Questionnaire Analysis: Foundations
This page introduces the model Askalot uses to analyze a questionnaire: what an item, an answer, a domain, a precondition and a postcondition are, and which questions the analysis answers about them. It explains the ideas in words and examples. The formal definitions, notation and proofs are in the peer-reviewed paper presented at ICAI 2026 (13th International Conference on Applied Informatics), proceedings pp. 227–243, doi:10.17048/icai.2026.227.
Overview
Section titled “Overview”Askalot treats a questionnaire as a formal system that can be analyzed for consistency, reachability and logical coherence. Design flaws are found automatically before a survey goes live, so respondents never meet an impossible situation or a question that can never be reached.
Why Formal Analysis Matters
Section titled “Why Formal Analysis Matters”Traditional survey platforms let designers build questionnaires that contradict themselves:
- Questions that can never be reached because their conditions are impossible
- Answer rules that no answer can satisfy
- Circular dependencies between questions
- Global inconsistencies where no valid way through the survey exists
Askalot’s analysis detects these issues before deployment.
Axiomatic Building Blocks
Section titled “Axiomatic Building Blocks”The model is built from a small number of parts. Each is described below; the paper gives the precise definitions.
An item is one element of the questionnaire that shows content to the respondent and may collect an answer. Items are numbered in the order they appear. There are four kinds:
- Informational items: comments or instructions with no answer.
- Scalar questions: a single question with one integer answer.
- Question groups: a set of related sub-questions that share logical properties.
- Matrix questions: a grid in which every cell shares the same answer domain, such as a rating scale.
Answer Variables (Outcome Variables)
Section titled “Answer Variables (Outcome Variables)”Every item that collects an answer has an outcome variable that holds the respondent’s answer. A scalar question’s outcome is one integer, a question group’s is a list of integers, and a matrix question’s is a grid of integers. Informational items have no outcome.
An assignment is one possible set of answers, one value for every outcome variable. A condition either holds or fails under a given assignment, and most of the analysis asks whether an assignment with some property exists.
Ranges and Base Domain
Section titled “Ranges and Base Domain”Each outcome variable has a domain: the values it may take. A domain is either a range with a lower and an upper bound, or a finite list of allowed values. Taken together, the domains of all items form the base constraint: the set of answer combinations that respect every item’s domain. Every other check is made against this base.
Preconditions and Postconditions
Section titled “Preconditions and Postconditions”Each item has a precondition: a condition on earlier answers that decides whether the item is shown. When the precondition holds for the answers given so far, the item is asked; when it fails, the item is skipped. Preconditions are what make a questionnaire branch.
Each item also has a postcondition: a condition that a valid answer to the item must meet. A postcondition is checked only when the item is asked, so a valid set of answers is one in which every item that is asked has its postcondition met. Postconditions keep answers consistent with each other and can narrow the acceptable values based on earlier answers.
Satisfiability Notation
Section titled “Satisfiability Notation”The analysis is phrased in terms of satisfiability. A condition is satisfiable when at least one assignment makes it true, and unsatisfiable when none does. A set of facts entails a condition when every assignment that makes all the facts true also makes the condition true. The SMT solver (Z3) answers satisfiability questions, and every check on the following pages is one or more of them.
Questionnaire Structure
Section titled “Questionnaire Structure”A complete questionnaire, in this model, is its items, their outcome variables, the domain of each outcome, the precondition and postcondition of each item, and the item where the survey starts.
Key Validation Questions
Section titled “Key Validation Questions”Given a questionnaire, the analysis asks:
- Reachability: Can each item ever be reached? (See Preconditions and Postconditions)
- Constraint feasibility: Can an item’s postcondition be met whenever the item is reached? (See Preconditions and Postconditions)
- Dependency cycles: Do the items depend on each other in a circle that prevents evaluation? (See Dependency Analysis and Cycle Detection)
- Global consistency: Is there at least one valid way through the whole survey? (See Comprehensive Questionnaire Validation)
The other pages in this section explain each of these questions.
Further Reading
Section titled “Further Reading”- Preconditions and Postconditions - Reachability analysis and constraint classification
- Dependency Analysis and Cycle Detection - Ensuring evaluation order and detecting circular dependencies
- Comprehensive Questionnaire Validation - Understanding different notions of questionnaire consistency
- Complex Types - Extensions for vector and matrix question types
This page is licensed under CC BY 4.0; attribute it to Askalot Survey & Research Kft.