- Construct validity
- Construct validity is raising and answering the question whether the aspects that are part to the meaning of a construct should be rated as equal.
- The origin of construct validity
- This is construct validity
- Related topics to construct validity:
- Having reliable and valid instruments in research is a necessity. Read how to make such instruments:
Construct validity is raising and answering the question whether the aspects that are part to the meaning of a construct should be rated as equal.
The origin of construct validity
Validity can be viewed as a rating score that lies between 0 and 100, in which 0 means that it is not at all clear what is meant and 100 means that there is no doubt about the meaning of what is being measured. To make a good rating you can take a closer look at two aspects: content validity and construct validity. In this section we will discuss construct validity. Elsewhere we discuss content validity.
Suppose you want to do a research in which ‘satisfaction with the place to live’ is an important issue. There are several ways to measure this, for instance:
Give on a 5-point rating scale how satisfied you are with
… the living room
… the doorway
… the kitchen
… the sleeping room
… the bathroom
… the garden
… the state your house is in
Now what is normally done with such items is computing a mean score over all items. Is that correct? Should the living room have equal weight as the garden? And if you calculate the average, the score is mainly determined by the parts of the house and the opinion about the garden hardly contributes to the total score about satisfaction with the place where you live. Do you think that is correct?
This is construct validity
Construct validity is arguing that the construct to be measured has been assembled correctly. There is no clear answer. Much can be said to weigh all components equally. However, all aspects can be assessed as very good, but if the house is poorly maintained with leaks everywhere, according to me the score should be very low. It could mean that constructing a variable called 'satisfaction with the place where you live' should be done differently than simply adding up the scores on the questions.
Make use of existing questionnaires and tests. Then you don’t have to argue about construct validity. If everyone else is measuring a construct in a particular, already standardised way, then why should you be so wise to do it differently? You must have good reasons for it.
Related topics to construct validity:
- Factor analysis