9 might be okay if the effect size is large enough. (If the study was pre-registered there should have been a power analysis to determine the sample size given the hypothesized effect size.)