5. Results
labelled importance of knowledge, was
A correlation matrix with descriptive
dropped from further analysis due to its
statistics for all variables is provided in
reliability (0.6385) being below the
Table 1 at the end of this paper, with items
acceptable threshold (0.70) suggested by
Q17 and Q18 negatively worded and
Nunnally and Bernstein (1994).
transformed for analysis. This matrix
reveals high correlation between items
To test the hypotheses H
1a
, H
1b
and H
1c
,
Q14-Q18, corresponding to the attitude
the means for the items that comprise the
measure (Fishbein & Ajzen, 1975) and
factors were calculated and analysis of
there was also noteworthy correlation
variance (ANOVA) performed (Table 2).
between items Q8-Q10, those that used
Using an alpha of 0.01 the results indicate
the word satisfaction within the item
a difference between the three groups of
phrase. Interestingly, little or almost no
question structure for the factor attitude
correlation was found for items Q5-Q7,
and satisfaction.
Table 2: ANOVA
Sum of Squaresdf Mean SquareF Sig.
Attitude Between Groups32.399 2 16.199 12.300 .000
Within Groups 114.577 871.317
Total 146.976 89
Satisfaction Between Groups8.727 2 4.364 4.288 .017
Within Groups 22.544 871.018
Total 97.272 89
The posthoc test of Tukey HSD
attitude are reported (Table 3) because
(alpha=0.01) was conducted for pairwise
there was no significant difference
comparison. Only the results for the factor
between the three question structure groups for the factor satisfaction.
Table 3: Tukey HSD Multiple Comparisons and Homogenous Subsets – Attitude
Multiple Comparisons Question (I) Question (J) Mean Difference (I-J) Std. Error Sig.
Tukey HSD Binary Open -1.1596* .29391 .000
Directed -1.3674* .29885 .000
Open Binary 1.1596* .29391 .000
Directed -.2078 .29647 .764
Directed Binary 1.3674* .29885 .000
Open .2078 .29647 .764
Homogenous Subset for alpha = 0.01
Subsets Question Structure N 1 2
Tukey HSDa,b Binary 30 3.0533
Open-ended 31 4.2129
Directed 29 4.4207
Sig. 1.00 .764
Means for groups in homogeneous subsets are displayed a. Uses Harmonic Mean Sample Size = 29.978 b. The group sizes are unequal. The harmonic mean of the group sizes is used. Type I error levels are not guaranteed. • Mean difference significant at the 0.01 level
Bạn đang xem 5. - AN EMPIRICAL STUDY OF THE IMPACT OF QUESTION STRUCTURE ON RECIPIENT ATTITUDE DURING KNOWLEDGE SHARING