2. Reliability

The reliability of a questionnaire is how consistently it will measure something. This is obviously a fundamental criteria to check when developing or deploying a questionnaire.

Questionnaire items, or questions should satisfy certain criteria to be reliable. It should be:

Stable
Consistent
And have equivalence

Select the icons below to learn what each of these concepts means.

Stability

A reliable questionnaire should give the same interpretation by respondents or the researcher when the questionnaire or its analysis is repeated (stability). There are two major types of reliability that you should look for:

Test/ retest reliability. Here, the administration of the questionnaire two or more times to the same participants at different points of time should yield the same results.
Inter-rater reliability. Here, two or more raters who code or score the same item response of a questionnaire should come up with the same results

Internal consistency

Internal consistency is a way of determining whether the question or questionnaire is measuring the factor it is meant to be measuring. There are different approaches that can be taken to this.

A Split-Half test involves splitting the same set of questions in two groups and administering each group to different group of participants to see the correlation among them.
Cronbach-alpha is a statistical procedure to measure consistency between similar questionnaire items when measuring certain concepts.

Equivalence

Equivalence is the extent to which two different questions or questionnaires addressing the same concept show agreement, or how researchers administering the same questionnaire to the same participants at different times collect the same results.

Sometimes the same question will be asked in alternate forms, such as using synonyms or response orders to see if there is correlation between the responses obtained.
Intra-rater reliability, whether the same individual can obtain similar results at two different time points may also be important here.