I have conducted the trials in the following table.
|
Actual Values |
Statistics Jargon |
||||||
| Trial |
r (g/ml) |
t (sec) |
T (F) |
r |
t |
T |
Q |
| 1 |
0.035 |
10 |
205 |
-1 |
-1 |
+1 |
2.5 |
| 2 |
0.075 |
10 |
185 |
+1 |
-1 |
-1 |
3 |
| 3 |
0.035 |
300 |
185 |
-1 |
+1 |
-1 |
4 |
| 4 |
0.075 |
300 |
205 |
+1 |
+1 |
+1 |
4.1 |
This study is a 2-factor partial factorial design, indicated as 23-1. The good people who write for Wikipedia tell me that this experiment design gives “main effects” but those may be confounded with two-factor interactions. Indeed, there is some reason to be concerned for the simple quality function evaluation in this situation—for example, I worry that row 1 and row 2 could easily produce similar effects. C/W is lower in 1, but the temperature is higher, so extraction should be similar to a higher ratio with lower temperature. In any case, if the result is not revealing, the experiment may be augmented with the remaining 4 cases in the factorial design.
The main effects plot shows how each parameter influences the result, though recognize that interactions may be masking main effects. The main effects plot for this experiment is shown below. From this it would appear that extraction time is the most sensitive parameter—to me a surprising observation. I had certainly expected the coffee/water ratio (cwrat) to be the dominant variable, but here it does not appear to be so. The negative correlation with temperature is very surprising, and indeed I am skeptical. Still, the variation appears to be quite small, and is likely due to two-factor interactions.
I would be inclined to fix the temperature and the C/W ratio at their midpoints, and vary the time alone; however, I believe the response hypersurface is more complex. I shall press forward with the response surface methodology on all three variables.
In case you want to see my data.
| Trial Order | Q | Date | Comment |
| 1 | 2.5 | 2/16/09 | Dishwatery taste and aroma. Very simple palate with almost no mouth feel. Equivalent to mediocre drip. |
| 2 | 3 | 2/17/09 | Not very aromatic and too simple. Has nice bitterness and some acidity but not well balanced between them. |
| 3 | 4 | 2/18/09 | Mildly aromatic. Good mouth feel with acid/bitter well balanced. Slight char flavor. |
| 4 | 4.1 | 2/19/09 | Too bitter. The aroma is marvelous being very rich and tantalizing. Mouth feel is creamy. If bitterness where attenuated this would be sublime. |
I wasn’t going to use fractions, but I decided I needed them. Note that trial 4 was assigned a quality of 4.1, which is really a note saying that this was a little bit better than trial 3, but not much.