Ricki is an educational psychologist. She wants to do a study examining whether or not a certain curriculum will help students learn math skills. In order to figure that out, she has to put together a math test that the students will take after they've been exposed to the curriculum.
Psychological measurement is the process of evaluating psychological traits, including cognitive skills, like math, and other traits, like depression or altruism.
Measurement is the cornerstone of psychological studies. Without measurement, there is no way to gather data in a study. Without data, there is no way to know if your hypothesis is correct. For example, Ricki might think that the curriculum will help students learn math, but unless she measures their math skills, she won't have the data to show whether it actually does help or not.
There are many tools used in psychological measurement. When looking at cognitive, or thinking, skills, tests are usually used to measure outcomes. IQ tests are examples of psychological tests that measure cognitive skills. So is Ricki's math test.
Let's look at two different ways of scoring tests: norm-referenced and criterion-referenced.
Ricki wants to know if her curriculum will help students learn math skills, and she's written a math test for the students to take. But, how should she determine what passing means?
A norm-referenced test scores a test by comparing a person's performance to others who are similar. You can remember norm-referenced by thinking of the word 'normal.' The object of a norm-referenced test is to compare a person's performance to what is normal for other people like him or her.
Think of it kind of like a race. If a runner comes in third in a race, that doesn't tell us anything objectively about what the runner did. We don't know if she finished in 30 seconds or 30 minutes; we only know that she finished after two other runners and ahead of everyone else.
So, if Ricki decides to make her test norm-referenced, she would compare students to what is normal for that age, grade, or class. Examples of norm-referenced tests include the SAT, IQ tests, and tests that are graded on a curve. Anytime a test offers a percentile rank, it is a norm-referenced test. If you score at the 80th percentile, that means that you scored better than 80% of people in your group.
Norm-referenced tests are a good way to compensate for any mistakes that might be made in designing the measurement tool. For example, what if Ricki's math test is too easy, and everybody aces it? If it is a norm-referenced test, that's OK because you're not looking at the actual scores of the students but how well they did in relation to students in the same age group, grade, or class.
But, norm-referenced tests aren't perfect. They aren't completely objective and make it hard to know anything other than how someone did in comparison to others. But, what if we want to know about a person's performance without comparing them to others?
A criterion-referenced test is scored on an absolute scale with no comparisons made. It is interested in one thing only: did you meet the standards?
Let's go back to our race scenario. Saying that a runner came in third place is norm-referenced because we are comparing her to the other runners in the race. But, if we look at her time in the race, that's criterion-referenced. Saying she finished the race in 58:42 is an objective measure that is not a comparison to others.
Tests that are pass-fail are criterion-referenced, as are many tests for certifications. Any test where there's a certain score that you have to achieve to pass is criterion-referenced. So, for example, Ricki could say that students have to get a 70% on her test to pass, which would make it a criterion-referenced test.
As we mentioned, criterion-referenced tests are good for giving an objective picture of how a person does. They are often seen as more fair than norm-referenced tests because how well the other people in the group do on the test doesn't affect your score.
But, it is difficult to measure certain things, and criterion-referenced tests run the risk of not giving a good picture of what people can and cannot do. Going back to Ricki, if her test is too difficult and the whole class fails, a criterion-referenced score will not take into account that the test was difficult.
So, in short, both criterion-referenced and norm-referenced tests have positives and negatives. A researcher has to decide which is better for their study: a measurement tool that offers information about how people do in relation to others or one that looks at non-comparative data of how students do.
Psychological measurement involves assessing psychological traits. There are two major ways to score assessments in measurement: norm-referenced tests look at how a person does in comparison to other people who are similar to them, while criterion-referenced tests look at a person's performance regardless of how others do. There are advantages and disadvantages to both.
After you've completed this lesson, you might be able to:
- Differentiate between norm-referenced tests and criterion-referenced tests
- Accent the advantages and disadvantages of both types of tests