Regression to the Mean

Publication Date :

Blog Author :

Edited by :

Table Of Contents

arrow

What is Regression to the Mean?

Regression to the mean (RTM) is a statistical phenomenon that indicates that if a random outcome of any occurrence or measurement is extreme in the first case, the second or later outcomes will be less extreme. That is, it will be close to the mean or center of the distribution.

Francis Galton first identified this regression to the mean fallacy in 1886. It can happen in various situations. It is more likely in the circumstances with multiple variables influencing the phenomenon, and less extreme ones follow extreme results. As a result, it contributes to data misinterpretation and the discovery of non-existent trends.

  • Regression to the mean (RTM) states that if one of the outcomes of an experiment is extreme, the next outcome will be closer to the population mean.
  • This bias causes false conclusions and errors in decision-making.
  • Entities can use critical thinking and conduct a randomized controlled trial (RCT) using an experimental group and control group to control the negative effects of regression to the mean.
  • A high blood sugar value was observed for an individual in the first test. Then measured again after treatment and observed a reduced value closer to the mean. However, if it occurred by chance but was interpreted as the result of treatment, then it is an regression to the mean example.

Regression to the Mean Explained

Regression to the Mean

Regression to the mean points to an extreme or unexpected outcome reverted to an average value if tested for the second time. Compared to the mean observed for the first group, the second group or observation set mean will be closer to the population's mean. If the sample selected from the population is nonrandom, also the first measurement is an extreme value, and the next measurement taken is imperfectly correlated or has a weaker correlation, the RTM effect will be more obvious and strong.

This phenomenon supports certain outcomes based on chance, not statistical analysis or calculation. Hence, Its occurrence has a negative impact when natural variation in repeated data is perceived as a real change. It is not uncommon when measurement closer to the mean follows extreme large or small measurements.

Examples

  • Consider values measured from people with high blood pressure or blood sugar. They are measured again after providing treatment and found that the mean value of the extreme group (people with high blood pressure or sugar) changes and is closer to the mean of the whole population. This result can make us believe that the treatment is effective. But this can be a case of regression to the mean because even if no treatment is given, the blood pressure or sugar can fall due to various other reasons like measurement error or biological variation and can be for a short duration.
  • Kevin is a poker player and is addicted to the game. When he started playing professional games, he had his first win on a Thursday, and on all other days, he failed. Kevin was thrilled and concluded that every Thursday was supposed to be his lucky day. Unfortunately, he couldn't repeat the success in another tournament conducted on a Thursday but won many other tournaments conducted on different days. It is a simple case of regression to the mean psychology; it states that any extreme event occurring at a random time is diminished by the second attempt, reversing back to mediocrity. It has happened by chance and not by statistical evidence, eventually following back to normality.
  • Consider another example. Consider two student groups, A and B, in a class containing equally talented students. The group is made to conduct competitive exams end of every term to promote healthy competition among students. Group A scored higher in the exam conducted at the end of the first term. Therefore, one can conclude that group A is better than group B academically and will continue succeeding in the next exams based on the first term exam result. If it is true, there would be no regression. Whereas, if luck played a part in winning group A wholly or partially, the significance of regression to the mean arises.

Regression to the Mean Bias

Regression to the mean is a widespread phenomenon in many studies resulting in false conclusions and errors in decision-making. It influences people or academics to consider or account for the output due to chance and believe that the event is returning to normality. Hence, it is vital to have the Knowledge of RTM to explain the experiment results segregating the biased interpretations effectively. For example, understanding how it applies to the financial markets can improve investment performance.

How to Avoid?

  • Be cautious at the design stage and analysis state to reduce the regression to the mean effect.
  • Conduct a randomized controlled trial (RCT) using an experimental group and a control group. The result exhibited by the control group gives information about regression to the mean impact.
  • Incorporate multiple baseline measurements.
  • Integrate critical thinking in all possible stages and use statistical methods apt to the scenario or event.

Frequently Asked Questions (FAQs)

What is regression to the mean definition?

This statistical phenomenon concept explains that the extreme measurement or outcome will be followed by a less extreme measurement closer to the mean value. For example, in healthcare, It leads to misinterpretation of output as a byproduct of treatment effect. However, in reality, it can be a result of chance.

What is regression to the mean in psychology?

In Psychology, it is explained as one of the selection phenomena causing bias in people. It is possible to detect its presence in different studies resulting in biased decisions affecting problem-solving abilities. Moreover, it makes people misinterpret the output or event that occurred due to chance.

What is an example of regression to the mean?

Francis Galton first discovered regression to the mean in 1886. As an example, consider his experimental study. Galton recorded the height of a specific number of children and their parents and computed the mean height of the parents. He discovered that when the parents' average height was higher than the population's mean, the children appeared shorter than the parents. Similarly, when the parents' average height was lower than the population's mean, the children were taller than their parents. Galton first described this phenomenon as regression to mediocrity, today known as regression to the mean.