ImpactMojo ImpactMojo
Membership

Data Literacy 101

Data sources, quality assessment & open data
ImpactMojo Workshop Series • Finding & Evaluating Reliable Information
90-105 Minutes

Workshop 2: Source evaluation & data quality

Target Audience: Development practitioners who need to find and evaluate data sources for reports, proposals, and evidence-based programming

Prerequisites: Workshop 1 or basic data reading skills

Materials Needed: Internet access, sample datasets, government reports

Learning objectives

By the end of this workshop, participants will be able to:

Part 1: The data source hierarchy - not all data is created equal

20 minutes

The same "fact" from different sources

Question: What percentage of Indian women participate in the labor force?

Source A: WhatsApp forward - "Only 12% of Indian women work - shocking!"

Source B: News article - "Women's labor force participation drops to 16.5% in 2022"

Source C: PLFS 2022-23 - "Labour Force Participation Rate for women: 37%"

Source D: ILO modelled estimate (World Bank WDI) - "Female labor force participation (ages 15+), 2022: 29.3%"

Why the differences?

  • Different definitions (work vs. labor force participation vs. employment)
  • Different age groups (15+ vs. working age)
  • Different time periods and survey methods
  • Urban vs. rural vs. total population

This is why source evaluation matters - the "same" statistic can vary dramatically based on methodology.

Data source reliability hierarchy

Unreliable Questionable Generally Reliable Highly Reliable
Unreliable sources
  • Social media posts without sources
  • Unverified WhatsApp forwards
  • Blogs without citations
  • Outdated reports (>10 years old)
  • Clearly biased advocacy materials

Use: Never for serious analysis

Questionable sources
  • News articles without data citations
  • Think tank reports with unclear methods
  • Corporate/marketing research
  • Small-scale studies without peer review
  • International rankings without transparency

Use: With extreme caution, verify elsewhere

Generally reliable
  • Government statistical reports
  • UN agency publications
  • Academic research with methods
  • Established NGO research reports
  • Well-documented survey data

Use: Good for most purposes, check currency

Highly reliable
  • Peer-reviewed academic papers
  • Government census/official surveys
  • International statistical databases
  • Replicated research findings
  • Meta-analyses and systematic reviews

Use: Gold standard for evidence-based work

Part 2: Quality assessment framework

25 minutes

The 8-point data quality checklist

Quality Dimension Key Questions Red Flags
1. Currency How recent is this data? Is it still relevant? Data >5 years old for rapidly changing topics
2. Coverage Who/what is included? What's excluded? Unclear sample, convenience sampling
3. Methodology How was data collected? What methods used? No methodology section, vague descriptions
4. Sample Size Is the sample large enough to be representative? Very small samples, no confidence intervals
5. Transparency Are limitations acknowledged? Data available? No access to raw data, missing details
6. Independence Who funded this? Any conflicts of interest? Industry-funded studies, undisclosed conflicts
7. Consistency Do findings align with other credible sources? Wildly different from established evidence
8. Precision Are uncertainty ranges provided? Margins of error? Overly precise claims, no confidence intervals

Source evaluation exercise (15 minutes)

Your Task: Evaluate three data sources using the 8-point checklist

Source A: "India Healthcare Report 2024"
  • Publisher: HealthConsult Pvt Ltd (consulting firm)
  • Sample: "500+ healthcare professionals across metros"
  • Key Finding: "85% of doctors report improved patient outcomes with new technology"
  • Methodology: "Online survey conducted in Q1 2024"
  • Funding: Not disclosed

Rate Source A (1-5 scale) on each dimension:

  • Currency: ___/5
  • Coverage: ___/5
  • Methodology: ___/5
  • Sample Size: ___/5
  • Transparency: ___/5
  • Independence: ___/5
  • Consistency: ___/5
  • Precision: ___/5

Overall Assessment: Would you use this source? Why or why not?

Source B: "National Sample Survey 75th Round: Health in India"
  • Publisher: National Statistical Office, Ministry of Statistics
  • Sample: "1,13,823 households across all states and UTs"
  • Time Period: July 2017 - June 2018
  • Methodology: "Two-stage stratified sampling, detailed in annexure"
  • Limitations: Ailments are self-reported, over a 15-day reference period (hospitalisation over 365 days)

Quick Assessment: How would Source B likely score compared to Source A? Discuss in pairs.

Part 3: The Indian data ecosystem

20 minutes

Key government data sources

National Statistical Office (MoSPI)

What: NSS, NSSO surveys, national accounts

Best For: Household consumption, employment, health

Frequency: Annual, quinquennial

Access: mospi.gov.in

Census of India

What: Population and housing census

Best For: Demographics, literacy, housing

Frequency: Every 10 years until 2011; the next census has a reference date of 1 March 2027

Access: censusindia.gov.in

ASER & AISHE

What: Learning outcomes, higher education

Best For: Education quality, enrollment trends

Frequency: ASER every two years since 2014; AISHE annual

Access: asercentre.org, aishe.gov.in

NFHS & HMIS

What: Health, nutrition, family welfare

Best For: Health outcomes, service delivery

Frequency: Every 3-5 years / Continuous

Access: dhsprogram.com, hmis.mohfw.gov.in

Agricultural statistics

What: Crop production, prices, farmer income

Best For: Rural livelihoods, food security

Frequency: Annual, seasonal

Access: desagri.gov.in (Directorate of Economics and Statistics, Ministry of Agriculture)

RBI & financial data

What: Financial inclusion, credit, SHG data

Best For: Economic indicators, banking

Frequency: Monthly, quarterly

Access: rbi.org.in, pmjdy.gov.in

Understanding Administrative Data Limitations:

  • Coverage gaps: May miss informal sector, marginalized groups
  • Reporting bias: Incentives for officials to over/under-report
  • Definitional changes: Categories may change over time
  • Time lags: Often 1-2 years behind current reality
  • Quality variation: Data quality varies by state/district

Case: the NREGA payment puzzle

A district collector reported that 95% of NREGA wages were paid within 15 days. But field visits found families still waiting months for payments. (MGNREGA was repealed from 1 July 2026 by the Viksit Bharat G RAM G Act, 2025; the lesson applies to any scheme's dashboard.) What is happening here? Each of the limitations above is a possible answer, and the only way to tell them apart is to check the dashboard against what people on the ground report.

