Discriminative measures- Can they be criterion referenced and NOT norm reference

[Anne Berry]

I know some tests are both norm and criterion referenced. My question is about a test that is a discriminative measure, but also criterion referenced such as the MABC-2 or the Developmental Coordination Disorder Questionnaire Revised? The chart made me think if it was a DISCRIMINATIVE measure it would have to also be norm referenced. Any tips on this?

[Helen Carey]

This is a great question! Most tests fall squarely into 1 camp, but sometimes there is some ambiguity. For example, the PDMS and BOT are clearly norm-referenced and discriminative due to why and how they were designed. By definition, a norm referenced test must have a comparison sample (normative sample of children considered “typically” developing) who were administered the test and who’s scores determined the test’s distribution of scores across a bell curve. Score output is then based on this distribution across a bell curve and scores are reported in terms of scaled/standard, percentiles, etc. You are correct that discriminative tests are typically norm referenced because we need the normative data to determine cut offs for categories such as average, below average, etc.

The MABC-2 is considered a discriminative, norm-referenced measure based on test design and score output; it was designed to discriminate mild motor delays. Scores are reported at standard scores and percentiles. There is also weak evidence to show that it is responsive to changes over time, which tends to be a stronger characteristic of criterion referenced tests.

There is probably a stronger case for the DCDQ’07 (Revised) to be considered a criterion referenced test because scores are reported as cut offs for probability of DCD instead of providing a scaled/standard score. However, the 3 factors within the test have data distributed on bell curves, which is more common for norm-referenced tests. The test was developed using samples of typically developing children and those with suspected DCD (learning, attention, and movement impairments), also a more common characteristic of norm-referenced tests. I think a major difference lies in the purpose of these tests, such as to rule in a condition (DCD) vs rule out a condition (gross motor delay using PDMS, BOT). So, I do think that in some cases during clinical practice a criterion-referenced test could discriminate between typical development and a suspected motor condition.

I should probably provide a bit more clarity in the study guide, but I think the primary take away is that you should understand the purpose of the test (based on development) as that is a key factor in selecting outcome measures in clinical practice. Using tests for how they were not intended reduces the validity of testing results. For example, most norm-referenced, discriminative tests don’t have evidence of responsiveness, so probably should not be used to measure change in performance over time.

I hope this helps! Please let us know if you have any further questions.

[Anne Berry]

This does help!

[Anne Berry]

So is the MABC 2 norm referenced and criterion referenced?

[Helen Carey]

Hi,
I fear that I have created some confusion, so sorry! The Outcome Measures study guide indicates in the purpose column that the MABC-2 is a discriminative measure, but I see now that I inadvertently put that it is also criterion-referenced in the Administration column (which is an error). In the administration column, it should indicate that it is norm-referenced (because items were developed using a normative sample of children). The test is definitely a norm-referenced, discriminative measure by design. If a therapist chooses to use the test for multiple administrations over time to measure change, they would be using it as an evaluative measure, although that does not mean that the tool becomes criterion-referenced (it remains norm-referenced).

I apologize for the confusion and will correct the study guide!

Helen

[Author]

Posts