List Experiment Question

A list experiment (item-count technique) estimates how common a sensitive behavior or attitude is without asking the respondent to admit that item directly. Control-list respondents see a short set of innocuous items and report how many apply. Treatment-list respondents see the same items plus the sensitive one, and also report a count. The difference in mean counts estimates prevalence.

Domandata does not need a special list-experiment question type. You build it with two list texts, random assignment in Flow, and one numeric count question. The export carries the assigned list plus the count.

Control list (three items): donated to a local charity this year; signed a petition; attended a public meeting. Treatment list: those three plus “I voted in a way I would not tell a pollster.” The question is “How many of the following apply to you?” with answers 0–3 on the control arm and 0–4 on the treatment arm. Analysis compares mean counts by assigned arm — not by which items the respondent named.

1. Step 1: Make two blocks. In Block Builder, add a control-list block and a treatment-list block. Each block holds a content block with the item list, then the same count question (short answer number or multiple choice 0–N).

   Step 1: Add list blocks

2. Step 2: Write the lists. Keep innocuous items identical across arms. Add the sensitive item only on the treatment list. Use a short answer or multiple choice count so the export is a single numeric field.

   Step 2: Add the count question

3. Step 3: Randomize the arm in Flow. Add a condition set, name the arms Control list and Treatment list, use random assignment (even/balanced if the sample is small), and route each arm to its block. Preview both paths.

   Step 3: Assign the list

4. Step 4: Check the export. Submit one test response per arm. Confirm the file has the assigned condition and the count, then document the list wording in the codebook.

   Step 4: Export assigned list and count

• Assigned list arm (control vs treatment).
• The count response, with a stable variable name.
• Any screener or attention check used as an exclusion rule.
• The exact list wording in the codebook, because the items never appear as separate columns.

For assignment mechanics see How Randomization Works. For cell size, see Sample Allocation Across Condition Arms.