Question

How can a T-test be useful and very applicable in social science? As you discuss T-tests, note the several different kinds of t-tests (independent samples, paired-samples, etc.) and in what research context (experiments, medical trials, etc.) they may be more appropriate?

Answer 1

T-tests can be very useful and very much applicable in social science since it used to compare means between two groups and also comparing the population mean with the sample mean when the sample size is small.

There are different types of t-tests namely, one sample t-test for comparing the population mean with the mean obtained from the sample, independent sample t-test used when we have two groups of samples which are independent of each other and paired sample t-test or dependent samples t-test where the group of samples are dependent.

One sample t-test is used when we perfrom a medical trial. For example:- Researchers want to try a new drug that dramatically increases weight loss. It is known that the average person loses two pounds per week. In a random sample of 20 people through the use of this new drug, participants lost on average three pounds per week. Is the difference between the population mean because of the new drug or is it because of sampling error.

Independent sample t-test is used when we perform an experiment. For example:- Is the difference in the mean police confidence(mean confidence in the police) between men and women statistically significant.

Paired sample or dependent sample t-test is used when we perform a medical trial. For example:- A treatment was given to ten patients and readings of HB(hemoglobin) readings were measured before and after the treatment was administered to the patients. Can we conclude that the hemoglobin percentage increased after the treatment was given to the patients.