I have designed a small survey for my work place of 2000 people. 200 people were selected based on gender (Male female), age group (4 25-year age groups) and working grade (a,b,c).
Out of my sample of 200, 120 surveys were returned completed.
I was recommended to first create a design weight, and weight the achieved sample (120) back up to my desired sample (200) and then weight this result back up to the population (2000).
I have consulted the following https://bookdown.org/jespasareig/Book_How_to_weight_a_survey/design-weights.html which expressed a design weight can be calculated by 1/prob of the observation being chosen.
This is my data:
This is my attempt at creating design weights for the survey sample and achieved sample for the groups:
I need to create weights which reflect gender * age group * grade.
Could anyone recommend any sources that would explain how to calculate the design weight, and explain the weight process above (achieved->sample->population) in more detail?