A transformed score is a modified version of a raw score that has been mathematically adjusted to allow for easier comparison or interpretation.
What Is a Transformed Score?
In social science research, it is often important to make data easier to understand and compare. One common way to do this is by changing raw scores into transformed scores. A raw score is the original, unmodified number someone gets on a test or measurement. But raw scores can be hard to compare across different people, groups, or tests. That’s where transformed scores come in.
A transformed score is created by applying a mathematical rule to a raw score. This rule helps standardize the results or make them fit a new scale. For example, if one test is scored out of 50 and another out of 100, researchers may transform both sets of scores to a 0–100 scale. This makes it easier to compare results from different tests or groups.
Researchers use transformed scores in many types of studies, including those in psychology, education, sociology, and political science. These scores help make patterns in data clearer and help others understand research findings.
Why Use Transformed Scores?
To Allow Comparisons
One of the biggest reasons for transforming scores is to make fair comparisons. Raw scores can be hard to compare when different tests have different scoring systems. If someone scores 30 on one test and someone else scores 75 on another, it’s hard to say who did better without knowing more. But if both scores are transformed to a common scale—like percentiles or z-scores—it becomes easier to see who performed better compared to others.
To Fit a Desired Scale
Sometimes researchers want to present results on a specific scale. For example, some researchers transform test scores to fit a 0 to 100 scale because it is familiar to most people. Others might use a scale from 1 to 5 if they are comparing the scores to survey responses. This kind of transformation helps different pieces of information “speak the same language.”
To Improve Interpretation
Raw scores can be hard to interpret, especially if the test is complex or if the scoring system is not well known. By transforming scores into something more familiar—like a grade (A, B, C) or a percentile—researchers help others understand what the numbers mean. This is helpful for teachers, policymakers, or anyone reading a report.
Common Types of Transformed Scores
Standard Scores (Z-scores)
A z-score shows how far a raw score is from the average, measured in standard deviations. The average (or mean) has a z-score of 0. A score one standard deviation above the mean has a z-score of 1.
Z-scores help researchers see how unusual or typical a score is. For example, in a psychology study, if someone’s memory test score has a z-score of -2, that means they performed worse than most people. This could be a sign that further investigation is needed.
T-scores
T-scores are another kind of standard score. They are like z-scores but are usually easier to understand because they don’t include negative numbers. T-scores are scaled so that the average is 50 and each 10 points represents one standard deviation.
These are used often in clinical psychology. For example, many behavior rating scales use T-scores. A T-score of 70 or higher might signal a possible behavioral problem that needs attention.
Percentile Ranks
A percentile rank tells you the percentage of people who scored below a certain score. For example, if someone’s test score is in the 90th percentile, that means they did better than 90% of the people who took the test.
Percentile ranks are common in education. They help students and parents understand where a child stands compared to others in the same grade.
Normalized Scores
Sometimes raw scores are transformed so that they fit a normal distribution—a bell-shaped curve where most scores are in the middle and fewer are at the high and low ends. This helps researchers apply statistical tests that assume the data are normally distributed.
For example, if researchers are studying voting behavior and the survey results are skewed (more people picked extreme answers), they might use a transformation to “normalize” the scores. This makes the results easier to analyze using standard methods.
Scaled Scores
Scaled scores are used when tests have multiple versions or forms. These scores adjust for differences in test difficulty. For instance, if two versions of a math test have slightly different questions, a scaled score puts them on the same level.
Large testing programs like the SAT or GRE often use scaled scores. This helps ensure that no test version gives an unfair advantage or disadvantage.
How Are Transformed Scores Calculated?
The way a score is transformed depends on what the researcher wants to achieve. But most transformations follow a few simple steps.
- Find the Mean and Standard Deviation: For standard scores like z-scores or T-scores, researchers first calculate the average score and how spread out the scores are.
- Apply a Formula: Each type of transformed score has its own formula. For example:
- Z-score = (Raw Score – Mean) divided by Standard Deviation
- T-score = (Z-score multiplied by 10) plus 50
- Percentile rank = (Number of scores below the raw score divided by total number of scores) multiplied by 100
- Double-Check for Accuracy: Researchers often double-check their transformations to avoid errors, especially when the scores will be used in important decisions.
Examples from Social Science Research
In Education
A school district gives reading tests to students at the start and end of the year. Raw scores show improvement, but they vary a lot. Some classes started with high scores, and others started low. To make fair comparisons, researchers transform the scores into percentiles. This shows how much each student improved compared to the national sample.
In Psychology
A clinical psychologist gives a personality test to a client. The raw score on one scale is 68. To interpret it, the psychologist transforms it into a T-score. It turns out the T-score is 75, which is well above average. This suggests that the person has more extreme traits in that area and might need support.
In Political Science
Researchers study political knowledge among voters in two regions. One group takes a 20-question quiz, and another takes a 15-question quiz. To compare results, the researchers transform all scores to a 0–100 scale. This allows them to compare average political knowledge across both regions.
In Criminology
A study looks at self-control levels among youth involved in criminal behavior. A survey includes several questions, each with different point values. Researchers transform the raw scores into z-scores so they can compare them with self-control scores from a national sample. This helps show how the group differs from the general population.
Cautions When Using Transformed Scores
Misinterpretation
Transformed scores can be helpful, but they can also be misleading if people don’t understand what they mean. For example, someone might think a T-score of 70 means the person got 70% of questions right. But that’s not how T-scores work.
Loss of Detail
Sometimes, transforming scores hides important details. For example, percentile ranks don’t show how far apart scores are. Two students could be in the 90th and 95th percentiles, but the difference in raw scores might be large or small. Without the raw scores, it’s hard to tell.
Overuse
Researchers should not overuse transformations. In some cases, raw scores are more appropriate, especially when they tell a clear story. Transforming just for the sake of it can make results harder to follow.
Conclusion
Transformed scores are a key part of social science research. They help researchers compare data, make results easier to understand, and present findings in a clear way. By using methods like z-scores, T-scores, percentile ranks, and scaled scores, researchers can turn raw numbers into meaningful information.
Still, it’s important to use these tools wisely. Not every score needs to be transformed, and readers must understand what the transformed scores really mean. When used carefully, transformed scores help bring clarity and fairness to research in education, psychology, political science, sociology, and more.
Glossary Return to Doc's Research Glossary
Last Modified: 04/01/2025