XLSTAT: Cochran's Q test in Excel

 

Background

Cochran's Q test is a statistical test that is used to identify if there are differences on a dichotomous dependent variable between more than two groups.

 

Dataset for Example

The attached data come from a survey where teenagers were asked whether they like the packaging of a computer game (responses: Yes or No). The goal is to determine whether there is a significant difference between packaging options to inform a production decision.

The hypotheses are:

• Null hypothesis (H₀): There is no difference in preference between the packaging options.

• Alternative hypothesis (H₁): There is a significant difference in preference between the packaging options.

 

Setting Up

• Open XLSTAT.

• Click on the Nonparametric tests menu and select the Cochran’s Q test feature:

• Once you've clicked the button, the dialog box appears.

In the General tab:

• Select the data on the Excel sheet. The Data are in the B to E columns.

• Activate the Multiple pairwise comparisons option in order to identify which packaging (one or more) are responsible for rejecting the null hypothesis in case we reject it.

• Click on the OK button to launch the computations.

• The results are displayed on a new Excel sheet.

 

Interpreting the Results

To begin with, a Summary Statistics chart is displayed. It shows counts and percentage of each category (Yes/No) by the columns selected (Packs). Also, a stacked bar chart is displayed to visualize the distribution of the data.

Then the results of the Cochran's Q test are displayed with a brief interpretation.

We see that the null hypothesis is rejected if we use a significance level of 0.05. As a reminder, choosing a significance level of 0.05 is the same as deciding that we want to be right in 95% of cases (and wrong in 5% of cases) when rejecting H0.

For the decision makers, that means that choosing a packaging at random or using a manager’s preference would not make sense as long as the sample of teenagers has been well chosen. We now want to identify if one or more packaging is responsible for rejecting H0. The following results give us the answer:

We see that Pack 3 and 1 are different, while Pack 2 and 4 are not statistically different. Therefore, by reviewing the statistics Pack 3 seems to be the best option so far.

 

Related Article

Difference Between Paired and Independent Samples