Geography Reference
In-Depth Information
Table 13
.
1 Survey statistics
Local cluster
Number of i rms
and local institutions
(network node)
Questionnaires
submitted
Questionnaires
collected
Redemption
rate (%)
Bormio
91
91
85
93
Cortina
61
61
30
49
Pila
50
50
24
48
Total
202
202
139
69
structural indicators and graphic network representation (George and Allen, 1989;
Nieminen, 1974), calculated by dif erent algorithms. A questionnaire was submitted to
all the main local tourist actors, including i rms - travel agencies, hotels, ski schools and
so on - and local institutions (see Table 13.1).
3
The questionnaire included a specii c question for each variable of analysis: knowledge
sharing, relational-based trust, competence-based trust.
4
The questionnaires were devel-
oped to map all the ties across network nodes. Each question followed the same proto-
col: a list of all nodes (local actors) was presented to the respondent as a list of all possible
'simple l ag' answers (0 = no relation with the local actor; 1 = relationship with the local
actor). These data were coded in symmetric adjacency matrices (Nieminen, 1974). The
symmetry of these matrices was based on the notion of reciprocity: if node 'A' declared
that it exchanged information or knowledge with node 'B', the relation was included in
the dataset only if node 'B' also declared that it exchanged knowledge with node 'A'.
This property is fundamental for ei cient interaction when inter-organizational relations
are based on informal agreements (Sutter and Kocher, 2004). The same criterion was
used for data regarding the one-to-one trust relations (relational and competence-based),
operationalized respectively according to the traits outlined in section 2. The net graphs
in Figures 13.2, 13.3 and 13.4 show all network ties - in terms of knowledge sharing -
among all nodes in the three considered clusters. These graphs rel ect the results of ques-
tionnaires where the respondents were asked to identify all the local counterparts (other
nodes in the network) with whom they regularly exchanged knowledge and business
experience over the medium and long period.
The degree of centrality of each node has been calculated with the following algorithm
(Nieminen, 1974):
n
n
c
j
5
(
(
I
ji
*
W
i
)
2
I
ji
*
W
i
)
2
c
j
Z
0
Å
a
i
51
i
1
C
j
= centrality of node '
j
' (where 0 <
C
j
< 1)
I
ji
= ef ective connection of node '
j
' with node '
i
' (1 = activated; 0 = not activated)
W
i
= node's weight (where 0 <
W
i
< 1)
Then, based on the theoretical assumptions mentioned in the previous sections, the
degree of relational reputation (
RR
) and competence-based reputation (
CR
) of each
node has been calculated as an aggregate of the trust relation of all network members in
respect of that specii c node:
