Volume 28, Number 3
2003

A Study of Supervisor and Employee Perceptions of Work Attitudes in Information Age Manufacturing Industries

Md. Shafiqul Azam
Pro-Tech Search, Inc.
Paul E. Brauchle
Illinois State University

Abstract

The self-perceived work ethic of industrial employees in information jobs (N=304) and non-information jobs (N=277), and employees' work ethic as assessed by their supervisors, were examined using the Occupational Work Ethic Inventory (OWEI). A Principle Components Analysis yielded four factors (Teamwork, Dependability, Ambition and Self-Control) based on factor loadings equal to or greater than 0.45. A MANOVA analysis revealed significant differences between self-perceived work attitudes of information and non-information employees and their supervisors, and indicated that the self-perceived work attitudes of information and non-information employees were different. Hierarchical Multiple Regression Analysis revealed that female information employees possessed better work attitudes than their male counterparts. No relationships were found to exist between the demographic variables of gender, age, education and length of service for non-information employees.

Introduction

While the life cycle of most animals tends to focus on the three dimensions of foraging food, sleeping, and reproducing, typical activities in humans contain one more dimension: work. Work is composed of tasks performed to add values to the products in exchange for monetary remuneration and psychological compensation (Wang, 1996).

Throughout history, work has been viewed and interpreted in various ways, sometimes with a very negative attitude - especially toward physical labor (Anthony, 1977; Hill, 1996a, Maywood, 1982; Tilgher, 1930). The Protestant Work Ethic evolved in the sixteenth century, based on the teachings of Luther and Calvin. This view, based on attitudes and beliefs that supported hard work, gradually became secularized and woven into the norms of western culture (Lipset, 1990; Rodgers, 1978; Rose, 1985).

During the industrial revolution in the eighteenth century, machines gradually replaced manual labor and individual control over the quantity and methods of personal production was diminished (Gilbert, 1977). In the early twentieth century, various management models were developed to motivate workers in industries with the aim of increasing productivity and generating more profits for the owners. Most of these models were early authoritarian styles of management such as Taylor's scientific management and McGregor's Theory X. Later, other more humane approaches, such as the behaviorist approach and McGregor's Theory Y, were developed (Gaither, 1996).

Regardless of their style of management, employers have always placed great importance on employee attitudes (Natriello, 1989). The social competencies (sociability, responsibility, self-management, integrity, and self-esteem) are the ones most in need of improvement (Smith, Jones, & Lane, 1997). For example, Beach (1982) cited research indicating that fully 87% of persons losing their jobs or failing to be promoted were found to have "… improper work habits and attitudes rather than insufficient job skills or knowledge" (p. 69).

The challenging nature of work itself also affects the relationships between workers and their jobs. According to Toffler (1981), a transition from the Industrial Age to the Information Age occurred in the mid 1950s when white-collar and service workers for the first time outnumbered blue collar workers. Today, the information and services sectors jointly control more than 80% of the U.S. economy (Huitt, 1997). Information age jobs, in contrast to industrial age jobs, are high-discretion positions which require considerable thinking and decision making on the part of workers (Miller, 1986). People who work with information often must set their own schedules, work with little or no supervision, and deal with frequent changes in work procedures. With jobs such as these requiring more self-direction and selfmotivation, good work attitudes and a positive work ethic seem more important than ever before.

Good work attitudes are often mentioned as attributes that employers want their employees to have, but these attributes are often hard to find. Even though various programs such as technical preparation, apprenticeships, Partners in Education projects, curriculum review, mentoring programs, and employability certification have attempted to address the problem, employers still complain that they are unable to find a dependable work force (Hill & Petty, 1995). Studies conducted by Custer and Claiborne (1992 & 1995) found that both vocational educators and employers gave more emphasis to employability skills than technical skills and basic skills. Their employability skills cluster includes all the ingredients of a good work ethic (i.e., good work habits, attitudes, and interpersonal skills).

The concept of work ethic relates to the desirable work attitudes expected from potential employees. Employability skills or positive affective work attitudes are not job-specific, but are skills which cut horizontally across all industrial and vertically across all jobs from entry level to chief executive officer (Sherer & Eadie, 1987).

Researchers have long sought to identify affective characteristics considered desirable for working people. Early attempts include those of Beech, Kazanas, Sapko, Sisson, and List (1978). They identified 63 affective work competencies that were considered important by industry leaders and educators and clustered them into 15 categories. Other studies that have attempted to identify personal-social work competencies include those of Petty, Kazanas, and Eastman (1981) and Brauchle, Petty, & Morgan (1983). The basis of their research was the Affective Work Competencies Inventory (AWCI) developed by Kazanas (1978). From the original AWCI, Brauchle, Petty, & Morgan (1983) developed the Work Attitudes Inventory (WAI). Building on this line of research, Petty (as cited in Hill & Petty, 1995), identified 50 work ethic descriptors which became the Occupational Work Ethic Inventory (OWEI).

The work attitudes of an employee can be measured by administering an appropriate instrument to the employee. However, in a work place, the supervisor often is the rater and may use work attitudes as a major component of that rating. Employee self-ratings may be unreliable and invalid for a number of reasons. On the other hand, the supervisor's rating of an employee's work attitudes seems to have value because employee attitudes that affect working environment may be more visible to the supervisor than technical skills. Even with a supervisor making the assessment, a faulty rating may result if the employee is not on good terms with the supervisor.

Brauchle (1979) studied the relationship between trainee and supervisor perceptions of trainee work attitudes. The results of this study suggested that self-perceptions of trainees' work attitudes did not match the perceptions of their supervisors. However, this study, as well as many other investigations about work attitudes and values, investigated these constructs in industrial workers and supervisors and did not differentiate between information employees and noninformation employees.

Several studies have reported differences in work ethic between males and females (Furnham & Muhiudeen, 1984; Hall, 1990, 1991; Hill, 1996b, 1997; Hill & Rojewski, 1999; Petty & Hill, 1994; Wollack, Goodale, Witjing & Smith, 1971). However, Tang (1989) found no relationship between gender and work ethic. Using the OWEI, Minton (1997) found a significant difference between employers' expectations of work ethic and the self-perceived work ethic of secondary school students. In terms of the relationships between education and work attitudes, some researchers (Goodale, 1973; Wollack et al., 1971) found positive relationships, while others (Aldag & Brief, 1975; Buchholz, 1978; MacDonald, 1972) found no relationship or negative relationships (Tang & Tzeng, 1992). Petty (1995) found significant positive differences in OWEI measured work ethic for the age group 36-55 over the other four age groups studied. Hatcher (1995) conducted a study to identify differences in perceived work ethic between apprentices and instructors categorized by job titles, job specialization, years of full-time work experience, and the year of participation in the apprenticeship program. He found significant differences in levels of work ethic between instructors and apprentices. Hill (1996b) found differences in work ethic perception between vocational students and workers in two OWEI dimensions. In a recent study, Boatwright and Slate (2000) obtained stronger work ethic values for females, persons aged 20-24, and college students than for others. However, there appears to be little or no research that focuses specifically on the relationship between work attitudes as perceived by employees and as perceived by their supervisors where the individuals studied are information employees and non-information employees.

