One useful alternative to significance testing is to examine the size of the relationship in terms of effect sizes, using point estimates of effect size and confidence intervals in individual studies, and using meta-analyses to bring together the findings of a range of studies