Uneingeschränkter Zugang

Reaching for Unique Resources: Structural Holes and Specialization in Scientific Collaboration Networks

Michał Bojanowski
Michał Bojanowski
 und 
Dominika Czerniawska
Dominika Czerniawska
   | 30. Juli 2020
Journal of Social Structure's Cover Image
Über diesen Artikel
Vorheriger Artikel
Nächster Artikel
Zitieren
Article Category: research_article
Online veröffentlicht: 30. Juli 2020
Seitenbereich: 1 - 34
DOI: https://doi.org/10.21307/joss-2020-001
Schlüsselwörter
© 2019 Michał Bojanowski; et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Figure 1:

Collaboration as a process of pooling resources.
Collaboration as a process of pooling resources.

Figure 2:

Burt’s redundancy in egocentric networks.
Burt’s redundancy in egocentric networks.

Figure 3:

Implications of structural holes and specialization hypotheses. (A) The more redundant the alters are, the more similar resources they contribute. Ego acquires similar resources from Actors 1, 2, and 3 because they are redundant (B) specialization only takes place in a tightly collaborating group (0-1-2-3). Ego acquires different resources from actors 1, 2, and 3 because they specialize within the group of 0, 1, 2, and 3.
Implications of structural holes and specialization hypotheses. (A) The more redundant the alters are, the more similar resources they contribute. Ego acquires similar resources from Actors 1, 2, and 3 because they are redundant (B) specialization only takes place in a tightly collaborating group (0-1-2-3). Ego acquires different resources from actors 1, 2, and 3 because they specialize within the group of 0, 1, 2, and 3.

Figure 4:

Collecting data on collaboration networks.
Collecting data on collaboration networks.

Figure 5:

Collaboration network.
Collaboration network.

Figure 6:

Network of resource flows.
Network of resource flows.

Figure 7:

Estimates of fixed effects (dots) and Bootstrap confidence intervals (bars) from random intercept models for similarity of resources acquired by ego. Reference category for ego’s degree: PhD. Confidence intervals are based on 1,000 draws. Detailed results can be found in Appendix A1 in Tables A.1 and A.2.
Estimates of fixed effects (dots) and Bootstrap confidence intervals (bars) from random intercept models for similarity of resources acquired by ego. Reference category for ego’s degree: PhD. Confidence intervals are based on 1,000 draws. Detailed results can be found in Appendix A1 in Tables A.1 and A.2.

Figure 8:

Effect plot of pairwise redundancy on resource similarity based on Model 3 assuming average ego-network size, alters from different departments, of different degree, and ego with a PhD degree.
Effect plot of pairwise redundancy on resource similarity based on Model 3 assuming average ego-network size, alters from different departments, of different degree, and ego with a PhD degree.

Basic information about the respondents in the sample

ID City Gender Scientific degree Branch of science Network size (degree)
201 A Male PhD Technical sciences 10
202 A Male PhD Social sciences 7
203 A Male Professor Technical sciences 9
204 A Male PhD Technical sciences 8
205 A Female PhD Life sciences 4
206 A Male PhD Natural sciences 4
207 A Female PhD Social sciences 10
208 A Female PhD Humanities 5
209 B Female Habilitated PhD Life sciences 9
210 B Male PhD Humanities 8
211 B Male Habilitated PhD Life sciences 10
212 B Female PhD Social sciences 10
213 B Male Professor Natural sciences 7
214 B Female PhD student Natural sciences/Engineering 4
215 B Female Habilitated PhD Life sciences 10
216 B Male Habilitated PhD Natural sciences 10
217 B Male Professor Technical sciences 10
218 C Female Habilitated PhD Humanities 10
219 C Male PhD Engineering 10
220 C Male Habilitated PhD Natural sciences 6
221 D Male PhD Engineering 9
222 D Female Habilitated PhD Life sciences 10
223 D Female Habilitated PhD Humanities 7
224 D Male PhD Humanities 10
225 D Male PhD Natural sciences 10
226 D Female Habilitated PhD Natural sciences 10
227 D Male Professor Humanities 9
228 D Male Habilitated PhD Life sciences 7
229 D Female PhD student Social sciences 6
230 D Female PhD student Technical sciences 7
231 E Female Habilitated PhD Social sciences 7
232 E Female PhD Social sciences 11
233 E Female PhD student Technical sciences 6
234 E Female PhD Social sciences 10
235 E Male PhD Natural sciences/Engineering 6
236 E Female Habilitated PhD Natural science 9
237 E Female PhD Natural science 7
238 F Male Professor Technical sciences 15
239 F Male Professor Technical sciences 9
240 D Male PhD student Humanities 7

Pairwise redundancy scores for the example collaboration network.

2 3 4 5 6 7 8 9
1 1 0 2 2 0 0 0 0
2 0 0.5 0.5 0 0 0 0
3 0 0 0 0 0 0
4 2 0 0 0 0
5 0 0 0 0
6 1 0 0
7 0 0
8 0

Random intercept models for alter contributions. Null model: σ U 2 = 0.01 , σ R 2 = 0.053 , AIC=–56.91.

Variable Model 1 Model 2 Model 3
Estimate SE Estimate SE Estimate SE
Random effects
Ego 0.08 – 0.08 – 0.08 –
Residual 0.05 – 0.05 – 0.05 –
Fixed effects
Intercept 0.17 0.08 0.15 0.08 0.07 0.08
Egos degree: PhD hab. 0.06 0.04 0.06 0.04 0.06 0.04
Egos degree: MA 0.03 0.06 0.03 0.06 0.02 0.06
Egos degree: professor 0.07 0.04 0.07 0.04 0.07 0.04
Network size -0.01 0.01 -0.01 0.01 -0.01 0.01
Alters same degree 0.11 0.01 0.10 0.01 0.11 0.01
Alters same department 0.07 0.02 0.05 0.02 0.05 0.02
Pairwise redundancy – – 0.01 0.00 – –
Pairwise redundancy (0<P<1) – – – – 0.12 0.02
Pairwise redundancy (P>1) – – – – -0.01 0.01
Model statistics
AIC -121.51 – -127.58 – -150.47 –
df 9.00 – 10.00 – 11.00 –

Directions of effects of redundancy on similarity of acquired resources under H1 and H2.

Hypothesis Effect of redundancy on similarity of resources acquired by ego
H1: Structural holes +
H2: In-group specialization −

Model comparison statistics for alter contributions.

Model Df AIC Deviance x 2 x 2 df p-value
Null model 3 -56.91438 -62.91438 – – –
Model 1 9 -121.50963 -139.50963 76.595259 6 0.000
Model 2 10 -127.57490 -147.57490 8.065262 1 0.005
Model 3 11 -150.46507 -172.46507 24.890172 1 0.000

Descriptive statistics of variables used in the analysis.

Variable Min. Mean SD Max.
Level: Alter-alter
Alters same degree 0 0.306 0.461 1
Alters same department 0 0.233 0.423 1
Pairwise redundancy 0 1.829 2.408 8
Similarity of alter contributions 0 0.161 0.246 1
Level: Ego
Egos degree: MA 0 0.100 0.304 1
Egos degree: PhD 0 0.400 0.496 1
Egos degree: PhD hab. 0 0.225 0.423 1
Egos degree: professor 0 0.275 0.452 1
Network size 5 9.350 2.293 16
eISSN:
1529-1227
Sprache:
Englisch
Zeitrahmen der Veröffentlichung:
Volume Open
Fachgebiete der Zeitschrift:
Sozialwissenschaften, andere
Zeitschrift RSS Feed