This chapter presents the findings of the study. It reports the results of the various statistical analysis procedures used to test the proposed hypotheses. It also provides an explanation and a brief discussion of each result. In addition the chapter contains the findings obtained from a series of Post-hoc interviews conducted in order to gain a deeper understanding of the unexpected results of the quantitative inquiry.
Descriptive Statistics
Table 4.1 presents the results of the descriptive analysis. Table 4.2 presents the means and standard deviations of the major variables of the study.
Table 4.1
Results of Descriptive Statistics
Note. EC $1 is equivalent US $0.37
Feature Categories Number (%)
Treatment Questionnaire A 67 33
Questionnaire B 68 34
Average monthly salary of main breadwinner in EC dollars
Under $1,000 29 14
$1,001 - $3,000 144 71
$3,001 - $5,000 23 11
Over $ 5,000 7 3
54 Table 4.2
Results of Descriptive Statistics (means and standard deviations)
Note: Work experience and socioeconomic status were represented by four categories.
Religious commitment, work values, misuse of resources and misuse of information were measured using a 5-point likert scale. Social desirability was scored from 0 to 10.
There are several noteworthy results from the descriptive analysis. As indicated in table 4.1 most of the respondents were female. This is a true reflection of the population since there are more female students enrolled in tertiary level institutions than male (stats.gov.lc). The average amount of work experience among participants is approximately two (2) years. More than half (144) of the participants are from lower middle income families whereas only seven (7) of them are from high income families. The level of religious commitment among members of the sample is relatively low (mean 2.78 std.
1.206). The work values which are ranked highest among this population are security, economic returns and achievement. The work value which ranked the lowest is associates.
Responses suggest that respondents were generally unlikely engage in misuse of resources Variable N Minimum Maximum Mean Std. Deviation
Work Experience 203 1 4 2.38 1.117
55
(mean 2.70 std .893) and very unlikely to engage in misuse of information (mean 1.77 std .751).
Correlational Analysis
The results of correlational analysis of all variables are reported in table 4.3. There are several variables that display significant correlations. The demographic variables had the following significant correlations. Gender has a negative significant correlation to economic returns and achievement. This suggests that within the sample and possibly the population, women have higher value for economic returns and achievement than men. Work experience shows a positive significant relationship to social desirability and security.
Therefore suggesting that as the amount of work experience a respondent acquires increases so does his or her value for job security. This may be due to the fact that people with more work experience are most likely on the higher end of the sample age range and may have more responsibility (family and/or financial commitments) and so they place greater value on security. Religious commitment is negatively correlated with the work value of independence.
The control variable, social desirability, is negatively correlated to both misuse of resources and misuse of information. This suggests that participants with a high level of social desirability are less likely to admit to the fact that they would engage in CWB. This is not a surprising result since most forms of CWB are considered socially undesirable and so respondents are likely to under report their tendency or propensity to engage in those behaviors.
Work values also showed some interesting correlations. Economic returns positively correlates with all other work values. Independence is positively related to security and achievement. Associates is positively related to security and achievement is negatively correlated to both misuse of resources and misuse of information. This suggests that a
56
person who values achievement is less likely to engage in both types of CWB. The only component of group norms which appears to be related to CWB is amount, which shows a significant negative correlation to misuse of resources.
57 Table 4.3
Results of Correlational Analysis
Variable 1 2 3 4 5 6 7 8 9 10 11 12 13 14
1 Gender
2 Work Experience -.045
3 Socioeconomic status .090 .119
4 Religious commitment -.072 .078 .008
5 Social Desirability -.057 .079 -.017 -.023
6 Economic Returns -.207** .140* -.068 -.085 -.126
7 Independence -.124 .085 .010 -.162* -.126 .385**
8 Associates .069 -.033 -.053 .059 .092 .150* .110
9 Security -.087 .148* .051 .079 -.041 .639** .273** .153*
10 Achievement -.191** .060 -.051 -.003 .001 .438** .248** .099 .463**
11 Consistency (group norms) .012 .018 .005 -.166* -.011 .103 .038 .034 .131 .146*
12 Amount (group norms) -.003 .078 .056 .154* .164* -.051 -.090 -.007 -.041 -.082 -.498**
13 Misuse of Resources .004 .043 -.009 -.003 -.243** .038 -.094 .069 .106 -.158* .033 -.189**
14 Misuse of Information .034 .033 -.047 -.085 -.215** .004 .005 -.104 -.099 -.207** -.093 -.120 .427**
** Correlation is significant at the 0.01 level (2-tailed). * Correlation is significant at the 0.05 level (2-tailed).
58
Hypotheses Testing Effect of Demographics on CWB
The hypotheses (1a-1d) dealing with the effect of demographics on CWB were tested using T-test procedures (hypothesis 1a), linear regression (1b) and one-way ANOVA (1c and 1d).
