🔎 What This Paper Asks
Most studies of candidate race assume that the way race is signaled—by a label/name or by a photo—does not meaningfully change results. That assumption goes largely untested. This study asks whether signaling candidate ethnicity with ethnic labels and names produces different effects than signaling ethnicity with ethnically identifiable photos and names.
🧪 How Ethnicity Was Signaled and Tested
A nationally representative sample of white voting-age citizens participated in a modified conjoint experiment designed to compare two treatment modes directly:
- One condition signaled candidate ethnicity using ethnic labels and names.
- The other condition signaled ethnicity using ethnically identifiable photos paired with names.
The design isolates whether the mode of racial/ethnic signaling alters voters’ choices, rather than assuming mode differences are negligible.
📌 Key Findings
The experiment finds clear, robust differences by treatment mode:
- Treatment-mode effects are large in substantive terms and statistically significant.
- Effects from label-plus-name signals differ meaningfully from effects produced by photo-plus-name signals.
- These observed mode differences are not consistent with existing theoretical accounts about how racial or ethnic cues should operate.
⚠️ Why This Matters
The results challenge a common methodological assumption in experimental work on race and voting and show that choice of signal can shape conclusions. This inconsistency with current theories indicates a need for new theoretical development and further empirical tests to understand how and why different modes of signaling race/ethnicity produce distinct effects.
