Sociation Today

Who Stole My Lunch? Gender Differences in Workplace Discrimination and Theft by Employees

by

John Casten
Longwood University

Introduction

    Although the relationship between gender discrimination and crime has been a focus of feminist criminology for some time, historically, it has not been a focus of "main stream" criminology. Some strain theory research, however, has recently turned to focus on sex differences in criminal behavior (e.g., Broidy 2001; Dingwell 2001; Eitle 2002; Langton 2004; Leeper-Piquero and Sealock 2004; Morgan 2006; Rebellon et al. 2009; Slocum, Simpson, and Smith 2005). Robert Agnew and other General Strain (GST) theorists predicted the strain resulting from problems of discrimination may induce negative affective states and the coping response of criminal behavior (Agnew 2006; Brezina 1996; Broidy 2001; Broidy and Agnew 1997: Eitle 2002; Mazarolle and Piquero 1998). Strain theory will act as the theoretical foundation of this study for that reason.

     Discrimination in the workplace based on gender may be correlated with theft by employees because discrimination is likely to be seen as unjust, high in magnitude, or associated with low social control (Agnew 2001: 343), all of which can lead to negative affective states and the coping mechanism of theft. In this sense, low social control refers to harsh treatment by others (supervisors or coworkers) in the workplace, which leads to reduced commitment or attachment to the workplace (Agnew 2001). Research has addressed the effects of general economic pressures or financial need on the part of the offender (Hollinger and Clark 1983; Mustaine and Tewksbury 2002) and the disjunction between how workers perceive they should be compensated and treated versus their actual pay and experiences on the job (Greenberg 1990; Hollinger and Clark 1983; Huiras et al. 2000). There has not, however, been much quantitative research examining discrimination as a cause of strain-induced workplace theft (Agnew 2001). One exception is David Eitle's work (2002). Eitle's (2002) research suggests there is some evidence that strain contributes to female criminality in the workplace through stratification. In his study, women who were involved in crime were more likely to report experiencing major forms of gender-based stratification than those who were not (28 versus 10 percent) (Eitle 2002: 435).

     This work examines student employee attitudes toward theft in the workplace and its relationship to attitudes toward perceived gender discrimination. Attitudes of students toward perceived gender discrimination and the willingness to commit theft in the workplace, labeled here as intentions to steal, are used as opposed to actual discrimination and theft data, due to the relatively limited work experience of the sample. This limited work experience, however, may provide sufficient information on workplace deviance since malfeasance is more likely to be committed by those with little stake in an organization occupying marginal jobs (Huiras, Uggen, and McMorris 2000). Indeed, age is a well-documented predictor of deviance "with younger workers more likely than older workers to engage in workplace deviance" (Huiras et al. 2000: 246). Attitudes on theft by employees have been studied in the field of business and psychology for some time. Conversely, criminology has focused mainly on the causes and implications of white-collar crime and rarely examines the attitudes of lower level employees and discrimination. Additionally, those studies that have been done tend to use a combined measure for strain and/or discrimination, where this study separates strain measures to examine individual differences among strains. The switch in approach is understandable, since Agnew (1992) suggested combined measures should be used. But later recanted somewhat (2001), recommending the use of separate measures. This research, then, is necessary to fill those voids. The hypotheses are: 1) strains resulting from perceived sexual harassment, the "pay gap", and the "glass ceiling" will positively and significantly predict intentions to steal, 2) the perceived negative affective states of anger or depression will positively and significantly predict intentions to steal, and 3) significant differences in perception will exist between males and females on gender discrimination and intentions to steal.

Unique Aspects of the Study

There are several unique aspects of the study. Each is deserving of some comment. A unique aspect of the study is the use of separate discrimination measures. Combined measures tend to underestimate the effects of perception scales; in this case, perceived objective strain on intentions to steal.

Another important aspect of the study was the testing of hypotheses on separate male and female samples. This allows for examination of psychosocial variables on different groups as well as z-scores between males and females. Both of these aspects add to the body of knowledge in this area, which is lacking in criminology.

Sample

     This study, with the approval of an IRB, used a sample of 400 students at a major mid-Atlantic University with the understanding that it was voluntary and anonymous. The data were collected using a survey distributed during the spring session of 2011 in 13 different Criminal Justice and Sociology introductory classes. Introductory classes were predicted to have a greater diversity of students in terms of majors and gender than upper level courses. Prior to the administration of the survey, it was pre-tested with a student sample of 25 (10 males and 15 females) to ensure reliability of the measures and that there were no language or data collection issues. Several wording and organizational issues were identified and corrections were made. The final survey consisted of 99 questions, soliciting responses that measured a variety of demographic factors, factors associated with social and psychosocial controls, perceptions of workplace discrimination and intentions to steal.

     This study used a vignette technique to capture attitudes toward perceived gender discrimination and intentions to steal by employees. This technique utilizes a series of fictitious scenarios based on real-life examples designed to elicit the attitudes of the respondents who may or may not have experienced these events (Finch 1987; Jasso 2006; Schoenberg and Ravdal 2000).

     Of the 400 students who participated in the study, surveys from 361 students were completed and returned. To satisfy the assumptions of regression analysis, the data were checked for outliers, normality, and linearity. Frequencies and boxplots were then run on each of the variables to check for outliers. Nine cases were identified as outliers and dropped from the analysis by case number. An additional nine cases were identified as having coding errors, which were then corrected, resulting in a total sample size of 352. The sample size was further reduced to 315 with the exclusion of those students who had never worked.

