mom I on duo «mo the In. about Io W lot-hum If I a. loud «hook III" "woman! and my to ma Io! (tool nod/o: pol-Ivy. I moo m o! It. Nor-om lot ”- m": “nu..." Y. unhouon woo. and um an I III in all Io Mm whiny. n D a m o B POI muons: Mao-mm-..» 11¢ a o my D D . no I: 0 1m 0 O 5...... HT...“- 15... W :r—_: V, D a on. ...... D D con-nun; VII 0 D loco-I'M: a“: :3: '0; ‘0 IV. 3. DAY ”VI" 0' IIJOIIIF‘I'Y: autumnal-on. J th-fidflz DY. ON» "I. IICIITIFICATION: ‘de.~m Icon-{yIn!IndommulfioMme‘m-‘MIMDIWM UV. OM; IuIHuno «I'onnom. fiance-u w ill.“- w ' -u ”1’ I09 (In-I III” 132 APPENDIX C DOL INTAKE FORM APPENDIX C DOL INTAKE FORM II. ...-n. .22.. :3 .l 5411. .25 8...... u... . . Illrl ID. ID. ID. ..I! an. . it... ll .33: :8 D D D ...—U. 19310. .- ..:II-Itft-a... 0.4.3 .33 1:! 3.39.... .325 5.1.8... u .6 35qu ago. 3 20:50:53 18.5.... 2 «36:3 a 18:3 3 .2295 o £35.10. a .5365: 9 34... a... : gnu 4 ...1 16D. oxDu {Du 3D~ ID. ID. {Du 7. ‘laiD. ....D. ....D. ....D. 8»D. ....D. :»D. ... ..IisisgD. ...D cggu 9325 ...: iris-SID cram. 53!. 5.3.8.0: .3898 :39... 5:0 $09.22...qu ...... (v.39 2: 3 32.. 8 59.3.0 3 39335 a :55! u. out}. = 8: 55.58 :9. 9:; 3.5.2. o. Inigo} a gills-i... .. ...D. ...-9- .31.26 n o: .u I. a 8. u . s! u S» . a:Dn 8.0a 8.0. 3.1.3.28.“ . I’m. 8pm. Sum. uopm. m .3. D. .... D. ....D. ‘ng. 95!; 2 9.3. 9.2qu So 5536 3253 . 0:0. {>0— 5 83:3 —|— — — _L - ioxuguo ...ISSS ill-.32 8.6.5.! 2:45 9:93. an 1.8: ...-I: ... 83:529. :4. .0 3.: 5.2308 2 1.1: 5.59. 2 85.385 2 35(1! 2 9.33 : I‘ll. $1.1.an {Du ...Du gnu. 3.9245130. ratio-ID. ED. ....D. ...D. m .260» €12.28... n I’lzlzdxiioD. . logoD .9... o 2.3.21... FD. .1153}... D... 328... a SIoBBJDu 3.533 523.33 {coitus-33.8: ...-1.8.0. Ira-.3950. ...—U. in....°2u¢‘!.!uoafiD. whole 2 c... t 39688 .n 2:45 5.3. .801. 2 3:: .833 u. 3.1... 29.333 .. Bozo... : 1.8.3260 voila-«D. :5..an .02.. BI. o-oIbD 8.9:}...0 gin-uliiuu EDn I‘ll—U. octagon 282133.120» inlatI..1.oDu B In lotyDa 33D. .i!!.ou-ID~...II.D. B $3.3. s 3...... 3:5 iii... .5. mm BBB » . 2:85.! .o 9. 2 3.1. 5! a .- Sol! 3::- 1522 .. sin-:55 . .35 9...: .. in o. a? o EB ......- o { - u I): u . .illf .. B Is . 3:836 #500 o u I... a .343 u ll l i [.l. to In. . I11" ii loo. 3 on filo-t. ~13.- h :23: 622.8 “5.8 2355.58 3...:- so. .306. can... of... 5...... 30:9 u ">480 003 IIEIOO 30.; E U!»‘ Q 5E8 B :3 5‘42; 0 lg! n DHDHD g A Isbigo‘uoo ogiaon 02,30— (.0— >3 :2 0 >5 nwpn ‘ Econ. ux¢whz_ wxh >I Owhflglou w. 0.. DMD lggaé n (:5! n no. a ism. .m. B ore-.250 ID. :3» g ply>OJ|Iu o0 31!- 133 APPENDIX D ARCHIVAL DATA CODEBOOK o HCMHm H mm wH Axoon o: u N mm» H H xumHn xomnov HonEmE nsHo non H mm mH on u N wow u H COHumucmHHo nsHo QOn ooocmuud H mm oH mOHono 0: NH xCMHQ oUHoso mmmm I 0000 cum ouucm3 xuo3 mo mmwa H mmImm MH ouHonu o: «H xcmHQ 00Hono mmmm I 0000 ocm ooucmz xuo: mo mmwa H mmImH NH pooch moon coHummsuoo mom oUHono ammo I oooo umH caucus xuoz mo mama H mHImH HH 0C .III N m@% H H HMO H VH OH 0: n m mm» H H coHumuuommcmue H mH m 0c n m mm» H H mcoHumuHEHH nuHmom H NH w nonuHo u m wEHu Hst u m oEHu ummm n H oouHmoU once xuoz H HH 5 oomoHu u m m>Huum u H msumum mHHm H 0H m o xcmHm H m m am I o quEsc DH m.umMHHm> H th v mm I 0 Homes: DH m.uooou H wIm m o xcmHm H v m mwv I 0 Monaco DH ucoHHU H mIH H Honesz monasz quEsz ouoomm casHou wHQMHum> xoommoou dbda H¢>Hmumd D xHozmmmé 134 135 o: u m mm> n N £2 n H oo>onEm n v nocuo n m omonmEm Iuoocs n N omonmecs u H susom n m Hoonom cH u N on n H Hoonum :mH: whom I o Hoonum ann omuwHQEOU n m uoomouo n N ucmpsum H H mm I 00 mm I oo ucoumm mHmCHm n v HmsoH>HQCH acoocomooucoc u m uwnEmE >HHEmm umsuo u N NHHEww ucoumm N cH ucoumm n H mm I 00 mm I oo ooumummwm n m ooouo>Ho n v oozooH3 n m onch u N ooHuuoE n H mmHoumm \momsmou n v Hocuo u m ucoonou u N .m.D u H UHCMQmH: n m cmHmc u v o>Humc n m xomHm n N ouH£3 n H meEom u N onE u H mmuoo mum—”HMO @000 msumum mouom HDOQMH msumym Hoonow msumum HmcoHumuscm woumeEoo momma Hmoanm mucmocomoo mo Honasz msumum NHHEom NHHEmm cH Hones: I mnucoa m >HHEmm CH Nunez: I mQHCOE m msumum HmuHum: mHnmcoNHuHU macho 0chum xom mmfl A.ucoov o xHozmmm< H mv hv 0v mv vvlmv NvIHv 0v mmlmm hmlmm mm vm mm Nm HmIOM om mN wN hN ON mN vN MN NN HN ON 0H mH NH 136 o: n N mox n H o mam I o H ooo wooH m>oom u v wOOHIom u m HHmHH ”mmIHH u N suuw>oo u H HOOH m>onm u v HOOHIom u m HHmHH wmmIHn u m suum>om u H mmmmm I 00000 mmmmm I 00000 womnm moou coHummoouo xuwcu mmmm I 0000 o w o w mm I mm n 00 I oo o: n N mm> u H o: n N mo> n H mmm I 000 uonsmnxo u m ucmEHmHo o no: u N acmEHmHo n H on u N mm> n H on n N mm> n H mm. mcH>HmomH uoc ass I mHnHmHHm xcmHm umnasc DH ucwHHU H0359: oumu xcmHm mnucoe o I msumum oHEocoom mSUCOE m mDUMUm UMEOCOUW ooNHHmsccm mLuCOE m I oEOUCH NHHEmm omNHHmsccm mzucoe m I mEOUCH NHHEmm AcoHumpHHm> How“ non demolcoc umMH mo mooo coHummsouo noH Hemo Icoc ummH mo mmmz NHusom mxmm3 ON mo mH I ©o>onEwCD mxwm3 NH HO OH I om>onEmca UoNOHmEocs mxooz oucmHsmcH ucmE>onEwca H0>QH I'- m omNOHQEmoH\mmo>MH n < ooHHuou \mcHomumms\onuHcsuuommo now A. UCOUV O xHQmemd HHNNN r-‘lr-Ir-IIH H mm vm mmIHm om mhlhh mm mm which molmw vaHw ooIhm mm mm vaNm Hm om mv bv mo mv vv mv Nv Hv og mm mm hm mm mm mm vm mm Nm Hm 137 mm o: n mow dBmUICOZ n dfimo o: n mmx o: n mow N N N N on n N mo> N N N m oc u mo> o: u mom 0: n mow Honuo u ooHnm HmHoomm u N mum EmcuwH> on n N mw> o: u N mmN on n N mo» NuHHHnmmHo mo>H>H5m n N ucoEwHHuou o: n N wow 0: n N mm» o: n N mm» oc n N mo> o: u N mo» mH r-Il Hm r-I H H r-I'U H H I-‘l r-I P! H H H O O v-Ir-II-Ir-I wouonEoo mQSI3oHH0m mo Monasz xcmHm ooHHouco demo uoEHom moumum unmEHHoucm umcumw wmm3 NHMEHHN ucoonmu Eumm noncommo nmmeoEon ooumHmmHo Hoxno3 Eumm oocommom\ucmumHz mcmeomm smHHmcm oouHEHH cmuoum> NH msumum cmuoum> oommucm>ommHo NHHmoHEocoom UwQQMOHocmm mm>u umz3 .