Contingency Questions: The Secret to Smarter Survey Design

Contingency Questions Explained: How Skip Logic Creates Better Surveys

Contingency questions are survey questions asked only when a respondent's previous answer meets a condition: "Have you returned a product this year?" filters who sees "What was the reason for the return?". Skip logic and branching are the mechanisms that implement them, routing each respondent through only the questions that apply: the difference between a questionnaire that interrogates everyone and one that converses with each person, built on the flow principles from questionnaire design.

Quick Answer

Contingency questions in 20 seconds:

• Definition: Questions shown only if a prior answer meets a condition: the filter-and-follow-up pair
• The vocabulary: Contingency questions = the design concept; skip logic / branching = the software mechanism that routes respondents
• Why they matter: Relevance (nobody answers what doesn't apply), shorter perceived surveys, cleaner data, honest base sizes
• The build sequence: Map paths → Draw the flow → Define conditions → Set defaults → Test every route → Monitor base sizes
• The trap: Every branch shrinks its base: a contingency question seen by 60 people produces a chart, not a finding
• The rule: Route on behaviour where you can, ask filters where you must

Introduction

Every survey without routing commits the same quiet offence: it asks non-buyers about their purchase experience, non-returners why they returned, and childless respondents about their children's snacks: forcing everyone through everyone else's questions, and forcing fake answers from people whose honest answer was "this doesn't apply to me."

Contingency questions fix this at the design level, and skip logic fixes it at the software level: same idea, two vocabularies, one discipline. This guide covers both: what contingency questions are and how they differ from the mechanism that delivers them, 20 filter-and-follow-up pairs you can field as written, the six-step process for building branching that doesn't collapse, the mistakes that quietly break studies, and the base-size math that most branching guides never mention.

Why This Topic Matters for Brands

Routing is where questionnaire quality compounds or collapses:

• Irrelevant questions manufacture fake data: A non-returner forced through return questions doesn't skip them: they answer them: fiction, recorded as findings
• Perceived length is the real length: A 60-question instrument that shows each respondent 25 relevant questions completes like a 25-question survey: routing is the only way to build depth without paying for it in drop-off and respondent quality
• Bases are budgets: Every branch spends sample: the follow-up seen by 12% of 400 respondents rests on 48 people: branching decisions ARE sample-size decisions
• Personalisation reads as professionalism: Respondents notice when a survey listens: relevance is the cheapest engagement upgrade in survey design
• Broken routes are invisible failures: A mis-wired skip pattern doesn't error: it silently shows wrong questions to wrong people, and the data looks fine

What Are Contingency Questions?

A contingency question is a question whose appearance is contingent on a previous answer. It always exists as a pair:

• The filter question: Establishes the condition: "Have you purchased a mattress in the past 12 months?"
• The contingency question(s): Shown only to qualifiers: "Which brand did you purchase?" "What almost made you choose differently?"

The structure does three jobs at once: it protects respondents from irrelevance, protects data from forced fake answers, and defines honest analytical bases: the "brand purchased" chart is now explicitly a chart of recent buyers, not of everyone.

Contingency Questions vs Skip Logic: Concept vs Mechanism

The two terms describe the same practice from different altitudes:

• Contingency questions is the research-design term: the WHAT: conditional questions and the logic of who should see them
• Skip logic / branching / survey logic is the implementation term: the HOW: the routing rules configured in survey software that skip, branch, and pipe respondents through the paths
• The practical distinction that matters: Design the contingencies on paper first (who needs to be asked what, and why), then implement them as skip logic: teams that start in the software build routing around tool features instead of research needs

Related mechanisms in the same family: skip patterns (jumping past blocks), display logic (showing items conditionally), piping (inserting earlier answers into later wording: "You said you bought Brand X: how satisfied were you with it?"), and disqualification logic (the screener's polite exit).

20 Contingency Question Examples (Filter → Follow-Up)

Purchase and ownership (1-5)

1. Filter: Have you purchased a mattress in the past 12 months? → If yes: Which brand did you purchase?
2. Filter: Do you currently own an air fryer? → If yes: How often do you use it? If no: Have you considered buying one?
3. Filter: Have you bought protein supplements in the past 3 months? → If yes: Which format do you prefer: powder, bars, or ready-to-drink?
4. Filter: Did you compare prices before your last purchase? → If yes: Which platforms did you compare on?
5. Filter: Have you ever bought this category as a gift? → If yes: What occasion was it for?

