Hi all
I'm currently working on a survey validation project. As a novice in this area, I have a question about female vs male respondents for the purpose of survey validation. Does it matter that there are many more female than male respondents? At this stage it is approximately 80% female, so a very large discrepancy.
FYI: Analysis will be done using Dima’s (2018) 6-step R-based psychometrics protocol for scale validation, reference:
Dima, L. (2018). Scale validation in applied health research: tutorial for a 6-step R-based psychometrics protocol. Health Psychology and Behavioral Medicine, 6, 136-161.
Cheers,
Candice
If the survey is addressed to both sexes (male and female), it is important that there is a balance. If the survey concerns women, many more women than men will be desirable. In the event that the survey is not related to gender, 80% female and 20% male, it seems to me that the gap is very large, in my opinion.
The context of the question is somewhat unclear. First, the content or topics of the survey are not stated. Second, it appears that more women than men are responding to the survey. If this is due to possible response bias, e.g. that the survey topics are of greater interest to women than men, or some other factor, this could be an issue with respect to validity. If the differential is related to the availability of subjects, this would not seem to be a bias issue.
Another issue is the actual number of respondents for doing the analyses. So long as the sample of males is of sufficient size it should not matter that the female sample is larger.
If gender is relevant (directly or not) for the topic of the survey, then it matters.
Thanks for your responses to my question. Here is some further clarity. We are validating a survey exploring social determinants of health in a general population sample (i.e., we have recruited broadly with no specific population groups in mind). As the main avenues for recruitment has been community groups, local councils, etc, it appears that many more women than men are electing to respond. We have 255 responses so far.
According to the estimated proportion of both sexes in the "TARGET POPULATION" OBJECT OF SUCH SURVEY
Of course, this factor influences the results (classic foundations of scientific methodology). However, the psyche as an object of research is very complex (multifactorial), so in each case the influence will be different. The magnitude of the differences between the answers of women and men may be due to the purpose of the survey, age, profession, income, etc. In some areas, the differences will be very significant, in others less significant.
It is also very important how general and universal result you need. If you are going to extrapolate the result in general to humanity, then your sample is, of course, irrelevant.
In addition, if the sample is large enough and random - for example, 1200 people, of which 300 are men and 900 are women, then experts consider the difference irrelevant. If the sample is small, then randomization is needed according to characteristics that are important for the study. However, it still seems to me that even in a large sample, such an advantage matters. But how important it is depends on the purpose (topic) of the survey).
I can advise you specifically: If you have a sufficient number of people in the sample, test your hypothesis about the difference of opinion between men and women by comparing them in your study. This is the simplest solution.
- Badges
- Science topic
- Similar topics
- Statistics
- Survey