I am doing a study on welders on their KAP on occupational health. I have to get verbal consent from employers before I administer my questionnaire to welders. If the employer does not give consent, how do I report/calculate the non-response rate? Usually 4-5 welders work under an employer.
Let ...T = Total response; Ta = Total approached; R = Response; Rn = Non-response; Rv = valid response; and O = Out of scope. Then the Response rate is given by:
(1) R = Rv / Ta - O
If the total response is:
(2) T = R + Ra
Then the non-response is:
(3) Rn = T - R
Note that T = 1 and that R + Rn = 1 or 100%.
Avinav -
It may be adequate and clearly understood if you just kept a weighted employer nonresponse rate and a nonresponse rate for the 'eligible' welders. However, if the tendency for an employer to allow welders to respond seems related to number of employees, you may want to break your data into groups based on size (number of employees). You may want to see if response propensity is related to other characteristics as well.
Your study seems to relate to the concept of "nonignorable nonresponse," which you may want to research. (Note that "ignorable" nonresponse is not really "ignorable.")
Cheers - Jim
- Badges
- Science method
- Similar topics
- Mathematical Sciences
- Statistics
- Applied Statistics
- Survey Design