>uHusowm HmHOOm mCH>HmowH «H quusowm HmHoom msH>kuom mxmmz NH HO OH umMH «m >cm mcH>Hmomm UHHnsm Honuo mcH>Hmuom Hmm mcH>Hmomm 00mm mCH>Hmumm ..ucoo. o xHozmmm< NNNNNNNNNN NNNN NN NNNN QOHImOH QOH MOH NOH HOH OOH mm mm mm mm mm vm mm Nm Hm om mm mm mm mm no mm mm v0 mm No Ho 00 mm mm mm mm mm vm mm Nm Hm Om mv mv 138 >umuom6mu umo> Homo umoum m N o mmm I 000 >mp I Coco: m uCouxm meow on n N HHm um uoc u H mmm I 000 owNHprQCmICOC n N UmNHonnsm u H O: u N mm» H H mmm I 000 who» I who I nuCOE umo> I >mo I CuCOE u N quCmEumm u H HMConmuuo n m HHsm u N uumo n H mmmm I 0000 0C n N mm» H H umo> I >mo I CuCoz Hones: oumu meHm COHHMCHEuou I mmeCH oumo CoHumCHEHmB msI3oHH0m Co mH a mooH>uom mo CoHumsHm>m mmoumoum CH mumo mo Hones: n msIonH0m I oxmuCH noH omuHuHmnsm ICOC .m> omNHUHmnsm quENOHmEm CH umwuouCH noH umuHu ou mNmu mo umnECC u mCHuHmum I mxmuCH mumo oxmuCH now mo mumo mCHuumum Numuomsou no UCoCCEHom HmConmooo HHsm\uHmm ooou\msumam CoHummsuoo UoNOHmEm msI3oHH0m ummH mo mama H.ucoov o xHozmmmC NNNN N NNNN NNNN 00H vaIomH mvHIva ovHIHvH ovH mMHINMH QMH mMH VMHINMH HMHImNH mNHIONH mHH mHH hHHIvHH MHH hHHIhOH mm Nm Hm om mm mm mm on m5 vb mm Nb Hm on mm mm APPENDIX E WORKSHEET INSTRUCTIONS APPENDIX WORKSHOP PRCXIEDURES The most important thing to remember when transferring data from the client files to the worksheet is accuracy and comprehensiveness. The manner in which you organize your data collection is up to you, as long as all the necessary information is recorded. The following is a list of required procedures which you must conduct for each file. After this list is a recarmended method of organizing your data collection in the files. These methods can be modified in response to personal preference or situational demands. 1) CROSS VALIDATION: Mlld‘l of the information we will be collecting is recorded in more than one location. Whenever possible check at least one alternate source to determine the validity of your information. This is definitely required for CEI'A STATUS, TERMINATION DAT'E, INTAKE DATE, PHONE NUMBER, NUMBER OF FOLIDN-UPS AND CLIENT ID NUMBER. All Of these items can be cross validated by checking the BLACK BOOK. 2) INCONSISTENCIES: Whenever you encounter discrepant entries check with Joe. If possible we will clear up the problem during the evening. If not the SES staff will clear up the problem the next day. WHENEVER YOUARE INDOUBTCHECKORASK! 3) MISSING DAT'A: airing the first few days check with me if you care across a file which is missing any informationwhich you are supposed to code. The SES staff will help us to locate the missing information. In the event that we cannot find the information or in those cases where the information requested is not applicable (e.g., Info on starting date of new job for clients who haven't found a jdo yet) we code zeroes wherever there is missing data. 4) HCMEWORK: It's to everyone's advantage to minimize the amount of time that we need to spend collecting this data. The sooner we have all of the data on the worksheets the sooner we no longer have to Spend two evenings a week downtown. Therefore several of the coding Operations are designed so that you can complete them at hone. Occupa- tion codes and calculation of days from intake to tenninaticn/follow—up/ job starting date can all be determined at hone. This means that the coding process translates into four discrete steps: A) Transfer data from file to worksheet B) Compute special codes at hcme C) Transfer data from worksheet to OPSCAN coding sheets D) Verify. 5) RECOMVIEINDED PRCXZEDURES: a) Before beginning each evening fill out a pile of worksheetswith your name, number and the date. You might also want to check If the files are Active or Closed. b) Select a pile of files frcm your designated drawer. Check that each file has not been coded before. 139 140 c) Record CLIENT ID on both pages of the worksheet! d) Find and record the information on the WHITE SLIP. If no job preference is indicated then check the EMPLOYABILIT‘Y DEVELOPMENT FORM to see if a preference is noted there. If so record this on the work- sheet. Check ccdesheet for appropriate codes. There will be a slightly different code for job preferences taken from the Employability form. e) Find and record the information on the APPLICATION/ELIGIBILITY FORM. This is the Social Security information located at item 25C. Record this information in columns 90 and 91. f) Find the INTAKE FORM and record all the necessary data. REMEMBER that ENROLLMENT STATUS (COL 102) is actually CETA STATUS. g) CETA STATUS: Check white slip and VERIFY with Dept. of Labor letter. h) locate the mum-UP FORIIB. Count the number of times a follow-up has been conducted. Find the most recent follow-up and check to see if a job has been found. If yes then record all the required informa- tion. If no then skip to the items dealing with Interest In Employment and Evaluation of the Program(Ql8) . REMEMBER TO ALSO CHECK JOB STATUS IN THE BLACK BOOK. i) You should now have all of the information obtainable from the client's file. Close the file and put your coding number on the front of the file folder. Go on to another file until you've completed at least 5 files. After you have gotten to this point go to the BLACK BOOK for the rest of the data for those files which you have now partially completed. j) BLACK BOOK: There are three main operations which take place with the black book: 1) Job Club Info 2) Termination Status and (3) Verification of Data.. 1) Job Club Info— The job club info is shown in the black book in either a two or a one column field. Regardless of the type of field the same codes were used to indicate whether a client has participated in the job club. If there are no marks then the client has not been involved in either the orientation or the job club. If only a single check mark is present then the client has attended an orientation, but has not attended the job club. If a check mark with a line through it is found then the client has attended both an orientation and a job club meeting. 