Table 2:

Means of Disordered Eating Behaviors Episodes Over the Previous 3 Months for Total Sample, Athletes, and Non-Athletes, and Results from T-Test with Adjusted Permutations Q-Values.

	Total Sample			Results from T-Test with Adjusted Permutation Q-Values
Disordered Eating (DE) Behavior	Mean (SD)	Range	Respondents Reporting at least 1 Episode of DE Behavior % (n)	Athlete Mean (SD)	Non-athlete Mean (SD)	t-value	q-value
Binge Eating	9.69 (18.72)	0–270	74.85 (17,903)	8.38 (16.44)	9.91 (19.08)	4.46	<.001
Vomiting	3.41 (15.01)	0–300	26.05 (6,231)	3.08 (13.04)	3.47 (15.32)	1.39	.211
Diuretics/Laxatives	1.76 (8.50)	0–170	16.95 (4,053)	1.37 (7.03)	1.82 (8.72)	2.92	.006
Excessive Exercise	6.27 (14.97)	0–150	48.52 (11,596)	10.97 (19.59)	5.46 (13.86)	20.27	<.001
Fasting	5.21 (12.58)	0–150	50.70 (12,068)	5.17 (12.54)	5.23 (12.58)	0.24	.826

q-values (p-values adjusted for multiple tests using False Discovery Rate correction and, in this case, were adjusted using permutation methods) and associated means (SDs) that are bolded are significant results.

All means, SDs, and ranges represent sample after using Winsorizing technique where data points above >3 SDs are converted to the value at 3 SD above the mean, thereby limiting the effects of spurious outliers. Given skewed data distributions and a large portion of the sample reporting 0 episodes for each disordered eating behavior, we used the mean and SDs for the subset of respondents who reported at least 1 episode of the respective behavior and converted the data points above 3 SDs to equal the highest data point not considered an outlier.