I'm using Minitab 15,

I have created a design and now want to enter the response data that has been collected. For each treatment I have over 100 values for the ONE response I am measuring. I see these as repeats, hope I'm right.
So, I have 70 treatments, each with 100 measures of the response.


Factor1 Factor2 Reponse(Run1) Response(Run2)
Treatment1 34 56 6 7
Treatment2 45 67 8 4
..
Treatment70 24 45 4 5


My problem is getting them into the model. I can paste them into the worksheet but don't understand how to get it to take them as multiple measure of the same response variable. It keeps taking them as multiple different reponse variables.

Why so many measurements for each setting of the factors? Perhaps you have a large intrinsic variation and want to average out the random variation. But then you might want to work on that challenge before trying to optimize the settings.

As far as I know, Minitab will always treat each column of responses as a separate variable - but only 25 maximum. Minitab help says:

"Enter up to 25 numeric response data columns that are equal in length to the design variables in the worksheet. Each row will contain data corresponding to one run of your experiment. You may enter the response data in any column(s) not occupied by the design data. The number of columns reserved for the design data is dependent on the number of factors in your design. If there is more than one response variable, Minitab fits separate models for each response."


One simple option would be to calculate a few extra columns from the 100 data points - like the average and st dev - perhaps max and min as well, or 10th %tile and 90th %tile. Then you could use these new summary columns in the analysis. You would see how the factors affect the average and how they affect the spread.



Also, 70 different trials seems like a very big design, especially of you really only have two factors. Typically, you are fitting a quadratic curve to the data. It would be overkill to have 70 different values for just two variables and still try to fit just a quadratic surface. 70 trials would be more typical of a design with 5-7 factors.


Tim F

I'm using Minitab 15,

I have created a design and now want to enter the response data that has been collected. For each treatment I have over 100 values for the ONE response I am measuring. I see these as repeats, hope I'm right.
So, I have 70 treatments, each with 100 measures of the response.


Factor1 Factor2 Reponse(Run1) Response(Run2)
Treatment1 34 56 6 7
Treatment2 45 67 8 4
..
Treatment70 24 45 4 5


My problem is getting them into the model. I can paste them into the worksheet but don't understand how to get it to take them as multiple measure of the same response variable. It keeps taking them as multiple different reponse variables.

You need to stack the Response from each run into a single variable called response, and then create a new variable called run that would tell you the run that each response is associated with. You then have a factor called run in your model that quantifies the run to run variability (or the variability between the measurements)