2) Terminaticns-CIECK the FAR RIGHT COIIJMN to see if the client has been terminated. If yes then enter the last follow-up date recorded in the black book as the Termination date on the Worksheet. If the Client's Status is Active code zeroes. 141 3) Verifications- Cross validate all the required items listed above. Take special note of JOB STATUS. IF A JOB IS NOTED FRGVI A FOLLOW-UP EARLIERTHANTHENDSTRECENTONEFRGVITHEFIIETHENCHECXWITH JOE! k) RETURN THE FILE 'IOTHE FILE CABINETANDBEGIN AGAIN. 1) AT HOME-- 1) Find codes for all occupations listed on worksheet. 2) Compute Intake — Starting (WWW? if no job), Intake - Follow-up, Intake - Termination. 3) CODE DATA ONTO OPSCAN---CHECK YOUR OODEBCDK CUDING INSTRUCTIONS Welcome to the wonderful world of coding data. There are several coding conventions that you need to be familiar with before we can begin. 1) RIGHT JUSTIFY** This means if you have a four column field 1234 but only three numbers to enter into the field you enter the numbers so that they fill all the space in the right area of the field. One method which we will use that simplifies right justifying is to always fill all the available columns. 80 in the case above we would put a [3 before our three digit number. By doing this we insure that all four columns are filled and that the data is right justified. Examples: Suppose we want to code today's date onto a six cclurm field. The codebook has allowed six columns, two for the month, two for the day and two for the year. Today's date is April 15, 1982. Therefore we would enter the following numbers on the coding sheet: {341582. Note that I placed a zero in front of the 4 in order to insure that I right justified the date. fi4=mcnth, 15=day, 82=year. APPENDIX F INTERVIEW INSTRUCTIONS 10. APPENDIX Interview Procedures If phone is answered but client isn't home, try to find out the best time to call. If R balks at the intro statement, reiterate that we are doing this for the Senior Employment Service, that we are contacting everyone who was a client between October 1980 and December 1981, and that all information will be kept confidential. The ability to predict lob search success simply means improving our ability to help older jobseekers find jobs. Probe, if necessary, to be able to code job title. When asking for previous jobs , always: a) Use the dates as reference points; b) Work backwards through time; c) Remember the dates when you flip the page; d) Get a 20-year work history. If R reports last job as being 6 years ago, Probe to see if R has worked during the last 6 years. Housewife for the last 10 years is a job. Skip dates, wage and satisfaction questions. Ask if R worked outside the here during or before this period. If you make a mistake - correct it. If you find out that the job you've just asked all the detailed questions on is not the most recent job, then go back and ask the detailed questions on the correct job. The R will not know that you've made a mistake. They'll think that what ycfiTre doing is rcutIne. For the 5-point scales, if an item is not applicable, code 0. For instance, if the person was self—employed then supervision and promotion will be not applicable. If the person worked alone, e.g. , a roof contractor, then co—workers would be not applicable. Reasons for leaving: If health was the reason a person quit, code "health." If a person quit because they didn't like the job, code "quit." The number of months unemployed refers to the time period between leaving their last job and their enrollment at SES. However, if the person retired from their last job, probe to find out whether they were actually looking for work during that perch . If a person retired in October 1980, came to SES in October 1981 but only began looking for a job in July 1981, then they were only unemployed for July, August, September, and October - 4 months . 142 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 143 Attitudes toward work . a) Read entire scale at least cnec; b) Remind R to answer as if this was intake date; c) If the person seems to be answering as they feel now, keep reminding them; d) If the R always uses one type of answer, then probe with choices. Health Staterents: Check with the R if his/her responses seem inconsistent . They may be misunderstanding the statements . The number of times during lst month that a letter was sent out = total number of letters sent out. Job Club - if yes, probe whether they attended one meeting other than the orientation meetings. If R says they wanted to use club but couldn't because they were terminated, then record that as answer for A but don't ask A. Employment Status: a) First ask if R is employed or unemployed, then probe to see if they're still looking or not looking; b) If retired, rpcbe to see if working or looking for a job. Job Preferences: We want to know if the types of jobs they were looking for when they first came to SES have subsequently changed. Therefore, if they say that they have, ask only for the new preferences. If person is unemployed/not looking or retired, ask both why they stopped locking and why they're still unemployed. If