Skills

Quantitative Social Research

10 weeks · 0 milestones

Conduct a real survey study (minimum 40 respondents) or secondary data analysis using a named real dataset (BHPS, GSS, ANES, NLSY, or equivalent). Document your methodology, produce your analysis output (SPSS, R, Stata, or Python), and write a sociologically-framed interpretation of your results. Proof is your methodology record, analysis output, and written interpretation plus a documented review by a sociologist or social statistician who challenges your operationalisation of key variables — specifically, how your concepts were measured and whether the chosen dataset is representative for your question. Your written response is a required part of the proof.

Part of: Social Sciences, Law & Humanities · Proves for: UG

Milestone map

3 milestones

Develop a research question and select an appropriate quantitative design

4 weeks

Develop a specific social science research question that can be addressed with quantitative methods using existing data — a question about income inequality and social outcomes, migration patterns and economic effects, political participation and demographic factors, or similar. Identify an appropriate design (cross-sectional survey analysis, longitudinal analysis, comparative analysis across units) and locate a real existing public dataset. Pre-register the research design on OSF before any analysis begins.

Proof required

Research design document (500+ words) with OSF pre-registration link, covering: research question with theoretical rationale, dataset source and access method, study design and justification, key variables with operational definitions, planned analysis, and ethical considerations.

What gets checked

Research question is specific and testable with existing data
OSF pre-registration link present and timestamped before analysis begins
Key variables are operationally defined — not just named

Common mistakes

Pre-registering after analysis — the registration must be timestamped before any analysis
Using a dataset without verifying its variables match the research question — check variable availability before committing to the design
Operational definitions that are ambiguous — each variable must have a precise definition and a specific measurement approach in the dataset

Resources

Foundationstart here

What a verifier looks for

Confirm OSF pre-registration link is present and timestamped before any analysis
Verify dataset source is accessible
Check operational definitions — confirm each variable has a precise definition and measurement approach

You'll sign in first, then come straight back here.

Conduct the analysis and produce a transparent data report

5 weeks

Conduct the pre-registered analysis using R, Python, or STATA. Report results transparently: all pre-registered analyses (including null results), descriptive statistics for key variables, primary analysis results with effect sizes and confidence intervals, and any deviations from the pre-registered analysis plan with reasoning. Deposit data, code, and materials on OSF or a public repository for reproducibility.

Proof required

Data analysis report (600+ words) covering: descriptive statistics, all pre-registered analyses with results (effect sizes and confidence intervals), any deviations from pre-registered plan with reasoning, and OSF or repository link for data, code, and materials.

What gets checked

All pre-registered analyses are reported — including null results
Effect sizes and confidence intervals are reported alongside p-values
Data, code, and materials are deposited with an accessible link

Common mistakes

Reporting only statistically significant pre-registered analyses — all pre-registered analyses must be reported
Reporting p-values without effect sizes
Not depositing code — reproducibility requires the analysis code, not just the data

Powstik Guide

A transparent analysis report with code and data deposited.

Steps

Write your analysis as a script, not by clicking.
Run every pre-registered analysis, including nulls.
Report effect sizes with confidence intervals.
Log each deviation from the plan and why.
Deposit data, code and materials together.

Template

Descriptives (key variables):
| Pre-registered analysis | Result | Effect size (CI) |
|-------------------------|--------|------------------|
Deviations and reasons:
Repository link (data + code):

What gets sent back

Only significant results reported.
p-values without effect sizes.
No deposited code.

Resources

Foundationstart here

Depthgo deeper

Masteryfor the dedicated

What a verifier looks for

Check all pre-registered analyses are reported — compare report to pre-registration
Verify effect sizes and confidence intervals are present
Access repository link — confirm data and code are deposited

You'll sign in first, then come straight back here.

Defend quantitative analysis under social researcher challenge

1 week

Present the quantitative research design and analysis to a social science academic, quantitative methods researcher, or data scientist with social science experience. The reviewer must probe design and analysis — 'your design assumes no selection bias here — is that defensible?', 'you deviated from the pre-registered analysis — does the deviation produce different conclusions from the pre-registered specification?'. This Q&A satisfies the ADVERSARIAL VERIFICATION RULE.

Proof required

Q&A notes (250+ words) documenting the reviewer's credentials (social science academic, quantitative methods researcher, or data scientist with social science experience), at least three specific challenges to design or analysis with responses, and any updates committed to.

