Social Science Research Assistant

$30–50/hrrate as listed

Write authentic social science research tasks — survey analysis, qualitative coding, literature synthesis, statistical work — that frontier AI models are then scored against. Each task ships with realistic research artifacts and a 35+ item rubric covering methodology, execution, and interpretation.

Summary, eligibility guidance and screening preparation are prepared by Sidequest from the platform's posting. Check the source listing before applying.

The work

You build evaluation items that look like a real week in a research shop, not a methods textbook. That means starting from a plausible research scenario in your discipline, then producing the artifacts it would generate: a survey instrument with its known flaws, a partially coded interview corpus, SPSS or Stata output with ambiguities in it, a stack of source documents that don't quite agree. You then specify the defensible methodological path through the material and write a rubric — typically 35 or more discrete items — that separates a competent analysis from one that merely sounds competent.

The rubric is where most of the difficulty lives. Graders need line items granular enough to catch a model that picks the right test for the wrong reason, mishandles missing data, over-reads a small effect, or produces clean-looking qualitative codes that don't hold up against the transcripts. Expect asynchronous back-and-forth with project leads and reviewers, and expect revision requests on early submissions while you calibrate to the spec.

What the screen looks for

micro1 runs an AI-led interview before any task work. It probes for specifics rather than credentials in the abstract: which instrument you designed and what you'd change about it, how you established intercoder reliability and what you did when it was poor, which software you actually run and what you do in it. Follow-up questions go a layer deeper than the first answer, so vague methodological language tends to surface quickly. No AI or ML background is required or expected.

Logistics

Fully remote, contractor, asynchronous — no fixed hours or standing calls.
Pay is output-based, per accepted task; the $30–50/hr band reflects observed effective rates and is not guaranteed, since throughput depends on your workflow and revision cycles.
A weekly minimum submission volume applies.
Hiring moves fast: roles often fill within 48 hours, with first tasks expected 24–48 hours after onboarding.
The work is deliberately detailed and at times tedious — artifact construction takes real time.

What they look for

Methodological rigor
Source synthesis
Rubric fidelity
Research mastery

Before you apply

Would you pass this screen?

Drawn from what this listing actually requires. Answered in your browser — nothing is sent to us, and we do not store your answers.

Do you hold a Bachelor's degree or higher in sociology, economics, psychology, political science, or a closely related social science field?
Degree in social science discipline · required
Do you have a minimum of 2 years of hands-on experience designing survey instruments, coding qualitative data, conducting literature reviews, or working with statistical software such as SPSS, R, Stata, or Python?
Research experience with surveys and data · required
Are you proficient in written English and comfortable authoring technical research content, including detailed rubrics and methodological documentation?
Proficient written English · preferred
Are you available to start your first tasks within 24–48 hours of completing onboarding and to maintain a reliable submission schedule as a contractor?
Remote work availability · required
Have you used at least one statistical software package (SPSS, R, Stata, Python, or equivalent) for data analysis?
Statistical software experience · preferred

0 of 5 answered

Included with this listing

Interview prep — 12 questions

Matched to what this platform's screening actually measures. Answer each out loud in 60–90 seconds: claim → concrete example → implication.

What it tests: Can they articulate the reasoning behind methodological choices in survey construction, or do they describe surveys as a checklist of items? This role demands someone who understands the theory-to-measurement pipeline, not just mechanical task completion.
A strong answer: A strong answer names a specific research context, explains why certain question formats (Likert, open-ended, branching) matched the construct being measured, and mentions a concrete decision (e.g., reverse-coding, avoiding leading language, pilot testing) that improved validity or reliability.
Where people lose it: Candidates often describe what they put in a survey without explaining *why*—'I asked about income, education, and beliefs' without discussing item validity, response bias, or how the design served the research aim. This signals they follow templates rather than reason through methodology.