Experience and satisfaction (6-10)

1. Filter: Have you contacted customer support in the past 6 months? → If yes: How satisfied were you with the resolution?
2. Filter: Rate your delivery experience (1-5). → If 1-2: What went wrong with the delivery?
3. Filter: Have you returned a product bought online this year? → If yes: What was the main reason for the return?
4. Filter: Did the product match its online description? → If "not really": What was different from what you expected?
5. Filter: How likely are you to recommend us? (0-10) → If 0-6: What would need to change for you to score us higher?

Behaviour and usage (11-15)

1. Filter: Do you use quick-commerce apps? → If yes: What do you order most often? If no: What stops you?
2. Filter: Do you follow a fitness routine? → If yes: Does your nutrition change on workout days?
3. Filter: Have you tried a new snack brand in the past month? → If yes: What made you pick it up?
4. Filter: Do you research products on YouTube before buying? → If yes: At what stage: before shortlisting or before final purchase?
5. Filter: Has anyone else in your household used this product? → If yes: How did their experience compare to yours?

Screening and demographics (16-20)

1. Filter: Do you have children under 12 at home? → If yes: Who decides which snacks they get?
2. Filter: Are you the primary grocery decision-maker? → If no: [Route to influence questions instead of purchase questions]
3. Filter: Do you experience regular back or body discomfort? → If yes: [Route to the pain-and-support module] If no: [Skip to comfort preferences]
4. Filter: Have you moved homes in the past 2 years? → If yes: Did the move trigger any furniture or appliance purchases?
5. Filter: Do you currently pay for any OTT subscriptions? → If yes: How many? → If 3+: Which one would you cancel first?

Example 20 shows a nested contingency: a follow-up contingent on a follow-up: powerful, and the exact place base sizes go to die: more on that below.

How to Build Survey Branching: The 6-Step Process

1. Map the respondent paths on paper first: List the distinct respondent types the study serves (buyers, rejectors, lapsed users) and what each must be asked: the contingency design before any software opens
2. Draw the flow: A one-page diagram: filters as diamonds, question blocks as boxes, every arrow labelled with its condition: if you cannot draw it, respondents cannot survive it
3. Define conditions precisely: "If Q3 = Yes" is clean: "if respondent seems interested" is not a condition: every branch rule must be answerable by the data the survey has already collected at that point
4. Set the default route: Every filter needs an explicit path for every possible answer, including "Don't know" and skipped items: orphaned respondents are the most common wiring failure
5. Test every path, not the happy path: Walk each respondent type through the live survey: buyers, non-buyers, the 0-6 scorer, the "None of these" selector: broken routes hide in the branches nobody previewed
6. Monitor base sizes during fieldwork: Track how many respondents each contingency question is actually reaching: a follow-up running at n=40 needs a quota boost or an honesty asterisk before anyone charts it

Common Mistakes

1. The base-size trap: Branching until key follow-ups rest on unreadable cells: every contingency question inherits the 100-per-cell rule from sample size calculation: design branches WITH the sizing math, not after it
2. Orphaned routes: A filter answer with no defined destination: the respondent stalls or sees nonsense: the "Don't know" path is forgotten most
3. Over-branching: Fifteen paths where three respondent types exist: complexity that multiplies testing burden and breaks silently: branch for analysis needs, not for elegance
4. Interrogating the skip: Asking non-buyers why they didn't buy, then what would change their mind, then when they might reconsider: the skip was supposed to be a skip: two questions for non-qualifiers, maximum
5. Leading routes: Routing only dissatisfied respondents to improvement questions builds a complaints file, not a finding: sometimes the contrast group needs the question too: route for comparison, not just relevance
6. Untested nested logic: Follow-ups on follow-ups (example 20) fail quietly when any level's condition is mis-wired: nesting doubles the testing, every time
7. Version drift across waves: Editing branch conditions mid-tracker changes who answers what: the trend line breaks without a single question's wording changing

Best Practices

• Design on paper, implement in software: The flow diagram is the instrument: the tool is just the printer
• Two contingency levels maximum for consumer surveys: Nesting beyond filter → follow-up → one refinement multiplies failure surface faster than insight
• Write the base into every chart title: "Reasons for return (among the 22% who returned, n=112)": the honesty that contingency design makes possible: use it
• Pipe for warmth, sparingly: Inserting the respondent's earlier answer ("You mentioned Brand X...") lifts engagement: overused, it reads as surveillance
• Give every skip a soft landing: Non-qualifiers routed forward gracefully, never told they "failed": the respondent experience is part of the panel's health
• Pre-register which contingent findings are decision-grade: Decide before fielding which follow-ups must hit readable bases: those get quota protection: the rest are texture

PulseAI Research Insight: The Best Filter Question Is One You Never Have to Ask

Every filter question in the bank above is a claim: "Have you purchased...?", "Do you experience...?": answered honestly by most, gamed by some, and inflated by a few. Branching built on claimed filters inherits their softness: the routing is only as true as the answer that triggered it.

