Changes in hypothesis testing: Better statistics for better decisions
Are you already subscribed?
Login to check
whether this content is already included on your personal or institutional subscription.
Abstract
Hypothesis testing and interpretation of results in psychology - based almost entirely on p-values - is a widespread practice. This contribution to the Italian psychological research community has two purposes: 1) to present the limits of and misunderstandings caused by mechanical and de-contextualized use of the Null Hypothesis Significance Testing approach; and 2) to provide guidelines for evaluating the relevance and replicability of results with accuracy, efficacy and clarity. We emphasize the importance of: a) presenting results so that observed effects are illustrated with appropriate descriptive statistics necessary for potential meta-analyses, and b) controlling for power of the statistical test, using confidence intervals, and presenting indices of effect sizes. Scientific progress will be achieved more effectively by shifting emphasis from statistical significance toward interpretation of the meaning and relevance of results.
Keywords
- Statistical significance
- practical significance
- effect size
- replication
- statistical thinking