Sentiment & Citizen-Experience Analysis
How PPRI handles citizen perception data, complaint evidence, media discourse, and the distinction between administrative statistics and genuine public opinion.
Terminology note: The phrase "sentimental analysis" is treated here as citizen sentiment / public-perception analysis, not an unsupported claim that PPRI has run a nationally representative natural-language sentiment model.
Data Types
1. Survey Sentiment
Direct survey responses from structured instruments:
- Transparency International Pakistan NCPS 2025/2023 — corruption perceptions and reported experiences
- Historical citizen satisfaction surveys — around the local-government/devolution period
These are treated as the strongest form of citizen-experience evidence because they use defined sampling, standardised questions and measurable confidence intervals.
2. Administrative Grievance Signals
Institutional complaint and redressal data from:
Administrative grievance data is not a representative public-opinion survey. It reflects people who used complaint channels and therefore measures revealed administrative demand and friction, not population-wide sentiment.
3. Media Discourse
Pakistani media is useful for:
- Identifying recurring public arguments
- Capturing regional grievances
- Identifying political claims
- Finding historical quotations
Media rhetoric is not treated as public opinion unless backed by survey evidence.
Analytical Rules
PPRI research never claims:
| Prohibited claim | Why it is wrong |
|---|---|
| "Pakistanis feel X" based only on a newspaper column | Media is not a representative sample |
| "Most people are corrupt/bribed" based on perception survey percentages | Perception ≠ audited reality; reported ≠ universal |
| "Resolved complaint = satisfied citizen" | Administrative status ≠ independent satisfaction verification |
| "Social-media sentiment = national public opinion" | Platform users are not representative; algorithms shape visibility |
Required precise language
- "Respondents reported…"
- "The portal records…"
- "The survey found…"
- "Media discourse frequently frames the issue as…"
Recommended Future PPRI Text-Analysis Project
For a future primary-research project, PPRI should collect a clearly defined corpus:
- Citizen Portal complaint texts where legally and ethically accessible
- Anonymised consultation submissions
- Ombudsman complaint descriptions where available
- Pakistani news reporting on provincial/local-government reform
Classify text into actionable themes instead of a simple positive/negative score:
The codebook and sampling method must be published alongside any results.
Related PPRI Research
This methodology informs the following research papers: