Data Analysis

Following the collection and inputting of the data, the next step is to see whether the difference between the mean scores is statistically significant. If you are using a simple pre- and post-test design, a simple related (repeated-measures) t-test may suffice. However, if other variables are of interest, eg gender, and if a control group has been employed, then Analysis of Variance (ANOVA) could be utilised to compare the pre- and post-test scores for males and females and for the different groups (mixed ANOVA). As recently indicated by Supplee et al. (2013), policy-makers have moved on from asking ‘what works?’ to asking the question, ‘what works for whom?’. An examination of moderated effects can help to refine theory, target interventions, and tailor interventions more appropriately to the needs of a specified group (Rothman, 2013). Multi-factorial ANOVA enables a researcher to test for potential moderation effects. More recently, Multi-Level Modelling (MLM) has begun to be employed within the field of Criminology and in evaluation research more generally. As noted by Baumer and Arnio (2012), this technique is well-suited to criminology data since often a researcher is examining individuals nested within one or more higher level units (e.g. schools or communities); it can also deal with repeated measures data. Owing to the clustered nature of the data, data from the present study were analysed using multilevel models, with two levels of clustering (within classes and within schools). MLM is also advantageous in this case because it is well suited to dealing with missing data it helps to correct for any bias induced by attrition.

This has been an overview of the types of research design employed within the social sciences that you might consider using in your own research. There are many sources of information available to support you further, such as the Sage Handbook of Criminological Research Methods.