HYPOTHESIS TESTING

Using SPSS for Windows

1. INTRODUCTION
In order to pass this aspect of your course, you need to understand the "mechanics" of the statistical processes as taught in your Introductory Statistics course, and be able to apply them to the practical example at hand. This means that you have to "think" about what your data means. You have to "think" about suitable analyses for the level of measurement of your variable (nominal, ordinal, interval, or ratio). You then have to be able to interpret the statistical results in the context of the given problem.

When requesting SPSS procedures:

a. Consider the results that you are expecting, for example, what will be the point of the results attained via a Regression Analysis, or a Means Analysis? Ask if this procedure is the best way of solving your informational need.

b. ALWAYS consider the level of measurement of your variable (nominal, ordinal, interval or ratio). There are varying opportunities for statistical analysis, and varying limitations, according to the level of measurement.

c. Always do a "reasonableness" check on your results, for example, if you ask SPSS to do a DESCRIPTIVES analysis on the variable, GENDER (Male = 1 Female = 2) it will happily provide you with a Mean of say, 1.586 and a Standard Deviation of say, 0.694. However unless you are dealing with a population of transvestites - this result is silly!!!

2. GENERAL PRINCIPLES OF HYPOTHESIS TESTING

The following generic procedures can be followed with ALL forms of hypothesis testing. By practicing these guidelines carefully, you will encounter less difficulty when performing the multitude of hypothesis tests possible.

Notation: The Null and Alternate hypotheses are framed in the notation of the relevant population parameter. Refer following table.

Population Parameter and Sample Statistics