Beyond government sources

International sources:

  • World Bank Open Data: Cross-country comparisons
  • UN Data: SDG indicators and progress tracking
  • OECD Development Database: Aid flows, development finance

Research organisations:

  • Centre for Monitoring Indian Economy (CMIE): Employment, COVID impact
  • PRICE (People Research on India's Consumer Economy): Household surveys
  • Centre for Social and Economic Progress (CSEP), New Delhi: Policy research

Part 4: Problem set - source selection & quality assessment

20 minutes

Research scenario: Data source selection

Background: You're writing a report on "Digital Financial Inclusion in Rural India" for an international development agency. You need credible, recent data to support policy recommendations.

Problem 1: Source identification (8 minutes)

Research Questions to Answer:

  1. What percentage of rural Indians have bank accounts?
  2. How has digital payment adoption changed post-COVID?
  3. What are the main barriers to financial inclusion for women?
  4. Which states have made most progress on financial inclusion?

For each question, identify:

  1. Best primary data source: _____
  2. Alternative/secondary source: _____
  3. Potential data limitations to mention: _____
  4. How recent should the data be: _____

Available Sources: NFHS-5, Financial Inclusion Insights Survey, RBI Annual Report, NABARD All India Rural Financial Inclusion Survey, State-wise PMJDY data, Digital India reports

Problem 2: Quality assessment challenge (7 minutes)

Source under Review (fictional report and publisher, hypothetical figures): "India Digital Payment Growth Report 2024" by Paisarekha Insights, an invented firm used only for this exercise

Key Claims:

  • "Digital payments grew 400% in rural areas since 2020"
  • "76% of farmers now use UPI for transactions"
  • "Women's digital payment adoption lags by only 5%"

Available Information:

  • Sample: 2,500 respondents across 12 states
  • Method: Phone interviews (mobile numbers from telecom databases)
  • Funding: Supported by major fintech companies
  • Response rate: Not mentioned
  • Sampling method: Not clearly described

Critical Assessment:

  1. Identify 3 quality red flags:
    • _____
    • _____
    • _____
  2. What sampling bias might exist with phone interviews? _____
  3. How might funding by fintech companies affect findings? _____
  4. Would you cite this report? How would you caveat it? _____
Problem 3: Data triangulation (5 minutes)

Scenario: You find conflicting statistics on women's bank account ownership:

  • Global Findex 2021: 78% of women have accounts
  • NFHS-5 (2019-21): 79% of women have accounts they use themselves
  • PMJDY Dashboard (pmjdy.gov.in, September 2026): 56% of PMJDY accounts belong to women (33.1 crore of 59.4 crore)

Resolution Strategy:

  1. Why might these numbers differ? (Consider definitions, time periods, samples)
  2. Which would you use for what purposes?
  3. How would you present this in your report?

Part 5: Applying data literacy in your daily work

10 minutes

From data consumer to data champion

The Data Literacy Mindset: Move from "What do the numbers say?" to "What story are the numbers telling, who's telling it, and what are they leaving out?"

Practical applications

Program planning

  • Use baseline data to set realistic targets
  • Disaggregate data by gender, caste, region
  • Question official statistics - triangulate with ground reality

Monitoring & evaluation

  • Design indicators that capture change for marginalized groups
  • Use multiple data sources to validate findings
  • Present the story behind the numbers

Advocacy & communication

  • Use data to build compelling narratives
  • Always provide context and comparisons
  • Acknowledge data limitations transparently

Learning & adaptation

  • Track trends over time as well as single snapshots
  • Use data to ask "why" and "so what"
  • Involve communities in interpreting data

Reflection & resource building

10 minutes

Personal data source library

Individual Task (7 minutes):

Create your go-to source list for your work area:

  1. Your main focus area: _____ (e.g., education, health, livelihoods)
  2. Top 3 most reliable sources for this area:
    • _____
    • _____
    • _____
  3. One source you used to trust but now question: _____
  4. Biggest data gap in your field: _____
  5. One new source you want to explore: _____

Quick Share (3 minutes): Exchange source recommendations with colleagues in similar fields

Quick reference guide

Data quality checklist

Before using any data, ask:

  • □ Who collected this data and why?
  • □ How recent is the data?
  • □ What's the sample size and methodology?
  • □ Who or what is missing from this data?
  • □ How do the numbers compare to other sources?
  • □ What assumptions are built into the indicators?

Red flags for misleading data

  • Dramatic headlines with no context
  • Charts with manipulated scales
  • Percentages without absolute numbers
  • Correlation presented as causation
  • Cherry-picked time periods
  • No mention of data limitations

Summary

Treat every source with some scepticism. Ask who collected the data, how and why, and what might be missing, and check an important figure against more than one source before you rely on it.

Source evaluation resources

Fact-Checking Organizations:

Data Quality Guides:

Indian Government Data Portals:

Next Steps in ImpactMojo: