Dichotomous question

"Did you make a purchase from us in the past month?" - Yes / No. "Your gender?" - M / F. Two mutually exclusive options, no third one. This is a dichotomous question - the simplest type of closed-ended question.

In SurveyNinja you build it with a "Single choice" element and two options. Fast, clear, handy for screening and demographics. But it forces a choice where there is nuance - "not sure" or "other" may be appropriate. Below: when a dichotomy is justified and when to add a third option.

A dichotomy gives clean proportions: 60% yes, 40% no. Convenient for branching: "if yes - go to question 5". The risk: people pick the "closest" option when none of them fits. For unambiguous facts (gender, fact of purchase) a dichotomy is natural. For opinions and evaluations you often need a middle ground.

Two options, a third one is not allowed

Dichotomous question - a closed-ended question with two mutually exclusive answer options. Yes/no, male/female, agree/disagree. The respondent must choose one. The result is a binary variable. Used for facts (gender, having experience), screening (screening), and branching in survey logic. Easy to analyze - proportions across two groups. The risk: a forced choice where the answer is not black-and-white.

In short: "pick one of two" - with no third option.

When a dichotomy is natural

Facts: gender (if no non-binary option is needed), "did you make a purchase", "do you have children", "are you employed". Screening: "do you use product X" - if not, the survey may end for you. Demographics and filters are the typical area.

Logical branching: "Did you buy in the past month?" - yes → a block about the purchase, no → a different block. In SurveyNinja this is set up through logic jumps. A dichotomy simplifies conditions: if the answer is A - go here, if B - go there.

The risk of a forced choice

The question "Do you agree with the statement?" - Yes / No. But some respondents feel "both yes and no" or "not sure". With no third option they are forced to choose. The data gets distorted: part of the "not sure" group ends up in "yes" or "no" at random. For opinions and attitudes you often add "Not sure" or a scale (agree - disagree with gradations).

A dichotomy for opinions is a blunt instrument. For facts it works well.

In SurveyNinja: single choice with two options

A dichotomous question is a "Single choice" element with two options. You add the options and, if needed, change the wording (Yes/No, M/F, Agree/Disagree). You can make it required - the respondent has to choose. For screening you tie it to logic: if "No" - jump to a thank-you screen or a different block. More details are in the element settings and logic jumps.

Answers in reports show up as proportions for each option. The export is a single column with two values. Handy for cross-tabulations: yes/no by gender, age, region.

Order of options: yes/no or no/yes

The primacy effect means the first option is chosen more often. For neutral questions the order can matter. Alternating (half the respondents see yes/no, half see no/yes) reduces the influence - but it requires randomization. For facts (gender, fact of purchase) the influence is smaller. For opinions it is worth taking into account.

Dichotomy vs an agreement scale

Instead of "Agree / Disagree" you can use a 1-5 scale or a Likert scale. More nuance, but slower. A dichotomy is faster. The choice: do you need gradation, or is a binary split enough? For screening and filters a dichotomy is sufficient. For measuring the intensity of an opinion, a scale is better.

The link to a scale in the broad sense: a dichotomy is a special case of a scale with two gradations. The minimal scale.

Common mistakes

A dichotomy for opinions with no "not sure". "Do you agree?" - Yes/No. If part of the audience is undecided - add a third option or switch to a scale.

Mixing two things in one question. "Did you buy it and did you like it?" - that is two questions. Yes/no will not cover "bought it but did not like it". See double-barreled question.

A forced dichotomy for complex topics. "Do you support the company policy?" - Yes/No. Some want "partly". If it matters not to lose them - give them a scale or three options.

The third option: "Not sure" and "Other"

A strict dichotomy has only two options. But sometimes "Not sure" is added - then it is no longer a pure dichotomy but a three-way choice. For screening "not sure" is usually not needed - you either used the product or you did not. For opinions it is appropriate. "Other" with a text field is rare for a dichotomy; if a custom option is needed, the question is probably not really dichotomous.

Case: screening before a product survey

"Have you used product X in the past 3 months?" - Yes / No. No → "Thanks, this survey is for product users". Yes → on to questions about the experience. 1,000 visits, 400 answered "Yes", 600 - "No". The share of relevant respondents is 40%. The dichotomy filtered out the off-target audience in a single question. Without it you would have had to survey everyone and filter afterwards - extra load and distorted metrics.

Areas of application

Demographics (gender, having children, employed/not). Screening in panel surveys and B2B. Survey logic - branching on a binary answer. Filters in reports - "show only those who answered Yes". In quantitative surveys a dichotomy is a frequent building block: fast, clear, easy to analyze.

A dichotomous question - two mutually exclusive options (yes/no, m/f). For facts and screening it is appropriate. For opinions there is a risk of a forced choice. In SurveyNinja it is a "Single choice" element with two options, convenient for logic jumps.