Information employees are employees who perform information jobs. An information job is one which is characterized by: a) Comprehensive, open-ended tasks requiring high responsibility and critical thinking; b) Tasks which need little supervision and require active individual initiative; c) Tasks that require creative solutions to non-routine situations, deviations are handled by the lowest level of specialist; d) Continued improvement of performance is as important as completing tasks; and e) Integrated work processes; increased ownership of product and process by the individual (Law, Knuth, & Bergman, 1992).

Non-information employees are those who perform non-information jobs. A non-information job is characterized by: a) Narrowly defined tasks that require minor responsibility; b) Heavy supervision and passive order taking; c) Specific response to a limited number of possible problems with deviations from the norm handled by specialists; d) Task completion is more important than continued improvement of performance; and e) Specific tasks are independent of the purpose in the organization's overall operation (Law et al., 1992).

Research Design and Method

The purpose of this study was, therefore, to investigate (a) whether the type of job (i.e., information job versus non-information job) was related to employee work attitudes, (b) if there existed any difference between work attitudes as perceived by employees and as perceived by their supervisors, and (c) if there existed any relationship between employee work attitudes and demographic variables such as age, gender, level of education, and length of service. The following five null hypotheses were used:

H01: At the p≤ 0.05 level of confidence, there is no significant difference between the self-perceptions of work attitudes of industrial employees with information jobs and their work attitudes as rated by their supervisors.
H02: At the p≤ 0.05 level of confidence, there is no significant difference between the self-perceptions of work attitudes of industrial employees with non-information jobs and their work attitudes as rated by their supervisors.
H03: At the p≤ 0.05 level of confidence, there is no significant difference between the perceptions of work attitudes of industrial employees with non-information jobs and industrial employees with information jobs.
H04: At the p≤ 0.05 level of confidence, there is no significant relationship between the work attitudes of information employees and the variables of gender, age, level of education, and length of service.
H05: At the p≤ 0.05 level of confidence, there is no significant relationship between the work attitudes of non-information employees and the variables of gender, age, level of education, and length of service.

Population and Sample

The population for this study consisted of employees of manufacturing organizations in the central Illinois area. These employees are often at the lowest level in the organization (except for part time hourly employees), or at most one or two levels up in the organizational hierarchy. This study used the procedure of cluster sampling without replacement (Parker, 1998) where each industry was treated as a cluster.

A list of all central Illinois' midsized manufacturing organizations employing 100 to 500 employees was prepared on the basis of the 2000 Illinois Manufacturers Directory (Manufacturers' News, Inc., 1999). Ninety-five midsized manufacturing organizations were listed from the central Illinois area. The total number employed in these organizations was 18,900.

The required sample size was calculated from the total employee population of 18,900 by using the formula by (Borg & Gall, 1979): n = N/(N(d)^2 + 1), where n = sample size; N = Total population (18, 990); and d = level of significance (0.05). Application of the formula resulted in a required sample size of 392 participants. The average response rate for this type of study was assumed to be about 33%. To achieve 392 respondents, a sample size of 1,176 participants (3 X 392) was targeted for this study. The average number of population members per cluster was: 18,990/95 = 199.89. The number of clusters (organizations) needed to represent the population was 1,176/199.89 = 5.88. Therefore, six organizations (clusters) were needed for this study.

Of the 95 manufacturing organizations in the list, six industries were selected randomly. The total number of employees in these six industries was 1,209, a number which exceeded the 1,176 participants needed for the study. We believed that this number was sufficient to represent the population and realized that the results can only be generalized to these respondents.

Power Analysis

Statistical power analysis is quite common in recent research studies. According to Wilkinson and the Taskforce on Statistical Inference (1999), "Because power computations are most meaningful when done before data are collected and examined, it is important to show how effect-size estimates have been derived from previous research and theory in order to dispel suspicions that they might have been taken from data used in the study or even worse, constructed to justify a particular sample size" (p. 596). Therefore, a prior power analysis was conducted in this study.

The sample size calculated earlier was verified by methods specified by Cohen (1988). Alpha was set to 0.05 and power to 0.8. Assuming the proportion of the multivariances of set Y (independent variables) accounted for by set X (dependent variables) to be 0.1 and with the two-group MANOVA design, the sample size was calculated as 302 participants. Therefore, the estimated respondent size of 392 seemed adequate to conduct the study.

In a similar way, the adequacy of sample size was checked for hypotheses 4 & 5. Using the formulas set forth by Cohen (1988) for use in multiple regression analysis, the required sample size was found to be 111 for an alpha of 0.5, a power level set at 0.8, assuming that four variables (age, gender, level of education, and length of employment) accounted for 0 .1 of the criterion variance in the population. The results implied that the estimated respondent size of 392 was more than adequate.

Instrumentation

The Occupational Work Ethic Inventory [OWEI] (Petty, 1995) was used to collect data from both employees and their supervisors. This is a 50-item selfreporting type instrument that uses Likert-type scaling to measure work attitudes. It has been used in industrial and vocational education settings and has been found to be highly reliable with various populations. Reported reliability coefficients (Cronbach's alpha) were: 0.95 (Hill, 1992), 0.95 (Petty, 1995), 0.90 (Hatcher, 1995), and in the present study 0.94 (employee responses) and 0.97 (supervisor responses). No predictive or concurrent validity studies have been reported; however, there appears to be evidence (Hill, 1992, Petty, 1995, Hatcher, 1995, Azam, 2002) for factorial validity. These studies used OWEI responses for educators, students, and workers in various industries. For this study the OWEI was used without modification to collect responses from the employees. However, a slightly modified OWEI was used to collect responses from the supervisors. In the modified OWEI for supervisors, items 3 (Years of full time work experience), 4 (Sex), 5 (Level of Education), 6 (Age), and 7 (Country of citizenship) were excluded because supervisors are not thought to provide accurate information on these items with respect to the employees they supervise.

Description of Research Procedure

A list of names and contact telephone numbers of the CEOs or highest-ranking officials of each of the 95 organizations was prepared based on the 2000 Illinois Manufacturers Directory (Manufacturers' News, Inc., 1999). Materials were sent to each of these individuals with a request to participate in the study. Because the sample was to be selected from this population of 95 organizations, materials included a request letter, information sheet, and a reply slip. Each industry was requested to select a contact person with whom the researcher could communicate to administer the instruments and to collect the completed instruments.

As noted earlier, a slightly modified version of OWEI was used to collect supervisor responses. A set of two questionnaires with the same serial number, (a) green - to be completed by the employees and (b) blue - to be completed by the supervisors, was printed and used for each employee-supervisor pair. Two kinds of instruction sheets, (a) "How to fill out the Survey Forms (FOR EMPLOYEES)" and (b) "How to fill in the Survey Form (FOR SUPERVISORS)," were also printed and given to supervisors. Supervisors were asked to identify employees as information or non-information employees according to the definitions printed on the instruction sheets, to indicate type of employee on both blue and green sheets, and to give the green questionnaires to the employees. On the instruction sheet used for employees, instructions were given on how to respond if they thought they had been wrongfully categorized as information or non-information employees by their supervisors. In only three cases were there differences of opinions between employees and supervisors about the information or non-information status of the employees. In these cases the researcher used the employees' self-designations. Of the 95 industrial organizations, 26 agreed to participate in the study (a response rate of 26.32%).

