Hello everbody!
I'm doing a survey experiment. I have 240 respondants, half of the group were treated with a politicians statement, half of the group were not treated (controlgroup).
First I was looking for a treatment effect. Now, I want to look for hetereogenity. Means I want to see if a special group of people (for example AfD-voters) responded more or less to the treatment.
My problem is, that due to the small number of respondants the treatment is not equally distributed.
This chart shows the party preference, split up into controll- and treatmentgroup:
So what I have here is that only 6 AfD-voters had a treatment, but twice as many had none. Same goes for CDU/CSU and SPD.
Does somebody know, how I can create weights so that the answers of some groups - for example the afd-voters with treatment - count more?
Thank you in advance!
I'm doing a survey experiment. I have 240 respondants, half of the group were treated with a politicians statement, half of the group were not treated (controlgroup).
First I was looking for a treatment effect. Now, I want to look for hetereogenity. Means I want to see if a special group of people (for example AfD-voters) responded more or less to the treatment.
My problem is, that due to the small number of respondants the treatment is not equally distributed.
This chart shows the party preference, split up into controll- and treatmentgroup:
So what I have here is that only 6 AfD-voters had a treatment, but twice as many had none. Same goes for CDU/CSU and SPD.
Does somebody know, how I can create weights so that the answers of some groups - for example the afd-voters with treatment - count more?
Thank you in advance!
Comment