Table 1.
Descriptive Statistics for Demographic Variables (N=315)

Variable	Sample %	Univ. %	Mean	SD	Min.	Max
Sex
0=Female	58.7	55.0
1=Male	41.3	45.0
Race
0=Other	43.8	43.3
1=White	56.2	56.7
Marital Status
0=Single	90.2	N/A
1=Not-Single	9.8
Age			22.14	5.29	18	55
University Age			25.3

The total sample (n = 315) ranged in age from 18 to 55 with 86% of the sample being under the age of 24. Females accounted for 58.7% (n = 185). Ninety percent (n = 284) of the sample indicated they were single and 56.2% (n = 177) indicated their race as White. Thirty three percent of the sample listed their race as African-American (n = 103), while 11% indicated their race as Hispanic, Asian, or Other (n = 35). Due to the small number of respondents indicating race as Hispanic, Asian, and Other, race was recoded as White (1) and Non-white (0) (Table 1).

Literature Review

Theories used to explain theft by employees have focused primarily on factors external to the workplace (e.g., differential association). But fewer attempts have been made to explain lower-level occupational crime from internal stressors. Agnew's General Strain Theory (Agnew 1992) provides a framework for this study as it specifically posits gender discrimination as an impetus for crime.

At GST's core, Agnew (1992: 50) hypothesized three "ideal" types, or categories, of strain dealing with the negative relationships with others. Specifically, Agnew suggested that strain induced crime can result from a failure to achieve socially valued goals (the disjunction between aspirations and actual achievements), the removal of positively valued stimuli (the loss of something or someone viewed as important to the individual), or the presence of negative stimuli (the stress caused by the deviant actions of others) (Agnew 1992). These items could include blocked opportunities or a discrepancy between what one perceives as fair or unfair. This particular brand of strain theory also observes that "females suffer from a range of oppressive conditions and that this oppression is at the root of their crime" (Broidy and Agnew 1997: 276). Additionally, males and females may experience similar levels of strain, but react differently to it (Broidy and Agnew 1997).

Theft by Employees

One of the more serious forms of employee deviance is theft. Hollinger and Clark (1983: 1) define theft by employees as "the unauthorized taking, control or transfer of money and/or property of the formal work organization perpetrated by an employee during the course of occupational activity which is related to his or her employment." Theft by employees can take many forms, including theft of cash, merchandise, services, time card fraud, confidential information, and false injury claims (Hollinger and Clark 1983). A growing form of theft involves manipulation of computer systems to make unauthorized purchases and false merchandise credits (Hayes 2008). These thefts, primarily committed by non-management personnel, are generally separate from and in addition to the myriad forms of white-collar crimes prevalent in business.

Despite fairly recent advancements in video surveillance and computer monitoring, theft by employees is still a growing concern. Hollinger (2011) notes, "the $35 billion in loss due to retail crime is…larger than any other form of property crime in the United States; larger than auto theft, bank robbery, burglary, and personal robbery." Theft by employees is still responsible for up to twenty five percent of those losses (Hayes 2013). In fact, a recent annual survey (Hayes 2013) of 23 retail chains with over 18,000 stores (approximately 1.85 million employees) indicated that one in every 40 employees were caught stealing in 2012 with over 50 million dollars recovered. It must be noted that the above figures represent only those employees who were caught and does not necessarily represent the entire scope of the problem.

Discrimination

Many forms of discrimination exist in the workplace. This study, however, focuses on the three most common, and arguably most studied, forms of gender discrimination; sexual harassment, the "pay gap", and the "glass ceiling." Sexual harassment is typically divided into two types; quid pro quo and hostile work environment. Quid pro quo harassment refers to "implicit or explicit efforts to make job-related outcomes conditional on sexual cooperation" (Glomb et al. 1999: 21). Hostile work environment consists of "offensive, misogynistic, and degrading remarks, behavior not intended to elicit sexual cooperation" (Glomb et al. 1999: 21). Due to its magnitude of strain, the vignette on sexual harassment involves only the quid pro quo type of harassment.

Sexual harassment is a violation of Title VII of the 1964 Civil Rights Act (Dye 2008), but whether the behavior is illegal or if it is simply unwelcome, it would be presumed to be stressful. Indeed, sexual harassment has been shown to be a strain on women similar to other job stressors in the workplace (Fitzgerald and Shullman 1993; Fitzgerald, Hulin, and Drasgow 1995; Swanson 2000). Conversely, the strain of sexual harassment imposed on males is relatively unknown due to a lack of study.

The other two types of discrimination have also been related in some way to strain. The "pay gap" between men and women is one of the most studied gender stratification issues in the workplace. Antidiscrimination legislation of the early 1960's, in particular the Equal Pay Act of 1963, started the process of closing the wage gap. Gender stratification in the form of differences in pay for the same work lessened, but did not disappear (Beeghley 1996; Kemp and Beck 1986). The median income for men was $40,798 and $31,223 for women (Barak et al. 2007: 74). Currently, women still make on average only 77 cents on each dollar made by a man.

In terms of the "pay gap" and negative affective states, Greenberg (1990) found that it is possible that pay reduction may lead to feelings of frustration and resentment, which motivate predatory theft. Likewise, recent research findings demonstrate that pay cuts are associated with negative reactions to organizational authorities. Negative actions also may be direct attempts to correct perceived pay inequity (Greenberg 2002).

The "glass ceiling" refers to the inability of women to break though middle-management jobs to the higher status positions in a company (Kerbo 2006). One explanation is that upper-level management qualities deemed to be essential typically mirror those important to males – assertiveness, aggressiveness, and independence (Kerbo 1996). Another explanation is that some women are assigned the "mommy track" preventing them from advancing further in the company. These attitudes and trends persist as male managers tend to promote "similar others" (Chatman et al. 1998).