in response to the 2nd question they say, "I just told you," then paraphrase what they just told you and ask "whether there are any other reasons . " Plans to Retire: If R says, "I have no immediate plans," record No. If the person is employed/not locking don't read "If you have any questions regarding your status at SES . . . " Always try to get as corplete an answer as possible. Always try to get an answer to every question. If R is in a hurry try to complete interveiw quickly — But don't skimp on comprehensiveness . If R cannot complete interview, try to re—schedule a specific time to complete the interview. IXDN'T BE INTIMIDATED -- GET‘ A COMPLETE INTERVIEW” 144 CODING PRQZEDURES Always code immediately after the interview is finished. Remember to enter all relevant information in the log immediately after the interview is finished. Where appropriate, right-justify, i.e., put zeros to the left of numbers to fill up the coding field. Missing Data: a) If you either forget to ask a question, or if the R will not give a response code: 1) 911 for Occupations 2) 999999 for Dates 3) 9 for S-point scales 4) 99.99 for wage/hour 5) 99 for 2 column/fixed choice questions (e.g., Reason you left jOb) 6) 9 for Grade in School 7) 999 for Number of Months unemployed 8) 9 for Yes/No 9) 99 for Retirement Age b) If a question does not apply to R, then code blanks. (e.g, if R = unemployed, leave columns corresponding to questions for Employed blank.) Annual Salary = Monthly Salary= weekly Salary Every 2 weeks Daily Salary a) a) Divide by 52 - weekly wage (w) Divide W'by 40 = hourly wage Divide by 4 = weekly wage (w) Divide w by 40 = hourly wage Divide by 40 hourly wage Divide by 80 hourly wage Divide by 8 = hourly wage * If Live-In, b) Divide by 12 = hourly wage REMEMBER -- AINAYS CODE ID NUMBERS AND CARD NUMBERS. Turn in completely coded interviews to Joe's office WITHIN 2 DAYS. Joe Bornstein: Home - 675-5615 Office - 353-5015 APPENDIX G CLASSIFIED INDEX OF INDUSTRIES AND OCCUPATION 1980 APPENDIX G Classified Index of Industries and Occupation 1980 003-037 043-199 203-235 243-285 303-389 403-407 413-427 433-469 473-499 503-549 553-599 613-617 633-693 694-699 703-799 803-859 863-889 999 Managerial and Professional Specialty Occupations/ Executives Professional Specialties Technicians and Related Support Occupations Sales Workers Clerical/Office Workers Private Household Occupations Protective Service Occupations Service Occupations Farming, Forestry, and Fishing Occupations Mechanics and Repairers Construction Trades Extractive Occupations (Oil/Mining) Precision Production Occupations Plant and Systems Operators Machine Operators Transportation Related Occupations and Handlers, Equipment Cleaners, Laborers Helpers, Occupation Not Reported 145 APPENDIX H SOCIOECONOMIC AND PRESTIGE SCORES FOR MAJOR OCCUPATIONAL GROUPS, 1970 CENSUS CLASSIFICATION SOCIOECONOMIC AND PRESTIGE SCORES FOR MAJOR OCCUPATIONAL GROUPS, (Stevens and Featherman, APPENDIX H 1970 CENSUS CLASSIFICATION 1981) Socioeconomic Scores Occupational MSE12 Group Professional 68.63 Managerial 51.07 Sales 42.30 Clerical 31.99 Crafts 25.63 Operatives 18.24 Transport 20.37 Laborers 15.99 Farm owners and 22.29 managers Farm laborers 14.19 Service (exe. private 20.81 household) Service (private 14.04 household) 146 APPENDIX I INTERVIEW SCHEDULE APPENDIX I INTERVIEW SCHEDULE Client Information Sheet Client ID# Client Name Client Phone # Interviewer ID# Marital Status Intake Date *AFTER INTERVIEW IS CODED REMOVE THIS INFORMATION SHEET FROM THE QUESTIONNAIRE AND FILE IT SEPARATELY. 147 148 SES Telephone Interview Date Client ID (1-3) Intake Date (1-3) (4) BLANK Interviewer ID (5-6) (5-6) "Good Morning/Afternoon/Evening Mr./niss/Mrs. My name is and I'm calling for the Senior Employment Service. ATEe Senior Employment Service would like to improve the services which it provides to the community. In order to do this we are conducting a research survey with Michigan State University for the purpose of improving the ability to predict job search success. We would appreciate it if you would help us by answering a few questions. Please understand that you are under no obligation to provide this informa- tion and that as always anything you say will be kept strictly confidential." (7) BLANK Work History: 1. Let' s begin by talking about some of the jobs you've held. What was the last job you held before coming to the Senior Employment Service in . PROBES (What was that job? What was your position?) a) Occupation Code (8-10)(8-10) When did you finish ' that job? Finishing Date / / (ll-16) (ll-16) mo day yr When did you start that job? Starting Date / / (17-22) (17-22) mo day yr b) What was the hourly pay rate that you were receiving when you left this job? (23-26) (23-26) (If they can't remember wage/hour ask them to estimate.) *(use two decimal places) . 149 Client ID Intake Date c) Now I'd like to ask some questions about your satisfaction with different aspects of this job. Please indicate whether you were: 5 4 3 2 1 Very Satisfied Neither Dissatisfied Very Satisfied Dissatisfied How would you rate your satisfaction with VS S N D VD l) The type of work you did on the job. . . . . . . . . . . 5 4 3 2 l (27) 2) The supervision you received from your employers. . . . . . . . . . . . . . . .5 4 3 2 1 (28) 3) The pay you received. . . . . . . . . . 5 4 3 2 1 (29) 4) The opportunity for Promotion .. . . .. . . . . . . . . . . 