Table 2. ITT-analysis and ITT-adjusted-for-treatment (AT) analysis of the cognitive learning performance.

Parameter	Reference	Script	GbEl	difference
N	117	63	81
gender ratio (M : F)	ND	24 : 39	36 : 45
Median	17	27	29	2
mean (SD) [SE]	18.3 (4.13)	26.0 (3.99)	28.6 (3.53)	2.66^*^ [0.63]
CI mean	17.5 - 19.0	25.0 - 27.0	27.9 - 29.4	1.42 - 3.909
mean AT [SE]		26.0	29.1	3.12^**^ [0.7]
CI mean AT		25.1 - 26.8	28.2 - 30.0	1.8 - 4.5

Data are shown for the reference group (without specific training) as a baseline measurement and the script group (control) vs. the GbEl group (intervention). Performance was measured with a 34-point single choice test.

The last column shows the performance differences between the GbEl group and the script group. The relevance of the differences should be interpreted considering the baseline value of untrained students from the first column. An increase of 2.66 points from script-based instruction to GbEl is about 35 % of the increase of 7.7 points from baseline to script.

The ITT-AT analysis pronounces the differences between the groups taking into account which students actually received the training. The differences between the groups are highly significant and relevant (7.8 % resp. 9.2 % increase). The Cohen's d effect size is 0.71 for the ITT analysis and 0.89 for the ITT-AT analysis.

^*^: t-test p < 0.001

^**^: Wald test p < 0.001