Fundamentals of Compensation Quantitative Methods

Hypothesis

Both types of statistical tests mentioned previously have a hypothesis. Confusing to many, statisticians test propositions to disprove what they wish to prove. First, they state that no difference exists.

This is the "null hypothesis," and this is what is tested. Refer to http://davidmlane.com/hyperstat/logic_hypothesis.html for more information on hypotheses.

Examples of hypothesis

For each of the instances below, state the null hypothesis:

  • Do selection techniques favor males?
  • Do men perform better than women?
  • Is a benefits communication film effective?

Null hypotheses:

  • Testing that a selection technique favors males:
    "Results show that percentages of those selected are equivalent."
  • Testing that men perform better than women:
    "No difference exists between average levels of performance."
  • Testing that a benefits communication film is effective:
    "The film has no effect."

Memory Jogger

What is the null hypothesis for the following:

"Are promotional opportunities different for protected and non-protected groups?"

Continue