Data Analysis

A total of 1,575 green-colored (employee) and 1,575 blue-colored (supervisor) instruments were used to collect employee and supervisor responses on work attitudes of 1,209 employees. A total of 492 (40.70% response-rate) completed instruments were returned by the employees and 633 were returned by the supervisors. Of the 492 employee responses, 454 were used in the analysis and of the 633 (52.36% response-rate) supervisor responses, 581 were used in the analysis. The rest were discarded because they contained missing values.

Of the total 581 responses received from supervisors, 304 responses were for information employees and 277 were for non-information employees. These responses were further sorted to match each employee's response with that of his or her supervisor. This procedure yielded 245 matched responses from supervisors and information employees. In a similar manner 169 matched responses from supervisors and non-information employees were obtained. Responses for the 11 negatively worded items - stubborn, tardy, irresponsible, depressed, devious, selfish, negligent, apathetic, rude, hostile, and careless - were reversed for consistency in the analysis.

Factor Analysis

In regression analysis, it is very important to maintain a large number of degrees of freedom, and degrees of freedom are lost as the number of variables increase (Korth, 1975). Because the number of dependent variables was very large (50), the available degrees of freedom were small, making the result of a multiple regression analysis to test hypotheses 4 and 5 difficult to interpret. Moreover, the instrument items (based on employee responses) were found to be highly correlated. In such a situation the interpretation of results is very difficult and likely to be spurious. In view of this, and because the type and the location of the population of the present study was different from previous factor analytic studies on the OWEI, an exploratory factor analysis was conducted to identify the factors represented by the 50 items of the OWEI for the respondents in this study. The analysis was conducted by using responses from information and non-information employees but not supervisors.

Using the SPSS 10 data reduction technique, a Principal Component Analysis (PCA) was used to analyze the dimensionality of the 50-item OWEI. Semantic analysis enabled the factors to be identified as follows: Teamwork, Dependability, Ambition, and Self Control. Because of the complexity of obtaining regression-based factor scores for supervisor responses, factor scores were computed by a simple but effective method recommended by Kerlinger (1973). For this study, a loading of 0.45, which represents 20% of the variance, was used to include a variable in a factor. Table 1 lists the factors and loadings.

The factor "Teamwork" represents those items of the OWEI where a higher score may indicate that a person is more comfortable in a team environment than those with lower scores. The factor "Dependability" represents those items of the OWEI where a higher score may indicate that a person is more dependable in the workplace than a person with a lower score. The factor "Ambition" represents those OWEI items where a higher score may indicate a more ambitious person than one who has a lower score. The factor "Self Control" represents those OWEI items where a higher score may indicate a person's greater self-control capability than a person having a lower score. These variables and their values were used in the multiple regression analysis.

Distributions of factor scores (Teamwork, Ambition, Dependability, and Self-Control) were found to have unacceptable negative skewness. This negative skewness was converted to positive skewness by subtracting each value from the largest score in that distribution + 1. For example, if the value of a particular score is 50 and if the largest score in the distribution is 60, then the modified score will be (60+1) - 50 = 11. These positively skewed distributions were then subjected to square root transformations as suggested by Tabachnick and Fidell (1983).

Table 1
Factor Loadings of Four Factors in Descending Order of Loading Size
Factors

Teamwork Dependability Ambition Self Control

Item Loading Item Loading Item Loading Item Loading

Friendly .77 Reliable .67 Persistent .61 Deviousa .70
Pleasant .74 Dependable .67 Resourceful .60 Hostilea .69
Courteous .72 Effective .64 Dedicated .59 Rudea .61
Loyal .66 Careful .62 Enthusiastic .56 Selfisha .59
Considerate .65 Following   Initiating .53 Depresseda .56
Helpful .64      directions .62 Devoted .52 Negligenta .50
Likeable .64 Following   Perceptive .50 Emotionally  
Cooperative .63      regulations .61 Productive .49      stable .47
Cheerful .63 Honest .59 Ambitious .47 Carelessa .45
Devoted .56 Accurate .51 Efficient .46    
Dedicated .50 Independent .50 Persevering .45    
Enthusiastic .46 Efficient .50        
Patient .46 Adaptable .50        
    Ambitious .46        
    Perceptive .45        

aNegative items were reversed for scoring.

A MANOVA was used to test hypotheses 1, 2, and 3. Multiple regression analysis was used to test hypotheses 4 and 5. For the MANOVA tests, Pillai's criterion was used over the other three commonly used MANOVA test statistics because, according to Olson (1976) and Tabachnick and Fidell (1983), it may improve the robustness of the test. As effect sizes given by SPSS output on MANOVA do not represent multivariate effect sizes, multivariate effect sizes were calculated by the method given by Stevens (2002), which suggested 0.25 as small, 0.5 as medium, and greater than 1 as large effect sizes. Seven test plans were devised for testing hypotheses 1 & 2, and four test plans were devised for testing hypothesis 3. These test plans were devised to compare the results of MANOVA with various combinations of test variables. The test variables were: a) number of dependent variables (50 OWEI items / 4 factors), sample size (unequal / equal and matched sample size), treatment to scores (untransformed / transformed / transformed and with outliers removed). The authors believed that this endeavor may present an opportunity to test MANOVA robustness indirectly.

Findings

Hypothesis 1

To test null hypothesis 1 "At the p ≤ 0.05 level of confidence, there is no significant difference between the self-perceptions of work attitudes of industrial employees with information jobs and their work attitudes as rated by their supervisors," seven MANOVA tests were conducted with seven different combinations of test variables (Tables 2 through 5). The test plans were designed with different combinations of dependent variables, matched/unmatched samples, presence or absence of univariate outliers, and transformed or untransformed scores. Each of the seven tests revealed statistically significant differences between the selfperceived work attitudes of industrial employees with information jobs and their work attitudes as rated by their supervisors (Tables 2, 3, 4, and 5 ), leading to rejection of the first null hypothesis. Effect sizes were medium to large.

