Geography Reference
In-Depth Information
28.730. We also i nd an increase in the centralization of the network from 0.602 to 0.717,
which means that peripheral actors in the network are more strongly connected to actors
in the centre and/or less connected to other peripheral actors. The overall clustering coef-
i cient slightly decreases from 1.228 in the i rst period to 1.200 in the second period. At
the same time, the average distance between actors decreases from 2.003 to 1.879.
Summarizing, we i nd increasing cohesion, which is interpreted as a stronger focus
on core competencies or technologies, where the i elds of activity of the central actors
become increasingly important for the whole network. With respect to the increased
connectedness of the
pr
-network we presume that the increasing potential to cooperate
goes hand in hand with this change over time. Before we go on to analyse this relation in
more detail, let us dwell a little bit more on the development of technological overlap and
consider actors entering and exiting the network.
As we recall from Table 17.1, the number of actors changes between P1 and P2,
with only 32 actors being part of the network in both periods. We thus ask whether the
observed increasing technological cohesion can be related to the following three groups
of actors - entering, exiting, and permanent innovators. Calculating mean degrees for
these dif erent groups, we distinguish between relations within the group and between
the groups. The resulting values and tests for the signii cance of dif erences between the
mean degrees are given in Table 17.4. Since we cannot assume a normal behaviour of
the degrees of the network members, we perform the nonparametric Mann-Whitney or
Wilcoxon rank sum test to examine our above reasoning. The tests are performed by
column, that is, we compare exiting innovators of the i rst period with the entering inno-
vators of the second period and so on.
Regarding the dif erent roles of exiting innovators and entrants, we observe a stronger
Table 17
.
4
Technological overlap: mean degree (within and between blocks)
Within
Between
Between
Within
P1
Exit
Exit and
permanent
Permanent
and exit
Permanent
N
107
107
32
32
Mean degree
8.729
6.738
22.531
15.125
Standard deviation
(5.984)
(4.187)
(22.361)
(14.914)
P2
Entry
Entry and
permanent
Permanent
and entry
Permanent
N
157
157
32
32
Mean degree
18.191
7.592
37.250
19.563
Standard deviation
(16.196)
(4.825)
(38.756)
(17.629)
Signii cance of dif erence between mean degrees
Mann-Whitney
Wilcoxon rank sum test
W
5854
7906.5
101.5
130
p
- value
0.000
0.208
0.002
0.011
Note:
One-sided tests are performed with H
0
as no dif erence between samples and H
1
in the direction of the
observed dif erences.
