Skip to content

Factorial vignette experiments

Factorial vignette experiments combine the realism of a survey with the causal-inference power of a randomised experiment — short structured scenarios with experimentally varied attributes, widely used in sociology, public health, social policy and management research. Random assignment of attribute levels across respondents lets you identify the marginal effect of each attribute on the evaluation, which a non-experimental observational survey cannot do.

tickStat is a survey and experiment platform built by academic researchers, for academic researchers. Vignette experiments reuse the choice-experiment design infrastructure (full and fractional factorial designs, balanced randomisation across respondents) but present each scenario in single-card rating form — slider or Likert — instead of a forced choice between alternatives.

Why tickStat for vignette experiments

  • Full and fractional factorial designs without leaving the platform. Specify attributes and levels; the platform generates the design (full factorial when feasible, fractional when the design space is too large) and assigns scenarios to respondents using balanced randomisation so every level appears with the right frequency.

  • Slider or Likert response, not just choice. Vignette evaluation is a rating task, not a choice. tickStat exposes both continuous (slider) and ordinal (Likert) response formats, so the right scale for your construct is one click away.

  • Within- and between-subject designs. Decide how many vignettes each respondent rates (between-subject as N=1, within-subject as N>1), and the platform handles balancing across respondents so the marginal-effect estimates remain unbiased.

  • Analysis-ready data export. Per-respondent scenario plus rating, attribute levels coded, ready for OLS, multilevel or fixed-effects regression in R, Stata, SPSS or your tool of choice — no manual reshaping.

  • Reproducible by design. Factorial design rules, randomisation seeds, attribute coding — every setting is stored with the survey definition. Reproducing the experiment means re-running the same design, not chasing a screenshot of an online wizard.

  • Multilingual fielding in 14 languages. Run the same vignette design across countries with a single survey definition.

  • GDPR and EU AI Act compliant. Important for academic researchers running publication-grade studies under the current regulatory regime.

What's on this page

Below: how to configure the vignette attributes and levels, choose the response scale (slider or Likert), control the design size with full or fractional factorial, and export the resulting respondent × scenario × rating data for regression analysis.

If you are new to the platform, start with the Getting started guide.

Experimental vignettes — short, structured scenarios in which several attributes are systematically varied across respondents — are a workhorse method in sociology, public health, social policy, environmental economics and management research. Vignette experiments combine the realism of a survey with the causal-inference power of a factorial experiment: by randomly assigning attribute levels to scenarios, the researcher can identify the marginal effect of each attribute on the respondent's evaluation.

tickStat implements vignettes as a configuration mode of the Choice question infrastructure, reusing the same experimental design machinery (full and fractional factorial designs, balanced randomisation across respondents, blocking) that underlies the Discrete Choice Experiment module — but presents each scenario in single-card rating form rather than as a forced choice between alternatives.

When to use it

  • You want to estimate how respondents value a set of attributes that combine in a scenario (a job offer, a policy package, a conservation programme, a clinical-care pathway).
  • You want a more naturalistic stimulus than a pure choice card — vignettes can include narrative prose, images and contextual framing.
  • You want a Likert-type or continuous rating outcome rather than a binary choice.

How it works in tickStat

Configuration is similar to a discrete choice experiment with a single alternative:

  • Define attributes (the dimensions varied across vignettes) and their levels.
  • Generate the experimental design (full factorial, fractional factorial, or upload your own design matrix).
  • Choose the visualisation method: slider (continuous rating from 0 to 100) or button (Likert-style discrete scale).
  • Use the standard scenario template — placeholders are filled with the assigned levels for each respondent.

Each respondent sees a sequence of vignettes drawn from the design, with attribute-level combinations randomised so that the design is balanced across the sample.

To enable vignette mode in the question configuration, activate the "Visualisation Slider Vignettes" checkbox and set the "Generation method design" field to "Vignettes".

Captured data and analysis

For each vignette shown to each respondent, tickStat records the full vector of attribute levels and the rating given. The data exports cleanly into a long-format dataset suitable for:

  • Linear, logistic or ordinal multilevel regression with attribute dummies as predictors and respondents as a random effect (the standard analysis for factorial vignette studies).
  • Conjoint-style attribute-importance decompositions.
  • Heterogeneous treatment effects via interactions between attribute levels and respondent characteristics.