5 4 3 2 1 (30) 5) The peOple you worked with (co-workers). , , , , , , , , , . ,5 4 3 2 1 (31) d) What was the reason that you left this job? 1. laid off (32-33) (32-33) 2. quit 3. fired 4. retired 5. temporary job finished 6. moved 7. Other 8. Healtfi_—‘— (REMEMBER--ONLY 20 YEARS) 2. O.K. What job did you have just before you worked at (Name of last job). a) Occupation Code (34-36)(34-36) Finishing Date / / (37-42) (37-42) mo day yr Starting Date _[ / (43-48) (43-48) mo day yr *(Continue to ask these questions until you have completed a work history for the last 20 years. Use finishing date as a reference point. If reSpondent cannot give a specific date try to get approximate year in which they finished.) USE SPECIFIC FINISHING DATES 150 Client ID Intake Date 3. 4. Occupation Code (49-51) (49-51) Finishing Date / / (52-57) (52-57) mo day yr Starting Date / / (58-63) (58-63) Occupation Code (64—66) (64-66) Finishing Date / / (67—72) (67-72) mo day yr Starting Date / / (73-78) (73-78) mo day yr (79) Blank (80) 1 (81-33)ID (84) Blank Occupation Code (85-87) (BS-87) Finishing Date / / (88-93) (BB-93) mo day yr Starting Date / / (94-99) (94-99) mo day yr Total number of jobs held in past 20 years (100-101) Thinking back what would you describe as being your major job or occupation? (Over entire worklife) (Probe: Cite previously mentioned occupations only if they have difficulty answering) Occupation Code (102-104) (102-104) (If R has no major occupation, code 899 .) What was the last grade that you finished in school? (105) (105) . Less than 7 years (0 to 6) . Junior High (7 to 9) . Partial High (10 to 11) . High School Graduate (12) . Partial College (1 to 3) . College Grad (standard 4yr_) . Graduate Professional HNw-bmmfl 151 Client ID Intake Date O.K. Now I'd like to ask you some questions that concern the time period when you first came to the Senior Employment Service back in (month-year) . 10. What was the primary reason that you were looking for work? (106) (106) (DO NOT SPECIFY ITEMS) 1. Financial 2. Keep Busy 3. Feel Useful 4. Enjoy Working 5. Other 11. How long had you been unemployed before coming to SE8? (107-109) (107-109) (If retired PROBE when they began looking) Total number of months (code 0 for less than one month-- Right Justify) 12. I'm going to read you some statements that concern how you felt about work back in (intake date) . I'd like you to tell me how much you Agree or Disagree with each of the statements. Please answer with one of the following choices: +++++S Strongly Agree (If you probe--give R more than 4 Agree one alternative) 3 Neither *(Remember--Answers are to be about 2 Disagree how they felt--Not how they feel) 1 Strongly Disagree *(Let R choose a pole and then determine the strength of their feeling) SA A N D SD 1. If I had enough money to live comfortably I would not work..5 4 3 2 l (110) 2. It really bothers me to be unemployed....................5 4 3 2 l (111) 3. I need a job to feel useful...5 4 3 2 1 (112) 4. I don't care if I find a job this year.................5 4 3 2 1 (113) 5. I would take any job..........5 4 3 2 l (114) 6. I would only take an interesting job...............5 4 3 2 1 (115) 7. Money is the most important reason for working............5 4 3 2 1 (116) 8. I am a good worker............5 4 3 2 1 (117) 9. I have never received satisfaction from a job.......5 4 3 2 1 (118) 10. When I don't have a job I ~ get bored.....................5 4 3 2 1 (119) 11. Finding a job is very important to me...............5 4 3 2 1 (120) _ 152 Client ID Intake Date 13. Now I have a few questions about how your health effected your ability to work. Please answer as if this was when you first came to the Senior Employ- ment service back in . We will use the same choices. (READ CHOICES). Strongly Agree Agree Neither Disagree Strongly disagree WNW-hm SA A N D SD . My health is excellent. 5 4 3 2 l (121) .... 2. I cannot use public trans- portation because of my health. 5 4 3 2 l (122) 3. I have difficulty climbing stairs. 5 4 3 2 1 (123) 4. I find it hard to sit for long periods of time. S 4 3 2 l (124) 5. I often go for long walks outdoors. 5 4 3 2 1 (125) 6. My health is poor. 5 4 3 2 l (126) 7. I do my own shopping. 5 4 3 2 l (127) 8. Because of my health I could not work a full-time job. 5 4 3 2 l (128) 9. My health interferes with my ability to look for a job. 5 4 3 2 l (129) 10. I am strong enough to work at a part-time job. 5 4 3 2 1 (130) 14. O.K. now I'd like to ask some questions regarding the time you spent looking for a job during the first month you were in the program. During the first month what was the approximate number of times that you: 1. telephoned prospective employers (131-132) 2. sent a letter out to an employer (133-134) 3. went to a job interview (135-136) 4. visited the SES office (137-138) 5. visited the MESC office (139-140) W (Michigan Employment Security Commission) 153 Client ID Intake Date 15. Did you participate in the Job Club? PROBE (Attended at least one meeting) (Orientation meetings don't count) (141)l. IF YES....ASK B (141) 2. IF NO....Ask A)Why did you decide not to use the Job Club? (Probe: Did you dislike particular aspects of Job Club; personal situation?) + B) How satisfied are you with the way in which the Job Club helped you? (142) 5 4 3 2 1 Very Very Satisfied Satisfied Neicher Dissatisfied Dissatisfied 154 I Client ID Intake Date FOLLOW-UP QUESTIONS (If you know their status verify by asking them directly.) 