Hypothesis Testing
Message posted by Natalka Junyk on June 7, 2000 at 12:00 AM (ET)

Does anyone know exactly what the general purpose of Hypothesis Testing is and some examples of where it is applied.

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Hypothesis Testing
Message posted by Vern Myers on June 8, 2000 at 12:00 AM (ET)

Here's an example of a real-world type of use: Let's say that we have made a slight change in a chemical or manufacturing process. Shortly thereafter, we notice that the average value of a desirable characteristic seems to have improved. Hypothesis test is a scientific way of answering the question, "Is the difference between the new average value and the historical average value real, or could it be due to the variation inherent in the random sampling process?"

Typical steps:
1. State two hypotheses, Ho(called the "null hypothesis"), and Ha(called the "alternate hypothesis") Ho may be a statement like "the mean equals 25" and Ha might be "the mean is not equal to 25."
2. Select a significance level, "alpha."
3. Calculate the sample statistics.
4. Calculate the test statistics.
5. Determine the acceptance region and critical region of the test statistics.
6. Either accept or reject Ho.

z-tests and t-tests are used to test hypotheses about means. Chi-square tests and F-tests are used to test hypotheses about variance.