Measures

Dependent Variable

The key dependent variable used in this study was intentions to steal. It was assumed that most employees have not stolen from work and would not be included in an analysis of actual crime. This does not mean, however, they would not steal given the right impetus (e.g., gender discrimination) or opportunity. Indeed, since research in this area became prominent in the 1970's, attitudes have been shown to be a reliable indicator of actual propensity (Azjen and Fishbein 1977; Bolin and Heatherly 2001). Intentions to steal were therefore examined in an attempt to capture attitudes of all respondents, even those without theft experience.

Both dependent and independent variables were derived from questions relating to survey vignettes. Vignettes have been used successfully in prior studies to measure strain, anger, and assault (e.g., Mazerolle and Piquero 1997), and are recommended by Agnew (2001: 347) in this type of study when actual theft data are not available. There were three separate vignettes used in this analysis. The first part of each vignette pertained to the independent variables and covered the results of perceived gender discrimination in the workplace; sexual harassment, the "pay gap", and the "glass ceiling". Each vignette was written in the first person "you" format, to elicit a more personal response. The second part of the vignette, pertaining to the dependent variable, asked if theft was justified as a result of the unfair treatment. The respondent was asked how likely it would be for him or her, based on the treatment, to engage in theft behavior: "Under the circumstances, I think keeping it [20 dollars] would be justified". The responses were measured on a four-point range from 1 to 4 (1= never; 2= somewhat; 3=definitely; 4=very likely) then dichotomized for logistic regression (Table 2).

Key Independent Variables

The objective strain measure was the level of perceived strain of each vignette. Objective strains are those that are usually disliked by most members of a particular group (Agnew 2001). For example, most would agree that sexual harassment happens in the workplace and would be considered stressful. The person representing the respondent in the vignettes was subjected to forms of workplace discrimination. The scenarios were designed to correspond to inequity, focusing on GST's disjunctions between just and fair outcomes and actual outcomes. The sexual harassment vignette represented the presence of negative stimuli, the "pay gap" vignette represented the absence of positive stimuli, and the "glass ceiling" vignette represented goal blockage. The questions were gender-neutral and answered by both males and females. Respondents were then asked to rate, on a scale of 1 to 10 the level of truthfulness of the following statement: "You would find [the situation] stressful."

Measures of the perceived negative affective states of anger and depression are also key independent variables. Anger and depression are typical emotional responses for males and females under stress (Agnew 2001). The perceived anger and depression variables in the vignettes were measured on a scale of whether or not the respondent would feel a negative affective state in that situation. They were scored on a 10 point range using the question "On a scale of 1 to 10 with 1 being low to 10 being high, indicate how this scenario would make you feel if it actually happened" and based on the statements "I would feel angry, [depressed]" (refer to Table 2 for statistics). Vignettes have been used for this purpose in previous studies to some positive effect (e.g., Mazerolle and Piquero 1997, 1998; Morgan 2006; Rebellon et al. 2009). Dingwell (2001) used vignettes to measure state anger toward inequity in the workplace. He found that state or situational anger was not only positively related to the strain of injustice, but also was more robust when the respondent scored high on the dispositional anger scale. Likewise, Morgan (2006) found similar results when addressing depression.

Table 2:
Descriptive Statistics for Dependent,
Independent and Control Variables

Variable	Mean	SD	Min.	Max	N
Vignette 1 Sexual Harassment
Objective Strain	.33	.472	0	1	315
Anger	.55	.498	0	1	315
Depression	.13	.341	0	1	315
Theft Intention	.37	.484	0	1	315
Vignette 2 "Pay Gap"
Objective Strain	.44	.497	0	1	315
Anger	.67	.471	0	1	315
Depression	.18	.389	0	1	315
Theft Intention	.45	.498	0	1	315
Vignette 3 "Glass Ceiling"
Objective Strain	.56	.497	0	1	315
Anger	.75	.435	0	1	315
Depression	.23	.422	0	1	315
Theft Intention	.42	.494	0	1	315
Controls
Delinquent peers	14.13	2.75	5	20	315
Family support	4.77	1.81	3	12	315
Impulsivity	8.47	2.47	0	16	315
Commitment	9.00	4.52	0	16	315

Control Variables

The effects of psychological and environmental conditions on decisions to engage in deviance has been well documented in both crime and psychology literature. This study controls for many of the often-cited "causal" factors for crime and deviance in an effort to isolate the effects of strain and discrimination; namely: social control, differential association, self-control, support, and demographic factors (Table 2).

Social control and differential association.

     As noted by Agnew, Brezina, and Wright (2002) social control and differential association are necessary controls when utilizing strain theory to avoid overestimating the effects of strain. According to Agnew "The failure to do so may cause us to overestimate the effect of the strain measures, because strain is frequently correlated with social control and the social learning of delinquency" (Agnew et al. 2002: 51).

