|
Analyze Factorial DesignGeneral Full Factorial Designs - Analysis of Variance Table |
Use the p-values (P) in the analysis of variance table to determine which of the effects in the model are statistically significant. If you have any interaction effects in the model, you need to interpret them first because a significant interaction will influence how you interpret the main effects. To use the p-value, you need to:
- if the p-value is less than or equal to a, conclude that the effect is significant.
- if the p-value is greater than a, conclude that the effect is not significant.
Example Output |
Analysis of Variance
Source DF Adj SS Adj MS F-Value P-Value Model 9 1.11278 0.123642 6.59 0.007 Blocks 1 0.00500 0.005000 0.27 0.620 Linear 4 1.00556 0.251389 13.41 0.001 ExCoat 2 0.65778 0.328889 17.54 0.001 Alloy 2 0.34778 0.173889 9.27 0.008 2-Way Interactions 4 0.10222 0.025556 1.36 0.328 ExCoat*Alloy 4 0.10222 0.025556 1.36 0.328 Error 8 0.15000 0.018750 Total 17 1.26278 |
Interpretation |
For the metal parts data, the analysis of variance table shows the following:
The p-value of 0.328 for the exterior coating by alloy type interaction is not less 0.05. Therefore, there is no significant interaction effect. That is, there is no evidence that effect of exterior coating on corrosion resistance depends on the type of alloy.
- ExCoat: the p-value of 0.001 for the effect of exterior coating on corrosion resistance is less than 0.05. Therefore, you conclude that there is a significant effect. That is, there is a difference in the mean corrosion resistance among the three different coatings. Use the table of means to look at this difference.
- Alloy: the p-value of 0.008 for the effect of alloy on corrosion resistance is less than 0.05. Therefore, you conclude that there is a significant effect. That is, there is a difference in the mean corrosion resistance among the three alloy types. Use the table of means to look at this difference.
The p-value of 0.620 for blocks is not less than 0.05. Therefore, there is no evidence of a significant block effect. That fact that the data was collected in two blocks (each block represents a different day) does not have a significant effect on the corrosion resistance of the metal part.
You may want to refit the model excluding the non-significant interaction and block terms.