We have a research paper that we want to get published in a journal. The paper is some simple statistical analysis of a survey we made. One of the questions in the demographics section is "What is your profession?" and you are given 4 options plus "Other" where you can write whatever you want. In some cases, people wrote things in "Other" that qualified as one of the other 4 predetermined possible answers. Is it bad practice to change the results to refelct this? Can we do it and simply add a disclaimer that this was done and document what changes were made? Is it better if we just leave it as is?
3
-
I'd think you'd want to have an absolutely pristine version of your survey responses (so, without any "adjustments"), and, then, another one that was "rationalized" by you, and certainly explain what you did... But, I am not at all expert in standards for any such area of research.– paul garrettJul 21 at 21:57
-
At the risk of making the question overly specific, can you share the four options (it's not obvious to me what four options for "what is your profession" would be) and an example of something that you would conform?– Azor Ahai -him-Jul 22 at 0:18
-
These are specific to healthcare. The options are medic, nurse, technician, administrativion, other– The BoscoJul 25 at 2:25
Add a comment
|
1 Answer
I think that something like the below table and note would be a reasonable way of approaching the issue. If you share your raw data, this would be easy to indicate.
| Profession | n |
|---|---|
| Job 1 | 17* |
| Job 2 | 10** |
| Job 3 | 8 |
| Job 4 | 5 |
| Other | 10 |
* Includes two "Other" responses that aligned with "Job 1". ** Includes one "Other" response that aligned with "Job 2".
That said, if you have only four options, you should apply your domain knowledge to understand why someone might not have selected the presupplied option but wrote something that looks to you like the same answer. (For example, a "nurse practitioner" isn't a subset of "nurse.")
answered Jul 22 at 0:25