What gets checked

Reviewer is a social science academic, quantitative methods researcher, or data scientist with social science experience
At least three specific design or analysis challenges are documented
At least one analysis update is committed to

Common mistakes

Reviewer validates without probing design assumptions
Not being able to explain deviations from pre-registered analysis under challenge
No updates after a genuine challenge to analytical choices

Powstik Guide

Your analysis, challenged by a social researcher.

Steps

Prepare a reason for every analysis choice.
Note what your design cannot rule out.
Book a social science academic or methods researcher.
Ask for three challenges to design and analysis.

Template

Challenge session log
Date:            Duration:
Reviewer name:
Reviewer role and credentials:

Challenge 1 (a design or analysis choice):
  What they asked:
  My answer:
  Did it hold?
Challenge 2:
  What they asked:
  My answer:
  Did it hold?
Challenge 3:
  What they asked:
  My answer:
  Did it hold?

Update I am committing to:

What gets sent back

A reviewer who doesn't probe assumptions.
Deviations you can't explain.
No update after the session.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

Confirm reviewer is a social science academic, quantitative methods researcher, or data scientist with social science experience
Review Q&A notes — confirm at least three specific design or analysis challenges are documented
Check that at least one analysis update is committed to

You'll sign in first, then come straight back here.

Part of

Sociology & Social Research →

Related outcomes

Milestone map

3 milestones

Develop a research question and select an appropriate quantitative design

4 weeks

Proof required

What gets checked

Research question is specific and testable with existing data
OSF pre-registration link present and timestamped before analysis begins
Key variables are operationally defined — not just named

Common mistakes

Pre-registering after analysis — the registration must be timestamped before any analysis
Using a dataset without verifying its variables match the research question — check variable availability before committing to the design
Operational definitions that are ambiguous — each variable must have a precise definition and a specific measurement approach in the dataset

Resources

Foundationstart here

What a verifier looks for

Confirm OSF pre-registration link is present and timestamped before any analysis
Verify dataset source is accessible
Check operational definitions — confirm each variable has a precise definition and measurement approach

You'll sign in first, then come straight back here.

Conduct the analysis and produce a transparent data report

5 weeks

Proof required

What gets checked

All pre-registered analyses are reported — including null results
Effect sizes and confidence intervals are reported alongside p-values
Data, code, and materials are deposited with an accessible link

Common mistakes

Reporting only statistically significant pre-registered analyses — all pre-registered analyses must be reported
Reporting p-values without effect sizes
Not depositing code — reproducibility requires the analysis code, not just the data

Powstik Guide

A transparent analysis report with code and data deposited.

Steps

Write your analysis as a script, not by clicking.
Run every pre-registered analysis, including nulls.
Report effect sizes with confidence intervals.
Log each deviation from the plan and why.
Deposit data, code and materials together.

Template

Descriptives (key variables):
| Pre-registered analysis | Result | Effect size (CI) |
|-------------------------|--------|------------------|
Deviations and reasons:
Repository link (data + code):

What gets sent back

Only significant results reported.
p-values without effect sizes.
No deposited code.

Resources

Foundationstart here

Depthgo deeper

Masteryfor the dedicated

What a verifier looks for

Check all pre-registered analyses are reported — compare report to pre-registration
Verify effect sizes and confidence intervals are present
Access repository link — confirm data and code are deposited

You'll sign in first, then come straight back here.

Defend quantitative analysis under social researcher challenge

1 week

Proof required

What gets checked

Reviewer is a social science academic, quantitative methods researcher, or data scientist with social science experience
At least three specific design or analysis challenges are documented
At least one analysis update is committed to

Common mistakes

Reviewer validates without probing design assumptions
Not being able to explain deviations from pre-registered analysis under challenge
No updates after a genuine challenge to analytical choices

Powstik Guide

Your analysis, challenged by a social researcher.

Steps

Prepare a reason for every analysis choice.
Note what your design cannot rule out.
Book a social science academic or methods researcher.
Ask for three challenges to design and analysis.

Template

Challenge session log
Date:            Duration:
Reviewer name:
Reviewer role and credentials:

Challenge 1 (a design or analysis choice):
  What they asked:
  My answer:
  Did it hold?
Challenge 2:
  What they asked:
  My answer:
  Did it hold?
Challenge 3:
  What they asked:
  My answer:
  Did it hold?

Update I am committing to:

What gets sent back

A reviewer who doesn't probe assumptions.
Deviations you can't explain.
No update after the session.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

Confirm reviewer is a social science academic, quantitative methods researcher, or data scientist with social science experience
Review Q&A notes — confirm at least three specific design or analysis challenges are documented
Check that at least one analysis update is committed to

You'll sign in first, then come straight back here.