I had design an experiment and it is a 2k matrix approach, 2^4(4 factors,2levels) x 3 replications = 48 runs. Total output is my response in study. The range of output falls in between 5000-8000. The analysis using minitab shows that there is no significance factor among each combination and individual factors. Any way I could change the situation ?
Dear Goi,
From the observation of your tabulation, it is evident that 'B' factor, combinations of C*D and B*C*D are significant. but the values of P are less.
You need to once again check your variables and their levels.
importantly are they (selected variables and their interactions) really have any influence on output responses?
Finally i don't think replications will have any effect on ANNOVA. It may be useful other analysis like repeatability etc.
Good luck
with regards,
P.Thirumal
You can have a problem of statistical power. The model seems two complex.
Are you having any clues about what factors are influencing the output?
hi, may be you can put alpha=0.1 but most important I believe is to clean your model
I mean that all non signifcant interactions (all 4 and 3 factors interactions can be removed, then 2 factors interactions if p-value>0.1, then main effects ...) then you will have a parcimonious model and better p-values
regards
yes! any interaction (or effect) with a p-value > 0.1 will be removed , step by step
In your case remove first ABCD, then BCD... as long as p-values are >0.1
regards
françois
yes, we eliminate non statistically significant factors (those with pvalue>0.1)
regards
you must remove the analysis less meaningful interactions and will return to perform the ANOVA.
You can change the analysis setting to ignore residuals, by then you can get significant model terms.
remove the non-significant interactions analysis to have more degrees of freedom for error and analyze again
- Badges
- Science method
- Similar topics
- Mathematics
- Statistics
- Statistical Modeling
- Analyses of Variance
- ANOVA