Why do we all occasionally need a refresher on assessment testing?

A standardized score looks precise. That does not mean our interpretation should be.

Consider a score of 84. It falls more than one standard deviation below the mean, making it somewhat unusual compared with the normative sample. It may also fall just below a cutoff that matters for eligibility, placement, or access to services.

But is an 84 meaningfully different from an 85? Usually, not in the way that the numbers might suggest

.

Scores do not become fundamentally different simply because they fall on opposite sides of a classification boundary. An 84 and an 85 reflect nearly identical performances, and their confidence intervals would almost certainly overlap. Treating one as meaningfully impaired and the other as reassuringly typical gives the cutoff more weight than the test can support.

At the same time, that single point may have very real consequences. If 85 is the threshold for a service, admission, or accommodation, the difference between 84 and 85 can change what happens next.

That is the tension: a difference can be psychometrically trivial and practically consequential at the same time.

Context also matters. If the referral question concerns eligibility, we may need to discuss the cutoff explicitly—including the uncertainty surrounding the obtained score. If the referral question concerns emotional functioning, attention, or behavior, a one-point distinction in cognitive performance may deserve far less emphasis.

Confidence intervals help us communicate that uncertainty. A score is an estimate of performance, not an exact measurement of an unchanging trait. The confidence interval provides a range of plausible values after accounting for measurement error. Depending on where that range falls, it may cross classification or eligibility boundaries—and that deserves careful explanation.

This does not mean that standardized scores are meaningless. It means they must be interpreted rather than merely reported.

Good assessment requires us to ask:

• How unusual is this performance?

• Is the difference statistically reliable?

• Is it clinically meaningful?

• Does it change the decision we are trying to make?

• How much certainty does the test actually allow?

• 
  

The number matters. The cutoff may matter. But neither tells the whole story.

Need a refresher on standardized scores, confidence intervals, and the judgment required to interpret them responsibly? Join me for our next assessment refresher workshop.