Question 872:

Asked on December 7, 2010

1

Answer:

No answer provided yet.The t-test is used to compare two continuous variables. For example, it can be used to compare the number of aggressive outbreaks, minutes in time-out or ratings of aggressive behavior.

It is ideally suited for smaller sample sizes (especially below 30).  An alternative test statistic is the z-statistic, which is also used to compare two continuous variables. The z-statistic however assumes the population standard deviation is known. In applied research, such as behavioral research, population standard deviations are often unknown and have to be estimated from the sample. This is fine as long as the sample size is reasonably larger (above 30 or so). The standard deviation tends to be underestimated with smaller sample sizes and is a poorer estimate of the unknown population  standard deviation

The t-test uses the t-statistic which is just like the z-statistic except that it takes into account the sample size when estimating the population standard deviation. What's more, the t-statistic can be used to compare continuous variables even for very small sample sizes (for example 5 children in each group).

A t-test is performed on the difference between the averages between two groups or between two time periods between two groups. For example, the average number of aggressive outbreaks in a control group compared to the  average number of aggressive outbreaks in a new treatment group. For more information on the mechanics of a t-test see the t-test calculator .

Since it is often difficult to obtain a large sample of human subjects for an experiment in things like aggressive behavior moderation, the t-test is an indispensable tool in understanding whether treatments are better than chance occurrences.