My team conducts a survey study. We invite aged 35-50 to answer questionnaire 1 and aged 51-65 to answer questionnaire 2. We employ random sampling by using apartment as a unit. When we go to the apartment to do the survey, we note there are two members fulfill our criteria with aged 49 and 52. Should we invite them to participate our study? Or just anyone of them?
When we go to the apartment to do the survey, we note there are two members fulfill our criteria with aged 49 and 52. They belong to two different group of sampling in our study. Should we invite both of them to participate our study or just only one of them? Would you leave the reason on representative bias and indicate some reading?
Thank you very much for your help.
What is the goal?
Why these age brackets. Why no 35-45 and 55 to 65? Including ages 50 and 51 will decrease the apparent differences in outcomes (assuming that survey 1 and survey 2 outcomes can be compared in some way).
Does it make any difference if there are males and female?
Does it make any difference if there is one person in category 1, versus two people in category 1, versus a person in category 1 and one in category 2, versus two people in category 2, versus one person in category 2?
To Ette, If I choose apartment as a cluster, is there a problem of over-representative?
To Timothy, This problem occurs in any kind of grouping (age brackets). You arise a good question about the difference between gender, and the various number of people in a unit.
Wish we could have more discussion on it.
Thank you
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Applied Statistics
- Survey Design
- Questionnaire