Author: Trial Design

A trial phase doesn't store a fixed list of trials — it generates them from a design. The Design tab (inside Author, for a selected trial phase) is where you describe that design: the factors that vary, and a pipeline that crosses, samples, derives, and orders them into concrete trial rows.

Factors

Factors are the variables that define your conditions. Each factor has a name and a set of levels. The Stroop example has two:

• congruency — congruent, incongruent
• color — red, green, blue, yellow

Use Add factor to create one, then add levels with + level; rename a factor or level by clicking it. Factors are the raw material — the pipeline decides how they combine.

1. Add a factor

   Click Add factor, name it (e.g. load), and add its levels.

2. Cross your factors

   Make sure the pipeline’s Cross step lists the factors you want combined (see below).

3. Check the rows

   Watch the generated rows table — and the Rows dock tab — update as you edit.

The pipeline

The pipeline is an ordered list of steps that transforms factors into trial rows. Each step shows how many rows it produces, so you can see the design take shape. The Stroop Practice phase uses five steps:

The common step types:

• Cross — combine factors into every condition (the design grid).
• Sample — take some or all rows, with or without replacement, to hit a target trial count.
• Repeat — duplicate rows so each condition appears multiple times.
• Derive — add a computed column from an expression.
• Shuffle — randomize order (optionally with constraints).

Use Add step to insert a step; the Step preview under the pipeline shows the rows after any chosen step, so you can inspect intermediate stages.

Derived columns and the expression language

Derive steps compute new columns with expressions. They reference the current row’s factor values as row.<factor>, study-level parameters as params.<name>, and a set of helper functions. Two real examples from the Stroop study:

// Incongruent trials show a different color word than the ink color.
word = row.congruency === 'congruent'
  ? row.color
  : nth(params.colors, (indexOf(params.colors, row.color) + 1) % length(params.colors))

// The correct key for this trial comes from the response-key map.
correctResponse = params.responseKeys[row.color]

Helpers like nth(), indexOf(), and length() operate on parameter lists; standard JavaScript operators (===, ?:, %, []) are available too.

Reading the generated rows

Below the pipeline, the rows table shows the final output — one row per trial, with a column per factor and derived value (#, congruency, color, word, correctResponse). The same data is available any time in the Rows tab of the bottom dock. These columns are exactly what becomes available to bind into the screen.

Next