See term variable changed for wave i?
There has been a massive increase in the number of records for cseeterm and cseeterm2 in the chmain dataset being a -8 or -9 in wave 9 compared to wave 7. What has changed to cause this? Is this variable still useable?
#2 Updated by Stephanie Auty 8 months ago
- Private changed from Yes to No
- % Done changed from 0 to 80
- Assignee changed from Alita Nandi to Laura M
- Status changed from New to Feedback
Apologies for the delay in response.
If you look at the questionnaire, here: https://www.understandingsociety.ac.uk/sites/default/files/downloads/documentation/mainstage/questionnaire/wave-9/W9-questionnaire-consultation.pdf, you will see that cseeterm2 replaced cseeterm due to a script error where the looping was incorrectly applied. Individuals have either answered one of these or the other, as you can see if you create a cross tabulation of the two variables. The same applies to cseehol and recon.
Thank you, this doesn't really help the situation though. I had looked at cseeterm2 and it has even more missings and I cannot produce my statistic with it when well over half of responses for my sample are a -8 or -9. I need clarification as to whether this is the way it will be going forward as I produce an Official Statistic using it so it's a bit of a worry there being such a change - is this saying it was wrong before?
#4 Updated by Stephanie Auty 8 months ago
As respondents were asked either one or the other, you can combine the responses from both questions. If you do this there are only 6 missing responses. However, there are two things to be careful about: first, the response options seem to have changed and i_cseeterm2 has codes 8 and 9 ("several times a week" and "once or twice a week") which are replacing code 2 in i_cseeterm ("at least once per week").
The second is that the group of people who have been asked this question at Wave 9 is not the same group as in previous waves due to the looping working differently. We are currently working on resolving this issue and will get back to you as soon as possible.
#6 Updated by Stephanie Auty 8 months ago
I have some updated information about i_cseeterm and i_cseeterm2: this was not due to a script error, but because we decided that the response options needed to be changed. From Wave 11 onwards, only cseeterm2 will be asked, and as I said previously, if you would like to use them together you will need to combine codes 8 and 9 ("several times a week" and "once or twice a week") in i_cseeterm2 to be equivalent to code 2 in i_cseeterm ("at least once per week").
I will follow up with information about the looping shortly.
Thank you for this. I'm less concerned with the change in options and more the change in looping. That said, will the response options continue to be numbered in the odd way they are for wave 9? The 8 and 9 in the middle of the run, whilst I can see why it has been done, is a bit odd.
I'm in touch with other members of the USoc team separately about the looping issue so hopefully we can resolve it for some future waves.
I have a tangentially related question. We have used the seeterm variable as a proxy for relationship quality with the ex-partner. This was because the relex variable has quite a high proportion of unusable responses. Do you have any thoughts on what the 'inapplicable' responses mean in this case? they make up over a quarter of all responses.
#10 Updated by Stephanie Auty 8 months ago
The 8 and 9 codes for cseeterm2 will remain the same as in Wave 9 to avoid confusion when merging waves. The participants won't see those numbers, they are just for use in the data. The variables will also keep the '2' suffix to be clear that the question is not exactly the same as cseeterm.