CHAPTER FIVE Results
5.2 Regression analysis
Having determined that positive humour data collected during this research loads onto the
PsyCap construct by using a CFA, the next step was to investigate the relationship between
positive humour and each of the current PsyCap elements (hope, optimism, resilience and
self-efficacy) and their associations with work attitude, work performance and workplace ‘fun climate’. Data were analysed using the regression facility from the Statistical Package
for the Social Sciences (SPSS), Nie et al., (1975).
5.2.1 Testing the hypotheses
The hypotheses to be tested are listed in Chapter 4, Table 4.1. Linear regressions were
performed and the results are reported below. All the regression analyses reported in this
each labelled table is a composite of three output tables from SPSS. The significance level of α chosen for this current study is 0.05 (5%).
5.2.2 Linear regressions performed
The newly established variables Work Performance, Work Attitude, Fun Climate and the Supervisors’ Sense of Humour were all used in the linear regression analyses conducted.
Initially the relationship between Positive Humour and Work Performance, and Positive
Humour and Work Attitude was tested. The possible moderating effect of a Fun Climate on
these relationships was then examined as was the possible moderating effect of the Supervisors’ Sense of Humour. (H1.1 - H1.6).
Linear regressions were then used to explore the relationships between Positive Humour and
all the PsyCap components, hope, self-efficacy, optimism and resilience, and with PsyCap
itself (hypotheses H1:7 to H1:11). Regressions were also used to determine the relationships
between PsyCap and Work Performance, and PsyCap and Work Attitude for the collected
data and finally the relationship between PsyCap including Positive Humour and the two
selected workplace outcomes, Performance and Attitude (H1:12 to H1:15).
The Model Summaries, CoefficientsTable and Correlations Table for each PsyCap element
and the PsyCap construct, with PosHum as independent variables, are presented in
Appendices 9 and 10.
5.2.3 Team agreement on ‘Fun Climate’
The concept of a ‘fun’ workplace climate was supported by the CFA in which four items
were tested and, for the data collected, all loaded strongly on the latent variable labelled ‘Fun Climate’. The two hypotheses exploring the moderating effect that a fun climate might have
on workplace performance and attitude were not initially supported. This analysis was
initially conducted using the ‘fun climate’ data of all 290 participants. A subsequent retesting
of the data occurred after an In-group Reliability (rwg) analysis was conducted and data for
the teams for which there was no consensus about that team’s ‘fun climate’ were removed.
The results of the rwganalysis are shown in Appendix 11 in which the teams removed (i.e.
those with less than a 70 per cent agreement of their team having a ‘fun climate’) are
highlighted. Appendix 11 also shows the average score of each team supervisor’s own assessment of the workplace’s ‘fun climate’ alongside an average of the individual averages
within each group.
A simple comparison between these two columns shows that the majority (57 per cent) of the supervisors’ averages exceed the teams’ average scores. In six cases the difference is 1 or
greater (out of a possible 5) indicating that for those groups the supervisor’s perception of the
workplace as having a ‘fun climate’ is ≥ 20 per cent higher than the group’s average of the
individual averages. In three cases this was reversed with the group’s average of the
individual averages being greater than the group’s supervisor average score. However one of
these three was scored by a group eliminated by the rwganalysis effectively leaving only two
teams in which the group’s average of the individual averages was much greater than the
group’s supervisor average score.
A possible explanation for supervisors’ assessing their workplaces as being ‘fun’ places with
a higher score than the average given by their subordinates is the influence of self-report bias.
Research participants often respond to questionnaires in a way that makes them look as good
as possible (Donaldson and Grant-Vallone, 2002). Knowing the current research was
questions relating to the existence of a ‘fun climate’ within their workplaces with answers
they felt were appropriate rather than factual.
This tendency for survey participants to respond in socially desirable ways rather than
providing factual information was also reported by Moorman and Podsakoff (1992).
The pre- and post- rwganalysis scenarios are presented and in both cases the existence of a
fun climate within the workplace had no moderating effect on either work performance or
work attitude.
5.2.4 Analysis overview
Referring to the detailed results shown in Appendix 10, the R2 value for each regression is
given on the graph indicating how much variance in the variable is explained by the model.
The R2 value is a statistical measure of the closeness of the data to the fitted regression line.
The Significance Level (Sig.) for all data and the Pearson Correlation are tabulated for each
of the hypotheses to be tested. The Significance Level (Sig.) for all the PosHum data is less
than 0.05 indicating a strong presumption against the null hypotheses and therefore rejecting
them, supporting the hypotheses. The results were mixed, as shown in the Results Summary
below. The Pearson product-moment correlation coefficient was computed to assess the
relationship between PosHum and the dependent variables (being the construct PsyCap and
its individual elements of hope, self-efficacy, resilience and optimism) plus WA, WP and