In: Statistics and Probability
The table below shows the quantity of watches made by a company for different plant layout and shift times.
Weekly quantity of wrist watches produced in a factory |
|||
Shifts (hours) |
|||
Layout |
6 |
8 |
12 |
1 |
300 |
400 |
900 |
350 |
350 |
950 |
|
450 |
490 |
850 |
|
330 |
500 |
800 |
|
2 |
80 |
250 |
500 |
100 |
300 |
450 |
|
60 |
190 |
550 |
|
150 |
240 |
600 |
|
3 |
700 |
800 |
1200 |
600 |
900 |
1800 |
|
750 |
680 |
2000 |
|
800 |
720 |
2200 |
Let's do the two-way ANOVA in excel. Enter the data in excel as follows. Then go to Data -> Data analysis -> two-way anova with replication and enter the following:
The output is:
Anova: Two-Factor With Replication | ||||||
SUMMARY | Shift 6 | Shift 8 | Shift 12 | Total | ||
Layout 1 | ||||||
Count | 4 | 4 | 4 | 12 | ||
Sum | 1430 | 1740 | 3500 | 6670 | ||
Average | 357.5 | 435 | 875 | 555.8333 | ||
Variance | 4225 | 5233.333 | 4166.667 | 60371.97 | ||
Layout 2 | ||||||
Count | 4 | 4 | 4 | 12 | ||
Sum | 390 | 980 | 2100 | 3470 | ||
Average | 97.5 | 245 | 525 | 289.1667 | ||
Variance | 1491.667 | 2033.333 | 4166.667 | 36390.15 | ||
Layout 3 | ||||||
Count | 4 | 4 | 4 | 12 | ||
Sum | 2850 | 3100 | 7200 | 13150 | ||
Average | 712.5 | 775 | 1800 | 1095.833 | ||
Variance | 7291.667 | 9433.333 | 186666.7 | 326644.7 | ||
Total | ||||||
Count | 12 | 12 | 12 | |||
Sum | 4670 | 5820 | 12800 | |||
Average | 389.1667 | 485 | 1066.667 | |||
Variance | 72862.88 | 56990.91 | 368787.9 | |||
ANOVA | ||||||
Source of Variation | SS | df | MS | F | P-value | F crit |
Layout | 4053689 | 2 | 2026844 | 81.17901 | 3.8E-12 | 3.354131 |
Shifts | 3226106 | 2 | 1613053 | 64.60586 | 5.11E-11 | 3.354131 |
Interaction | 757244.4 | 4 | 189311.1 | 7.582273 | 0.000313 | 2.727765 |
Within | 674125 | 27 | 24967.59 | |||
Total | 8711164 | 35 |
If we take a look at the p-value column, we can see that that there is a significant difference in mean production between layout and shifts. Also, there is a significant interaction effect between both the variables. Hence, we can find out a layout & shift in which maximizes production. Let's take a look at each table for different layouts and observe the row which tells us the averages (marked in italics), we can see that the highest average is of Layout 3 and Shift with 12 hours having an average of 1800 (marked in bold).
Hence, layout and shift which maximizes production is Layout 3 and shift with 12 hours.
2. The difference between a one-way anova and two-way anova is that the former has only one independent variable and the latter has two or more independent variables. In this case, we have two independent variables, Layout and Shift, hence this is a two-way ANOVA. If there was only one independent variable, Shifts or Layout, it would have been a one-way ANOVA.