     There are equivalencies of social control in the workplace and social control in the general population. For example, parents, clergy, and teachers can be equated to managers; police can be equated to security and supervisors; and social rules, boundaries, and limitations can be equated to workplace policies and procedures. In this sense, low social control refers to harsh treatment by significant or important others in the workplace (supervisors, managers, or coworkers), which leads to reduced commitment or attachment to the workplace (Agnew 2001). The level of commitment to the workplace was therefore used as the low social control measure. In this study, individual levels of commitment were being examined rather than a single organizational culture. Questions measuring commitment were derived from the Organizational Commitment Questionnaire (OCQ) (Mowday, Steers, and Porter 1979). The fifteen question OCQ was developed to measure "three aspects of commitment: 1) a strong belief in and acceptance of the organization's goals and values, 2) a willingness to exert considerable effort on behalf of the organization, and 3) a strong desire to maintain membership in the organization" (Mowday et al. 1979: 226). The OCQ has been shown to be reliable (alpha's in the .81 to .93 range with a mean of .91; see Azjen 2001) in predicting levels of commitment in the workplace. Of the three types of commitment predicted by the OCQ, calculative commitment is the result of necessity for the purpose of sustaining life or lifestyle (Lee and Gao 2005). Only those 4 items from the questionnaire relative to calculative commitment were used. Calculative commitment may have the potential to be more influential on decisions to steal from the workplace due to the influences of opportunity, peer pressure, or other external factors rather than organizational culture. Additionally, though most research on commitment is done within a single work environment, place of employment is of less importance when examining calculative commitment as work is perceived as a means to an end. The wording of the questions indicates to the respondent to think of his/her own workplace. For example, "I have no allegiance to my current company and I could just as well be working for a different organization doing the same thing." Those 4 items were scored on a four point Likert-type scale ranging from 1 "strongly agree" to 4 "strongly disagree". Factor analysis produced one factor explaining 46.9% of the variance (KMO = .674; Chi Sq. = 135.64; df = 6; p = .000) and an alpha of .617. Factor loadings ranged from .590 to .784.

     In order to distinguish between social control and differential association measures, Agnew (1992) suggested only those variables that clearly indicate negative relations with others be used when controlling for differential association. The differential association measures were derived from a series of six questions relating to contact with or knowledge of delinquent or criminal others. They were scored on a four point Likert-type scale ranging from 1 "strongly agree" to 4 "strongly disagree". Factor analysis produced one factor (KMO = .767; Chi Sq. = 290.04; df. =15; p = .000) explaining 39.5% of the variance with an alpha of .643. The alpha was improved to .678 by removing one question. Factor loadings ranged from .581 to .722.

Social support.

Agnew and White (1992) also cite the importance of the presence of social support in the decision to commit crime. A lack of support may lead to the decision to use crime as an adaptation to strain. Alternatively, those with adequate conventional support should better handle strain in a non-deviant manner. For this reason, the availability of social support must be a control.

For this study, a version of the social support appraisal scale developed by Vaux et al. (1986) was used. The scale measures the degree to which the respondents are involved with family, friends, and others. There were 6 questions, 3 negatively and 3 positively phrased, used from the larger set based on appropriateness to this study. They were measured on a 4 point Likert-type scale ranging from 1 "Strongly Agree" to 4 "Strongly Disagree." The negatively phrased questions were reverse coded prior to analysis. Factor analysis produced 2 factors explaining 63.89% of the variance. (KMO = .781; Chi Sq. = 520.98; df. = 15; p = .000). The factors were split between support from family (alpha = .619) and support from friends (alpha = .774). The friend scale was not used in this study as it correlated highly with deviant peers, resulting in support being negatively viewed by respondents. Factor loadings for family support ranged from .524 to .854.

Self-control.

Recent studies have noted the importance of self-control in theft related behaviors (e.g., Langton, Leeper-Piquero, and Hollinger 2006) and the importance to differentiate low self-control from situational strain. For example, Langton, Leeper-Piquero, and Hollinger (2006) found that fraudulent statements on resumes were correlated with both low self-control and intentions to commit white-collar crime among college students. For this study then, a version of the well-tested Low Self-Control Scale (Grasmick, Tittle, Bursik, and Arneklev 1993) was used to assess the respondent's level of self-control. In this case, however, only the impulsivity section of the scale will be used. As noted in Arneklev, Grasmick, and Bursik (1999: 327), the impulsivity component is strongly correlated with the other five components, causing the authors to remark, "Is low self-control, to a large extent, simply impulsivity?" Self-control was measured by 4 questions using a 4 point Likert-type scale ranging from 1= "strongly agree" to 4= "strongly disagree." Factor analysis produced one factor (KMO = .723; Chi. Sq. = 419.82; df = 6; p = .000) explaining 60.38% of the variance and an alpha of .775, which is consistent with Arneklev et al. (1999). Factor loadings ranged from .713 to .841.

Analytic Strategy

     All analyses were produced using Statistical Package for Social Sciences (SPSS) with the "enter" method. To investigate differences across gender, analyses on the data were reproduced for males and females separately with z-scores. While some loss of statistical power was inevitable with this method, it allowed for a detailed examination of males as a separate group and facilitated the use of z-score comparisons. The population for each group was greater than 30 with known variances. The analysis included models derived from the dependent, independent, and control variables. The analyses were conducted using logistic regression, which does not assume a normal distribution and is more robust to skewed data. There was no improvement with the data through Log 10 or square root transformation. Although some variation is lost, the question of inclusion or exclusion is still answered.

     The dependent variables of intentions to steal were regressed on the independent variables of perceived objective strain and negative affective states for their respective vignettes: specifically, whether theft is perceived to be justified through the logged odds of group membership.

     Agnew (2001) noted that past research on strain had a tendency to construct strain measures using combined scales (e.g., Mazerolle and Piquero 1997 summed the scores of all of the strain measures). But that this method had pitfalls due to the difficulty of teasing out those strains that affected the likelihood of crime from those that did not. In light of these types of issues, Agnew (2001) recommended separating strains in future research. This study uses separate measures for discrimination types for that reason.

