Question 305:

Asked on August 2, 2008

Tags: Discrete Data , Continuous Data , Null Hypothesis

1

Answer:

No answer provided yet.

You need to know if the data is continuous or discrete. For example, is the data measured in time (seconds, minutes) dollars, weight, height, voltage. The data is continuous if you can divide the data into smaller units that still make sense. So 1 minute, 1/2 a minute .00056 minutes all are meaningful. If you data is discrete it is not decomposable into smaller units, such as pass/fail data, number of errors.

Assuming your data is continuous, you could use the 2-sample t-test, which is one of the most common ways to test the difference between 2 samples. With this test you get a p-value which indicates the probability there is a difference, and your .05 significance level indicates you'd only be comfortable concluding there is a difference if the p-value was less than or equal to .05.  Your test statistic would then be the t-statistic.