| Sign In to gain access to subscriptions and/or personal tools. |
Using Quasi-variance to Communicate Sociological Results from Statistical ModelsUniversity of Stirling, vernon.gayle{at}stirling.ac.uk
University of Stirling, paul.lambert{at}stirling.ac.uk This article introduces a sociological audience to `quasi-variances' as a solution to the `reference category problem'.The reference category problem is associated with the interpretation of the effects of categorical explanatory variables within statistical models, and is especially relevant to sociological applications, where categorical explanatory variables are very common. This article presents a selection of examples (using multiple and logistic regression) to illustrate and comment on quasi-variance calculations for sociological models. In addition, the article is augmented with online materials provided by the authors, which aim to help social researchers practise and apply this technique using the popular data analysis software packages SPSS and Stata.The authors conclude that quasi-variance methods offer an attractive and practicable solution to the reference category problem that can, and should, be routinely operationalized by sociological researchers.
Key Words: categorical variables quasi-variance reference category regression models statistical models
Sociology, Vol. 41, No. 6,
1191-1208 (2007) |
|||