JMP Problems

[thefizwidget]

I am extremely unfamiliar with the program JMP. Could anyone show me how to find the answers to problems like these using the program?



1. What is the mean of the difference scores obtained in the sample?
2. What is the obtained value of t?
3. What is the obtained p value for t?
4. What are the correct degrees of freedom for the t-statistic?
5. What is the correct statistical decision?
6. What is the probability of your having made a TYPE I Error?
7. What is the probability your having made a TYPE II Error?
8. What is value of the Wilcoxon for this set of data?
9. What is the associated p-value of the Wilcoxon?
10. Based on the Wilcoxon, should you reject the null hypothesis:
11. What is the primary threat to the internal validity of this study?
12. What could be done to improve this study (e.g. what changes in methodology)?
13. Is the assumption of normality reasonable?

[WeeG]

If I understand you properly, then here is some of the things you want to do:

1. Analyze --> Distribution --> Choose the difference variable, and press on the "Y,Columns" button ---> OK --->you'll find the mean under moments.

2. I don't understand what "t" you are looking for. If you mean a statistic for a one sample t test, then in the window of the "Distribution", you need to press on the lower red arrow and then choose "test mean", enter a specific hypothesis and you'll get your t.

3. Same as 2.

4. Same as 2.

6. I think it's 5% unless you change it, I didn't find where.

7. In order to find you can press on the red arrow of "Test Mean = Value", then you choose power animation, it will show it to you.

8,9. You'll see it when you do the t test, you'll have an option to add Wilcoxon by clicking once on the box, you'll see.

13. on the same window ("Distribution"), when you open the menu with the red arrow, you will get "Normal Quantile Plot", press on it and according to the graph decide if it's normal or not (you can also produce a teat for normality, choose a normal fit and then in the sub menu you can choose a goodness of fit test).

Play with it a bit, you'll find it, JMP is easy after you learn to use it, it got a different style then other software.

[thefizwidget]

When running the "test mean", what do I enter for the hypothesized mean? Do i use the default 0 or do I use the actual mean (6.9). Pardon my ignorance and thanks in advance.

[WeeG]

when you make a one sample t-test, you check if the mean of the population equals some number. in the "test mean" you need to specify your hypothesis, in other words, you need to tell JMP what exactly is it that you want to check.
In your case, of difference between two variables, it makes lot's of sense to enter 0 in that field. do you understand why ?