Final answer:
The number that measures how well an item differentiates between high and low capability on a construct is called the item's discriminative power. It is related to effect size measurements such as Cohen's d and is crucial for determining the effectiveness of test items.
Step-by-step explanation:
The computed number for how well an item distinguishes between people high and low in a construct is often referred to as the item's discriminative power or discrimination index. This measurement allows educators and psychologists to see which items on a test are effective at differentiating between those who have mastery of the content and those who do not. For instance, in a series of test scores, such as 80, 68, 20, and 92, we could look at how each question discriminates between higher and lower achievers. Specifically, we refer to effect size as a way to measure the strength of the relationship between two variables. For example, Cohen's d is a measure of effect size that would categorize an effect size of 0.834 as large, suggesting a significant difference between two groups being compared.
A related concept is the confidence interval, which in educational testing can provide a range in which we are a certain percent confident (for instance, 95%) that the true average lies. If 95% of confidence intervals constructed contain the true average, this supports the reliability of the testing instrument.