Leroy Metze

Significance tests and their interpretation: An example utilizing published research and ω2

The issue of the interpretation of significance tests is addressed. An arugment is presented that some measure of association such as ω2 should be provided as an interpretation/decision-making aid for scientific consumers and journal editors. Published research articles were examined regarding use of measures of association and the relationship between sample size and the amount of variance shared by the independent and the dependent variable. The results indicated that no articles reported measures of association and that many published studies are based upon small degrees of relationship between the independent and dependent variables. A change in report-writing and journal-editing practices is suggested.

The Protestant ethic effect in a multichoice environment

The Protestant ethic effect was investigated using eight rats that learned to barpress for food in a four-choice situation involving: (1) a functional bar and food dish where a pellet was dispensed with each barpress, (2) a nonfunctional bar and food dish which contained no free food during training, (3) a nonfunctional bar, and (4) a foodcup which contained no free food during training. After a stable rate of responding was observed, free-food pellets were added, and the choice behavior of the animals was observed. The results indicated that the rats preferred to barpress to obtain food in the presence of other food-getting or manipulation activities.