The Simpson's Paradox Checker

Check whether this aggregate result flips when I break it down by group.

The finding: [DESCRIBE THE OVERALL RESULT]
The subgroups in the data: [DESCRIBE — E.G. BY SEGMENT, REGION, PERIOD]

Provide:
1. Whether the overall pattern could reverse within subgroups (Simpson's
   paradox)
2. A worked illustration of how that can happen with my kind of data
3. What to check in my data to catch it
4. Which view — aggregate or segmented — is the right basis for my
   decision
5. The confounding variable most likely at work

Protect me from acting on an aggregate that hides the opposite truth in
every subgroup.