Construct validity addresses the extent to which your empirical measure effectively addresses the theoretical domain to which it is related.  Thus, the only way that you can meaningfully address construct validity is to track the extent to which your measure is aligned with expectations that derive from a specific theory, i.e., does your measure align with the theoretical construct?  This can be seen as the central concern of validation, and the end toward which other validation strategies are directed.  If our other approaches to validation yield good results, we typically assume that it is likely that the measure that we are using or have developed must validly reflect the construct we are seeking to address.

Dear Miss De,

Please understand that we need to be very clear about the concept construct first of all. I has come to us from the world of Psychology where initially people tried to measure a phenomenon or its severity by hearing the symptoms and trying to understand where do the totality of these symptoms point to. For example someone with depression could tell about being fatigued, anxious, hopeless etc. and these symptoms could point towards depression or anxiety, but we can't determine the severity of depression just by hearing the symptoms because we need to have a tool to measure it as closely as possible. Here we trying to measure the construct called depression. In psychology, a construct is a skill, attribute, or ability that is based on one or more established theories. Constructs exist in the human brain and are not directly observable. For example, though you may know a person is smart by the way they speak and what they say, you cannot directly observe intelligence. You can tell someone is anxious if they are trembling, sweating, and restless, but you cannot directly observe anxiety. You also cannot directly observe fear or motivation. They are all complex, abstract concepts that are indirectly observed through a collection of related events.

Construct validity refers to how well a test or tool measures the constructs that it was designed to measure. In other words, to what extent the statements we are using in the scale to measure  depression are actually measuring it. Here, it is important to keep two things in mind, one is that the items in the measurement scale measuring a latent construct should have sufficient inter-item correlation that it they should converge to measure the latent variable. This is what we refer to convergent validity,  and secondly, if there are different latent constructs in a model, then their should be sufficient difference between those constructs too. This is what we refer to discriminant validity. Using SEM, we can assess these two types of validity by examining the measurement model. I have assessed both these in the following published paper of mine. 

You can visit the following site for further details about construct validity. Hope it helps

http://jalt.org/test/bro_8.htm

Anant Jyoti

Dear Suman i aggred with Anantjyoti Badgaiyan

best regards

I use a simpler definition when I teach research methods. Construct validity is simply observing what you would expect to observe based on past research.

Cheers, Doug

Nazia Zabin Memon

Definition of Construct Validity:

There are a number of different measures that can be used to validate tests, one of which is construct validity. Construct validity is used to determine how well a test measures what it is supposed to measure. In other words, is the test constructed in a way that it successfully tests what it claims to test? Construct validity is usually verified by comparing the test to other tests that measure similar qualities to see how highly correlated the two measures are. For example, one way to demonstrate the construct validity of a cognitive aptitude test is by correlating the outcomes on the test to those found on other widely accepted measures of cognitive aptitude.