Check whether a question steers an answer

Label whether a question's wording favors, disfavors, or stays neutral toward a proposed answer.

You need to check leading questions or answer pressure in a survey, interview, or evaluation prompt.

Explore this recipe interactively · Source and implementation guide

Use question-leading in TypeScript

Install with npm install jev-recipes. Requires Node.js 22.9 or newer and ES modules. Set TYPESAFE_API_KEY in your server environment for live calls, which send input to TypeSafe and use API quota. See the installation guide.

import { questionLeading } from 'jev-recipes/question-leading';

const result = await questionLeading({
  "question": "Surely you agree the new layout is easier to use, don't you?",
  "proposedAnswer": "The new layout is easier to use."
});
console.log(result);

Input contract

Field	Type	Needed
question	string	Required
proposedAnswer	string	Required
context	string	Optional
minConfidence	number	Optional

Full input and result schemas

{
  "input": {
    "$schema": "https://json-schema.org/draft/2020-12/schema",
    "type": "object",
    "properties": {
      "question": {
        "type": "string",
        "description": "One question, including any framing that the respondent sees."
      },
      "proposedAnswer": {
        "type": "string",
        "description": "One candidate answer or position whose treatment by the question should be assessed."
      },
      "context": {
        "type": "string",
        "description": "Supplied context needed to interpret references and the candidate answer."
      },
      "minConfidence": {
        "type": "number",
        "minimum": 0,
        "maximum": 1
      }
    },
    "required": [
      "question",
      "proposedAnswer"
    ]
  },
  "result": {
    "$schema": "https://json-schema.org/draft/2020-12/schema",
    "type": "object",
    "properties": {
      "model": {
        "type": "string"
      },
      "usage": {
        "type": "object",
        "properties": {
          "input_tokens": {
            "type": "integer",
            "minimum": 0,
            "maximum": 9007199254740991
          },
          "output_tokens": {
            "type": "integer",
            "minimum": 0,
            "maximum": 9007199254740991
          }
        },
        "required": [
          "input_tokens",
          "output_tokens"
        ],
        "additionalProperties": false
      },
      "status": {
        "type": "string",
        "enum": [
          "ready",
          "review"
        ]
      },
      "verdict": {
        "type": "string",
        "enum": [
          "favors",
          "disfavors",
          "neutral",
          "unclear"
        ]
      },
      "confidence": {
        "type": "number",
        "minimum": 0,
        "maximum": 1
      },
      "probabilities": {
        "type": "object",
        "propertyNames": {
          "type": "string",
          "enum": [
            "favors",
            "disfavors",
            "neutral",
            "unclear"
          ]
        },
        "additionalProperties": {
          "type": "number",
          "minimum": 0,
          "maximum": 1
        },
        "required": [
          "favors",
          "disfavors",
          "neutral",
          "unclear"
        ]
      }
    },
    "required": [
      "model",
      "usage",
      "status",
      "verdict",
      "confidence",
      "probabilities"
    ],
    "additionalProperties": false
  }
}

Saved example result

This hand-authored response demonstrates the contract. It is not a model accuracy measurement. Run it without an API key: npx jev-recipes demo question-leading.

{
  "model": "demo-fixture",
  "usage": {
    "input_tokens": 0,
    "output_tokens": 0
  },
  "status": "ready",
  "verdict": "favors",
  "confidence": 0.97,
  "probabilities": {
    "favors": 0.97,
    "disfavors": 0.01,
    "neutral": 0.01,
    "unclear": 0.01
  }
}

Evaluation evidence

Fixture only

No verified live accuracy measurement is available. Evaluate representative cases before using this decision in your workflow.

Use the evaluation guide to measure this decision on your own labeled cases.

Limitations

Labels directional wording relative to one proposed answer, not author intent, factual correctness, or observed persuasion.
Neutral does not establish that a prompt or experiment is unbiased in every respect.
Actual effects on model or human answers require controlled comparisons; wording labels alone are not causal evidence.

Related recipes

outcome-framing: Use outcome-framing to label gains and losses in wording about a specified outcome.
question-assumption: Use question-assumption to check whether a question takes a specific claim for granted.
query-specificity: Use query-specificity to check whether an information need is focused and unambiguous.
claim-stance: Use claim-stance to label the response's expressed position after a question has been answered.
tone-check: Use tone-check to assess draft wording against caller-supplied writing criteria.