Jeongmil Munhanggaebal (Precision Item Development)

A precision survey-item design GPT that supports construct mapping, content validity checks, cognitive interview simulation, and reliability prediction in one workflow

Overview
Version
v1.0.0
Created
2025-12-14
Updated
2025-12-14
survey-designquestionnairemeasurementpsychometricsvalidityreliabilitycognitive-interviewhr-analyticsresearch-assistant
jeongmil-munhanggaebalprecision-item-devsurvey-item-audit
Key functions
  • Derive measurement constructs and sub-dimensions (factors) from items
  • Audit content validity (double-barreled items, ambiguity, leading/biasing wording, inappropriate terms) and propose revisions
  • Simulate cognitive interviews (virtual respondents’ answers and rationales) to detect interpretation variance
  • Detect semantic redundancy across items and recommend merge/removal
  • Predict reliability (e.g., Cronbach’s alpha) via simulated data and flag problematic items
  • Deliver a structured final item set (item-construct-dimension mapping) with version comparisons
Technical details
_id
g-688c23ce47388191960e1f0ba68c8ab0
gpt_id
g-688c23ce47388191960e1f0ba68c8ab0
viz1
public
viz2
show_url
language
en
Other fields
additional_features
["Default scale: 5-point Likert (1=Strongly disagree ~ 5=Strongly agree)", "Default cognitive interview: 5 virtual respondents", "Default reliability prediction: 100 virtual respondents with item-level diagnostics", "Version tracking (before/after) and item-construct-dimension mapping table"]
example_commands
["Start precision item generation. Map these 12 items to constructs/dimensions, flag content-validity issues, and propose revised versions.", "Start precision item generation. Target: office workers aged 20–39. Create an 8-item job stress scale and show a 5-person cognitive interview simulation with rationales.", "Start precision item generation. Identify redundant items, propose merges, and predict Cronbach’s alpha using a 100-respondent simulation.", "Start precision item generation. Neutraliz
gpt_id
g-688c23ce47388191960e1f0ba68c8ab0
ideal_use_cases
["Review and refine existing survey items for validity, bias, and redundancy", "Design new scale items aligned to a construct-dimension framework and organize an item bank", "Pre-test wording via cognitive-interview-style simulation to anticipate misinterpretations or reactance", "Estimate reliability risks (low alpha, inconsistent items) before running a pilot", "Neutralize socially desirable or leading phrasing in Likert-style questionnaires (default 5-point)"]
limitations
["Reliability and response distributions are simulation-based estimates, not real survey results; empirical validation is required", "If the target population profile (role, industry, culture, language nuance) is unclear, item fit may degrade", "Sensitive domains (politics/discrimination/health, etc.) may require additional ethical/legal review beyond wording fixes", "Detailed internal mechanisms (procedural diagrams, algorithms, hidden rules) are not provided"]
target_users
["Survey and social science researchers (academic/public/private)", "HR/organizational diagnostics practitioners and consultants", "UX researchers and data analysts designing quantitative surveys", "Educational/assessment item developers"]