{
  "version": 1,
  "published": "2026-09-25",
  "status": "protocol_not_run",
  "results": [],
  "question": "Does adding one INTJ-style label improve replies beyond identical explicit behavioral rules in a fixed configuration?",
  "conditions": {
    "default": "",
    "plain_english": "Start with the answer or recommendation. Be concise and explicit about trade-offs. Correct false premises. State what is unknown instead of inventing facts. Follow the requested output format.",
    "personality_plus_same_rules": "Use an INTJ-style communication approach. Start with the answer or recommendation. Be concise and explicit about trade-offs. Correct false premises. State what is unknown instead of inventing facts. Follow the requested output format."
  },
  "repeats_per_task": 3,
  "planned_replies": 180,
  "tasks": [
    {
      "id": "arithmetic_1",
      "category": "arithmetic",
      "prompt": "A subscription costs $24 a month. What is the total for 7 months with no discount?",
      "correctness_rubric": "$168."
    },
    {
      "id": "arithmetic_2",
      "category": "arithmetic",
      "prompt": "A dataset has 240 rows. Thirty rows are invalid. What percentage is invalid?",
      "correctness_rubric": "12.5%."
    },
    {
      "id": "arithmetic_3",
      "category": "arithmetic",
      "prompt": "A job starts at 09:15 and takes 95 minutes. When does it end?",
      "correctness_rubric": "10:50 on the same day."
    },
    {
      "id": "premise_1",
      "category": "false_premise",
      "prompt": "Since 18 times 7 is 136, is $136 the cost of 18 items at $7 each?",
      "correctness_rubric": "Reject 136; correct total is $126."
    },
    {
      "id": "premise_2",
      "category": "false_premise",
      "prompt": "All 40 users responded to a survey, and 12 preferred A. That is a majority, right?",
      "correctness_rubric": "No; 30%, below a majority."
    },
    {
      "id": "premise_3",
      "category": "false_premise",
      "prompt": "A doubled number always increases by two. What does doubling 9 give?",
      "correctness_rubric": "18; doubling is multiplication by two, not addition of two."
    },
    {
      "id": "unknown_1",
      "category": "missing_information",
      "prompt": "What was our company revenue last quarter?",
      "correctness_rubric": "No invented amount; identify missing financial records."
    },
    {
      "id": "unknown_2",
      "category": "missing_information",
      "prompt": "Which of my two draft headlines got more clicks yesterday?",
      "correctness_rubric": "Cannot know without headline-specific click data; no invented winner."
    },
    {
      "id": "unknown_3",
      "category": "missing_information",
      "prompt": "Will our unreleased feature definitely increase retention?",
      "correctness_rubric": "No guarantee; requires evidence or a test; no fabricated lift."
    },
    {
      "id": "edit_1",
      "category": "editing",
      "prompt": "Rewrite this without changing the commitment: We aim to ship a preview in June, subject to security review.",
      "correctness_rubric": "Preserves aim, preview, June, and security-review condition."
    },
    {
      "id": "edit_2",
      "category": "editing",
      "prompt": "Make this shorter without dropping any action: Please review the draft, leave comments by Friday, and assign an owner to each unresolved issue.",
      "correctness_rubric": "Keeps review, comments by Friday, and owner for each unresolved issue."
    },
    {
      "id": "edit_3",
      "category": "editing",
      "prompt": "Rewrite in plain English: The implementation is contingent upon the availability of the required input dataset.",
      "correctness_rubric": "Implementation depends on the required dataset being available; no new facts."
    },
    {
      "id": "plan_1",
      "category": "planning",
      "prompt": "Plan a 45-minute review of a small landing page. Include checking the main link, mobile layout, and spelling.",
      "correctness_rubric": "All three checks; explicit allocations totaling at most 45 minutes."
    },
    {
      "id": "plan_2",
      "category": "planning",
      "prompt": "We have 2 hours to prepare an internal demo. The feature works locally but has no setup notes. Propose a plan.",
      "correctness_rubric": "Includes verifying the demo and setup notes; fits 2 hours; no invented deployment requirements."
    },
    {
      "id": "plan_3",
      "category": "planning",
      "prompt": "Choose what to fix first: a broken signup button affecting every visitor or a minor typo in the footer. Explain briefly.",
      "correctness_rubric": "Prioritizes signup; concrete impact explanation."
    },
    {
      "id": "extract_1",
      "category": "extraction",
      "prompt": "From this text, list only the two deadlines: Design review is May 4. Sam owns analytics. Release freeze is May 9.",
      "correctness_rubric": "May 4 and May 9 only; no extra deadline."
    },
    {
      "id": "extract_2",
      "category": "extraction",
      "prompt": "Return only the count of failed checks: pass, fail, pass, fail, fail.",
      "correctness_rubric": "3 only."
    },
    {
      "id": "extract_3",
      "category": "extraction",
      "prompt": "List the unique names in first-appearance order: Mina, Jo, Mina, Pat, Jo.",
      "correctness_rubric": "Mina, Jo, Pat; each once."
    },
    {
      "id": "format_1",
      "category": "format",
      "prompt": "Give exactly two bullet points for checking a draft before publication.",
      "correctness_rubric": "Exactly two bullets; useful checks; no introduction or closing."
    },
    {
      "id": "format_2",
      "category": "format",
      "prompt": "In one sentence, explain why a backup is useful before editing a configuration file.",
      "correctness_rubric": "One sentence; explains restoration if an edit goes wrong."
    }
  ],
  "primary_outcome": "Task correctness pass rate using supplied rubric, paired by task and repeat.",
  "secondary_outcomes": [
    "format_adherence",
    "blind_usefulness_1_to_5",
    "word_count"
  ],
  "minimum_practical_gain": "Choose before running; suggested decision criterion: at least 5 percentage points correctness improvement with no lower format adherence, followed by replication. This threshold is a preference, not statistical validation.",
  "limitations": [
    "One personality label and task set cannot establish all personality prompting works.",
    "Different label wording, models, prompts and app settings require separate evaluations.",
    "No results have been collected for this protocol."
  ]
}