Finally, statistical validity –ensuring that the appropriate statistical tests are used– was considered when the research instrument was being developed, when the statistical methods were decided upon, and when the data were collected. Threats to statistical validity were tackled mainly by carefully operationalising the variables under study and by choosing the statistical analysis method that had been previously recommended in the literature –that is, multiple regression analysis with main and interaction effects (e.g. Hartmann and Moers, 1999; Schoonhoven, 1981; Venkatraman, 1989).
đang được dịch, vui lòng đợi..
