Full Factorial Analysis

DOE-Full factorial

Graphical Analysis
-Run Chart


We see that variation over time looks like constant because points in mR-Chart remain between limits.
On the other hand, the process perse looks like constant over time because results in X-Chart remain between limits.
From this analysis we can see that long-run variation is not different than short-run variation and the process isstable over time.
Moreover, all difference in data could be caused by noise and there are not unexpected points.

-ANOG

In this analysis we can see that C and FC are driving results.
The factorsthat have most effect in result are:
* Brochure Main Factor
* Order Card Front/Brochure Interaction

-Statistically Significance
From data, we can calculate the standard error of effects.Our sample size is 120,336 (15042*8) and our success cases are 4638 (sum of orders) so, our success rate is 0.038542. With those data and formula shows in document we have:
Standard erroreffect=4×0.038542×0.961458120,336=0.001110
So, an effect is significant if its estimate has absolute value greater than 2*Standard Error
2×Standard Error=0.0022197
Now, we can decide if some practical effectreally is important or only was by chance.

-Main Effect Plot

* Effect F: 0.0393-0.0378=0.0015
* Effect B: 0.038-0.039=0.0004
* Effect C: 0.037-0.04=0.003
From this analysis, we cansee that although factors F and C have effect in response (steep slope), only C is statistically significant.
When Dr Phil is in front of brochure, causes a 0.3% less orders when Kelly Ripa is infront of brochure.
In other words, different persons in front of brochures have important effect over number of orders received. The other factors (F and B) have not statistical significance so, couldbe produced by chance.
In conclusion, among main effects, we can say that only Brochure has important effect over results.

-Interaction Plot
Order Card Front / Order Card Back (FB) Interaction...