spq-research-design

Research Design (spq-research-design)

SPQ accepts experiments, surveys, and observational/interpretive work, but is demanding about each. The

design must credibly connect the social-psychological argument (spq-theory-building) to evidence about

the structure–individual link. This skill is tradition-aware: pick the section that matches your work

and defend it against the strongest alternative explanation.

When to trigger

• Specifying an experimental setting, a survey/measurement plan, or a fieldwork/interview design
• A reviewer questioned causal claims, construct validity, case/site selection, or a confound
• Justifying why your design adjudicates the rival account from spq-literature-positioning
• Deciding how the design operationalizes the social-psychological mechanism

Experimental (group processes, status, exchange — the lab tradition)

• Standardized experimental settings. State the setting (e.g., status/expectation-states paradigm,

exchange networks) and how the manipulation realizes the theoretical construct.

• Manipulation / standardized-setting checks; randomization; attention checks; attrition.
• Inference: pre-specify primary outcomes; correct for multiple comparisons; power/MDE; appropriate

models for nested (group/dyad) data.

• Generalization: be explicit about what a lab effect does and does not license about real settings.

Survey / secondary-data (social structure and personality)

• Measurement first. Validate the social-psychological constructs (identity salience, mastery,

status, sentiment); report reliability; show results aren't an artifact of a scaling choice.

• Structural variables measured and theorized, not just controls — the structure–individual link is

the point.

• Inference for complex designs: survey weights/clustering for GSS/PSID-type data; multilevel models

for individuals nested in contexts; sensitivity to unobserved confounding for any causal claim.

Observational / interpretive (symbolic interaction)

• Site / case selection justified by analytic logic (what is this a case of?), not convenience.
• Evidence and disconfirmation: state what observations would have challenged the analytic claim;

document how interaction, accounts, or fieldnotes support it.

• Reflexivity and access: position of the researcher, consent, and how meaning is interpreted.

The adjudication test (SPQ-specific)

For the single strongest rival explanation, write one sentence: *"If the rival were true rather than

my argument, the data would look like ___; instead they look like ___."* For experiments this is the

manipulation contrast; for surveys, the confound ruled out; for interpretive work, the alternative

reading. If you cannot, the design does not yet identify the contribution.

Design stress ledger

Use a design stress ledger before committing to the analysis plan:

| Tradition | Stress test |

|-----------|-------------|

| Experiment | What manipulation failure, demand effect, group-composition imbalance, or dyadic dependence would overturn the status/process claim? |

| Survey / secondary data | Which omitted structural variable, measurement-invariance failure, weighting choice, or contextual clustering rule could flip the conclusion? |

| Observational / interpretive | Which negative case, deviant interaction, or access/reflexivity concern would force a narrower interpretation? |

For each row, write the planned diagnostic and the interpretation if it fails. SPQ reviewers are

comfortable with different methods, but they expect the method's limits to be explicit. A design that

names its own failure mode usually reads stronger than a design that implies no failure mode exists.

Execution bridge (StatsPAI / Stata MCP)

Estimate and audit the design, don't only describe it. Full map:

[execution-with-mcp](../../../shared-resources/empirical-methods/execution-with-mcp.md). SPQ spans lab/survey experiments and observational work; randomization inference and mediation done right matter for the experimental lane.

• detect_design → recommend → fit with as_handle=true → audit_result.
• Observational causal claims: staggered DiD (callaway_santanna / sun_abraham +

bacon_decomposition + honest_did_from_result); IV (effective_f_test +

anderson_rubin_ci); RDD (rdrobust + mccrary_test).

• Experiments: randomization-based inference, romano_wolf for many-outcome

family-wise control, and mediate for mediation (not naive controlling-away).

• Sensitivity: oster_delta / sensemakr for observational claims.

Report the effect size in interpretable units; route the full battery to the

appendix/supplement. A run end-to-end (synthetic data, real returns) is in the

[JF execution walkthrough](../../../Journal-of-Finance-Skills/resources/worked-examples/02-execution-walkthrough.md).

Anti-patterns

• A lab effect over-generalized to real-world structure with no caveat
• Treating structural variables as nuisance controls rather than theorized causes
• Convenience site selection dressed up as theory-driven
• "Causal" language on a cross-sectional survey that only supports association
• A design that cannot distinguish your social-psychological mechanism from the leading alternative

Output format

【Tradition】experiment / survey-SSP / observation-interpretive
【Estimand or analytic claim】what is identified/shown about the structure–individual link
【Key assumption(s)】and how each is defended
【Rival ruled out】the adjudication sentence
【Measurement / setting validity】constructs validated or setting standardized? [Y/N]
【Robustness/sensitivity】planned checks
【Next】spq-data-analysis

Supplementary resources

• [../../resources/external_tools.md](../../resources/external_tools.md) — experimental, survey, measurement, and CAQDAS tooling by tradition
• [../../resources/official-source-map.md](../../resources/official-source-map.md) — SPQ scope and methods breadth