Other Commonly Used Statistical Procedures

In this chapter, we explore some other commonly used but less ‘traditional’ statistical procedures. While these procedures are commonly reported in behavioural and social research, they tend not to be well-covered in standard statistical texts. Procedures discussed and illustrated include: reliability analysis & classical item analysis (useful for assessing measurement quality); data screening & missing value analysis (useful for preliminary explorations looking for anomalous data patterns); confidence intervals (useful for assessing the precision of statistical estimates); bootstrapping and jackknifing (useful for estimating errors associated with statistic estimates where traditional methods are not available or do not work); time series analysis (useful for understanding data patterns over time, with or without an intervention); confirmatory factor analysis (useful for evaluating theorised factor structures); structural equation models (useful for evaluating theorised causal models); and meta-analysis (useful for exploring data patterns evident in samples of published, and occasionally unpublished, research).