A Local-State Government Spatial Data Sharing Partnership

A Local-State Spatial Data Sharing Partnership Model to Facilitate SDI Development
correlations which may provide a useful insight into the final results of the regression
analysis. Finally, the 13 factors were used as predictor variables in regression analysis to
model the single outcome factor. This identified which of the grouped factors contributed
more positively to the successful outcomes of the partnership.
6.4.2 Factor Analysis
The variables used in the questionnaire included a range of measurement types including
continuous numeric values (e.g. number of properties), descriptive ordinal/internal values
(e.g. Likert scale – agree, strongly agree) and categorical or nominal values. The
categorical or nominal variables were not suitable for factor analysis and were therefore
not utilised. Prior to the factor analysis the continuous numeric variables and the ordinal
Likert variables were transformed to numerical interval classes between 0 and 5.
Factor analysis is a well documented technique that assists in identifying clusters of
variables that may be logically grouped into a smaller set of these variables which have
common underlying constructs or factors. Factor analysis generally can be applied when:
1. the range of variables being analysed are at least of an ordinal level of
measurement;
2. the variables are normally distributed;
3. the relationship between variables is reasonably linear;
4. the sample is at least 100;
5. there are more participants than variables and extracted factors.
188
(Brace et al. 2006, p. 310)
Once the appropriate transformations had been completed, the data from the LGA
responses satisfied these criteria.
The factor analysis was undertaken using the standard principal component analysis
method to reduce the total number of independent variables from 36 to 13 grouped factor
components. Table 6.4 lists the 13 new components (factors) and the original 36
independent variables with their factor loadings. A full listing of the factor analysis is
given in Appendix 6