Chi square analysis is best applied in statistical research seeking to compare the observed frequencies against anticipated frequencies. It is mainly applied in cases involving categorical data. The assessment of different categories of respondents brings about categorization of data which makes the chi-square analysis method the best choice such analysis (Estes, 1991). A good example of a behavioral study that would apply the use of the chi-square model of statistical analysis would be a research seeking to find whether there are differences between female and male university students with regard to the choice of their major fields in career. The choices to be compared would for example categorize the courses into elective versus quantitative courses (Estes, 1991).
Notably, the variables whose frequencies are to be observed are course and gender, both of which are categorical best fit for such a statistical analysis. This introduces the mutual exclusivity that is characteristic of chi-square analyses. The acquired data is nominal in nature and questions arising from it can best be answered by frequency data analysis which is made possible by the chi-square method. The proposed study is thus best handled by chi-square analysis because it is generates nominal, categorical data which seeks to analyze frequencies.
Pearson coefficient research studies are best suited for research cases in which an individual is seeking to find out whether there is a correlation between two variables. This usually works with quantitative data which is paired in nature (Estes, 1991). The method measures the effect size which seeks to show the strength of relation between two variables. A good example under behavioral studies would be the measure of the relationship between parents’ drug abuse and potential child engagement in substance abuse if the child is brought up by a substance abusing parent. The variables generated are paired in nature and thus fit for the Pearson coefficient research. The correlations seek to find causal relationship of phenomenon and thus any behavioral studies seeking to find causal relationships in behavior are always fit for this statistical method (Estes, 1991).
Estes, K. W. (1991),. Statistical models in behavioral research: Routledge Publishers