Table 2
Test Plans 1 and 2 for Hypothesis 1 Including Summary of MANOVA results, Using All Untransformed 50 Items in the OWEI
Test Details Test Plan 1 Test Plan 2

Independent Variable    
  Quantity 1 1
  Name (s) Perception Perception
 
 
 
Levels
 
 
2 ( Information employee
Self-perception and
Supervisor Perception)
2 ( Information employee
Self-perception and
Supervisor Perception)
Dependent Variables Untransformed Untransformed
  Quantity 50 50
  Names (s) As appeared in the OWEI As appeared in the OWEI
Number of Cases    
  Employee 265 244
  Supervisor 304 244
Matched? No Yes
Summary of Results    
  Pillai's Trace    
       Value 0.480 0.477
       F 9.548 7.969
       Sig. < 0.001 < 0.001
       Effect Size 3.69 3.63


Table 3
Test Plans 3 and 4 for Hypothesis 1 Including Summary of MANOVA Results, Using the 4 Untransformed Factors in the OWEI
Test Details Test Plan 3 Test Plan 4

Independent Variable    
  Quantity 1 1
  Name (s) Perception Perception
 
 
 
Levels
 
 
2 ( Information employee
Self-perception and
Supervisor Perception)
2 ( Information employee
Self-perception and
Supervisor Perception)
Dependent Variables Untransformed Untransformed
  Quantity 4 4
 
 
Names (s)
 
Teamwork, Dependability
Ambition, Self Control
Teamwork, Dependability
Ambition, Self Control
Number of Cases    
  Employee 265 244
  Supervisor 304 244
Matched? No Yes
Summary of Results    
       Pillai's Trace    
            Value 0.321 0.312
            F 66.669 54.763
            Sig. < 0.001 < 0.001
            Effect Size 1.89 1.81


Table 4
Test Plans 5 and 6 for Hypothesis 1 Including Summary of MANOVA Results, Using the 4 Factors with Transformed Scores in the OWEI
Test Details Test Plan 5 Test Plan 6

Independent Variable    
  Quantity 1 1
  Name (s) Perception Perception
 
 
 
Levels
 
 
2 ( Information employee
Self-perception and
Supervisor Perception)
2 ( Information employee
Self-perception and
Supervisor Perception)
Dependent Variables Transformed Transformed
  Quantity 4 4
 
 
Names (s)
 
Teamwork, Dependability
Ambition, Self Control
Teamwork, Dependability
Ambition, Self Control
Number of Cases    
  Employee 265 244
  Supervisor 304 244
Matched? No Yes
Summary of Results    
       Pillai's Trace    
            Value 0.178 0.179
            F 30.56 26.279
            Sig. < 0.001 < 0.001
            Effect Size 0.87 0.87


Table 5
Test Plan 7 for Hypothesis 1 Including Summary of MANOVA results, Using the 4 Transformed Factor Distributions Without Outliers in the OWEI
Test Details Test Plan 7

Independent Variable  
  Quantity 1
  Name (s) Perception
 
 
Levels
 
2 ( Information employee Self-perception and
Supervisor Perception)
Dependent Variables Transformed and outliers removed
  Quantity 4
  Names (s) Teamwork, Dependability, Ambition, Self Control
Number of Cases  
  Employee 265
  Supervisor 301
Matched? No
Summary of Results  
       Pillai's Trace  
            Value 0.186
            F 32.054
            Sig. < 0.001
            Effect Size 0.92

The ANOVA yields most reliable results when dependent variables are uncorrelated and sample (group) sizes are equal (Tabachnick & Fidell, 1983). Of the seven Test Plans, only Test Plan 4 (Table 3) satisfies these conditions. The data in Table 6 reveals significant follow-up ANOVAs for Test Plan 4 along with F-values in descending order. The Bonferroni method (Barker & Barker, 1984) was used to control Type I error.

Table 6
Summary of Test Results of Follow-up ANOVAs (Hypothesis 1)
  Variable (Factors) scores F-value
Test Plan Sig. at p < 0.0125 (in descending order)

Test Plan Dependability 70.34
4 Ambition 55.40
  Teamwork 43.31

The data in Table 6 indicates that three of the four follow-up ANOVAs were significant, with the exception of the factor Self Control. Null hypothesis 1 was rejected for the factors Dependability, Ambition, and Teamwork. Self-perceived work attitudes of Information employees were different from their supervisors' perceptions of their work attitudes on those three factors.

Hypothesis 2

To test null hypothesis 2, "At the p ≤ 0.05 level of confidence, there is no significant difference between the self-perceptions of work attitudes of industrial employees with non-information jobs and their work attitudes as rated by their supervisors", seven MANOVA tests were conducted with seven different combinations of test variables (Table 7 through Table 10). These test plans were also designed with different combinations of dependent variables, matched/unmatched samples, presence or absence of univariate outliers, and transformed or untransformed scores.

Table 7
Test Plans 8 and 9 for Hypothesis 2 Including Summary of MANOVA Results for All the Untransformed 50 Items in the OWEI
Test Details Test Plan 8 Test Plan 9

Independent Variable    
  Quantity 1 1
  Name (s) Perception Perception
 
 
 
 
Levels
 
 
 
2 ( Non-information
employee Self-perception
and Supervisor
Perception)
2 ( Non-information
employee Self-perception
and Supervisor
Perception)
Dependent Variables Untransformed Untransformed
  Quantity 50 50
  Names (s) As appeared in the OWEI As appeared in the OWEI
Number of Cases    
  Employee 189 169
  Supervisor 277 169
Matched? No Yes
Summary of Results    
       Pillai's Trace  
            Value 0.532 0.523
            F 9.425 6.288
            Sig. < 0.001 < 0.001
            Effect Size 4.69 4.36

Table 8
Test Plans 10 and 11 for Hypothesis 2 Including Summary of MANOVA Results for the Four Factors in the OWEI, Using Untransformed Factor Scores
Test Details Test Plan 10 Test Plan 11

Independent Variable    
  Quantity 1 1
  Name (s) Perception Perception
 
 
 
 
Levels
 
 
 
2 ( Non-information
employee Self-perception
and Supervisor
Perception)
2 ( Non-information
employee Self-perception
and Supervisor
Perception)
Dependent Variables Untransformed Untransformed
  Quantity 4 4
  Names (s) As appeared in the OWEI As appeared in the OWEI
Number of Cases  
  Employee 189 169
  Supervisor 277 169
Matched? No Yes
Summary of Results  
       Pillai's Trace    
            Value 0.356 0.326
            F 63.759 40.236
            Sig. < 0.001 < 0.001
            Effect Size 2.29 1.92


Table 9
Test Plans 12 and 13 for Hypothesis 2 Including Summary of MANOVA Results for the Four Factors in the OWEI, Using Transformed Factor Scores
Test Details Test Plan 12 Test Plan 13

Independent Variable    
  Quantity 1 1
  Name (s) Perception Perception
 
 
 
Levels
 
 
2 ( Information employee
Self-perception and
Supervisor Perception)
2 ( Information employee
Self-perception and
Supervisor Perception)
Dependent Variables Transformed Transformed
  Quantity 4 4
 
 
Names (s)
 
Teamwork, Dependability
Ambition, Self Control
Teamwork, Dependability
Ambition, Self Control
Number of Cases  
  Employee 189 169
  Supervisor 277 169
Matched? No Yes
Summary of Results    
       Pillai's Trace    
            Value 0.24 0.202
            F 36.396 21.02
            Sig. < 0.001 < 0.001
            Effect Size 1.3 1


Table 10
Test Plan 7 for Hypothesis 2 Including Summary of MANOVA Results for All 4 Factors in the OWEI, Using Transformed Factor Scores with Outliers Removed
Test Details Test Plan 14

Independent Variable  
  Quantity 1
  Name (s) Perception
 
 
Levels
 
2 ( Non-information employee Self-perception and
Supervisor Perception)
Dependent Variables Transformed and outliers removed
  Quantity 4
  Names (s) Teamwork, Dependability, Ambition, Self Control
Number of Cases  
  Employee 189
  Supervisor 266
Matched? No
Summary of Results  
       Pillai's Trace  
            Value 0.276
            F 42.827
            Sig. < 0.001
            Effect Size 1.56

As can be seen, each of the seven tests (Tables 7, 8, 9 and 10), revealed statistically significant differences between group means. All of the effect sizes were large. Thus, null hypothesis 2 was rejected. There were significant differences between the self-perception of work attitudes of industrial employees with noninformation jobs and their work attitudes as rated by their supervisors. The results of follow-up ANOVA tests according to Test Plan 11 (uncorrelated variables and equal sample size between groups) are presented in Table 11.

