Struggling to Explain Statistics to Others?
Use Charts to Explain the Meaning of Statistical Tests
While it's easy to run an ANOVA or t-Test to get a p-value, it can be hard to explain to coworkers and leaders exactly what that means. Solution: Turn your data into a chart or two.
ANOVA or t-Test for Means
Using the boxwhisker.xls data provided with QI Macros on your PC at My Documents/QI Macros Test Data, we can use ANOVA to determine if the means (i.e., averages) are the same or diffierent.
The p-value (cell K13) is zero, so we can reject the null hypothesis that the means are equal. But all of the information give can cause confusion, so we can use the Box and Whisker chart to "show" the difference. Keyway2 is high; Keyway4 is low; and Keyway's 2 and 4 are very similar.
We could also use the Box and Whisker with t-Test data comparing two columns of data.
Levenes or F-test for Variances
We can use Levenes test to determine if the variances are the same or different:
The p-value (cell F9) is less than 0.05, so we can reject the null hypothesis that the variances are equal. But what does that mean to a non-statistician? Not much. So we can use the box and whisker plot above (distance between ends of whiskers) or turn this data into a control chart that shows the variation.
First we use QI Macros Restacking tool to stack columns A:D into one column:
At the prompt, answer 1.
Then insert a blank row between the data for each keyway and run an XmR chart. QI Macros will recalculate a new set of control limits whenever it detects a blank row in the data. The resulting XmR chart with separate control limits for each of the four keyways will show the differences in the averages as well as the wider variation in Keyway 2 and 4:
It's also easy to see that Keyways 2 and 4 have a much wider variation in the ranges.
Turn Your Statistics into Charts that Everyone Can Understand
Once you have a significant difference, it's easy to use QI Macros Box & Whisker or Control Charts to turn that difference into a chart that you can explain to anyone.