Interpretation of t-test results

The following paper discusses the relationship between t-statistics, obtained from testing differences between means, and correlations.

It begins by introducing t-tests used when evaluating the significance of a difference between two averages, provided the usual significance testing interpretation. The assessment of significance based solely on the t-test can be strengthened by drawing on the relationship between t-test results and correlations. Creating a decision-making triage based on the size of correlations is well understood by the research community and can provide a more stable basis for identifying strong and more meaningful t-test results, without strict reliance on statistical significance. 

In particular, the paper shows that converting a t-test result, significant with 95% confidence, into a measure of correlation will reveal that the t-test result is not necessarily an indication of a large, meaningful difference between averages. The relationship between t-test results and correlations will be beneficial in aiding the researcher when interpreting results. It will be especially helpful as a way to avoid overstating the importance and taking marketing actions from, an otherwise small difference between averages.

Research On Research: Interpretation of t-test results

Note: This paper originally appeared as part of the Market Facts, Inc, series of Research on Research reports.