Table 4.4 presents the results of the independent sample T-test which compared male and female respondents’ propensity to engage in both types of CWB. Results indicate that there is no significant difference between males and females with regard to the propensity to engage in both misuse of resources and misuse of information. This conclusion is also supported by the results of the correlational analysis (table 4.3) as well as hierarchical regression (tables 4.9 and 4.10). Therefore it can be concluded that hypothesis 1a is not supported.
Table 4.4
Results of Independent Sample T-test (gender)
Gender
N Mean Std. Deviation
t-test Sig
Misuse of Resources Female 136 2.70 .915 0.056 0.956
Male 67 2.70 .852
Misuse of Information Female 136 1.75 .756 0.476 0.635
Male 67 1.80 .745
Results of linear regression (table 4.5) show that religious commitment has no significant influence on both misuse of resources or misuse of information. This is also evident in the results of the hierarchical regression which are presented in tables 4.9 and 4.10. Therefore, the findings clearly indicate that Hypothesis 1b is not supported.
59 Table 4.5
Results of Linear Regression (religious commitment)
Variable Standardized Coefficients (Beta)
Misuse of Resources Misuse of Information
Religious Commitment -.003 -.085
R2 .000 .007
Adjusted R2 -.005 .002
***P<0.001, **P<0.01, *P<0.05
Results of one-way ANOVA (tables 4.6) indicate that there are no significant
differences between various levels of work experience and socioeconomic status regarding the propensity to engage in misuse of resources and misuse of information. These results are also supported by correlational analysis (table 4.3) and hierarchical regression (tables 4.9 and 4.10). Therefore it can also be concluded that hypotheses 1c and 1b are not supported.
60 Table 4.6
Results of One-Way ANOVA (work experience and socioeconomic status)
Sum of Squares df Mean Square F Sig.
Effect of Group Norms on CWB
The differences between the different factors of group norms (amount in group norms, consistency in group norms and no group norms [control group]) with regard to the propensity to engage in both misuse of resources and misuse of information was tested using one-way ANOVA. Table 4.7 presents the results of this analysis. The results indicate that there are significant differences between the groups namely; amount in group norms, consistency in group norms and no group norms (control group) for both misuse of resources and misuse of information.
Table 4.8 indicates where these between group differences lie. For misuse of resources,
61
there is a significant negative difference between amount in group norms (1) and the control group (3). This means that respondents exposed to the vignettes portraying amount in group norms reported a significantly lower propensity to engage in misuse of resources than the respondents who were exposed to vignettes portraying no group norms. Therefore respondents in the control group reported a significantly higher propensity to engage in misuse of resources. This result is the opposite of what was hypothesized, it seems that when faced with a situation where some members frequently engage in CWB (amount in group norms), respondents did not perceive there to be a group norm that supports this behavior and so were unlikely to engage. Moreover, it seems that respondents in the control group (not faced with descriptions of group norms) simply admitted to the fact that they were likely to engage in the listed behaviors. It is also important to note that there was no significant difference between amount in group norms and consistency in group norms for misuse of resources. Although it is not significant, the mean difference between consistency in group norms and amount in group norms is positive (.281), suggesting that the respondents who were exposed to consistency in group norms had a higher propensity to engage in misuse of resources then those exposed to amount in group norms.
For misuse of information, both amount in group norms and consistency in group norms showed a significant negative difference when compared to the control group. This means that respondents who were exposed to the two factors of group norms reported a lower propensity to engage in misuse of information than the respondents who were not exposed to any factor of group norms. Again, this is the opposite of what is hypothesized. It appears that when faced with group norms respondents are less likely to engage in misuse of information than those who were not presented with a group norm. It appears that both hypotheses 2a and 2b are not supported.
62 Table 4.7
Results of One-Way ANOVA Comparing Factors of Group Norms Sum of
Results of Scheffe’s Multiple Comparison Test for Inner Group Differences
Dependent Variable
Note. 1 =Amount in group norms, 2 = Consistency in group norms, 3 = no group norms
* P < 0.05
63
Effect of work values and moderating effect of group norms on CWB
Hypotheses 3 and 4 were tested using Hierarchical regression. The results (tables 4.9 and 4.10) of hierarchical regression reconfirm the fact that the tested demographic variables appear to have no influence on both categories of CWB. Social desirability shows a significant negative influence on both categories of CWB. This result is not unexpected and suggests that respondents may have underreported their propensity to engage in CWB because of the social undesirability of these behaviors especially those listed under misuse of information.
