turn on suggestions

Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

- JMP User Community
- :
- Discussions
- :
- Confidence Intervall DOE

Topic Options

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Apr 26, 2016 11:24 PM
(454 views)

Hello there,

I have some issues concerning the evaluation of my results of a Design of Experiment which I conducted.

To Confirm my Response-surface-model, I conducted 6 confirmation runs. To check whether my model is able to describe the system I need it's confidence interval. But there is a problem. First I used the profiler in the prediction-model-window. But those values seemed to be to small, so I went on "Save Columns" and I saved the Formulas for "Individuals Confidence Formulas".

Using them in the prediction profiler graph, the results were different from those in the fit model window (See picture below) . How is that possible?

My second question is, what is the difference between the two formulas "Mean Confidence Limit Formula" und "Individual Confidence Limit Formula"?

I hope anyone can help.

Alexander

(I'm not a native speaker so I hope you can understand my problem even if there are some issues concerning language)

3 REPLIES

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Apr 26, 2016 11:25 PM
(299 views)

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Apr 28, 2016 7:26 AM
(299 views)

The confidence interval is for the mean response. The prediction or individual interval is wider because covers both the uncertainty in the mean and the error variance of the individual observations.

Alternatively, you could use a one-sample t-test with your sample of 6 observations against your predicted or expected mean.

Learn it once, use it forever!

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

May 2, 2016 6:07 AM
(299 views)

Thank you for answering my question.

If i would use a t-test to compare my experimental results with my predicted value I would calculate my t-value using the standard deviation of my repeated measurements?

If I want to compare my experimental results (every confirmation experiment was measured three times to calculate the mean) with my predicted values is it better to chose the "confidence interval for the mean" or the "confidence interval of the individual measurements"?