Some commonly used phrases reveal methodological ignorance and misunderstandings. If you plan to use these phrases, check that they are used in your intended sense. Here are a few problematic examples.
| Phrase | Problem
|
| t-tests were used to "compare variables" | Statistical hypothesis tests are used to test hypotheses about a population using data from a sample representing this population, not to describe the characteristics of the sample itself.
|
| "independent samples" t-tests | Several different t-tests have been developed for use with independent groups, e.g. Student's, Satterthwaite's, Welch's, Prien's t-tests and Hotelling's T-test.
|
| normality was "assessed" using the Shapiro–Wilk test | A variable's distribution in the population can perhaps be tested using a statistical hypothesis test but not be "assessed" until the entire population has been observed, which is not possible for unobservable populations.
|
| "nonparametric data" | A nonparametric hypothesis may be tested using a distribution-free test, but the phrase "nonparametric data" is nonsense.
|
| "statistical difference" | All differences are statistical in some sense.
|
| "no difference" | No clinically relevant difference or no statistically significant difference?
|
| "significant difference" | Does significant refer to uncertainty (statistical significance) or to relevance (clinical significance)?
|
| "independently associated with" | The effect estimates from a regression model used for estimating causal effects can only be interpreted on the basis of assumed cause-effect relationships among the included variables. No regression model would be necessary if independence is assumed.
|