Validitas Kriteria

Summary

This lecture discusses 'validity criteria' in psychometrics, focusing on how new psychological tests or scales are validated by comparing them with established, reliable criteria. The two primary types of criterion validity covered are predictive validity, which assesses how well a test predicts future outcomes, and concurrent validity, which evaluates the correlation between a new test and an existing one measuring the same attribute. The importance of ensuring that the criteria are both relevant and reliable is emphasized, along with the use of statistical methods, particularly correlation, to measure the validity of psychological tools.

Takeaways

• 😀 Validity in psychological measurement is essential, and one key type is criterion validity, which assesses how accurately a tool measures what it is intended to measure.
• 😀 Criterion validity is obtained by comparing a new psychological scale with a relevant, reliable, and established criterion measure.
• 😀 The criterion used for validation must be relevant to the attribute being measured by the scale and reliable to ensure accurate results.
• 😀 The process of validating a tool using criterion validity involves statistical calculations, such as correlation coefficients, to determine how well the scale performs compared to the criterion.
• 😀 If the correlation between a new tool and the criterion is high (close to 1), the tool can be considered valid. A low correlation indicates invalidity.
• 😀 There are two main types of criterion validity: predictive validity and concurrent validity.
• 😀 Predictive validity measures how well a tool predicts future behaviors or performances, and requires a time gap between the administration of the scale and the observation of the outcome.
• 😀 Concurrent validity measures the correlation between the new scale and an established criterion measured at the same time. It does not require a time gap.
• 😀 For predictive validity, the correlation between scores on the scale and future performance should be close to 1 for high validity. If the correlation is low, the tool’s predictive validity is questionable.
• 😀 For concurrent validity, the correlation should ideally be above 0.3, with values closer to 1 indicating stronger validity. If the correlation is below 0.3, the tool's concurrent validity may be inadequate.
• 😀 Both predictive and concurrent validity rely on using reliable and relevant criteria for comparison. This ensures that the new tool’s validity is accurately assessed.

Q & A

What is criterion validity?

Criterion validity refers to the degree to which a test or measurement scale accurately predicts or correlates with a relevant criterion or outcome. It is an empirical method of validating a tool by comparing it with an established, reliable external criterion.

How does criterion validity differ from content validity and construct validity?

While content validity assesses whether a measurement tool covers all relevant aspects of a concept, and construct validity examines if the tool truly measures the intended construct, criterion validity focuses on how well the tool correlates with an external, reliable measure or outcome related to the attribute being measured.

What are the two types of criterion validity mentioned in the transcript?

The two types of criterion validity are predictive validity and concurrent validity. Predictive validity measures how well a test can predict future performance, while concurrent validity assesses how well a test correlates with an existing, reliable test or criterion measured at the same time.

What does it mean for a criterion to be 'reliable' in the context of criterion validity?

A criterion is considered reliable if it provides consistent, stable, and accurate measurements over time. This reliability is crucial for the validity of the comparison, as unreliable criteria can lead to inaccurate estimates and underestimation of the test's validity.

How can predictive validity be calculated?

Predictive validity can be calculated by correlating the scores from a newly developed scale (predictor) with a future outcome or performance (criterion). The higher the correlation, the stronger the predictive validity.

What role does time play in predictive validity?

Time plays a critical role in predictive validity because it requires a delay between administering the predictor (e.g., a test) and measuring the criterion (e.g., performance outcomes). This allows the predictor to demonstrate its ability to forecast future events or behaviors.

What is the significance of a correlation coefficient close to 1 in criterion validity?

A correlation coefficient close to 1 indicates a strong, positive relationship between the predictor and the criterion, suggesting that the test is highly valid in measuring what it intends to measure and accurately predicting future outcomes or performance.

How does concurrent validity differ from predictive validity in terms of timing?

Unlike predictive validity, which requires a time gap between administering the test and measuring the outcome, concurrent validity involves measuring both the test (predictor) and the criterion at the same time, allowing for immediate comparison and correlation.

What is the threshold for an acceptable correlation in concurrent validity?

In concurrent validity, a correlation coefficient of 0.3 or higher is considered acceptable, while a correlation above 0.5 is deemed strong. A correlation near 0 indicates poor validity.

Why is reliability important for both the criterion and the test in criterion validity?

Reliability is essential because an unreliable criterion or test can distort the correlation, leading to underestimation of the true validity. A valid test must be measured against a reliable criterion, and vice versa, to ensure accurate and consistent results.