Decision Involving Probability

Regression Fallacy

Ideally, researchers often use observational data to evaluate the efficiency of a specific program. Statistical regression is used to describe the tendency of extreme measurement to move closer towards the mean when repeatedly done. The regression effect is also used to explain why a test-retest situation in a bottom group shows improvement on the second test, while the top group indicates a mean drop. The statistical regression is also referred to as regression towards the mean and is considered as a rival explanation for a change in nonrandomized interventions such as case-control studies. Consequently, the regression effect is determined by how far the scores on a baseline test move away from a population mean. 

Averagely, baseline scores that tend to move further away from the mean often regress more upon retest as compared to baseline scores that are less far from the mean. Simply put, the mean stands for things going back to normal, and when results go back to normalcy, people often believe that their actions cause the change, when in fact it was causal. In control groups, the regression effect often assumes that something returned to normalcy due to corrective procedures that were initiated while the situation was abnormal. In this case, it fails to consider the probability of natural fluctuations and is frequently referred to as a regression fallacy. Consistent with the case study, Heather is often disturbed by migraines and always takes Tylenol, and the migraine goes away. After some time, Heather concludes that Tylenol is an effective remedy against a migraine and orders a case of bottles from Amazon. A migraine going away after Heather took the painkillers is more easily explained by regression towards the mean. However, assuming that only Tylenol that can cause the pain relief is fallacious. More specifically, because a migraine cannot be measured, regression, except a treatment effect would only cause the mean to lean towards the population mean of Heather going back to normal. The probability that regression towards mean was confounded with Tylenol, Heather should have been causing about ordering for a full case of this particular medicine.

Reducing the Impact of regression

Virtually, there are several methods used to reduce the effect of a regression. According to Weeks (2007), the most effective method to control a regression effect is through a random assignment to experimental groups, in a way that subjects who have extreme scores are randomly allocated among the study groups. In this case, to determine whether the regression fallacy has occurred, the additional information required would be whether Heather tried other painkillers to treat her migraine. Consequently, the use of a highly reliable test procedure is recommended to help reduce the measurement error within an occasion, thus preventing the regression effect. Weeks also note that it is important when using subjects with extreme, scores at baseline to carry out at least two baseline procedures set apart. In this case, in order not to fall victim to the regression fallacy, the researcher should have requested for information on whether Heather has tried other migraine relievers before concluding that only Tylenol was effective. Heather should also have considered experimenting with another individual to check whether the specific drug only worked with her or was effective for a migraine with everyone else.

ReferencesDouglas, L. W. (2007). The Regression Effect as a Neglected Source of Bias in Nonrandomized Intervention Trials and Systematic Reviews of Observational Studies. Evaluation & the Health Professions, 30(3), 254-265