Main Effect - Interaction Plots - 2 Level Factorial DOE with 18 Runs

A

Adebar

Hi all,

I have an (very) urgent problem:

I did a 2level factorial doe with 18 runs.
now i'm analyzing that whole thing which produces a huge amount of data. I have round about 190 maineffect and interaction plots which need to be interpreted. Is there any way to "translate" the information of those plots?

In the end I'd like to have a chart which represents all the data of the plots....

Thx for you quick help :bigwave:
 

Miner

Forum Moderator
Leader
Admin
The number of plots seems excessive for this size of experiment.

Try plotting graphs in multiple panes per page under the multiple graphs options, or use the Layout tool under the Editor menu (Note a graph must be selected to see this tool) to combine graphs on a page.
 
A

Adebar

hmm yeah also thought about the layout option but I need to integrate those plots in my thesis which means that I have to scale them down--> they gonna be to small and even so the reader has to have a look on every of the 176 charts.... spreadsheet would be perfect....

oh the amount of data is due to the fact that I did the doe with 12 different samples which are all correlated to each other......
 

Miner

Forum Moderator
Leader
Admin
Right click on the charts and select copy. You can then paste them into a spreadsheet.
 
A

Adebar

no I don't want to have the graph itsself in a spreadsheet, but the information of the graph.... some numbers which tell me the exact same thing as the graph....
 

Miner

Forum Moderator
Leader
Admin
You could always create tables that summarize the graphical content, but tables typically hide the relationships that graphs make clear.
 
A

Adebar

I know that I always loose information but in the end the reader understands more if there is a table instead of 176 graphs.... An those graphs will be in the annex of my thesis anyway.... But still I have to get the information of the graphs into a table
 

Statistical Steven

Statistician
Leader
Super Moderator
How many factors can you have in 18 runs with 2 levels? What does the ANOVA table tell you? Only present the significant 2 factor interactions. Do not try to present any graphs of 3 factor or higher interactions, since they are hard to interpret. KISS principle works best for me.
 
A

Allattar

I am missing something somewhere, 2 level factorial 18 runs... 190 main effect and interaction plots.

How many factors do you have?
Would I be amiss in assuming you have 2 centre points?
I cannot seem to see how you get so many charts, you must have a lot of factors to have that many graphs, but if it is 16 runs +2 centre points 9 factors or higher are resolution III and interactions should be confounded with main effects.
6-8 factors its resolution 4 and a lot of interactions are confounded and needn't be plotted.
4 -5 factors can show 2 way interactions unconfounded with main effects or each other.

Anyway another option is have you tried using the response optimiser instead?
 
Top Bottom