16. Which of the following best describes your present (143) employment status? (Read List) 1) Employed - still looking for other work (Q 18) 2) Employed - not looking (0 18) 3) Unemployed - still actively looking (Q 35) 4) Unemployed — want a job, but not looking (Q 36) 5) Retired (PROBE- if looking for work fit into appropriate category above, if not looking for work go to Q 36). ASK Q 17 17. How satisfied are you with your present employment status? (144) S , 4 3 2 1 very very satisfied satisfied neither dissatisfied dissatisfied IF EMPLOYED 18. What is your position: Occupation Code (145-147) (145-147) 19. What is your hourly wage? (148-151) (148-151) ‘wage/hour 20. Is the job l.Permanent(152) (152) 2.Temporary(Probe: if there is actual ending date or if job is permanent, but R only con- siders it as temporary) 21. Is it 1.Full time(153) (153) 2.Part Time 3.0ccasiona1 22. When did you start this job? Starting Date / / (154-159) mo day yr (160) 2 (161-163)ID (164)Blank I /¢L 155 Client ID Intake Date 23. 24. 25. I'd like to ask you some questions about your satisfaction with this job. Like before please respond to the statements with: S 4 3 2 1 very very satisfied satisfied neither dissatisfied dissatisfied How would you rate your satisfaction with: VS 8 N D VD l) The type of work you do in this ‘ (165) job 5 4 3 2 l 2) The supervision you receive from your employers 5 4 3 2 l (166) 3) The pay you receive 5 4 3 2 l (167) 4) The Opportunity for promotion 5 4 3 2 l (168) 5) The people you work with (co-workers) 5 4 3 2 l (169) How likely do you think it is that you will be at this job one year from now? (170) 5 4 3 2 1 very very likely likely neither unlikely unlikely If you leave your present job what would be the main reason? (171-172) . (171-172) 1) Found a better job 2) Laid off 3) Quit 4) Fired 5) Retired 6) Hoving 7) Health 8) Other 156 Client ID Intake Date 26. I'd like you to rate how similar the job you have now is to the types of jobs you were looking for when you first started your job search.(173) (173) 5 4 3 2 1 very very similar similar neither different different 27. Did you have any other jobs between the time you came to SES and the start of the job you have now? 1. Yes 2. No (174) (174) If yes ask Q 28 to 34 If No ask next appropriate question from Q 35 on. 28. What was your position: Occupation Code (175-177) (175-177) 29. What was your hourly wage wage/hour (178-181) 30. Was the job l.Permanent (182) (182) 2.Temporary (Note criteria for temporary job) 31. Was it l.Pull Time (183) (183) 2.Part Time 3.0ccasional 32. a) When did you finish: Finishing Date / / (184-189) b) When did you start mo day yr the job Starting Date / / (190-195) mo cay yr 157 Client ID Intake 33. 34. I'd like to know how satisfied you were with this job. 5 4 3 2 very satisfied satisfied neither How would you rate your satisfaction with: 1) 2) 3) 4) 5) VS The type of work you do in this job 5 The supervision you receive from your employers The pay you receive The Opportunity for promotion 5 The people you work with (co-workers) 5 What was the reason you left this job 1) Found a better job 2) laid off 3) Quit 4) Fired 5) Retired 6) Moved 7) Temporary job finished 8) Other 9) Health S N 1 very dissatisfied dissatisfied 2 1 (196) 2 1 (197) 2 1 (198) 2 1 (199) 2 1 (200) (201-202) 158 Client ID Intake Date + 35. If Employed Looking/Unemployed Looking] Are the jobs that you are looking for now the same or different from the jobs you were looking for when you first came to the Senior Employment Service. 1. Same 2. Different If different + 'What are your new job preferences? 1. 8. 3. Why did you change your preferences? 159 Client ID Intake Date + 36. If Unemployed/Not Looking or Retired What made you decide to stop looking for a job? 37. If unemployed What do you think are some of the reasons that you are still unemployed? PROBE (Please be as specific as possible.) (ASK ALL RESPONDENTS THE FOLLOWING QUESTION UNLESS RETImD) 38. Do you have any plans to retire 1. Yes (203) (203) 2. No How old will you be when you retire (204-205) Thank you very much for your time and cooperation (The information that you've provided will help us to determine how we can improve our services) Are there any questions you'd like to ask me before we say good-bye? If you have any questions regarding your status at SES please contact the office during business hours at 485-7900. Thanks again--Have a nice day. (206-211)Intake Date (212-229)Blank (230) 3 APPENDIX J ADMINISTRATIVE AGREEMENT APPENDIX J ADMINISTRATIVE AGREEMENT MICHIGAN STATE UNIVERSITY mwununst 00 nycmnocy EAST lANSlM. ~ MICHIGAN - “‘24 PSYCHUICXL" RESEARCH BL'II DI\6 ADMINISTRATIVE AGREEMENT BETWEEN TEE TRI/COUNTY SENIOR EMPLOYMENT SERVICE AND JOSEPH BORNSTEIN, OF MICHIGAN STATE UNIVERSITY This agreement is to specify the intention and.future relationship of Joseph Bornstein of the Michigan State University and the Tri County Senior Employment Service. The purpose of this relationship is to conduct research documenting: l) the types of older jobseekers serviced by the Tri County Senior Employment Service and 2) factors which predict client success at job seeking. In line with this intention the following agreement is specified: Joseph Bernstein will: 1. Provide any resources needed (i.e., personnel, materials, computer time) to carry out the research. 2. Take responsibility for developing and managing the research protocols. including supervising and training undergraduate students from Michigan State University to assist in collecting the data. 3. Provide the Tri County Senior Employment Service with monthly oral progress reports. 4. Provide the Tri County Senior Employment Service with a copy of the completed thesis. 5. Insure the confidential and anonymous handling of all information collected on individual clients. The Tri County Senior Employment Service will: 1. Provide access to client files. 2. Provide access to their offices after normal business hours in order to facilitate data collection. 