What Is Predictive Validity? | Definition & Examples
Predictive validity refers to the extent to which a measure or test accurately predicts future behavior, performance, or outcomes. It is considered a subtype of criterion validity and is often used in the fields of education, psychology, and employee recruitment.
By ensuring high predictive validity, researchers and practitioners can make more informed decisions and develop more effective interventions.
What is predictive validity?
Predictive validity is a type of validity that assesses the extent to which a measurement tool, scale, or model accurately predicts future events, behaviors, or outcomes.
This type of validity is concerned with identifying individuals who are likely to exhibit certain behaviors, forecasting the likelihood of certain events occurring, and detecting changes or trends in outcomes over time.
Predictive validity is used in a wide range of fields, including:
- Education: Predicting student achievement, graduation rates, or dropout rates.
- Healthcare: Predicting patient outcomes, disease progression, or treatment response.
- Business: Predicting employee turnover, sales performance, or customer loyalty.
- Psychology: Predicting personality traits, mental health outcomes, or behavioral tendencies.
Predictive validity example
The following example shows the use of predictive validity in education.
To assess the predictive validity of their new instrument (the survey), they use simple random sampling to recruit first-year students to fill it out. At the beginning of the next academic year, the researchers check how many of them dropped out.
The results show a low correlation between the survey outcomes and the drop-out rate. There were also issues with other types of validity, such as external validity (generalizability of outcomes) and construct validity (how well the test measures the concept it was designed to measure).
Therefore, the researchers decide to adapt their survey to create an instrument with better predictive validity.
Content validity vs predictive validity
Content validity refers to the degree to which a measure (e.g., a test, assessment, or survey) measures what it is intended to measure. In other words, content validity is concerned with whether the instrument is actually testing the knowledge, skills, or attitudes it claims to be measuring.
Predictive validity, on the other hand, refers to the ability of a measure to predict future behavior, performance, or outcomes. For example, a measure has high predictive validity if it can accurately forecast how well someone will perform in a specific context or situation.
If that same test is also a good predictor of university test scores for an algebra course, it would have high predictive validity as well.
Concurrent vs predictive validity
Concurrent validity refers to the degree to which a measure or assessment correlates with a known criterion or outcome (gold standard) that is measured at the same time.
Predictive validity assesses how well a measure predicts an outcome that will occur at a later time.
Concurrent validity and predictive validity are both subtypes of criterion validity, but they have a key difference:
- In concurrent validity, the test scores and criterion variables are assessed at the same time.
- In predictive validity, the test scores are measured before the criterion variables.
The teacher wants to know if the reading comprehension test scores from the first year can predict the final reading test score in the students’ final year of elementary school. They compare the reading test scores from the first year with those from the final year to determine whether the reading comprehension test has good predictive validity.
How to assess predictive validity
Contrary to some other types of validity, such as face validity, you can formally assess predictive validity. It consists of comparing the test’s outcomes to the test outcomes of an already established instrument (gold standard or criterion).
To ensure a test has high predictive validity, researchers investigate whether its outcomes correlate with the criterion variable using a correlation coefficient (e.g., Pearson’s r). The type of correlation coefficient depends on the data’s level of measurement (nominal data, ordinal data, interval data, or ratio data).
A correlation coefficient indicates the strength and direction of a relationship between variables. You can use these rules of thumb to interpret the correlation coefficient:
- r = -1: Perfect negative correlation
- r = 0: No correlation
- r = +1: Perfect positive correlation
A strong positive correlation supports the hypothesis for good predictive validity. It provides evidence that the instrument is a good predictor of what you’ve hypothesized. A negative correlation or no correlation at all means that the instrument has low predictive validity.
They compare the customer retention survey scores with the actual customer retention rates 1 year later.
- Survey A has a correlation of r = 0.18
- Survey B has a correlation of r = 0.79
This means survey B’s scores are a better predictor of customer retention than survey A’s scores. It has higher predictive validity.
Frequently asked questions about predictive validity
- Which type of interview has been shown to have the highest predictive validity?
-
The interview type with the highest predictive validity differs based on the goal of the interview.
- Generally speaking, a structured interview has the highest predictive validity.
- Unstructured interviews have the lowest predictive validity, especially in recruitment or job performance settings.
- Semi-structured interviews have adequate predictive validity but not as high as structured interviews.
Situational questions, work sample requests, and interview questions about past behavior are the best question types in the case of job interviews.
When designing job interview questions, make sure to minimize bias and to also account for other types of validity, such as construct validity and content validity.
You can use QuillBot’s Grammar Checker to make sure your interview questions are error-free.
- What is the difference between construct validity and predictive validity?
-
Construct validity assesses how well a test measures the concept it was meant to measure, whereas predictive validity evaluates to what degree a test can predict a future outcome or behavior.
