Choose Language Hide Translation Bar
Community Member

How to do a hypothesis test to check if a new data point matches model prediction?

I have fit a multivariate linear response model to my data set in JMP 10, with four input factors, and one output. I know how to use this formula to generate the predicted output mean, and the predicted upper and lower 95% confidence intervals for that output mean, for given values of the four input factors.

So now I have done a new experiment, and I get a measurement (output) from this experiment. I can tell by inspection that my measured output value (say, 29) does fall inside the 95% confidence intervals (say, 15 to 35) for the predicted output, based on the model. But I want to do a hypothesis test (t-test, I think?) to test the hypothesis that my mean (29) is equal to the predicted mean (27), and at what confidence level. (So, yes my new data point falls within the 95% confidence interval, but can I say I am 97% confident? 99% confident?).

I am not sure that I am using the correct terminology here, as I am fairly new to statistics, but any help would be appreciated. I am happy to provide additional information or clarification if needed. Thank you.

0 Kudos