The results support a direct relationship between work values and both categories of CWB. The work values which appear to have the strongest influence on misuse of resources are independence, security and achievement. Independence and achievement both negatively influence misuse of resources. These results support the hypothesized relationships and suggest that workers who value independence and achievement will be more unlikely to engage in misuse of resources. An unexpected result is the fact that security has a positive significant relationship to misuse of resources. It was hypothesized that this work value would be negatively related to CWB, however, the results indicate otherwise. The work value that has the strongest influence on misuse of information is achievement. It is shows a significant negative relationship and as a result, supports the proposed hypothesis.
The hypothesized moderated relationship between work values and group norms, on the other hand, received little support. None of the interaction terms had a significant influence on either categories of CWB. This suggests that both work values and group norms have a direct influence on CWB.
Hierarchical regression results also point to the influence of group norms on both forms of CWB. Although the influence is the opposite of what was hypothesized, in that
64
both of the group norms show negative influence on both categories of CWB. Amount in group norms shows a significant negative relationship to misuse of resources, moreover, its negative influence is much greater than that of consistency in group norms. This suggests that a person who observers only some members of a group frequently engaging in misuse of resources is more unlikely to engage in that behavior than a person who observes all the members of a group occasionally engaging in misuse of resources.
Both types of group norms are negatively related to misuse of information. This could be because misuse of information is influenced by factors other than behavior observed within the work group. Factors such as informational and procedural justice (Jones, 2009), attraction to group, norm strength as well as the leadership style of immediate supervisors (Ehrhart & Naumann, 2004) may exert more influence on misuse of information. For example, just observing colleagues lying to supervisors and withholding information for their own purposes might not be considered sufficient justification for a person to imitate those behaviors. However, if the person feels that procedures dealing with the dissemination of information is unjust or if the supervisor has an authoritarian style of leadership then he/she might believe that engaging in misuse of information is justified.
Altogether the independent variables of this study explain 14-15% (R2 model 4
= .139 and model 5 =. 148) of the propensity to engage in misuse of resources and 8% (R2
=.083) of the propensity to engage in misuse of information. Although both of these can be considered a very low percentage, there is still value in the findings. As far as misuse of resources is concerned, the effects of these variables are statistically significant (0.001) within that sample. This means that these variables do influence this form of CWB and attention should be paid (by managers, supervisors and researchers) to them. Also the results suggest that there are other factors and combinations of factors that have even greater influence on misuse of resources and misuse of information. The fact that these
65
variables do not significantly influence misuse of information, points to the fact that these behaviors are influenced by a different set of variables and that the antecedents of the various forms of CWB are not the same.
66 Table 4.9
Results of Hierarchical Regression for Misuse of Resources
Misuse of Resources
Group Norms (Consistency) -.065 -.073 -.077
Group Norms (Amount) -.192* -.208*** -.217***
Work Values
Economic Returns -.028 -.323
Independence -.172* -.030
Associates .112 .155
Security .250*** .409***
Achievement -.255*** -.103
Interaction Terms
Economic Returns x Consistency .127
Economic Returns x Amount .212
Independence x Consistency -.068
Independence x Amount -.176
Associates x Consistency .006
Associates x Amount -.087
Security x Consistency -.002
Security x Amount -.170
Achievement x Consistency -.110
Achievement x Amount -.159
R2 .002 .064 .091 .190 .241
Adjusted R2 -.018 .040 .058 .139 .148
∆R2 .062 .027 .100*** .050
F-value .105 2.683** 2.779** 3.725*** 2.597***
***P<0.001, **P<0.01, *P<0.05
67 Table 4.10
Results of Hierarchical Regression for Misuse of Information
Misuse of Information
Group Norms (Consistency) -.201** -.171* -.193*
Group Norms (Amount) -.176* -.183* -.190*
Work Values
Economic Returns .114 -.113
Independence -.014 -.107
Associates -.064 -.035
Security -.052 .164
Achievement -.219*** -.188
Interaction Terms
Economic Returns x Consistency .050
Economic Returns x Amount .243
Independence x Consistency .124
Independence x Amount .021
Associates x Consistency -.017
Associates x Amount -.010
Security x Consistency -.033
Security x Amount -.214
Achievement x Consistency .020
Achievement x Amount -.049
R2 .013 .062 .097 .146 .183
Adjusted R2 -.007 .038 .064 .092 .083
∆R2 .049 .035* .050 .037
F-value .635 2.585* 2.982*** 2.712*** 1.831**
***P<0.001, **P<0.01, *P<0.05
68
Table 4.11 provides an overview of the results of hypothesis testing. Overall, findings indicate that all but one of the proposed hypotheses was supported fully supported.
