How can I make the statistical significance of differences in performance apparent to readers?
Many report sponsors want to show whether differences across plans or providers are statistically significant. Others are concerned that when the number of survey respondents is very large, small differences may be statistically significant. Highly trained data analysts would argue that it is inappropriate not to report statistical significance levels. Unfortunately, we have yet to find language to explain statistical significance to the public, many of whom have limited quantitative literacy. It is your choice whether to present information on statistical significance in a report. An unobtrusive discussion of this issue will be skipped by the vast majority of lay people, but a few will dig right in. It is critical, however, to ensure that presentations of statistical significance do not make it more difficult for the typical report user to understand both text and graphics as quickly as possible. In public reporting, less is almost always more; most people will stop looking at a repor
Related Questions
- Why are the point differences between predicted achievement and actual achievement that are needed for statistical significance so much smaller with the WIAT-IIUK as compared to the original WIAT (WOND, WORD and WOLD)?
- May an insurer discriminate on the basis of sexual orientation where bona fide statistical differences in risk or exposure have been substantiated?
- What are the differences between USAV High Performance Programs?
