All hypothesis tests require that we compare the sample values of our statistics to critical values of the sampling distribution. The sampling distributions represent the range of sample values of our statistic that we can expect under the Null Hypothesis -- in this case, that our experimental factors do not have an impact on eye witness memory. The critical value of the F ratio provides a cut-off beyond which we are willing to conclude that our sample did not come from the null population, suggesting that we have observed a statistically significant experimental effect.