Not worth your time
Offering average savings (A “noteworthy” deal)
So good you should skip work and go shopping
The bar chart helped me to organize my data so it was easy to analyze, but it didn’t help me determine statistical significance. Sometimes bar charts are all you really need to make your data come to life, but there are other times when showing statistical significance is essential—especially when money-saving quality improvement projects are on the line.
To find if there was a statistically significant association between coupon favorability and day, I placed my most recent data into a Minitab two-way table and then performed a Chi-Square Test:
Analysis
You can see from the Minitab output that at an alpha level of 0.05, the p-value of 0.042 reveals that there is an association between coupon favorability and day. Statistical significance has been found!
Chi-Square tests are neat because they compare the observed distribution of your data to their expected distribution (in accordance with the null hypothesis). To identify outcomes with the greatest impact, we can look for the highest Chi-Square contribution values.
For example, the contribution of 6.477 tells me that I’ve received more “So good you should skip work and go shopping” coupons (22 of them) on Thursday than was expected (12.87 coupons). The Chi-Square results makes it easy to see that Thursdays are the best day to open e-mails that I consider “so good you should skip work and go shopping,” followed by Mondays and Tuesdays.
The coupon totals horizontally and vertically make it simple to track the percentages of each coupon type that I classified. I found that I ranked more coupons as “noteworthy” than any other category. And unfortunately, I found there were fewer “so good you should skip work and go shopping coupons” than any other category. I think I’m learning to outsmart those pesky retailers!
Thanks to Michelle Paret for her contributions to this blog!
Image licensed under Creative Commons Attribution ShareAlike 2.0.