Page 47 - Special Topic Session (STS) - Volume 4
P. 47
STS560 Haniza Yon et al.
Gender Main Effects. Gender DIF can occur regardless of the presence or
absence of any gender main effects on the estimated factor measures. In our
data, the traits for which some items showed DIF were not those with the
largest differences in average scores between men and women. Figure 2 below
shows the factor means for women (solid lines) and men (dotted lines), with
the factors arranged along the X-axis. The factors are ordered according to the
difference (MWomen–MMen) between the female and male means. Rather
surprisingly, women reported greater self-confidence and productivity but less
flexibility and innovation. These results should be interpreted cautiously, as
statistical tests for pairwise group differences (see Table 1) indicate that only
two of the differences reached statistical significance at p < .05.
Figure 2: Factor Means by Gender
55
Scale Score (M=50, SD = 12) 51
54
53
52
50
49
48
47
46
45
Men Women
4. Discussion and Conclusions
All our 15 factors showed acceptable model fit and reliability, indicating
that we succeeded in constructing a set of internally valid scales in accordance
with the one-parameter IRT model. A very small number of misfitting items
was detected, and further research is being conducted to understand the
issues involved.
However, no substantive gender biases were found, and we conclude
therefore that any gender differences in average observed trait scores reflect
genuine trait-level differences, rather than (say) item- or test-related artifacts.
We observed some surprising gender differences: the women in our
sample showed greater self-confidence, customer service orientation,
productivity, execution, and initiative than did the men. Conversely, men
showed greater flexibility, innovation, problem-solving and resourcefulness,
and entrepreneurship. These findings are not what one might expect in a
male-dominated business culture. Admittedly, these differences were not very
36 | I S I W S C 2 0 1 9