Data Methodology & Sources

Overview

CSSNorthstar's recommendations are built on 5 years of FPSC-authenticated exam data covering 800,000+ candidate records. This page explains our data sources, collection methodology, and analytical approach.

Data Sources

Primary Source: FPSC Examiner Reports

All data is sourced from official FPSC examiner reports published annually (2020–2024). These reports are public documents that contain:

  • Candidate pass rates by optional subject
  • Score distributions and percentiles
  • Allocation statistics by province and domicile quota
  • Gender-disaggregated allocation data
  • Year-on-year trend analysis

Secondary Sources

  • Official FPSC website and announcements
  • Published merit lists and posting details
  • Provincial government allocation data

Data Collection & Processing

1. Data Extraction

We manually extract key statistics from FPSC reports for all 45 optional subjects across 5 years:

  • Mean score per subject per year
  • Standard deviation (volatility measure)
  • Median and quartile distributions
  • Pass rates (score ≥ 40)
  • Maximum and minimum scores

2. Allocation Pipeline Analysis

For each subject in each province, we track:

  • Appeared: How many candidates attempted the subject
  • Cleared: How many passed (score ≥ 40)
  • Allocated: How many received postings in that subject
  • Funnel Ratio: Appeared → Cleared → Allocated conversion rates

3. Provincial & Quota Analysis

We disaggregate allocation data by:

  • Province/Region of domicile
  • Quota category (Punjab, Sindh, KP, Balochistan, FATA, Gilgit-Baltistan, etc.)
  • Gender

Analytical Framework

Risk Profile Calculation

Each subject receives a Risk Profile (Low / Moderate / High) based on:

  • Score Variance (SD): High SD = unpredictable scoring = higher risk
  • Year-on-Year Volatility: Do mean scores swing significantly? How does pass rate change?
  • Allocation Volatility: How many seats are available? Does this change year-to-year?
  • Candidate Assessment: Your assessed capability for that subject (from your assessment responses)
  • Timeline: How much time do you have to master the subject?

Allocation Probability Modeling

Your allocation probability is calculated using a weighted model that factors:

  • Subject Combination Performance: Historical allocation rates for your specific combination
  • Your Merit Position Estimate: Based on assessment scores and subject difficulty
  • Domicile & Quota Competitiveness: How many candidates from your quota compete for subject seats
  • Allocation Seats per Subject: Average seats available per subject in your province

Confidence Score

Each recommendation includes a confidence score (0–1.0) that reflects:

  • Data completeness (5 years of records increase confidence)
  • Subject volatility (stable subjects = higher confidence)
  • Sample size for your profile match (more similar candidates = higher confidence)
  • Consistency of allocation patterns (consistent = higher confidence)

Data Limitations & Assumptions

What We Don't Know

  • Individual candidate preparation levels or coaching quality
  • Exam question difficulty changes year-to-year (we infer from score distributions)
  • Allocation preferences or candidate bargaining behavior
  • Future FPSC policy changes or quota adjustments

Key Assumptions

  • Historical patterns will continue in the near term (2–3 years)
  • Your assessment responses accurately reflect your profile
  • CSS exam structure and optional subjects remain consistent
  • Allocation quotas remain roughly stable

Data Privacy & Usage

Your Data

Your assessment responses (academic profile, subject preferences, etc.) are:

  • Used only to calculate your personalized recommendation
  • Encrypted and stored securely
  • Not shared with third parties
  • Not used to train any models beyond improving CSSNorthstar accuracy

FPSC Data

All FPSC data we use is publicly available from official examiner reports. We comply with copyright and attribution requirements.

Methodology Validation

Accuracy Checks

We validate our methodology by:

  • Cross-referencing against published FPSC statistics
  • Backtesting recommendations against historical allocation data
  • Comparing our allocation probability predictions against actual outcomes

Continuous Improvement

As new FPSC data becomes available (new examiner reports), we:

  • Update our dataset immediately
  • Recalculate risk profiles and allocation probabilities
  • Notify users when recommendations change significantly

Questions About Our Methodology?

If you'd like to understand the methodology behind your specific recommendation in more detail, please reach out to [email protected]. We're happy to explain the data and assumptions behind any recommendation.

Last updated: May 2026 · Data current through 2024 FPSC examiner report