3. Provide explanation and clarification of record-keeping systems. 4. Allow a sample of their clients to be contacted for a brief telephone interview. i335ih Bernstein I ,l’ / ‘ ‘ . / , fizjfiill €;>it?aéét¢itf\~ *//'/’<;"8 9 Sue Hadden H" u a. "NH.H“ a S. (on! 'Vhdll’rru'lgnol\ Imnnlglnun 160 APPENDIX K INTERCORRELATIONS OF JOB SATISFACTION SCALE ITEMS (BEFORE UNEMPLOYMENT) APPENDIX K INTERCORRELATIONS OF JOB SATISFACTION SCALE ITEMS (BEFORE UNEMPLOYMENT) l 2 3 4 5 l 1.0 2 .47 1.0 3 .53 .42 1.0 4 .43 .61 .53 1.0 5 .51 .45 .49 .53 1.0 l. The type of work you did on the job. 2. The supervision received from your employers. 3. The pay you received. 4. The opportunity for promotion. 5. The people you worked with. 161 APPENDIX L INTERCORRELATIONS OF JOB SATISFACTION SCALE ITEMS (REEMPLOYED) APPENDIX L INTERCORRELATIONS OF JOB SATISFACTION SCALE ITEMS (REEMPLOYED) 1 2 3 4 5 l 1.0 2 .34 1.0 3 .58 .32 1.0 4 .54 .34 .50 1.0 5 .26 .45 .18 .44 1.0 1. The type of work you did on the job. 2. The supervision received from your employers. 3. The pay you received. 4. The opportunity for promotion. 5. The people you worked with. 162 APPENDIX M INTERCORRELATIONS OF EXTRINSIC 'WORK VALUE' SCALE ITEMS APPENDIX M INTERCORRELATIONS OF EXTRINSIC 'WORK VALUE' SCALE ITEMS If I had enough money to live comfortably I would not work. Money is the most important reason for working. 163 APPENDIX N INTERCORRELATIONS OF INTRINSIC 'WORK VALUE' SCALE ITEMS APPENDIX N INTERCORRELATIONS OF INTRINSIC 'WORK VALUE' SCALE ITEMS 1 2 3 4 1 1.0 2 .62 1.0 3 43 .38 1 o 4 .37 .36 49 1 o It really bothers me to be unemployed. Finding a job is very important to me. I need a job to feel useful. When I don't have a job I get bored. 164 APPENDIX 0 INTERCORRELATIONS OF HEALTH SCALE ITEMS .nofl oEHUIuHmm m um xHOB ou :mpocm vacuum Em H .OH .QOn m now xooH ow >uHHHQm >8 nuH3 mmuomuwucH nuHmmn >2 .m .nom oEHuIHHsm m xuo3 uoc pHpoo H nuHmms >E mo mmsmowm .m .mCHmmonm :30 >6 Op H .h .Hoom mH nuHmmn >2 .0 .mHOOUupo mxHMB mcoH How om cwumo H .m .oEHu mo mpoHumm mCOH How uwm on was: uH chH H .v .muflmum OCHQEHHU >uHsonme w>mn H .m .nUHmos >E mo mmsmoon coHumunommcmuu UHHQDQ mus uoccmo H .m .ucmHHmoxm mH nuHmwn >2 .H oo.H mm. Hm. Vm. ov. mm. mm. .ov. Hv. hm. 0H oo.H v5. mv. mm. mm. 0v. mv. mv. mv. m oo.H mm. mm. 0m. 0m. om. nv. mm. m oo.H mm. Hm. Om. mm. mm. Hm. n oo.H mm. vv. mm. nv. mm. o oo.H 5H. Hm. vH. mm. m oo.H vv. vm. vv. v oo.H mm. mm. m oo.H mm. m oo.H H 0H m m n o m v m m H mZmBH quum mhqdmm m0 mZOHfidqmmmoumMBZH O XHDmemd 165 APPENDIX P PERCENT AGREEMENT - ARCHIVAL ITEMS APPENDIX P PERCENT AGREEMENT - ARCHIVAL ITEMS (5 raters) Percent Items* Agreement Work Mode 90 Health 90 Transportation 90 Job Club Orientation 60 Job Club Member 90 Age 90 Sex 100 Marital 100 Ethnic 100 Education 100 Income 1 50 Income 2 70 Social Security 100 CETA Status 80 *Items correspond to items found on Intake Forms (see Appendices B and C). 166 APPENDIX R PERCENT AGREEMENT - INTERVIEW ITEMS #2 APPENDIX R PERCENT AGREEMENT - INTERVIEW ITEMS #2 (10 Raters) Percent Percent Percent Item* Agreement Item Agreement Item Agreement 8 100 119 100 153 90 11 100 120 100 154 80 17 100 121 100 165 100 23 100 122 100 166 100 27 100 123 100 167 100 28 100 124 100 168 100 29 100 125 100 169 100 30 100 126 100 170 100 31 100 127 100 172 100 32 100 128 100 173 100 100 100 129 100 174 100 102 100 130 100 175 100 105 100 131 100 178 100 106 100 133 100 182 100 107 100 135 100 183 100 110 90 137 100 184 90 111 90 139 100 190 90 112 90 141 100 196 100 113 90 . 142 100 197 100 114 90 143 100 198 100 115 90 144 100 199 100 116 90 145 100 200 90 117 90 148 100 201 100 118 90 152 100 203 100 206 80 *Item numbers correspond to column numbers on interview questionnaire (see Appendix I). 168 APPENDIX S PERCENT AGREEMENT - OCCUPATION CLASSIFICATIONS PERCENT AGREEMNT - OCCUPATION CLASSIFICATIONS APPENDIX S (10 Raters) Classification Percent Occupation Code Agreement Grant Administrator 005 100 Clerical 389 100 Practical Nurse 207 70 Security Director 415 80 Traffic Shipping Clerk 364 100 Forestry Supervisor 494 100 Housekeeper 405 100 Employment Specialist 027 90 Assembly Line Worker 785 100 Garbage Collector 875 90 169 REFERENCES REFERENCES Azrin, N.H., Flores, T and Kaplow, S.J. "Job Club: a group assisted program for obtaining employment," Behavior Research and Therapy, 1975, 13, 17-27. Becker, Gary 8. Investment in human capital: A theoretical analysis. J.P.E., Oct, 1962, LXX, 9-49. Bould, Sally, Unemployment as a factor in early retirement decisions. American Journal of Economics and Sociology, (Apr) 1980, V.39(2),—I23-136. Campbell,J.P. and Pritchard,R.D. Motivation theory in industrial and organizational psychology. In M. Dunnette (Ed.) Handbook of industrial and organizational psych- ology. Chicago: Rand McNally, 1976. Chan, Teresita and Donald G. Fowles. The older worker. U.S. Department of Health and Human Services, Office of Human Development Services. Oct, 1980, #81-20265. Ciscel, David H. & Tuckman, Barbara H. The peripheral work- er: CETA training as imperfect job socialization. Journal of Economic Issues, 1981, June, Vol £212), 489-500. Colledge, Maureen & Richard Bartholomew. The long termed unemployed: Some new evidence. Employment Gazette, Jan 1980, 9-12. Collins, Lora 8. Older Workers: Economic Road Maps. New York: The Conference Board, July 1978. Dunn, [Ennis J. Vocational rehabilitation of the older disabled worker. 'Journal of Rehabilitation, Oct/Nov/Dec, 1981, V. 4114), 76-83. Davis, Larry D. Life Cycle of Human Resources. 