One of the primary areas of focus for research and evaluation in public and nonprofit sectors is exploring the relationship between variables. For example, as a public administrator, you might be interested in determining the effectiveness of a public program. To do this, you would explore the relationship between the program and its intended outcome. Consider a program intended to reduce homelessness. You would need to reduce the number or percentage of citizens who lack suitable housing in order to be considered successful. Sometimes, you might need to consider additional variables in assessing the success of the homelessness program. In this case, measures of association and inferential statistics can help you sort out these types of relationships.
There are many types of inferential statistics (e.g., t-tests for independent samples means, t-tests for proportions, analysis of variance, etc.) and many types of measures of association (e.g., Lambda, Gamma, and Pearson’s r, etc.). A skilled researcher must have the ability to select the most appropriate statistic for a given situation. As noted in your course text, selecting the appropriate statistic is based on the purpose of the research, the measurement level of the variables, the measure’s sensitivity, and the researcher’s familiarity with the statistic. Therefore, there is no “perfect” statistic. Instead, you must determine the “most appropriate” statistic for your research.
Paying particular attention to the selection and use of inferential statistics and measures of association. Review the four criteria for selecting the best measure, keeping in mind the purpose for which you might use measures of association. Finally, think about which inferential statistics and measures of association you might use in your Final Evaluation Design (Final Project) to answer your research question. Use the four criteria to help you make a selection.