Mobile Edition
Print
Get the Mag
Weekly UpdatesResults
Descriptive statistics for the psychological contract scores in each of the four time periods are presented in Table 1. Because of missing data, all hypotheses were tested using the general linear modeling procedure for unbalanced ANOVA's. The results for Hypothesis 1 indicate a main effect for time on the relational component scores (F 71.76, df = 3, 770, p [less than] .001). These results are presented in Table 2. To examine the nature of this effect, pairwise comparisons were analyzed using Scheffe's Test (Scheffe, 1953). Significant differences were found between: (a) psychological contract scores for the 1950s and 1960s (M = 3.79) and the 1970s and 1980s (M = 3.37) and (b) between the psychological contract assessments for the 1970s and 1980s (M = 3.37) and the 1990s (M = 3.03). The magnitude of the difference between the 1990s and the Year 2000 and beyond (M = 2.93) was not significant. Thus, Hypothesis 1 generally was supported.
The second hypothesis predicted a significant interaction effect of time and generation on assessments of the relational component of the psychological contract. The interaction effect of time and generation was not significant (F = 0.81, df = 6, 770, n.s.; see Table 2). In other words, subjects viewed the psychological contract the same across time regardless of their generational membership. Thus, Hypothesis 2 was not supported.
The expected interaction of time and employment status on the relational component was examined in the third hypothesis. Results of the analysis of variance demonstrated a significant interaction (Time x Employment Status; F = 3.37, df = 3, 469, p [less than] .05). The results of this analysis are presented in Table 3. An examination of the means indicated, as predicted, that full-time employees perceived greater decreases in the relational component of the psychological contract over time than did part-time employees (see Figure I).
Given that a subject's employment status may have been influenced by his or her gender or age, a series of chi-square tests were conducted. Results showed no effect of gender on employment status ([X.sup.2] = 0.01, df = 1,120, n.s.). However, there was a significant relationship between age and employment status, such that younger workers were more likely to be employed part-time ([X.sup.2] = 55.73, df = 3,120, p [less than] .05). Additional analyses were conducted to identify whether age had confounded the employment status by time interaction or whether employment status alone had influenced the psychological contract scores over time. Results of these ANOVA's showed that age did not have a significant main effect on psychological contract scores (F = 1.79, df 3,453, n.s.), nor did it have a significant interaction with time on contract scores (F = 1.19, df = 9,453, n.s.). Overall, the above analyses support Hypothesis 3.
Discussion
The results of this study clearly reveal that perceptions of the relational component of the psychological contract have changed over time. The mean relational score was 3.79 for the 1950s and 1960s, 3.37 for the 1970s and 1980s, 3.03 for the 1990s, and 2.93 for the Year 2000 and Beyond. No differences in scores were found across the three generations of participants (students, parents, grandparents). The findings also suggest that part-time employees view the psychological contract differently than their full-time counterparts, in that full-time employees perceived a greater decrease in the relational component over time.
Although participants were not asked to make explicit value judgements with regard to changes in the psychological contract (i.e, whether the changes were good or bad), the relational component scores showed a significant decrease over time. This decrease suggests that the current relationship is characterized by lower levels of trust, support, respect, loyalty, and commitment. It seems reasonable to infer that the absence of these elements may be troubling for some employees. These results are consistent with research findings that suggest a changing and dynamic relationship between employees and their individual organizations (Cavanaugh and Noe, 1999; Robinson and Rousseau, 1994; Robinson et al., 1994; Rousseau, 1995).
The findings indicate that participants perceived a substantially greater decrease in the psychological contract during the earlier two time frames. The mean relational score decreased 11% between the 1950s and 1960s and the 1970s and 1980s, and decreased by additional 10% between the 1970s and 1980s and the 1990s. Analyses of the main effect of time on these scores showed that the only pairwise comparison which failed to reach significance was between perceptions of the psychological contract in the 1990s and the one that is anticipated in the Year 2000 and Beyond. The mean score decreased by only 3% between these latter time periods. A possible explanation for this finding is that the gap between these two time periods (about 10 years in total) is much shorter than the elapsed time between the initial two time periods (two decades each). Respondents may have perceived that there was insufficient time for the psychological contract to undergo dramatic changes.
