Statistical Significance ( P Values)

Empirical work within criminal justice and criminology often focuses on hypothesis or significance testing. Probability theory and statistical techniques are applied in order to test these hypotheses, which include the null and alternative hypothesis. Dichotomizing p-values into significant and nonsignificant is often arbitrary and may lead to the loss of important information. The small p-value provides strong evidence against the null hypothesis and supports the alternative hypothesis that antisocial peers indeed have an effect on an adolescent's future offending behavior. When considering an outright ban of significance testing, the American Psychological Association Task Force on Statistical Inference concluded it was better to present effect sizes in conjunction with p-values rather than banning significance testing altogether.