What’s Driving the Difference?

 

Two groups can look different. A model can help show which differences actually stand out.

​

Rather than comparing one characteristic at a time, the model considers multiple factors simultaneously. That helps separate patterns that still stand out from those that are mostly explained by other differences between respondents.

Use the tool to compare promoters with other respondents, then explore which factors appear most closely associated with promoter status.

 

What this demonstrates
Statistical modeling • Adjusted comparisons • Translating complex analysis • Association vs. causation

​

Data: 2025 Greater Boston Jewish Community Study
Created by Daniel Parmer, PhD

 

About this experiment:
A score of 9 or 10 is treated as a promoter. The model then examines characteristics including familiarity with CJP, perceptions of the Jewish community, charitable giving, volunteering, Jewish connection, income, and other demographic factors.

​

The percentages shown in the tool are adjusted estimates, not simple raw percentages. Imagine groups of respondents who are similar on the other characteristics included in the model but differ on the factor being explored. The percentages estimate how likely respondents in each of those groups are to be promoters.

Choose a measure and change its framing to see how those choices affect the impression the data create.

 

Want to see another experiment?