I implemented a survey (in Switzerland) using a survey firm (Intervista), who use a panel of survey participants.
My survey includes 1000 participants. For sampling, the following quotas were used:
- age x sex (15-29, 30-44, 46-49, 60-79) interlocked quota
- disproportional Area (3 areas of Switzerland, Ticino was sampled disproportional, instead of only 8% we used 20 %, I would like to fix this later using rake)
- a disproportional larger proportion of people having more money than CHF 100000 on their banking and savings account. (approx 15 % of people in Switzerland, we used a quota of 300 people out of 1000 participants).
I already did the calculations. However, I think I missed the strata when specifying the svydesign:
svydesign(ids = ~1,
data = Data2,
weights = NULL)
In a second step I used rake to weigh the data back, because I oversampled two groups.
I think I miss strata in the svydesign. However, this is all the information I have. To use strata in the svydesign function, I need a weights-vector. However, I did not get any weights from the survey firm. How can I fix this? Is there a way to calculate the weights without having further information? Even if I had the weights, how can I adjust for interlocked quotas such as age x gender?
Thank you very much!