Results

Bivariate Correlations

Bivariate correlations were conducted on the key determinant variables for males and females separately. Major significant results are discussed here without tables due to space limitations. The results for males showed only perceived strain associated with the "glass ceiling" scenario significantly correlated (r = .216; p = .014) with intentions to steal. In terms of emotion, both perceived anger and depression associated with the "pay gap" vignette were also significantly (r = .258; p = .003, r = .248; p = .004 respectively), correlated with intentions to steal for males. For females, perceived strain associated with the sexual harassment vignette was significantly (r = -.203; p = .006) correlated with intentions to steal. The "pay gap" and the "glass ceiling" vignettes produced no significant results for females. In addition, neither perceived anger nor depression was significantly correlated with intentions to steal for females. For both males and females and with all three vignettes, both impulsivity and peer associations were significantly correlated with intentions to steal. Interestingly, age and commitment were significant correlates in all three vignettes for females, but not for males.

Logistic Regression Models

A logistic regression analysis was run with the perceived objective strain variable and major control variables predicting intentions to steal for each type of discrimination measure associated with the individual vignettes. Although separate blocks were run for each vignette to examine the effects of perceived strain on intentions to steal and perceived strain, anger, and depression on intentions to steal, the relationships held with the addition of the control variables in each case. Only the full models, therefore, are reported here.

Table 3.
Full Logistic Regression Models of All Vignettes for Males (N = 130)

Vignettes Odds Ratios	Sexual Harr. Exp (B) B (SE)	Pay Gap Exp(B) B(SE)	Glass Ceiling Exp(B) B(SE)
Objective Strain	1.06 .06 (.11)	1.14 .13 (.12)	1.18 .17 (.13)
Anger	1.04 .04 (.09)	1.26 .23 (.13)	1.03 .03 (.16)
Depression	1.04 .04 (.08)	1.12 .12 (.07)	1.09 .09 (.08)
Age	.98 -.02 (.06)	.96 -.04 (.06)	1.00 .00 (.06)
White	.59 -.52 (.43)	.42* -.86 (.44)	.59 -.52 (.44)
Married	1.38 .32 (1.14)	1.49 .40 (1.14)	1.86 .62 (1.19)
Impulsivity	1.38** .32 (.10)	1.42** .35 (.11)	1.41** .35(.10)
Peer Association	.95 -.06 (.09)	1.06 .06 (.10)	.98 -.02 (.09)
Family Support	.94 -.06 (.13)	1.14 .13 (.14)	.90 -.11 (.13)
Commitment	.92 -.08 (.06)	.87* -.14 (.06)	.92 -.09 (.06)
Chi-SQ (df)	23.35 (10)	37.29 (10)	32.08 (10)
-2 Log	155.35	142.65	148.11
Nagelkerke R-SQ	.22	.33	.29
Significance	.01	.00	.00

** = p<.01; *= p<.05

Males.

As is shown in Table 3, most of the significant results for males were associated with the "pay gap" scenario. Indeed, none of the key determinant variables were significant predictors of intentions to steal. In the "pay gap" scenario, a unit change in race (White) and a unit change in commitment decreased the odds of intentions to steal by 58% and 13%, respectively (p = < .05). Respondents who indicated their race as White and who indicated higher levels of commitment were less likely to perceive theft as justifiable.

With respect to commitment, this result is consistent with literature on organizational behavior. A unit change in impulsivity increases the odds of intentions to steal by 38%, 42%, and 41%, (p = < .01) in all three vignettes respectively. This too would be an expected result based on current literature. The models were significant and explained between 22 and 33 percent of the variance (Table 3).

Females.

For females (Table 4), none of the key determinant variables were significant predictors of intentions to steal. This was consistent with the results for males. Interestingly, significant results were achieved with control variables different from those that were significant for males and on different discrimination types. For example, in the sexual harassment scenario, a unit change in age (older) decreases the odds of intentions to steal by 13% (p = < .05). Also in the sexual harassment scenario, a unit change in race (White) increases the odds of intentions to steal by 2 ¼ times (p = < .05). The commitment variable was also significant in both the "pay gap" and "glass ceiling" scenarios. A unit change in commitment decreases the odds of intentions to steal by 11% and 13%, respectively (p = < .05). The models were significant and explained between 18 and 19 percent of the variance.

Comparisons between Males and Females

As shown in Table 5, there were few significant differences between males and females. Impulsivity in the sexual harassment and "glass ceiling" vignettes did indicate, however, significant differences. In addition, the race (White) variable indicated significant differences between males and females.

Table 4
Full Logistic Regression Models of All Vignettes for Females (N = 185)

Vignettes Odds Ratios B (SE)	Sexual Harr. Exp (B) B (SE)	Pay Gap Exp(B) B(SE)	Glass Ceiling Exp(B) B(SE)
Objective Strain	.99 -.01 (.11)	1.07 .06(.11)	.90 -.01 (.11)
Anger	.96 -.05 (.16)	1.17 .16 (.18)	.87 -.14 (.16)
Depression	1.03 .03 (.06)	.95 -.05 (.06)	.98 -.02 (.06)
Age	.87 -.14 (.07)	.92 -.09 (.05)	.94 -.06 (.05)
White	2.20* .79 (.36)	1.80 .59 (.34)	1.77 .57 (.35)
Married	1.57 .45 (.77)	.72 -.32 (.62).	1.90 .64 (.70)
Impulsivity	1.05 .05 (.08)	1.15 .14 (.07)	1.10 .10 (.07)
Peer Association	.92 -.08 (.07)	.92 -.08 (.07)	.93 -.07 (.07)
Family Support	1.08 .08 (.09)	1.07 .07 (.08)	1.11 .11 (.09)
Commitment	.90 -.10 (.06)	.89* -.12 (.05)	.87* -.14 (.06)
Chi-SQ (df)	26.51 (10)	25.59 (10)	24.89 (10)
-2 Log	205.13	219.82	209.69
Nagelkerke R-SQ	.19	.18	.18
Significance	.00	.00	.01

** = p<.01; *= p<.05

This result is somewhat inconsistent with those achieved in other research (e.g., Broidy and Agnew 1997) in that males and females experience perceived strain in the same ways but respond differently to relieve that strain. The differences in this study suggest males and females may respond similarly to perceived strains, in that theft would not be justifiable, and may actually experience them differently as well. This is due to the effect of the different control variables within each group. Stated alternatively, males and females may have different experiences given the different social or psychosocial factors affecting each group.