PulseAI Research fields on Smytten's network of 30M+ active Indian consumers, where the highest-stakes filters are answered by behaviour before the survey begins:

• Routing on records, not claims: Category purchase, trial history, and usage are observed: respondents can be routed into buyer, lapsed, and rejector paths from verified behaviour: the filter question's job, done by data
• The base-size trap, structurally eased: Because contingent modules can be quota-filled from known behaviour (verified recent buyers, confirmed returners), follow-ups launch WITH readable bases instead of hoping fieldwork produces them: the branching math solved at sampling
• The proof in a branched study: The Mattress? More Like "Mat-Stress" report is contingency design at work: the pain-and-support module (89% of regular pain sufferers replaced early), channel-specific paths (offline buyers' priorities vs marketplace buyers'), and returner follow-ups (why 49.5% sent trials back): each a contingent read, each on a base deep enough to trust
• Claims still asked, now checkable: Where filters must be asked, stated answers meet the behavioural record: the routing layer with an audit trail

The design lesson: skip logic decides who gets asked what: sampling decides whether enough of the right people exist to ask: the two disciplines are one system.

How Brands Can Use This

1. Audit your current instruments for forced irrelevance: Any question a respondent could honestly answer "this doesn't apply to me" is a missing filter: the fastest questionnaire upgrade available
2. Adopt the paper-first rule: Flow diagram before software, every study: it converts branching from tool configuration into research design
3. Marry branching to the sizing math: For every planned contingency question, estimate its expected base (incidence × sample) at design stage: protect decision-grade follow-ups with quotas
4. Institute the every-path test: A named person walks every respondent type through the live survey before launch: broken routes are a pre-field problem or a post-field disaster
5. Standardise your filter modules: The same purchase-recency and category-usage filters, worded identically across studies: comparable bases across your whole research programme
6. Route on behaviour where the stakes are high: Purchase, trial, and return paths built on verified data rather than claimed filters: the principle connecting this page to the sampling stack and the wider consumer behaviour research methods discipline

Related Concepts

• Questionnaire design: The 8-step instrument architecture branching lives inside: flow is step 5
• Questionnaire question examples: 110 fielded questions: the blocks these routes connect
• Close-ended questions: The structured formats filters depend on: clean conditions need clean options
• Sample size calculation: The base-size math every branch spends against

FAQs

1.What are contingency questions in surveys?

Contingency questions are questions shown only when a respondent's previous answer meets a condition: they exist as filter-and-follow-up pairs, such as "Have you returned a product this year?" filtering who sees "What was the reason?". They keep surveys relevant, prevent forced fake answers, and define honest analytical bases.

2.What is the difference between contingency questions and skip logic?

They describe the same practice at different levels: contingency questions is the research-design concept (which questions should be conditional and for whom), while skip logic, branching, and survey logic are the software mechanisms that implement the routing. Best practice designs the contingencies on paper first, then configures them as skip logic.

3.What is skip logic in surveys?

Skip logic is the survey-software capability that routes respondents past questions that do not apply to them, based on their previous answers: skipping non-buyers past purchase-experience questions, or sending low scorers to a diagnostic follow-up. It is the implementation layer of contingency question design.

4.What is an example of a contingency question?

Filter: "Have you contacted customer support in the past 6 months?" Contingency: "How satisfied were you with the resolution?": shown only to those who answered yes. The pair protects non-contacters from an irrelevant question and makes the satisfaction chart an honest read of actual support users.

5.What is survey branching?

Survey branching routes different respondents down different question paths based on their answers: buyers into purchase-experience modules, rejectors into barrier questions, high scorers and low scorers into different follow-ups. It lets one instrument serve multiple respondent types while each person experiences only their relevant path.

6.How many branches should a survey have?

As many as the distinct respondent types your analysis needs, and no more: typically two to four main paths for consumer studies, with contingency nesting capped at two levels. Every branch spends sample: follow-ups need readable bases (100+ respondents for decision-grade reads), so branching design is inseparable from sample-size planning.

7.What is the difference between skip logic and display logic?

Skip logic routes respondents past entire questions or blocks based on prior answers: it controls the path. Display logic conditionally shows or hides individual items or answer options within a question: it controls the content. Both implement contingency design: skip logic at the flow level, display logic at the question level.

8.Do contingency questions improve data quality?

Yes, in three ways: they eliminate forced answers from respondents to whom questions do not apply, they shorten each person's experienced survey (protecting completion and attention), and they produce explicitly defined bases, so every contingent finding is honestly labelled as a read of the people it actually describes.