Since we didn’t choose “pool” option, the hypothesis test conclusion here is not fully correct, so that is marked out. In the output, we see the confidence interval. In the dialog box, we will put in the 99%, make sure the alternative hypothesis is “not equal”, and change the option to “Estimate the proportions separately.” We did not learn about Fisher’s exact test in our course, so do not quote this result.įind a 99% confidence interval for the difference of p1 and p2. The output also gives a second p-value from a different test that could be used on these data. The method we chose, of pooling the data to estimate the proportion for the standard deviation, is appropriate for a hypothesis test, but NOT for a confidence interval. The output gives a confidence interval, but that is marked out here because you should not use it. The output gives a very large p-value, which isn’t a surprise because the two sample proportions are very close. As that is the default value already in this box, we leave it. Our Ho says the two proportions are equal, so that says the “hypothesized difference” is 0.
Hypothesis tests and Confidence Interval calculations are carried out with the same command. We aren’t doing a confidence interval so the confidence level isn’t relevant. (Independent samples from two populations). Since this is a hypothesis test, we use the pooled proportion to estimate the standard deviation, so choose that in the dialog box. We will test Ho: p1 p2 versus Ha: p1 p2. (The others are done in a similar way to the corresponding situtations in 2-Sample-t.) In this handout, we will focus only on entering summarized data. Columns 4 and 5 have each sample in its own column.
The data is in C1 and the Group is in C2. Columns 1 and 2 have both samples in one column. Here is the worksheet with two different sets of data. There are three ways that the data can be read in: īoth samples are in one column Each sample is in its own column Summarized data
2-Proportion command in Minitab 17 for confidence intervals and hypothesis tests on two proportions Please read the material on the 2-sample-t command before reading this.
0 kommentar(er)