Table 5.
Z-scores for All Vignettes (n = 315)

	Sexual. Harassment	Pay Gap	Glass Ceiling
Objective strain	-.45	-.43	-1.06
Anger	-.49	-.32	-.75
Depression	-.10	-1.84	-1.10
Age	-1.30	-.64	-.77
White	2.34*	2.61*	1.94
Married	.09	-.55	.01
Impulsivity	-2.11*	-1.61	-2.05*
Peer Association	-.18	-1.15	-.44
Family Support	.89	-.37	1.39
Commitment	-.24	-.26	-.59

** = p<.01; *= p<.05

Discussion and Conclusion

     Similar to other studies using strain theory to examine discrimination, the hypotheses were not generally supported. But it did produce some interesting results from which to build future studies. For hypothesis one, the perceived strains of workplace gender discrimination did not predict intentions to steal. This may indicate that, at least in terms of theft by employees, the strain of gender discrimination and the corresponding emotional responses are not significant predictors of intentions to steal as posited by General Strain and other theories. Indeed, none of the models explained more than 33% for males and 19% for females. An interesting note, however, is that the explanatory power of the models was greater for males than for females.
In reference to non-significant results of hypothesis two, while gender discrimination in general is certainly seen as unjust, it may not be perceived with these respondents as high enough in magnitude or "recency" to elicit a negative emotional state that leads to a theft response. It may also be that the respondents have had limited exposure to this type of strain. Additionally, though unpredicted, males tended to show increases in perceived anger and depression while females tended to show decreases. This result may foster questions about socialization in the workplace and should be explored in further research.

     Hypothesis three was not supported for objective strains and intentions to steal. But contrary to other literature (e.g., Hoffman and Cerbone 1999; Hoffman and Su 1997; Mazerolle 1998; Mazerolle and Piquero 1998) this study did find significant differences between males and females when examining psychosocial stressors separately. These results illustrated that differences in the way strains are perceived between males and females depended on personal traits and the specific stressors within perceived "gender discrimination" for the college-age sample. This clearly demonstrates the need for separating strain types, examining gender differences, and examining gender discrimination as applied to males.

     Results pertaining directly to theory were mixed relative to current literature. For example, contrary to studies using combined measures, the importance of using separate measures was highlighted by some significant differences between males and females based on the type of strain. Males indicated the most significant results with the pay gap scenario. This is consistent with what Gasser, Flint and Tan (2000) and Wood and Lindorff (2001) would suggest, that males tend to expect their goals of promotion and advancement to be met in the workplace. As noted by Broidy and Agnew (1997: 279), "males tend to be focused on material success and extrinsic achievements, whereas females are more concerned with the establishment and maintenance of close relationships." An interesting pattern in terms of theory were the non-significant results for perceived objective strains and negative emotions for both males and females. Consistent with GST, these results, while seemingly unremarkable, suggested males and females do indeed experience perceived strain similarly.

Males and females did, however, differ on which demographic and psychosocial factors were significant contributors to their perception. Race (White) and commitment to the workplace were significant predictors for females, while impulsivity was the most frequent significant predictor for males. In addition, males and females differed on the type of perceived strain on which the control variables had the most effect. Unsurprisingly, females showed the most significant results with the sexual harassment scenario and males showed the most significant results with the "pay gap" scenario.
In sum, general indicators of crime potential, such as impulsivity and commitment, were significantly better predictors of intentions to steal than were perceived objective strains or negative affective states. The higher the indicated amount of impulsivity, the greater the likelihood theft would be perceived as justifiable. Conversely, the higher the indicated amount of commitment to the workplace, the lesser the likelihood theft would be perceived as justifiable. While the general patterns were the same for males and females, the two groups differed significantly on these two factors. This indicates the strength of social and psychosocial factors not only outweigh the effect of strain on the decision to steal, but also indicates they differ by gender.

Future Research/Policy Implications

    This study contributes to the literature in two major ways: 1) by suggesting the lumping of strains into "negative life events" or some other combined category may not be appropriate for a major strain like perceived discrimination, and 2) the separation of male and female respondents with group comparisons may yield interesting results in other gender discrimination studies.    Toward future research, a sample with an older mean age and a greater variety of job types, particularly in the blue-collar sector, may provide more substantial results.

    The results of the study suggest perceived discrimination is not a significant predictor of intentions to steal. Consistent with current literature, however, impulsivity and commitment were significant. Therefore, employers and businesses should continue the use of pre-screening inventories, or begin if not currently doing so, that screen for high impulsivity (low self-control) and commitment to the workplace in an effort to limit the potential for employee theft.

    Additionally, businesses should implement and maintain strategies that encourage growth and commitment to the organization through pay and promotional opportunities, as this seems to have a positive effect on reducing workplace deviance.

References

Agnew, Robert. 1992. "Foundation for a General Strain Theory of Crime and Delinquency." Criminology 30(1): 47-87.

Agnew, Robert. 2001. "Building on the Foundation of General Strain Theory: Specifying The Types of Strain Most Likely to Lead to Crime and Delinquency. Journal of Research in Crime and Delinquency 38: 319–361.

Agnew, Robert. 2006. Pressured Into Crime: An Overview of General Strain Theory. Los Angeles, CA: Roxbury.

