You said "before going to data collection", meaning you probably have a questionnaire at hand. One easy way to determine if the questionnaire is reliable and valid is to search the literature and see if the said questionnaire has already been validated in your intended population or country.
If not, then you have to pretest the questionnaire in a sample (at least 30, but the more the better) drawn from the population of interest. Administer the questionnaire to them, as you would in your intended "data collection" method.
Determine the Cronbach's alpha for reliability (target is >=0.70 score). Validity can be determined by a wide range of tests such as thr discriminant validity, convergent validity and construct validity. [You probably should not worry about these tests at this point...Get on with the pretest first, then ask how to do these tests later on.]
You said "before going to data collection", meaning you probably have a questionnaire at hand. One easy way to determine if the questionnaire is reliable and valid is to search the literature and see if the said questionnaire has already been validated in your intended population or country.
If not, then you have to pretest the questionnaire in a sample (at least 30, but the more the better) drawn from the population of interest. Administer the questionnaire to them, as you would in your intended "data collection" method.
Determine the Cronbach's alpha for reliability (target is >=0.70 score). Validity can be determined by a wide range of tests such as thr discriminant validity, convergent validity and construct validity. [You probably should not worry about these tests at this point...Get on with the pretest first, then ask how to do these tests later on.]
Agree with Mohammed! Would like to add a bit more that generally two types of validity in quantitative research- (i) internal validity: that indicates how accurately questionnaire is made by right items or questions, (ii) external validity that we usually called generalizability: that represents how we can use findings from one context to another.
Reliability- once you you have data in your hand you may determine reliability by Cronbach's Alpha. If the Cronbach's Alpha is > 0.7, the consistency and precision of the data is quite ok.