1973, Human Resources Institute: University of Utah. Fairweather, G.W. and Tornatzky, L. Experimental Methods for Social Policy ResearchJ New York: Pergamon Press, 1977. Gillaspy, Thomas R. Labor force participation of the older population: Toward 1990 -- A demographic approach. In Ragan, Pauline K. Work and Retirement: Policy Issues. California: University of Southern California Press, 1980. 170 171 Gordon, Margaret S. The older worker and hiring practices. Monthly Labor Review, Nov, 1959, V. 82(11), 1198-1205. Graney, Marshall J. and Doris M. Cottam. labor force non participation of older people: United States, 1980-1978. The Gerontologist, 1981, V. 21(2), 138-145. Gray, Denis O. A Job Club for Older workers and Retirees: An Egperimental Evaluation of Outcome and Process. Doctoral Dissertation, Michigan State UniversIty, 1980. Gunnar, Myrdal, The Political Element in the Development of Economic Theogy. Cambridge, Massachusetts: Harvard University Press, 1954. Hadden, Sue. Annual report of the Senior Employment Service. Tri-County Office On Aging: Lansing, MI, 1981. Hall, Arden & Johnson, Terry R. The determinants of planned retirement age. Industrial and Labor Relations Review, Jan, 1980, V33(2), 241—254. Harris, Lou and Associates Inc. Aging in the eighties: America in transition. A survey conducted for the National Council on Aging Inc., Nov., 1981. Hill, J.M.M. The Social and Psychological Impact of Unemployment. A Pilot Study, Tavistock Institute of Human Relations, London, 1977. Doc#2T 74 Kane, Rosalie, A., and Kane, Robert 1. Assessing the elderly: A practicalyguide to measurement. Lexington, Mass: Lexington Books, 1981. Kiker, B.F. The historical roots of the concept of human capital. Journal of Political Economy, Oct, 1966, Vs Yip 481-99. Kingson, Eric R. The early retirement myth: Whygmen retire before age 62. Report of the Select Committee on Aging. U.S. House of Representatives, Oct 1981 (No. 97-298). Lippman, Steven A. & John J. McCall. The economics of job search: A survey Part I. Economic Inquiry, 1976, June Maslow, Abraham Harold. Motivation and Personality, New York: Harper, 1954. 172 McConnell, Stephen. Alternative work patterns for an aging labor force. In Ragan, Pauline I. Work and Retirement: Policy Issues. California: University of Southern California Press, 1980. Morrison, Malcolm H. International developments in retirement flexibility. Aging and Work, Fall 1979, 221-234. Morse, Dean. The Utilization of Older Workers. Washington: National CommiEsion’fSr ManpowerIPolicy, Special Policy Report No. 33, March 1979. Nelson, Thomas C. The age structure of occupations. In Ragan, Pauline K. Work and Retirement: Policy Issues. California: University of Southern California Press, 1980. Nie,N.H., Hull,C.H., Jenkins, J.G., Steinbrenner, K., and Bent, D.H. Statistical package for the social sciences. New York: McGraw-Hill, 1975 Nunnally, Jim C. Psychometric Theory: 2nd Edition, New York, N.Y.: McGraw Hill, 1978. OECD, Promoting the Placement of Older Workers, 1966. OECD, Socio-Economic Policies for the Elderly, Paris, 1979. Pampel, Fred C. Changes in the labor force participation and income of the aged in the United States, 1947-1976. Social Problems, Dec, 1979, V. 21(2), 125-142. Porter, L.W., and Lawler, E.E. Managerial attitudes and performance. Homewood, Ill: Irwin-Dorsey, 1968. Reid, Graham L. Job search and the effectiveness of job-finding methods. Industrial and Labor Relations Review, 1972, V.2§_(4), 479-495. Rones, Philip L. The retirement decision: A question of opportunity? Monthly Labor Review, Nov, 1980, 14-17. Rosenfield, Carl & Scott Campbell Brown. The labor force status of older workers. Monthly Labor Review, Nov, 1979, 12-18. Schmitt, N., and McCune, J.T. The relationship between job attitudes and the decision to retire. Academy of Management Journal, 1981,_24(4), 795-802. 173 Schultz, Theodore W. Economic Research: Retrospect and Prospect -- Human Resources, New York: CoIfimbia University Press, 1972. Sheppard, Harold L. and Sara E. Rix. The Graying of Working America. New York: The Free Press, 1977. Smith, P.C., Kendall, L.M. and Hulin, C.L. The measurement of satisfaction in work and retirement; a strategy for the study of attitudes. Chicago: Rand McNally, 1969. Sobel, Irvin. Older worker utilization patterns: Human capital approach. Industrial Gerontology, 1972, (13) 6-280 Stagner, Ross. Propensity to work: An important variable in retiree behavior. Aging and Work (Summer) 1979, Stevens, G. and Featherman, D.L. A revised socioeconomic index of occupational status. Social Science Research, 1981, m, 364-395. Terborg, James R., George 5. Howard & Scott E. Maxwell. Evaluating planned change: A method for assessing alpha, beta and gamma change. Academy of Management, 1980, 109-120. U.S. Bureau of Census. General population characteristics, 1980. U.S. Department of Labor. Employment and training administration. Employment-Related Problems of Older Workers: A Research Strategy, 1979. Vandergoot, David, Richard Jacobsen and John WOrrall. New direction for placement practice in vocational rehabilitation In Placement for Rehabilitation: A Career Development Perspective. Editedgby David Vandergoot and John D. Worrall, 1979, Baltimore, University Park Press, 236 pages. Vroom, V.H. WOrk and motivation. New York: Wiley, 1964. Walker, J.W., and Price, K.F. Retirement policy formulation: A systems perspective. Personnel Review, 1976, 5, 39-43. 174 Wolozin, Harold. Early retirement and the older worker. Journal of Economic Issues, June 1981, V. XV (2), Reference Note Personal Communication with Denis Gray at Annual Gerontological Association Conference, Fall 1981. HICHIGQ IIIIIIIIIIIIIIIIT