11.03.2014 Views

thesis_Daniela Noethen_print final - Jacobs University

thesis_Daniela Noethen_print final - Jacobs University

thesis_Daniela Noethen_print final - Jacobs University

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Intergenerational Knowledge Transfer in Work Teams: A Multilevel Social Network Perspective<br />

ICC(1), we analyzed the distribution of variance for knowledge transfer across the different<br />

levels (see section 3.4.3. for the respective formula); 77% of the variance was due to<br />

differences between dyads “within a person”, 11% of the variance was due to individual<br />

differences within teams, and 12% of the variance was due to mean differences between<br />

teams. According to Bliese (2000), values for ICC(1), in our case equivalent to the amount of<br />

variance located at the two higher levels, normally lie between 0.05 and 0.30, so the values<br />

obtained here are within a normal range. This shows that knowledge transfer was influenced<br />

to a certain degree by characteristics of the source and the team, suggesting nonindependence,<br />

and calling for a multilevel analysis.<br />

Control variables. To rule out alternative influences on the dependent variable, we controlled<br />

for several factors in our analyses. First, as posited by social exchange theory (Blau, 1964;<br />

Molm, 1997) and shown by several empirical studies (Bock et al., 2005; Kankanhalli et al.,<br />

2005; Wasko & Faraj, 2000), knowledge transfer within a dyad should be largely influenced<br />

by reciprocity. Accordingly, the frequency with which the source receives knowledge from<br />

the recipient (knowledge reception) should influence the frequency of knowledge transfer<br />

from source to recipient.<br />

Knowledge reception was measured in equivalence to knowledge transfer by averaging<br />

the source’s rating concerning the frequency of seeking knowledge from the recipient, and the<br />

recipient’s rating concerning the frequency of sharing knowledge with the source. Thus, we<br />

again averaged two subjective ratings to receive a more objective measure. The correlation<br />

was, as before, 0.47. The variance of the resulting variable was distributed across the three<br />

levels in the following manner: 83% of the variance was located at the dyadic level, 5% at the<br />

individual level, and 12% at the team level. According to the sociometric approach described<br />

to measure knowledge transfer and knowledge reception, the data of each dyad A-B appear<br />

105

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!