JMP Tutorial:

Confidence Interval and Hypothesis Test for a Proportion

A random sample of n=150 Stat 201 students in Spring 09 revealed that 91 of them were born in Tennessee.

Create a JMP data table as follows. Notice that 91 + 59 = 150 = sample size.

Go to the **Analyze** menu and select **Distribution**:

Click the column **Born in TN?** and then click **Y, Columns**.

Click the column **Number** and then click **Freq**.

You Should see:

Click **OK**.

This will produce the following:

To create 95% confidence intervals, click on the red down arrow next to **
Born in TN?**
and select **Confidence Interval**, then **0.95**.

You should see the following additional output:

**Hypothesis Testing**

Click the red down arrow next to** Born in TN?** and select **Test
Probabilities**.

Suppose, prior to collection the data, someone thought that 50% of all Stat 201 students in Spring 09 were born in Tennessee. You disagree, but you are not sure if it is above or below 50% (i.e., a 2-sided alternative hypothesis).

Fill in the resulting dialog box as follows:

Click **Done**:

You should see the following output:

Notice that the Pearson p-value above is 0.009, and will match the p-value obtained using the z test statistic approach. In this case, for a 2-sided alternative, JMP uses the square of the z test statistic, called the Chi Square test Statistic. The two method are mathematically identical.

Note: If your data are "unaggregated", such as 150 rows of the following:

all the above instructions apply, except you place nothing in the **Freq**
text box, like this: