Evaluating the effectiveness of an evidence-based online training program for health professionals in eating disorders

Table 2 Pre and post evaluation of learning outcomes

	Pre-mean [SD]	Post-mean [SD]	Effect size	Z score
Unmatched (Npre = 1812, Npost = 1010)
Willingness	4.11 [1.08]	4.17 [0.92]	0.07	0.66 (p = 0.51)
Confidence	2.44 [1.00]	3.41 [0.82]	1.11	24.33 (p < 0.01)
Knowledge	2.67 [0.71]	3.73 [0.56]	1.70	33.88 (p < 0.01)
Skills	2.59 [0.73]	3.47 [0.64]	1.28	28.28 (p < 0.01)
Stigmatisation	1.63 [0.54]	1.45 [0.50]	−0.33	−9.29 (p < 0.01)
Attitudes	1.04 [0.54]	1.11 [0.57]	0.04	2.93 (p < 0.01)
Matched (Npre = 480, Npost = 480)
Willingness	4.10 [1.05]	4.17 [0.94]	0.06	0.89 (p = 0.74)
Confidence	2.40 [0.97]	3.42 [0.81]	1.09	17.19 (p < 0.01)
Knowledge	2.67 [0.66]	3.76 [0.55]	1.74	18.75 (p < 0.01)
Skills	2.57 [0.71]	3.45 [0.65]	1.29	18.24 (p < 0.01)
Stigmatisation	1.57 [0.49]	1.40 [0.47]	−0.38	−8.94 (p < 0.01)
Attitudes	1.13 [0.53]	1.15 [0.56]	0.13	0.56 (p = 0.74)

Notes: The unmatched sample is comprised of all participants who completed the in-built pre questionnaire or post questionnaire; and the matched sample is comprised of participants who completed both the pre and post questionnaires and could be matched using unique person ID identifiers. The relevant sample size is indicated in parenthesis. Wilcoxon-Mann-Whitney test and Wilcoxon signed rank sum test were performed, respectively, for unmatched and matched data to test the differences in the pre and post ratings reported by the participants. The Cohen’s d effect size was calculated. A Z score in bold indicates a statistically significance change at the 5% significance level. P-values were adjusted for familywise error rates using Holm-Bonferroni method. Estimates were rounded to 2 decimal places

ISSN: 2050-2974