Suppose a "social intelligence" test in a popular magazine...

Name: Problem 3, Chapter 5: Research Design Explained
Uploaded: 2021-06-03T04:30:13-08:00
Duration: 1 min 10 s
Channel: Emily Himsel
Description: Suppose a "social intelligence" test in a popular magazine had high internal consistency. What would that mean? Why would you still want to see whether the test had discriminant validity? How would you do a study to determine whether the test had discriminant validity?

Suppose a "social intelligence" test in a popular magazine had high internal consistency. What would that mean? Why would you still want to see whether the test had discriminant validity? How would you do a study to determine whether the test had discriminant validity?

Key Concepts

Internal Consistency

Internal consistency refers to the degree to which items within a test are correlated with one another, ensuring that they are all measuring the same underlying construct. A high internal consistency (commonly measured by Cronbach's alpha) indicates that the items are homogenous and reliably capture the intended concept. However, while it confirms that the test items are related, it does not establish that the test is measuring a unique or isolated construct.

Discriminant Validity

Discriminant validity is a component of construct validity that assesses whether a test is not inadvertently measuring a construct different from the one it intends to assess. It is important to demonstrate that scores on the 'social intelligence' test are not excessively correlated with tests measuring unrelated constructs. This helps to ensure that the test has a distinct identity and is not capturing overlapping or confounded traits.

Study Design for Assessing Discriminant Validity

To evaluate discriminant validity, a study should include the social intelligence test together with measures of theoretically unrelated constructs. By examining the correlations between the social intelligence test results and these other measures, researchers can determine if the correlations are low, which would support the test’s discriminant validity. A well-designed study may include a diverse sample and appropriate statistical methods (such as correlation or factor analysis) to ascertain that the test is not conflated with unrelated psychological constructs.

Please give Ace some feedback