Managerial Hypotheses Testabilities

Solution Preview

Hypothesis testing has been described by some as the use of statistics to determine the probability that a given hypothesis is true. Is this the case. Are real world managerial hypotheses testable? Do you think you could translate real world managerial problems into a testable framework allowing for the application of the methods we have been learning. Are there time constraints applicable to managerial decision-making? If so, how; if not, why not? Are there certain classes of problem more suitable than others? Construct an argument for your assertion and provide a real world example supporting your argument.

Hypothesis testing is a statistical tool that is used to interpret data. It is also called inferential statistics, because the results you get from the data can help you make decisions.

Let us go through the steps of experimental design first to see how we set up research studies. I will explain the steps in hypothesis testing along the way.

In research study, we want to have a clear direction on what we are studying. Lets give the example of training in a work force. Perhaps we want to investigate if one type of training reduced the number of errors made versus then the standard training that all employees receive. Therefore, in an experimental design, we need to have two groups:

The experimental (treatment) group - this is the group of people who will receive treatment - in other words, they would receive the new training program.

Now, if we just look at the results from the experimental group, we would have nothing to compare them to. This is where a second group comes into play.

The second group is called the control group. This group is used as a 'baseline' - they would receive the standard training.

This brings us to the first step in hypothesis testing: setting up hypothesis:

So our hypothesis would be:

Null hypothesis: There is no difference in number of errors made in both training groups
(a=b)
Alternative hypothesis: The new training program results in less errors then the standard program.
(a is not equal to b)

Next step ...