Table 11
Summary of Test Results Follow-up ANOVAs (Hypothesis 2)
  Variable (Factor) scores F-value
Test Plan Sig. at p < 0.0125 (in descending order)

Test Plan Ambition 85.10
11 Dependability 67.51
  Teamwork 55.45
  Self Control 17.58

It is evident from Table 11 that all four follow-up ANOVAs were found significant for hypothesis 2. Self-perceived work attitudes of non-information employees were significantly different from their supervisors' perceceptions of their work attitudes on all of the dimensions measured by the OWEI.

Hypothesis 3

To test hypothesis 3, "At the p ≤ 0.05 level of confidence, there is no significant difference between the perceptions of work attitudes of industrial employees with non-information jobs and industrial employees with information jobs," four MANOVA tests were conducted according four Test Plans (Table 12-13). The test plans were designed with different combinations of dependent variables, presence or absence of univariate outliers, and transformed or untransformed scores.

Table 12
Test Plans 15 and 16 for Hypothesis 3 Including Summary of MANOVA Results, Using Untransformed Scores for all 50 Variables in the OWEI
Test Details Test Plan 15 Test Plan 16

Independent Variable    
  Quantity 1 1
  Name (s) Perception Perception
 
 
 
 
Levels
 
 
 
2 (Information employee
Self-perception and Non-
information employee
Self-perception)
2 (Information employee
Self-perception and Non-
information employee
Self-perception)
Dependent Variables Untransformed Untransformed
  Quantity 50 4
 
 
Names (s)
 
As appeared in the OWEI
 
Teamwork, Dependability
Ambition, Self Control
Number of Cases    
  Employee 265 265
  Supervisor 189 189
Matched? No No
Summary of Results    
  Pillai's Trace    
       Value 0.152 0.033
       F 1.447 3.78
       Sig. 0.03 0.005
       Effect Size 0.74 0.14


Table 13
Test Plans 17 and 18 for Hypothesis 3 Including Summary of MANOVA Results Using transformed Scores and Transformed Scores with Outliers Removed for the Four Factors
Test Details Test Plan 17 Test Plan 18

Independent Variable    
  Quantity 1 1
  Name (s) Perception Perception
 
 
 
 
Levels
 
 
 
2 (Information employee
Self-perception and Non-
information employee
Self-perception)
2 (Information employee
Self-perception and Non-
information employee
Self-perception)
Dependent Variables
 
Transformed
 
Transformed and outliers
removed
  Quantity 4 4
 
 
Names (s)
 
Teamwork, Dependability
Ambition, Self Control
Teamwork, Dependability
Ambition, Self Control
Number of Cases    
  Employee 265 264
  Supervisor 189 189
Matched? No Yes
Summary of Results    
  Pillai's Trace    
       Value 0.03 0.03
       F 3.495 3.494
       Sig. 0.008 0.008
       Effect Size 0.13 0.13

As a result of the significant differences for Test Plans 15, 16, 17, and 18, null hypothesis three was rejected. There was a significant difference between selfperceived work attitudes of Information and Non-information employees. However, the effect sizes were not as large as those in the cases of hypothesis 1 and 2. None of the four Test Plans represented uncorrelated dependent variables and equal sample sizes between groups. Therefore, results of follow-up ANOVA tests were not appropriate.

Hypothesis 4

A multiple regression analysis was conducted to test null hypothesis 4, which states that at the p ≤ 0.05 level of confidence, there is no significant relationship between the work attitudes of information employees based on gender, age, level of education, and length of service.

Multiple regression analyses are extensions of bivariate regression analyses and are related to partial and semi-partial analyses. Multiple regression analysis can be termed as a specialization of the more general canonical correlation. In the simplest way, multiple regression analysis allows one to assess relationships between one dependent variable and more than one independent variable at one or more levels (Tabachnick & Fidell, 1983).

For better interpretability of multiple regression results, the number of dependent variables should be reduced. In this study, 4 factors had been identified from the 50 items of the OWEI by conducting an exploratory factor analysis. These factors were named as Teamwork, Dependability, Ambition, and Self Control and they were treated as dependent variables in the multiple regression analyses. Both untransformed and transformed factor scores were used in the analysis. The categorical independent variables of gender, years of full-time work experience, level of education, and age were coded into dummy variables with two, three, five, and five levels respectively. Multiple regression analysis was used to test the relationship between gender, age, level of education and length of service on work attitudes because it was believed that there these four variables might affect the criterion variable in unison and not independently.

The regression model applicable to this study was the fixed-effects model. This model assumes that values of independent variables are "fixed" because these are selected by the researcher rather than sampled from some population (Cohen & Cohen, 1983). The values of independent variables exhaust the set of all possible values of interest. For example, the variable gender has two values: male and female, and these two values exhaust all possible values for gender.

Hierarchical multiple regression analysis was conducted using SPSS 10 based on responses from information employees. The sampling plan was random, i.e., numbers of participants in different levels of an independent variable were not the same (Cohen & Cohen, 1983). Sets of independent variables were entered in terms of causal priority, which is required for conducting any hierarchical multiple regression analysis. It is important to note that no variable can be causally affected by one that appears after it. The order of assumed causal priority was gender, level of education, age, and years of full-time work experience. Using this order, level of education can not affect gender, age can not affect level of education (e.g., being older is not necessarily associated with having a higher level of education, but having more education does require the expenditure of time, which would make the employee older), and years of full-time work experience can not affect age.

The aim of the regression analysis was to predict the work attitudes as measured by each of the four factors from the variables:

In hierarchical regression, the effect size is a function of R2 and R2-Change (Cohen & Cohen, 1983). R2 accounts for the variance in the dependent variable that is contributed by the set or sets of independent variables and R2-Change accounts for the variance in the dependent variable contributed by the set of independent variables over and above the set or sets of independent variables entered earlier in the regression equation.

Table 14 shows the R2, R2-Change, effect size, and significance of the change in F-value for the relationship between gender and latent dependent variables. The dependent variables are the factors Teamwork, Dependability, Ambition, and Self-Control.

