By means of a randomized control trial (rct), you want to evaluate the impact of the triple p programme, a parenting intervention with the main goal of increasing parenting behaviour that is beneficial for the child. randomization worked out very well. however, quite a number of the parents in the treatment group actually never showed up to receive the treatment.
what happens if you simply drop treatment group parents from the analysis who did not receive the treatment?
can you get the causal effect of the treatment if you compare the means of treated and control group according to the actual treatment status?
is it a better idea to compare means of treatment and control group according to the assigned treatment?
when you perform the comparison from ( you see that children in the treatment group have on average a higher test score, which amounts to 45 points. interpret this finding.
do you have a better idea to analyse the existing data, if you are interested in the effect
of actually participating in the triple p programme? assume that 95% of the assigned treatment group and 5% of the assigned control group receive the actual treatment. how large is the effect of actual programme participation on the test score?
assume that before the randomization only 40% of the parents who were invited for participation in the rct actually agreed to participate. the other 60 % did not participate in the study at all and are, hence, part of neither the control nor the treatment group. does this hamper the validity of the study?