Choosing a panel provider for your DCE

Panel quality is the single biggest determinant of DCE data quality that is outside the researcher's direct control. Choosing and briefing your panel provider correctly is essential.

This article explains how to select a panel provider for DCE fieldwork, what specifications to include in your panel brief, and how to use independent quality checks to monitor data quality during fieldwork.

Ben White, 07.07.2026

The panel quality problem

Online access panels are the most common source of respondents for academic and commercial DCE research. They are fast, affordable, and can deliver large samples quickly. They are also subject to well-documented quality problems: professional survey takers, fraudulent respondents, bots, and panel members who do not match their claimed demographics.

The consequences of poor panel quality in a DCE are particularly severe. Fraudulent or inattentive respondents bias utility estimates in ways that are not always detectable through standard quality checks. A study with 20% low-quality responses may produce parameter estimates that have weak signal - or in pathological cases where fraud nears the majority, nonsense biased models.

What makes a quality panel provider

Quality panel providers invest in respondent verification, ongoing panel management, and fraud detection. Key indicators include: verification of demographic claims through multiple data sources; regular panel health checks that remove inactive or fraudulent members; transparent incidence rate data; and willingness to provide sample composition reports.

Good practise dictates independent quality controls at 20% interim review and completion. One unfortunately needs a dose of scepticism and be able to verify all the panel's claims independently and with evidence. Quality is an acceptance criterion, not a cost driver.

For specialist populations - patients, healthcare professionals, business decision-makers - standard panels are rarely appropriate. Specialist recruitment through patient advocacy organisations, professional associations, or targeted digital channels is required. But note - in some cases with low operational experise, e.g. PAG's, fraud can creep in through weak controls.

TLDR Quick links

SurveyEngine respondent sourcing

Request a sample estimate

Open a free account

Specifying and managing panel fieldwork

Step 1: Write a detailed panel brief. Include: target population definition with all eligibility criteria; sample size and quota specifications; survey length and completion rate target; fieldwork timeline; data quality requirements; and what happens at the interim review.

Step 2: Specify quality controls in the contract. Include: duplicate detection, bot detection, minimum completion time requirements, and your right to reject data that fails your quality criteria without paying for it.

Step 3: Set up an interim review at 20% of gross sample. At the 20% mark, review completion rates, consistency check pass rates, response time distributions, and the demographic composition of the sample. Flag any anomalies before the remaining 80% is collected.

Step 4: Run independent fraud checks. Do not rely solely on the panel provider's fraud controls. Run your own checks in SurveyEngine: response time monitoring, consistency checks, speeder detection, and the fraud detection toolkit methods.

Step 5: Reserve the right to reject and resample. Specify in the contract that you will accept payment only for respondents who pass your quality criteria. Agree the threshold before fieldwork begins.

Worked example - managing a multi-country DCE panel

A multi-country health preference study requires 300 respondents in each of five countries (Germany, France, UK, Spain, Italy). SurveyEngine manages the recruitment across five country-specific panel providers, each briefed with identical eligibility criteria and quality specifications.

At the 20% interim review, the Spanish panel shows an unusually high consistency check failure rate (31% vs 8–12% in other countries). Investigation reveals a translation issue with the consistency check design - the dominant alternative is not clearly dominant in the Spanish attribute label translation. The translation is corrected and the Spanish sample is restarted. The final study achieves consistency check pass rates of 88–93% across all countries.