Table 14
Summary of Hierarchical Multiple Regression Analysis for Predictor Gender Predicting Work Attitude Measures of Teamwork, Dependability, Ambition, and Self Control with Untransformed and Transformed Scores (N = 223)

Dependent Variables R2 R2-Change Effect Size Sig.F-Change

Teamwork        
  Untransformed Scores 0.047 0.047 0.0493 0.001
  Transformed Scores 0.06 0.06 0.0638 0.001
Dependability        
  Untransformed Scores 0.053 0.053 0.056 0.001
  Transformed Scores 0.066 0.066 0.0707 0.001
Ambition        
  Untransformed Scores 0.041 0.041 0.0428 0.002
  Transformed Scores 0.048 0.048 0.0504 0.001
Self Control        
  Untransformed Scores 0.045 0.045 0.0471 0.002
  Transformed Scores 0.062 0.062 0.066 0.001

Inspection of Table 14 lead to the rejection of null hypothesis 4, "at the p ≤ 0.05 level of confidence, there is no significant relationship between the work attitudes of information employees based on gender, age, level of education, and length of service" because of the existence of statistically significant relationships between each of the dependent variables and gender. It was recognized, however, that the presence of a statistically significant relationship may not necessarily indicate significance in a practical sense. Table 15 shows bivariate and partial correlations of being female (a level of gender) with information employee work attitude measures with untransformed and transformed scores.

Table 15
Bivariate and Partial Correlations of Being Female (Level of Gender) with Information Employee Work Attitude Measures with Untransformed and Transformed Scores
    Correlation Between
    Being Female And The
  Correlation Between Work Attitude Measure
Work Attitude Being Female And Work Controlling For All
Measure Attitude Measure Other Predictors

Teamwork    
  Untransformed 0.217 0.202
  Transformed -0.244 -0.244
Dependability    
  Untransformed 0.230 0.216
  Transformed -0.257 -0.242
Ambition    
  Untransformed 0.203 0.183
  Transformed -0.218 -0.197
Self Control    
  Untransformed 0.211 0.197
  Transformed -0.249 -0.233

A review of Table 15 indicates that the bivariate and partial correlations between being female and work attitude measures were positive for untransformed scores but negative for transformed scores. However, the negative bivariate correlation was not an error. This was an effect of transforming the negatively skewed distributions to positively skewed distributions. Therefore, the result suggests that female Information employees possess better work attitudes than their male counterparts.

Hypothesis 5

Application of multiple regression analysis to test hypothesis 5, "At the p ≤ 0.05 level of confidence, there is no significant relationship between the work attitudes of non-information employees and the variables gender, level of education, age, and length of service" revealed no statistically significant relationship between the work attitudes of non-information employees based on gender, level of education, age, and length of service. Null hypothesis 5 was, therefore, accepted.

Discussion

The OWEI is a self-reporting type instrument designed to obtain scores on work attitudes. The distributions of data collected from information employees, non-- information employees and their supervisors were found to be severely negatively skewed. Even though this negatively skewed distribution made multivariate analysis more difficult, this kind of distribution was anticipated. It is a general human tendency for people to rate themselves in a positive way. The other reason for high self-ratings on OWEI items may be attributed to the fact that these employees were probably hired over their competitors for good reasons and some of these reasons may be good work attitudes.

Judging or rating one's own self is always difficult and can be affected by one's state of mind at the time of filling out the self-rating instrument, in addition to other factors. Because the supervisors are the ones who judge the work performance of the employees under them, a supervisor's rating may appear to be more authentic than the employee's self-rating. While the supervisor rates an employee, he/she is more likely to evaluate the employee by making a comparison with another employee. This may not be the case when the employee rates himself/herself. However, this does not imply that the supervisors' perceptions of work attitudes of their employees are always correct. Employees having good relationships with their supervisors may get a higher rating and employees with poor or negative relationships with their supervisors may suffer from poorer ratings. In this study, supervisors' responses were found to be more evenly distributed than those of the employees. This statement is true for both original scores and untransformed scores on latent variables (factors).

Perceptions of work attitudes, as perceived by information employees and their supervisors, is significantly different with all combinations of test variables. Medium to large effect sizes confirm that the difference is indeed strong. Similarly, significant differences were found between perceptions of work attitudes by non-information employee and their work attitudes as perceived by their supervisors. In this case, the large effect sizes indicate a very strong difference. Moreover, statistically significant differences were found in self-perceived work attitudes of information and non-information employees at an alpha level of 0.05 (hypothesis three). Given the very small F-value and small to medium effect sizes, these differences may or may not have much practical significance.

For hypothesis 1, Test Plan 4 (Table 6) three significant follow-up ANOVAs on three latent variables (factors or dimensions) were presented in descending order of F-values. The differences in perceptions of information employee work attitudes as measured by these three latent variables, between information employees and their supervisors, contributed directly to the rejection of hypothesis 1 (significant MANOVA test). With an F-value of 70.34, the latent variable Dependability topped the list while with an F-value of 55.40 Ambition ranked second, and with an F-value of 43.31 Teamwork ranked third. These results mean that supervisors and information employees differed significantly in their perceptions of information employee work attitudes measured by these three latent variables (factors or dimensions). Mean scores obtained for the factor Dependability for employees and supervisors were 81.18 and 74.18 respectively. Similarly, mean scores obtained for the factor Ambition for employees and supervisors were 69.39 and 63.62 respectively. For the factor Teamwork, mean scores obtained for employees and supervisors were 89.73 and 82.79 respectively. It can, therefore, be said that supervisors rate information employee work attitudes measured by the above mentioned three latent variables (factors or dimensions) lower than the information employees rate themselves on work attitudes measured by the same three latent variables.

A gap between employers' expectations of employee work attitudes and the employees' assertion of their own work attitudes is possible for at least two reasons: (a) No set standard for ideal information employee work attitudes is available which could be taken as a benchmark and (b) selection of the levels (1 through 7) of work attitude measures (original OWEI items) were subjective.

The greatest concern for employers should, therefore, be of the scarcity of dependable information employees followed by the scarcity of ambitious information employees. The least serious concern is the scarcity of team-player information employees. There appears to be no scarcity of information employees who are emotionally stable, and not devious, hostile, rude, selfish, depressed, negligent, or careless.

Test Plan 11 (Table 11) listed four significant follow-up ANOVAs on four-latent variables (factors or dimensions) in descending order of F-values. The differences in perceptions of non-information employee work attitudes as measured by these four latent variables, between non-information employees and their supervisors, contributed directly to the rejection of hypothesis 2 via a significant MANOVA test. The results of the ANOVAs were similar to that for hypothesis 1 with the exception that a statistically significant difference was also obtained for the latent variable Self Control. However, the relatively small F-value for ANOVAs testing Self Control, though statistically significant, may be of little practical importance. In this case too, supervisors' ratings on non-information employees' work attitudes were lower than the non-information employees' self-perceived work attitudes. Mean scores obtained for factors Dependability, Ambition, Teamwork, and Self-control were 81.17, 68.77, 90.21, and 61.82 for employees and 70.87, 60.25, 80.09, and 58.09 for supervisors.

