Psychometric Assessment Applications

A comprehensive psychometric assessment toolkit implementing the NASA Task Load Index (TLX) and related workload measurement tools.

Overview

This project provides tools for conducting psychometric assessments, particularly focused on workload measurement using the NASA Task Load Index (TLX). The TLX is a widely-used tool for assessing the subjective workload experienced by users when performing tasks.

Features

• NASA TLX Implementation: Complete implementation of the NASA Task Load Index
• Raw TLX Scores: Unweighted average of six workload dimensions
• Weighted TLX Scores: Weighted scores based on pairwise comparisons
• Interactive CLI: Command-line interface for data collection
• Data Export: JSON export for analysis
• Statistical Analysis: Aggregate statistics across multiple assessments
• REST API: Production-ready API for remote assessment collection

NASA Task Load Index (TLX)

The NASA TLX measures workload across six dimensions:

1. Mental Demand: How mentally demanding was the task?
2. Physical Demand: How physically demanding was the task?
3. Temporal Demand: How hurried or rushed was the pace?
4. Performance: How successful were you in accomplishing the task?
5. Effort: How hard did you have to work?
6. Frustration: How insecure, discouraged, or annoyed were you?

Each dimension is rated on a scale of 1-20, where:

• 1 = Very Low
• 20 = Very High (Note: Performance is inverted - 1 = Perfect, 20 = Failure)

Installation

Basic Installation

pip install -r requirements.txt

Production Setup

# Install with production dependencies (gunicorn, postgres driver)
pip install -r requirements.txt psycopg2-binary gunicorn

Usage

Command-Line Interface

Run the interactive assessment tool:

python main.py

Programmatic Usage

from nasa_tlx import NASATLX

# Initialize system
tlx = NASATLX()

# Create assessment
result = tlx.create_assessment(
    task_name="User Interface Evaluation",
    participant_id="P001"
)

# Add ratings
tlx.add_rating(
    result,
    mental_demand=15,
    physical_demand=3,
    temporal_demand=12,
    performance=5,
    effort=14,
    frustration=8
)

# Calculate scores
tlx.calculate_scores(result)
print(f"Raw TLX Score: {result.raw_tlx_score:.2f}")

🏭 Production Deployment

Deployment Strategy

Deploy as a REST API service to collect assessments from web or mobile clients.

Docker Deployment

1. Dockerfile:

FROM python:3.9-slim
WORKDIR /app
COPY requirements.txt .
RUN pip install -r requirements.txt gunicorn
COPY . .
CMD ["gunicorn", "-w", "4", "-b", "0.0.0.0:8000", "api:app"]

2. Run Container:

docker run -d -p 8000:8000 \
  -e DATABASE_URL=postgresql://user:pass@db:5432/psychometrics \
  psychometrics-api:latest

Data Privacy & Compliance

• GDPR/HIPAA: Store participant IDs as anonymized hashes. Do not store PII (Personally Identifiable Information) in the assessment database.
• Data Retention: Configure automated deletion policies for raw data after the retention period expires.
• Access Control: Restrict database access to authorized researchers only.

Database Setup

For production, use PostgreSQL instead of SQLite:

CREATE TABLE assessments (
    id UUID PRIMARY KEY,
    participant_hash VARCHAR(64),
    task_name VARCHAR(255),
    ratings JSONB,
    scores JSONB,
    created_at TIMESTAMP DEFAULT NOW()
);

Monitoring

• Application Metrics: Track API latency and error rates.
• Business Metrics: Track number of completed assessments per day.
• Data Quality: Monitor for anomalies in rating distributions (e.g., all 1s or all 20s).

API Documentation

POST /api/v1/assessments

Submit a new assessment.

Request:

{
  "task_name": "Task A",
  "participant_id": "P001",
  "ratings": {
    "mental_demand": 15,
    "physical_demand": 5,
    "temporal_demand": 10,
    "performance": 2,
    "effort": 12,
    "frustration": 6
  }
}

Response:

{
  "id": "uuid",
  "raw_tlx": 8.33,
  "status": "success"
}

GET /api/v1/stats/{task_name}

Get aggregate statistics for a task.

Scoring Methods

Raw TLX Score

Raw TLX = (Mental + Physical + Temporal + Performance + Effort + Frustration) / 6

Weighted TLX Score

Uses pairwise comparisons to determine dimension weights.

References

• Hart, S. G., & Staveland, L. E. (1988). Development of NASA-TLX.
• NASA TLX Official Documentation

License

See main repository LICENSE file.