Understanding Convergent & Discriminant Validity

Convergent and discriminant validity are two subtypes of construct validity. This post explains the definitions of these types of validity, provides examples, and shows how the two differ from each other.

What is convergent validity?

Convergent validity shows whether a test that is designed to assess a particular construct correlates with other tests that assess the same construct. A French vocabulary test would have high convergent validity if candidates who took the test would get similar scores on different French vocabulary tests.

We can analyze convergent validity by comparing the results of a test with those of others that are designed to measure the same construct. If there is a strong positive correlation between the results, then the test can be said to have high convergent validity.

This goes some way towards proving that the test has good construct validity – in other words, showing that it measures what it is supposed to measure. To confirm this, the test would also need to show high discriminant validity (more on this below).

Example of convergent validity

The following scenario provides an example of determining convergent validity:

• A group of employees takes a series of five assessments, all of which are designed to test their abilities as salespeople through different types of questions.
• To establish the convergent validity of each assessment, the results are compared with those of the other assessments.
• The results of one test do not correlate with the results of the others – this assessment can be said to have low convergent validity.
• The results of the other tests all show a strong positive correlation with each other – these assessments all have high convergent validity.

What is discriminant validity?

Discriminant validity shows whether a test that is designed to measure a particular construct does not correlate with tests that measure different constructs. This is based on the idea that we wouldn’t expect to see the same results from two tests that are meant to measure different things (e.g. a math test vs a spelling test).

We can analyze discriminant validity by comparing the results of an assessment that measures one thing with those of a test that measures something else altogether. If there is no correlation between the scores, the test can be said to have high discriminant validity; a strong correlation would indicate low discriminant validity.

Example of discriminant validity

This example shows how to work out the discriminant validity of a test:

• At the end of the school year, a class of students take exams in all of the subjects they have studied.
• To establish the discriminant validity of the mathematics exam, the results are compared with those of the literature exam.
• No correlation is found between the results of these two exams, so they can be said to have high discriminant validity.

What is the difference between convergent and discriminant validity?

Convergent and discriminant validity are essentially two sides of the same coin: convergent validity requires a positive correlation between different tests that measure the same thing; discriminant validity requires there to be no correlation between tests that measure different things.

Although they are both subtypes of construct validity, it is possible for a test to have high convergent validity and low discriminant validity (or vice versa). For example, the results of a math test may have a strong positive correlation with those of other math tests (high convergent validity), but there could also be some correlation with the results of a spelling test (low discriminant validity).

The table below summarizes the similarities and differences between convergent and discriminant validity:

Why are convergent and discriminant validity often evaluated together?

Convergent and discriminant validity are subtypes of construct validity. To establish high construct validity, a test must be shown to have both high convergent validity AND high discriminant validity:

• The test results should show a strong positive correlation with those of other tests designed to measure the same thing (high convergent validity);
• Equally, the test results should show no correlation with those of tests designed to measure different constructs (high discriminant validity).

Example: using convergent and discriminant validity together

A group of employees take five different IT skills assessments. There is a strong positive correlation between the results of each assessment, indicating high convergent validity.

To establish discriminant validity, the test results are compared with those from a separate series of assessments designed to measure communication skills. There is no correlation between the results of the IT and communication skills assessments – a sign of high discriminant validity.

The analysis has shown that the IT skills assessments have both high convergent validity and high discriminant validity. This suggests that these assessments also have high construct validity (i.e. they are an appropriate way to measure IT skills).

Having established high construct validity, the overall validity of the IT skills assessments could then be confirmed by measuring other types of validity such as their criterion validity and content validity.

Next steps

This guide has explained convergent and discriminant validity, which are two subtypes of construct validity. To learn more, take a look at our post on how to test construct validity, which discusses the basics of construct validity and some examples of different tests.

Questionmark provides an online assessment platform that’s used by a wide range of organizations around the world. The results of tests are provided in a format that makes it easy to establish convergent and discriminant validity.