Pre-identifying protest respondents in environmental DCEs

Protest responses can be partially anticipated before the choice tasks begin. Pre-screening attitudes can improve data quality and inform your analysis strategy.

This article explains how to pre-identify likely protest respondents in environmental DCEs using attitude questions before the choice tasks, and how to use this information in data quality analysis.

Ben White, 07.07.2026

The pre-identification challenge

Protest responses are typically identified post-hoc using debriefing questions after the choice tasks. But by this point, protest data is already in your dataset. An alternative approach - pre-identifying likely protest respondents using attitude questions before the choice tasks - allows you to flag data for sensitivity analysis more robustly and to understand the relationship between prior attitudes and protest behaviour.

What predicts protest responses

Research on protest responses in environmental valuation has identified several strong predictors: distrust of the delivering institution (government, utility company, conservation body); principled objection to the payment vehicle (taxes, donations); scepticism about whether the described improvements would actually be delivered; and general opposition to monetising environmental goods.

Measuring these attitudes before the choice tasks gives you a pre-treatment measure of protest risk that is not influenced by the choice task experience itself. This is more reliable than post-hoc debriefing for separating genuine preferences from protest motivation.

TLDR Quick links

Environment resources

Protest responses article

Open a free account

Implementing pre-screening in SurveyEngine

Step 1: Add attitude questions before the choice tasks. Include 3–5 questions measuring: trust in the delivering institution, credibility of the payment vehicle, belief that the improvements would be delivered, and attitude toward environmental monetisation. Use 5-point Likert scales.

Step 2: Create a protest risk score. Combine the attitude responses into a composite protest risk score. In SurveyEngine, create a calculated variable that sums the responses (after reverse-coding where appropriate). Flag respondents above a threshold for sensitivity analysis.

Step 3: Do not exclude pre-flagged respondents from the choice tasks. Pre-screening is for analysis purposes only. All respondents should complete the choice tasks. Excluding pre-flagged respondents would bias your main estimates and may not be ethically appropriate.

Step 4: Compare pre-flagged and post-flagged identification. After fieldwork, compare the respondents identified as likely protesters by pre-screening with those identified by post-hoc debriefing. Agreement between the two measures strengthens the validity of both.

Step 5: Run models stratified by protest risk. Estimate WTP models separately for low, medium, and high protest risk groups. If WTP differs significantly across groups, this confirms that protest motivation is inflating the overall WTP estimate.

Worked example - wetland restoration pre-screening

A wetland restoration valuation study adds four pre-screening attitude questions: trust in the Environment Agency (very low to very high), belief that the described wetland improvements would actually be delivered (very unlikely to very likely), attitude to paying through council tax for environmental improvements (strongly oppose to strongly support), and general attitude to placing monetary values on nature (strongly oppose to strongly support).

Pre-screening identifies 22% of respondents as high protest risk (bottom quartile on combined attitude score). Post-hoc debriefing identifies 19% as protest respondents. Agreement between the two measures is 78% - pre-screening identifies most but not all protesters, and some pre-screened protesters give non-protest responses in the choice tasks. Both measures are reported in the study.