Another explanation for the less dramatic decline in contract scores in the latter time periods could be related to differences in perceived intent. As previously stated, much of the research examining the psychological contract has focused on perceived contract violations. This literature has distinguished between different types of violation (Morrison and Robinson, 1997; Rousseau, 1995). Morrison and Robinson (1997) identified the term reneging to describe a situation where employers break promises with employees because they are unwilling to fulfill them (even though they could). In contrast, disruption is a situation where an employer is willing but unable to honor its commitments with employees (Rousseau, 1995). Obviously, reneging causes more negative affective responses than disruption. It is possible that the emotions associated with perceived reneging led to larger changes in the contract during the earlier time periods. Beginning in the 1980s, organizations engaged in restructuring, re-engineering, and downsizing in an attempt to remain globally competitive (Cascio, 1993, 1998; De Meuse et al., 1994). As the decade unfolded, top management began to communicate more clearly to employees the competitive demands of the marketplace and the financial constraints in which they were operating. Perhaps, the noticeable decline in the contract scores during the 1970s and 1980s was due to perceptions of reneging; whereas, the smaller decline in the 1990s may be perceived as disruption, a less serious violation, due to a better informed work force.
Figure I illustrates a significant shift in perceptions of the psychological contract based on employment status. This shift occurred sometime between the late-1980s and early-1990s when full-time employees began to perceive a greater decline in such elements of the psychological contract as trust, loyalty, and respect than their part-time counterparts. The timing of the shift coincides with an era when downsizing was prominent among many American companies. It may have been a particularly difficult time for employees who had been asked to support empowerment programs such as TQM and self-managed teams. Full-time employees with ingrained expectations of loyalty, commitment, and life-time employment may have viewed downsizing as a more serious violation than part-time workers. Perhaps, the current contract is more in line with the goals of part-time employees (e.g., the transitory nature of employment).
Limitations and Suggestions for Future Research
There are a number of logistical problems that arise when one conducts longitudinal research (e.g., sample attrition, coordination difficulties). To track employee perceptions of the psychological contract over 50 years in a truly longitudinal fashion would have been virtually impossible. Consequently, a retrospective methodology was utilized in this study. The authors recognize that there are problems associated with this approach. A major problem in this study was that students were asked to make inferences with regard to time periods in which they did not live and/or did not possess any firsthand knowledge of the workplace. To ensure that the results for Hypothesis 1 were meaningful, we reran the analysis without the student group to determine if it altered the pattern of results. Again, we found a main effect of time on perceptions of the psychological contract (F = 54.44, df = 3,509, p [less than] .001). We also tested whether student recollections of the first two time periods were homogeneous when com pared to those subjects who actually experienced the time periods. Likewise, the findings indicated no significant differences in how the student and non-student samples recollected the 1950s and 1960s (t = 0.16, n.s.) and the 1970s and 1980s (t = -0.08, n.s.). It also should be noted that the results of the Scheffe's pairwise comparison tests revealed no significant differences in the responses by generation (students, parents, and grandparents) across the four time periods, as indicated by the lack of support for Hypothesis 2. Thus, the overall pattern of findings alleviates some of the concerns pertaining to the use of the retrospective paradigm in this study. Despite the justifiable concerns about the accuracy of retrospective designs, this study supports the contention of researchers who assert that these designs can be useful in identifying patterns indicative of dynamic processes (Howard, 1993; Leonard-Barton, 1990).
A second limitation was that the biographical items were not comprehensive enough to capture the subtleties of employment status which may have affected the results for Hypothesis 3. First, the participants only were asked about their current employment status in the 1990s. This presents a problem because the promises made to part-time and full-time employees may be different, and unless the participant was uniformly part-time or full-time during their entire career, their recollections may have been confounded. Secondly, the items made no distinctions among different types of part-time workers and did not ask the participants if they worked part-time by choice. Due to those shortcomings, we should interpret the findings with caution.
FEED