Table 4.11
Overview of the Results of Hypotheses Testing
Hypothesis Results
Hypothesis 1a: Males will generally display a higher propensity to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 1b: Employees with high religious commitment will have a lower propensity to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 1c: Work experience in negatively related to the propensity to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 1d: Employees of high socioeconomic status will have a lower propensity to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 2a: Consistency in CWB group norms will be positively related to the propensity of an individual to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 2b: Amount of CWB group norms will be positively related to the propensity of an individual to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 3a: A high value for economic returns is positively related to the propensity of an individual to engage in both misuse of resources and misuse of information
Not supported
(continued)
69 Table 4.11 (continued)
Hypothesis Results
Hypothesis 3b: A high value for independence is negatively related to the propensity of an individual to engage in both misuse of
resources and misuse of information
Partially supported
Hypothesis 3c: A high value for associates is positively related to the propensity of an individual to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 3d: A high value for security is negatively related to the propensity of an individual to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 3e: A high value for achievement is negatively related to the propensity of an individual to engage in both misuse of resources and misuse of information
Supported
Hypothesis 4a: A high value for economic returns when moderated by group norms is positively related to the propensity to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 4b: A high value for independence when moderated by group norms is negatively related to the propensity to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 4c: A high value for security when moderated by group norms is negatively related to the propensity to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 4d: A high value for associates when moderated by group norms is positively related to the propensity to engage in both misuse of resources and misuse of information
Not supported
Hypothesis 4e: A high value for achievement when moderated by group norms is negatively related to the propensity to engage in both misuse of resources and misuse of information
Not supported
70
Post- Hoc Interviews
In an attempt to explain the unexpected findings of this research as well as the large number of unsupported hypotheses, a series of post hoc interviews were conducted with questionnaire respondents. The semi-structured interviews were conducted using Facebook chat, Msn messenger or Skype. The following questions were used to guide the interviews:
1. What did you think about the questionnaire?
2. Why did you answer the way you did? What (thought process) guided your response?
3. What other factors, apart from the observed behavior of coworkers, would influence your decision to engage in the behaviors listed in the questionnaire?
4. (After a brief explanation of results) What do you think about the results?
Only seven interviews were conducted because many of the respondents who were contacted were either reluctant to participate or complained that they had filled out the questionnaire so long ago (approximately three to one month) that they couldn’t sufficiently remember the experience to provide useful responses to the first two questions. A copy of the questionnaire was sent to each participant prior to the interview in an effort to help them recall the experience. Table 4.12 provides a brief description of the participants in the order in which they were interviewed.
Table 4.12
Post –Hoc Interview Participants
ID Gender Age Questionnaire answered CH Male 24 Questionnaire A
KJ Male 20 Questionnaire B NH Female 22 Questionnaire B JD Female 24 Questionnaire C IM Female 22 Questionnaire B TA Female 23 Questionnaire C GL Male 19 Questionnaire A
71
Findings
In response to question one, four of the participants said that that questionnaire was not like any they have ever been exposed to before. Two of them said that they thought it was very “interesting” especially the scenarios or vignettes. Another interviewee said that the scenario adequately described some of what he had observed in his own work place. One of the interviewees said that the questionnaire was “tricky” and “hard to figure out” meaning that she had a hard time deducing its purpose. Similar sentiments were expressed by at least three other participants. One of them said that it was hard for her to imagine what she would really do if placed in the situation. Another stated that there was not enough information provided in the vignettes (he referred to it as scenario) to help him make a decision about whether to engage in the listed behaviors. Four of the interviewees said that they found the social desirability scale to be disconcerting, especially since there were only two options i.e.
true or false.
Responses to question one suggest several important issues that may help in explaining some of the unexpected results of the study. First, it is possible that in answering the questionnaire respondents were actively trying to decide what the purpose/motive/objective of the questionnaire was and what kind of responses were expected. Secondly, it is also suggested that the vignettes were not elaborate enough or did not portray the group norms factor in a realistic or relatable way. Finally, it appears that some respondents felt uncomfortable with the social desirability scale.
Responses to question one suggest several important issues that may help in explaining some of the unexpected results of the study. First, it is possible that in answering the questionnaire respondents were actively trying to decide what the purpose/motive/objective of the questionnaire was and what kind of responses were expected. Secondly, it is also suggested that the vignettes were not elaborate enough or did not portray the group norms factor in a realistic or relatable way. Finally, it appears that some respondents felt uncomfortable with the social desirability scale.