DATA TYPE: if the surveys were taken at different time, in addition to having the same questions asked, the data response type sought must also be looked at. The data type sought in both times must be of the same type. The types of data include: (i) quantitative, (i) ordinal and (iii) nominal. If the data type of the response sought are different in both times, the surveys could not be combined. However, if the respondents in both surveys were the same group, both survey may be treated as Z-test for 2-counts in Poisson distribution (data must be discrete). Various combinations follow:
(1) DIFFERENT SURVEYS - SAME QUESTION - SAME TIME PERIOD - SAME DATA TYPE: No problem. This is a normal case.
(2) DIFFERENT SURVEYS - DIFFERENT QUESTIONS: these surveys may not be combined. They are to be treated as different sets regardless of time when the surveys are taken because they deal with different issues.
(3) DIFFERENT SURVEYS - SAME QUESTIONS - DIFFERENT TIME PERIODS: These survey may be combined. In fact, for stimuli studies or opinion polls, this type of survey is used, i.e. collect survey at time T1 and then introduce stimuli and then collect the survey at T2. The purpose is to study the effect of the stimuli.
(4) SAME QUESTIONS - SAME RESPONSE DATA SOUGHT: If collected at the same time period, there should not be any issues. If collected at a different time, must test for homogeneity of data and compare results if there is a significant different before running further tests or engage any studies. Same response data means: quantitative, ordinal, or nominal for Y and X respectively. If there is a mismatch then they could not be combined, i.e. (YqXo)t1 and (YoXn)t2; these surveys are completely different and should not be combined.
(5) SAME QUESTIONS - DIFFERENT RESPONSE DATA SOUGHT: The two survey may not be combined even the same questions were asked, but the data type sought were different, i.e. (YqXo)t1 and (YoXn)t2.
(6) DIFFERENT QUESTIONS - SAME DATA SOUGHT: regardless of time, they may be combined with bias or error correction IFF the questions ask about the "same issue" but are phrased differently. If the respondents understood these questions to have "different meaning" or about "different issues" then they may not be combined.
(7) DIFFERENT QUESTIONS - DIFFERENT DATA SOUGHT: regardless of time, this type of survey combination must be analyzed separately; they should not be combined.
I suggest you that you checked the methodology Servqual methods. It moteoda assess the quality of services. In the literature, there are many publications on the subject, including empirical. I can also assist you if you need it.
I like Paul's response, very comprehensive. You want to be sure that whatever you do, your concepts are very comparable; or are standardized for comparisons, reliability, and/or validity's sake.
You get to check the research outcome, research terms, and the research statement first other then just to collected. Do not forget to validated it as well
If your separate questionnaires were filled-up by the same sampled respondents, you can use the single SEM model. If the different questionnaire filled-up by different respondents, you can't (from your statements, not sure filled-up by the same respondents)
However, you might want to do more literature review through various theoretical frameworks to see how different concepts / constructs are related or potentially related or previous research gaps are identified e.g. tourism & community etc. Once the constructs nailed down, then search for / adapt their respective survey questionnaires. Meaning all questionnaire used must tied to certain constructs you want to measure & test. Otherwise, even though different questionnaire filled-up by the same sampled respondents, they might not be useful for research outcomes / knowledge contribution. All the best.