---
title: Mirubiki vs. a traditional US panel — where synthetic testing fits
description: Synthetic pretesting is not a replacement for a fielded US panel. It catches the questionnaire defects that a panel would only reveal after you paid for the field.
---

# Where synthetic testing fits, and where it does not.

A US panel measures a population. Mirubiki tests the instrument before you pay for that measurement. The two answer different questions, and the confusion between them is what makes synthetic research look unreliable.

## Same questionnaire, two jobs

### What Mirubiki does

Finds items that do not measure what they claim, wording that pushes an answer, scales that collapse, and routing that loses a segment. A run costs minutes.

### What a fielded panel does

Produces the number you report: incidence, share, margin of error, on a sample drawn from the population you are addressing.

### Where they meet

Pretest with Mirubiki, fix the instrument, then field the corrected version. The panel stops paying for mistakes the questionnaire already contained.

## What we do not claim

Mirubiki does not replace a probability sample, and does not produce a margin of error. It is a pretesting instrument: its output is a diagnosis of the questionnaire, not an estimate of the population.

The US panel is calibrated on public survey data. Where the data does not support a segment, the answer is an abstention, not a guess.

## Try it on a questionnaire you already fielded

The clearest test is a study whose results you know. Run the same items through Mirubiki and compare the diagnosis with what the field actually returned.

[Run a test survey](/login) [How the panel is built](/us/personas-synthetiques)