Agnew, Robert, Timothy Brezina, and John Paul Wright. 2002. "Strain, Personality Traits, and Delinquency: Extending General Strain Theory." Criminology an Interdisciplinary Journal 40(1): 43-72.

Agnew, Robert and Helene Raskin White. 1992. "An Empirical Test of General Strain Theory." Criminology 30(4): 475-499.

Arneklev, Bruce J., Harold G. Grasmick, and Robert J. Bursik, Jr. 1999. "Evaluating the Dimensionality and Invariance of 'Low Self-Control'." Journal of Quantitative Criminology 15(3): 307-331.

Azjen, Icek. 2001. "Nature and Operation of Attitudes." Annual Reviews Psychology 52: 27-58.

Azjen, Icek and Martin Fishbein. 1977. "Attitude-Behavior Relations: A Theoretical Analysis and Review of Empirical Research." Psychological Bulletin 84(5): 888-918.

Barak, Gregg, Paul Leighton, and Jeanne Flavin. 2007. Class, Race, Gender, and Crime: The Social Realities of Justice in America. Lanham, MD: Rowman and Littlefield Publishers, Ltd.

Beeghley, Leonard. 1996. The Structure of Social Stratification in the United States. 2nd ed. Boston, MA: Allyn and Bacon.

Bolin, Aaron and Linette Heatherly. 2001. "Predictors of Employee Deviance: The Relationship Between Bad Attitudes and Bad Behavior." Journal of Business and Psychology 15(3): 405-418.

Brezina, Timothy. 1996. "Adapting to Strain: An Examination of Delinquent Coping Responses." Criminology 34: 39–60.

Broidy, Lisa and Robert Agnew. 1997. "Gender and Crime: A General Strain Theory Perspective." Journal of Research in Crime and Delinquency 34(3): 275-306.

Broidy, Lisa. 2001. "A Test of General Strain Theory." Criminology 39(1): 9-35.

Chatman, Jennifer A., Jeffrey Polzer, Sigal Barsade, and Margaret Neale. 1998. "Being Different Yet Feeling Similar: The Influence of Demographic Composition and Organizational Culture in Work Processes and Outcomes." Administrative Science Quarterly 43(4): 749-780.

Dingwell, Heath. 2001. "Criminal Reactions to Inequity: A Test of General Strain Theory." Ph. D. Dissertation, Washington State University. UMI 3023574 Proquest Information and Learning Company.

Dye, Thomas. 2008. Understanding Public Policy. 12th ed. Upper Saddle River, NJ: Pearson Publishing.

Eitle, David J. 2002. Exploring a Source of Deviance-Producing Strain for Females Perceived Discrimination and General Strain Theory." Journal of Criminal Justice 30(5): 429-442.

Finch, J. 1987. "The Vignette Technique in Survey Research." Sociology 21: 105-114.

Fitzgerald, Louise F. and S. L. Shullman. 1993. "Sexual Harassment: A Research Analysis and Agenda for the 1990's." Journal of Vocational Behavior 42: 5-27.

Fitzgerald, Louise F., Charles L. Hulin, and Fritz Drasgow. 1995. "The Antecedents and Consequences of Sexual Harassment in Organizations." In Job Stress In A Changing Workforce: Investigating Gender, Diversity, and Family Issues. Pp. 55-73. Edited by G. Keita and J. Hurrell Jr., Washington, DC: American Psychological Association.

Gasser, Michael, Nichole Flint, and Rowena Tan. 2000. "Reward Expectations: The Influence of Race, Gender, and Type of Job." Journal of Business and Psychology 15(2): 321-329.

Glomb, Theresa M., Liberty J. Munson, Charles L. Hulin, Mindy E. Bergman, and Fritz Drasgow 1999. "Structural Equation Models of Sexual Harassment: Longitudinal Explorations and Cross-Sectional Generalizations." Journal of Applied Psychology 84(1): 14-28.

Gottfredson, Michael and Travis Hirschi. 1990. A General Theory of Crime, Palo Alto, CA: Stanford University Press.

Grasmick, Harold G., Charles Tittle, R. J. Bursik Jr., and Bruce Arneklev. 1993." Testing the Core Empirical Implications of Gottfredson and Hirschi's General Theory of Crime." Journal of Research, Crime, and Delinquency 30: 5–29.

Greenberg, Jerald. 1990. "Theft by employees as a Reaction to Underpayment Inequity: The Hidden Cost of Pay Cuts." Journal of Applied Psychology 75(5): 561-568.

Greenberg, Jerald. 2002. "Who Stole the Money, and When? Individual and Situational Determinants of Theft by Employees." Organizational Behavior and Human Decision Processes 89: 985–1003.

Hayes, Read. 2008. "Strategies to Detect and Prevent Workplace Dishonesty." CRISP Report, ASIS International,
(http://www.asisonline.org/foundation
/noframe/research/library/
ASIS_Workplace_Dishonesty_crsprpt.pdf. (accessed March 9, 2009).

Hayes International Report. June 2013."Theft Surveys", http://www.hayesinternational.com/
thft_srvys.html. (accessed November 29, 2013).

Hoffmann, John P. and Felice Gray Cerbone. 1999. "Stressful life Events and Delinquency Escalation in Early Adolescence." Criminology 37: 343–373.

Hoffmann, John P. and Susan Su. 1997. "The Conditional Effects of Stress On Delinquency And Drug Use: A Strain Theory Assessment Of Sex Differences." Journal of Research in Crime and Delinquency 34(1): 46-78.

Hollinger, Richard. 2011. "2011 National Retail Survey." University of Florida, http://www.clas.ufl.edu/users/
rhollin/srp/nrssfinalreport2011.pdf . (accessed November 29, 2013).

Hollinger, Richard and James Clark. 1983. Theft By Employees. Lexington, MA: Lexington Books.

Huiras, Jessica, Christopher Uggen, and Barbara McMorris. 2000. "Career Jobs, Survival Jobs, and Employee Deviance: A Social Investment Model of Workplace Misconduct." The Sociological Quarterly 41(2): 245-263.

Jasso, Guillermina. 2006. "Factorial Survey Methods for Studying Beliefs and Judgments." Sociological Methods and Research 34(3): 334-423.

Kemp, Alice A., and E. M. Beck. 1986. "Equal Work, Unequal Pay: Gender Stratification within Work-Similar Occupations." Work and Occupations 13(3): 324-347.

Kerbo, Harold R. 2006. Social Stratification and Inequality. 6th Ed. Boston, MA: McGraw Hill.

Langton, Lynn. 2004. "Can General Strain Theory Explain White-Collar Crime? A Preliminary Investigation." Master's Thesis, University Of Florida.

Langton, Lynn, Nicole Leeper-Piquero, and Richard C. Hollinger. 2006. "An Empirical Test of the Relationship Between Theft by employees and Low Self-Control." Deviant Behavior 27(5): 537-565.

Lee, Keun S. and Tao Gao. 2005. "Studying Organizational Commitment with the OCQ in the Korean Retail Context: Its Dimensionality and Relationships with Satisfaction and Work Outcomes." International Review of Retail, Distribution and Consumer Research 15(4):375-99.

Leeper -Piquero, Nicole and Miriam D. Sealock. 2004 "Gender and General Strain Theory: A Preliminary Test of Broidy and Agnew's Gender/GST Hypotheses." Justice Quarterly 21: 1125-158.

Mazerolle, Paul. 1998. "Gender, General Strain, and Delinquency: An Empirical Examination." Justice Quarterly 155-91.

Mazerolle, Paul and Alex R. Piquero. 1997. "Violent Responses to Strain: An Examination of Conditioning Influences." Violence and Victims 12: 323–43.

Mazerolle, Paul and Alex R. Piquero. 1998. "Linking Exposure To Strain With Anger: An Investigation of Deviant Adaptations." Journal of Criminal Justice 26(3): 195–211.

Morgan, 2006. "The Role of Negative Emotion in General Strain Theory." Ph.D. Dissertation, Emory University. UMI 3234040, Proquest Information and Learning Company.

Mowday, Richard T., Richard M. Steers, and Lyman W. Porter. 1979. "The Measurement of Organizational Commitment." Journal of Vocational Behavior 14(2): 224–247.

Mustaine, Elizabeth and Richard Tewksbury. 2002. "Workplace Theft: An Analysis of Student-Employee Offenders and Job Attributes." American Journal of Criminal Justice 27(1): 111-127.

ODU statistics, http://www.odu.edu/ao/ira/
Fall10Updates/EnrDemo/enrolfilefall10.html
(accessed July 6, 2011).

Rebellon, Cesar J., Nicole Leeper-Piquero, Alex R. Piquero, and Sherod Thaxton. 2009. "Do Frustrated Economic Expectations and Objective Economic Inequity Promote Crime?: A Randomized Experiment Testing Agnew's General Strain Theory." European Journal of Criminology (6): 47-71.

Sampson, Robert J. and John H. Laub. 1997. Crime in the Making: Pathways and Turning Points Through Life. Cambridge, MA: Harvard University Press.

Schoenberg, Nancy E. and Hege Ravdal. 2000. "Using Vignettes in Awareness and Attitudinal Research." International Journal of Social Research Methodology 3(1): 63-74.

Slocum Lee Ann, Sally S. Simpson, and Douglas A. Smith. 2005. "Strained Lives and Crime: Examining Intra-Individual Variation in Strain and Offending In A Sample Of Incarcerated Women." Criminology 43(4): 1067-1110.

Swanson, Naomi G. 2000. "Working Women and Stress." JAMWA 55(2): 76-79.

Vazsonyi, Alexander T., Lloyd E. Pickering, Marianne Junger, and Dick Hessing. 2001. "An Empirical Test Of A General Theory Of Crime: A Four-Nation Comparative Study Of Self-Control And The Prediction Of Deviance." Journal of Research In Crime and Delinquency 38: 91-131.

Vaux, Alan, Jeffrey Phillips, Lori Holly, Brian Thomson, Deirdre Williams, and Doreen Stewart. 1986. "The Social Support Appraisals (SS-A) Scale: Studies of Reliability and Validity." American Journal of Community Psychology 14: 195-219.

Wood, Glenice and Margaret Lindorff. 2001. "Sex Differences in Explanations for Career Progress." Women in Management Review 16(4): 152-162.

A Member of the EBSCO Publishing Group
Abstracted in Sociological Abstracts
Online Indexing and Article Search from the
Directory of Open Access Journals (DOAJ)

Return to Home Page to Read More Articles

Sociation Today is optimized for the Firefox Browser

The Editorial Board of Sociation Today

Editorial Board:
Editor:
George H. Conklin,
 North Carolina
 Central University
 Emeritus

Robert Wortham,
 Associate Editor,
 North Carolina
 Central University

Board:
Rebecca Adams,
 UNC-Greensboro

Bob Davis,
 North Carolina
 Agricultural and
 Technical State
 University

Catherine Harris,
 Wake Forest
 University

Ella Keller,
 Fayetteville
 State University

Ken Land,
 Duke University

Steve McNamee,
 UNC-Wilmington

Miles Simpson,
 North Carolina
 Central University

William Smith,
 N.C. State University