The question of whether questionnaires measure what they should be measuring is a classic definition of validity, which should be tested according to hypotheses about how that measure will correlate with other accepted measures in the field.

In the case you describe, if the measure in question is indeed intended to capture positive feelings, then it should be highly correlated with other, well-proven measures of that same concept. Simply asserting that a proposed measure matches a given concept is only a claim for "face validity."

It depends on the extent that:

1. They are valid and reliable, and

2. The respondents are giving accurate and not "self-flattery" responses 

Hossein

Yes, agree with all the earlier responses. Just wanted to add that social desirability bias and other potential response biases are also important to factor into account. Social desirability bias is a response bias describing the tendency of survey respondents to answer questions in a manner that will be viewed favorably by others. It can take the form of over-reporting "good" behavior/feelings or under-reporting "undesirable" behavior/feelings.

Dear All

I agree with David L Morgan, I think Mr. Morgan provided reasonably answer.

Regards, 

I agree with all the above answers.

Regards

SM Najim

I think the Researcher has to explain the value of the survey especially the likely social benefit.  In this way, the respondent will see the value of his responses.

To add to the discussion: there is a certain extent of unknown associated with the actual measure. Even valid tests may or may not be measuring the exactness of your construct (as you state, knowledge versus sense of a particular emotion). The use of multiple scales can assist in parsing this relationship and in the review of your findings. 

You may need to further explore your findings in questionnaire survey by interviewing some key respondents. Your interview will give further insights on the issue of interests. 

I can comment in general terms to say that the reliability and validity of participants' responses in questionnaires (and interviews etc) have their origins in the quality - clarity, soecificity and non-ambiguity - of each question that (hopefully) characterise well established research instruments used in your field. The highest quality questionnaires should, in my opinion, contain a number of questions, randomly located in the questionnaire,  that evaluate the same phenomenon, thereby increasing researcher confidence that they are measuring what they're seeking to measure. The bigger challenge relating to your question about the extent to which questionnaires measure what a researcher wants them to measure is faced when there is a need to develop a questionaire for his/her unique line of enquiry, no easy matter.

Validity might be the question here. In any case responses/outcomes from questionnaires are dependent on clarity of the questions asked. Perhaps more important is the need for the Researcher to ensure the yardsticks for measurement fit the context and are objective;  and also respondents find it easy to understand and relate to the questions asked.

Yes, the questionnaire had been used historically to solicit primary data. To guarantee that the questionnaire will ultimately satisfy the research objectives, then its design is very important to achieve validity and reliability. In my view, even merits of adopting or adapting an existing questionnaire depends on the research objectives. The test is whether the researcher and respondents are aligned while decoding the questions and responses.

I guess I'll provide the controversial answer here: It is actually quite likely that for a topic like the one you describe, the semantic relationships between constructs will drive the majority of the correlations you get.  The better the validity measures are, the more likely you have semantic relationships, unfortunately. Here is a link to a project with some paper showing this: https://www.researchgate.net/project/The-semantic-theory-of-survey-response-STSR

In addition to Prof Morgan's answer, you should include in any write up an account of how you established validity in the ways that he and Lynne Johnstone have described.