Which inferential statistic would be appropriate for determining whether there was an effect on memory due to varying levels of noise exposure (quiet vs. medium vs. loud)?