The results of this study also revealed that information employee work attitudes, as measured by the OWEI, relate significantly with gender (Tables 12 and 13) and female information employees are more likely to endorse a positive work ethic than their male counterparts. This finding was consistent with other studies (Boatwright & Slate, 2000; Hall, 1990, 1991; Hill, 1997; Wayne, 1989; Petty & Hill, 1994; Wollack, Goodale, Witjing, & Smith, 1971). However, the same was not true when the employees are non-information employees. In this case, gender could not predict good or poor work attitudes.

Cohen & Cohen (1983) mentioned 0.02 as small effect, 0.15 as medium effect, and 0.35 as large effect for behavioral studies. For this study (hypothesis 4), effect sizes varied from 0.0471 to 0.0707. These fall into the category of small effect size. This finding generally agrees with the small to medium effect sizes reported by Hill (1997).

This study demonstrated that age (age groups), levels of education, and years of fulltime work experience are of little practical significance in predicting work attitudes of Information and non-information employees. It also showed that noninformation employee work attitudes can not be predicted by gender.

Conclusions

To the extent that the data and findings of this research study were valid and reliable, the following conclusions were drawn:

Supervisors perceive and rate Information employee work attitudes differently than the Information employees perceive and rate their work attitudes by themselves. The information employees rated themselves higher on 45 of the OWEI items than their supervisors rated them and the supervisors rated the information employees higher on 5 OWEI items than the employees rated themselves. Information employees rated themselves higher on all four OWEI factors than their supervisors rated them on those same factors.

Non-information employees perceive their work attitudes differently than their supervisors do. Non-information employees rated themselves higher on 45 of the OWEI items than their supervisors rated them. The supervisors rated the non-information employees higher on 5 OWEI items than the employees rated themselves. The Non-information employees also rated themselves higher on all four factors of the OWEI than their supervisors rated them on those same factors.

Information and Non-information employees do not perceive their work attitudes similarly. The information employees rated themselves higher on 45 of the OWEI items and the non-information employees rated themselves higher on 23 items. One item was a tie. Each group had higher means in two OWEI factors. Information employees had higher means for Dependability and Ambition and noninformation employees had higher means for Teamwork and Self-Control.

In terms of the factors measured by the OWEI, there were differences in the strength of disagreement of responses across the groups. The strongest disagreement between employees (both information and non-information) and supervisors on the perceptions of employee (both information and non-information) work attitudes was obtained for the work attitudes dimension Dependability. The second strongest disagreement between employees (both information and non-information) and supervisors on the perceptions of employee (both information and non-information) work attitudes was obtained for the work attitudes dimension Ambition. The third strongest disagreement between employees (both information and non-information) and supervisors on the perceptions of employee (both information and non-information) work attitudes was obtained for the work attitudes dimension Teamwork.

Recommendations

We suggest that employers explore the following areas to minimize the gap between employers' expectation of employee work attitudes and employees' assertion of their own work attitudes:

Both information and non-information employees perceived their own work attitudes differently than their supervisors perceived those work attitudes. This situation may lead to miscommunication between the supervisor and worker if the worker does not clearly understand the supervisor's expectations. To minimize the likelihood of misunderstandings, we recommend that supervisors discuss with workers their expectations of worker job performance, both technical and nontechnical, and provide frequent feedback on worker performance in both technical and non-technical areas.

The process of seeking and obtaining ISO 9000 certification may help bring work expectations into better alignment between workers and supervisors. In a manufacturing concern, the ISO requirement that job duties and tasks be clearly, precisely, and completely described can provide a foundation for a clearer understanding of supervisor's work expectations and can ensure that those expectations are reasonable and are related to the effective conduct of the job. We think that future research may show the merits of the ISO 9000 process for clarifying job expectations for both supervisors and workers.

Because information and non-information employees perceive their work ethic in very different ways, we recommend additional research to ascertain whether this is a function of the type of job duties each must perform or a function of the type of persons who are attracted to information versus non-information jobs. If the differences in work ethic perception prove to be a function of the nature of job duties workers can be recruited according to the work ethic they possess insofar as it is consistent with the demands of the types of jobs that are open.

This study did not consider the contribution that appropriate training interventions may have upon employee work attitudes. The contribution of training interventions to important employee work attitudes, values and habits should be explored in subsequent research.

The reasons for the differences between the employers' and workers' perceptions of the workers' work ethic is not clear. We suggest additional research to ascertain the causes of differences between employers' expectations of employee work ethic and the employees' perceptions of their own work ethic.

For an information job position, when there is a tie between a male and a female applicant, the female applicant may be given preference for hiring because she may be more likely to have desirable work attitudes.

References

Aldag, R. & Brief, A. (1975). Some correlates of work values. Journal of Applied Psychology, 60, 757-760.

Anthony, P. D. (1977). The ideology of work. Great Britain: Tavistock.

Azam, M. S. (2002). A study of supervisor and employee perceptions of work attitudes in information age manufacturing industries. Unpublished master's thesis, Illinois State University, Normal.

Barker, H. R., & Barker, B. M. (1984). Multivariate analysis of variance (MANOVA). University, AL: The University of Alabama Press.

Beach, D. P. (1982). A training program to improve work habits, attitudes, and values. Journal of Epsilon Pi Tau, 8 (2), 69-74.

Beech, J. P., Kazanas, H.C., Sapko, J. Jr., Sisson, K. A., & List, R. (1978). Necessary work values, habits, and attitudes: A final report. (Report No. 1368). Jefferson City, MO: Missouri State Department of Education.

Boatwright, J. R. & Slate, J. R. (2000). Work ethic measurement of vocational students in Georgia. Journal of Vocational Education Research, 25(4), 503-531.

Borg, W. R. & Gall, M. D. (1979). Education research: An introduction. New York, NY: Longman.

Brauchle, P. E. (1979). Self and supervisor perceptions of affective work competencies in CETA trainees: A comparative study (Doctoral dissertation, University of Missouri-Columbia, 1979). Dissertation Abstracts International, 41, 2026A.

Brauchle, P. E., Petty, G. C., & Morgan, K. R. (1983). The factorial validity of the affective work competencies inventory. Educational and Psychological Measurement, 43(2), 603-609.

Buchholz, R. A. (1978). An empirical study of contemporary beliefs about work in American society. Journal of Applied Psychology, 63, 219-227.

Cohen, J. (1988). Statistical power analysis for the behavioral sciences. (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers.

Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Erlbaum Associates, Publishers.

Custer, R. L., & Claiborne, D. M. (1992). Critical skills cluster for vocational education. Journal of Vocational Education Research, 17(4), 15-40.

Custer, R. L., & Claiborne, D. M. (1995). Critical skills cluster for vocational education: The employers' perspective - A replication study. Journal of Vocational Education Research, 20(1), 7-33.

Furnham, A. & Muhiudeen, C. (1984). The protestant work ethic in Britain and Malaysia. Journal of Social Psychology, 122, 157-161.

