Separating Analysis from Advocacy
What you'll learn: How to distinguish between presenting objective evidence and making value-based recommendations in data science work.
As a data scientist, you wear two hats—but they're fundamentally different. Analysis means presenting what the data shows: trends, correlations, predictions, uncertainties. Advocacy means recommending what should be done based on values, priorities, and trade-offs that go beyond the numbers.
The Core Distinction
Analysis answers: "What does the data tell us?" Advocacy answers: "What should we do about it?"
For example, your model might show that Policy A reduces costs by 15% but increases wait times by 20%, while Policy B does the opposite. That's analysis. Recommending Policy A because "efficiency matters most" is advocacy—it injects your (or your organization's) values into the decision.
Why This Matters
When you blur these lines, you risk:
- Losing credibility: Decision-makers can't tell where facts end and opinions begin
- Hiding trade-offs: Your values might mask important considerations others prioritize differently
- Undermining trust: Stakeholders may feel manipulated rather than informed
Practicing Separation
Present findings clearly: "The data shows X, Y, and Z outcomes under different scenarios." Then, if asked for recommendations, be explicit: "Based on the priority of cost reduction, I would suggest..." Make your assumptions and values visible.
Think of yourself as a mapmaker, not a tour guide. You can show all the routes, highlight obstacles, and estimate travel times—but the traveler chooses the destination based on what they value.
Key Takeaway: Strong data science clearly separates objective findings (analysis) from value-based recommendations (advocacy), making your assumptions transparent and empowering decision-makers to apply their own priorities.