Cogburn, Addom, and Mwangi – Gender in Global ICT Governance
Page 29 of 50
To prepare for data analysis, both the qualitative data and quantitative data were cleaned and
screened. For the quantitative data, we used SPSS to run frequency distributions to identify missing
data, non-plausible values, and checked for skewness, kurtosis and for normality of distributions.
We also explored the relevant variables in the dataset for multicollinearity and singularity through
correlation matrices.
The qualitative data was cleaned using N6, and focused on the removal of images and file
transfer encoding language within the qualitative dataset (primarily the e-mail archive). Organization
of the qualitative data yielded 38 text files, 148 free notes, and 116 tree nodes. The qualitative data
analysis strategy is based on grounded theory (Strauss & Corbin, 1990; 1998) and is analyzed using
an axial coding schema developed out of the theoretical framework above. Appendix A presents the
initial coding schema used in the content analysis. Table 1 illustrates the data on which the content
analysis was performed. Further cleaning was certainly possible (e.g., removal of redundant html
coding) but did not affect the quality of the data, and only served to make the files larger than
necessary.
ii
LIMITATIONS
We have developed a research design that is most appropriate to the phenomena under
examination. Also, we have triangulated our findings by bringing together qualitative and
quantitative approaches. However, all social science research has limitations. While our survey is
well designed and was pilot tested internationally, and across the sectors relevant to the study
(government, private sector, civil society), it is still limited to those participants in the WSIS process,
and more specifically, to those WSIS participants that provided their e-mail address to the secretariat
in registering for the Summit. To address this limitation, we have compared the study sample both
to the sampling frame and to the population of all registered WSIS preparatory delegates on key
demographic variables (i.e., gender, region, organizational type), and found that only organizational
Convention