Gaither, N. (1996). Production and operations management. Belmont, CA: Duxbury Press.

Gilbert, J. B. (1977). Work without salvation. Baltimore: The John Hopkins University Press.

Goodale, J. (1973). Effects of personal background and training on work values of the hard-core unemployed. Journal of Applied Psychology, 57, 1-9.

Hall, G. S. (1990). Work attitudes of traditional and non-traditional technical community college students. Unpublished master's thesis. The University of Tennessee, Knoxville.

Hall, G. S. (1991). Do older college students have different attitudes about work as compared with younger traditional students? Tennessee Education, 21(2), 27-29.

Hatcher, T. (1995). From Apprentice to Instructor: Work ethic in apprenticeship training. Journal of Industrial teacher Education, 33(1), 24-45.

Hill, R. B. (1992). The work ethic as determined by occupation, education, age, gender, work experience, and empowerment. (Doctoral dissertation, The University of Tennessee, Knoxville, 1992). Dissertation Abstracts International, 53, 2343.

Hill, R. B. (1996a). History of work ethic [On-line] Available: http://www.coe.uga.edu/~rhill/workethic/hist.htm [Retrieved December 1,2001].

Hill, R. B. (1996b). Work ethic differences in vocational education students and fulltime employed workers. Journal of Vocational Education Research, 21(3), 13- 29.

Hill, R. B. (1997). Demographic differences in selected work ethic attributes. Journal of Career Development, 24(1), 3-23.

Hill, R. B., & Rojewski, J. W. (1999). Double jeopardy: Work ethic differences in youth at risk of school failure. Career Development Quarterly, 47(3), 267-279.

Hill, R. B., & Petty, G. C. (1995). A new look at selected employability skills: A factor analysis of the occupational work ethic. Journal of Vocational Education Research, 20(4), 59-73.

Huitt, W. G. (1997). Success in the information age: A paradigm shift [On-line]. Available: http://www.valdosta.edu/~whuitt/psy702/context/infoage.html [Retrieved on December 1, 2001].

Kazanas, H. C. (1978). Affective work competencies for vocational education. Columbia, MO: University of Missouri. (ERIC Document Reproduction Service No. EJ 075 733).

Kerlinger, F. N. (1973). Foundations of behavioral research. New York, NY: Holt, Rinehart and Winston, Inc.

Korth, B. (1975). Exploratory factor analysis. In D. J. Amick, & H. J., Walberg (Eds.), Introductory multivariate analysis (pp. 113-146). Berkeley, CA: McCutchan Publishing Corporation.

Law, C., Knuth, R. A., & Bergman, S. (1992). What does research say about schoolto-work transition? [On-line]. Available: http://www.ncrel.org/sdrs /areas/stw_esys/7sch2wrk.htm [December 1, 2001].

Lipset, S. M. (1990). The work ethic - then and now. Public Interest. Winter 1990, 61-69.

MacDonald, A. P. (1972). More on the protestant work ethic. Journal of Consulting and Clinical Psychology, 39, 125-127.

Manufacturers' News, Inc. (2000). 2000 Illinois Manufacturers Directory. Chicago, IL : Author.

Maywood, A. G. (1982). Vocational education and the work ethic. Canadian Vocational Journal, 18(3), 7-12.

Miller, W. F. (1986). Emerging technologies and their implications for America. USA Today, 115, November 1986, 60-65.

Minton, L. W. (1997). The impact of a workplace readiness course on the selfperceived work ethic of students (Doctoral dissertation, North Carolina State University, 1997). Dissertation Abstracts International, 58-06A, 2147.

Natriello, G. (1989). What do employers want in entry-level workers? An assessment of the evidence. (National Center for Education and Employment Occasional Paper No. 7). New York: Columbia University.

Olson, C. L. (1976). On choosing a test statistic in multivariate analysis of variance. Psychological Bulletin, 83, 579-586.

Parker, M. (1998). Sampling with replacement and sampling without replacement. [On-line]. Available: http://rene.ma.utexas.edu/users/parker/sampling/woreplshort.htm [Retrieved December 1, 2001].

Petty, G. C. (1995). Adults in the work force and the occupational work ethic. Journal of Studies in Technical Careers, XV(3), 133-140.

Petty, G. C., & Hill, R. B. (1994). Are women and men different? A study of the occupational work ethic. Journal of Vocational Education Research, 19(1), 71-89.

Petty, G. C., Kazanas, H. C., & Eastman, R. M. (1981). Affective work competencies of workers, supervisors, and vocational educators. Journal of Vocational Education Research, IV, 55-71.

Rose, M. (1985). Reworking the work ethic: Economic values and socio-cultural politics. London: Schocken.

Rodgers, D. T. (1978). The work ethic in industrial America, 1850-1920. Chicago: The University of Chicago Press.

Sherer, M., & Eadie, R. (1987). Employability Skills: Key to success. Thrust, 17(2), 16-17.

Smith, B. N., Jones, C., & Lane, J. (1997). Employers' perceptions of work skills. Business Education Forum, 51(4), 11-17.

Stevens, J. (2002). Applied multivariate statistics for the social sciences. Mahwah, N.J. : Lawrence Erlbaum Associates.

Tabachnick, B. G., & Fidell, L. S. (1983). Using multivariate statistics. New York, NY: Harper & Row.

Tang, T. L. (1989). Effects of work ethic and task labels on task performance. The Journal of Psychology, 123(5), 429-438.

Tang, T. L. & Tzeng, J. Y. (1992). Demographic correlates of protestant work ethic. The Journal of Psychology, 126(2), 163-170.

Tilgher, A. (1930). Homo faber: Work through the ages (D.C. Fisher, Trans.). New York: Harcourt Brace.

Toffler, A. (1981). The third wave. New York, NY: Bantam.

Wang, G. T. (1996). A comparative study of extrinsic and intrinsic work values of employees in the United States and Japan. New York: The Edwin Mellen Press, Ltd.

Wayne, F. S. (1989). An instrument to measure adherence to the Protestant ethic and contemporary work values. Journal of Business Ethics, 8, 793-804.

Wilkinson, L., & Task Force on Statistical Inference.(1999). Statistical methods in psychology journals: Guidelines and explanations. American Psychologist, 54, 594-604. [On-line] Available: http://www.apa.org/journals/amp/ amp548594.html #c13 [Retrieved December 1, 2001].

Wollack, S., Goodale, J. G., Witjing, J. P., & Smith (1971). Development of the survey of work values. Journal of Applied Psychology, 55, 331-338.

The Authors

Md. Shafiqul Azam can be contacted at Pro-Tech Search, Inc., 400 Chatham Road, Suite 201, Springfield, IL 62704. Office Phone: (217) 793- 2790.

Paul Brauchle is Professor and Director, Bureau of Training and Development, and Associate Director, The Center for Mathematics, Science & Technology, Department of Technology, Illinois State University, Room 210K Turner Hall, Campus Box 5100, Normal, IL 61790-5100. Phone: (309) 438-2696. E-mail: pebrauc@ilstu.edu.

by Radiya Rashid