Orgstud methods

Methods & Research Design (orgstud-methods)

When to trigger

• You are choosing between a qualitative/process design and a quantitative one
• The design is chosen but its rigor and transparency are not yet defensible to OS reviewers
• A qualitative study lacks a sampling logic, immersion account, or trustworthiness safeguards
• A quantitative study leads with the estimator instead of the organizational mechanism it reveals

Method follows the theoretical question — and OS leans qualitative

At OS, no method is privileged in principle, but the journal's center of gravity is qualitative, ethnographic, process, and historical research, and such work is genuinely first-class here — not a tolerated minority. The non-negotiable is that the design fits the question (see orgstud-theory-development) and is executed with craft. A sophisticated estimator cannot rescue a thin theory, and a single immersive case can carry an OS paper if the theoretical insight is deep — a different bar from venues where a clean identification design is itself treated as the contribution. OS reviewers ask, above all, does this design let you see the organizing process you claim to theorize?

Branch A — Qualitative / process / ethnographic / historical

Use for how/why organizing unfolds: emergence, becoming, contestation, meaning, identity, institutional dynamics.

• Theoretical (not convenience) sampling. Cases/sites/informants chosen to illuminate the process or construct; state the logic — polar/extreme cases, theoretical replication, revelatory case, longitudinal window.
• Access and immersion. Specify duration, depth, and your role (participant vs. non-participant); for ethnography, time in the field; for historical work, the archive and its limits.
• Triangulated data sources. Interviews (count, who, when, guide), observation, internal/archival documents, secondary sources — and how they corroborate.
• Process design. If the contribution is a process model, build in the temporal leverage: real-time and/or retrospective data, event sequences, turning points (Langley's process strategies — narrative, temporal bracketing, visual mapping — are the standard idiom).
• Trustworthiness. Credibility, transferability, dependability, confirmability: member checks, prolonged engagement, audit trail, investigator triangulation, negative-case analysis.
• Reflexivity. State your standpoint and how it shaped access and interpretation — expected at a European, critically-aware journal, not optional.

Branch B — Quantitative organization theory

Use for whether/how much/under what conditions across many cases — welcome at OS when it does organization-theoretic work.

• Sample and unit of analysis justified by the theory (organizations, fields, events, dyads, individuals nested in units).
• Identification in service of theory. Be explicit about the causal claim and its threat (panel FE, event-history/survival, matching, natural experiments, DiD with modern staggered-adoption caveats). At OS, identification is a means to a theoretical end; a flawless quasi-experiment that yields no new understanding of organizing is still rejected. Lead with the mechanism, not the estimator.
• Measurement validity. Operationalizations defended; multi-item reliability; common-method bias addressed if same-source.
• Multilevel structure. If the theory is cross-level, use appropriate models and justify aggregation.

Either branch

• The design must let you see the mechanism / process, not just the endpoints.
• Pre-empt the obvious alternative explanations at the design stage, not only in robustness.
• Plan the data-to-theory link now — it feeds orgstud-data-analysis and orgstud-tables-figures.

Execution bridge (StatsPAI / Stata MCP)

For the empirical / causal lane, estimate and audit rather than only specify. Full map: execution-with-mcp. Organization Studies is largely qualitative/theoretical; use the chain below only for its quantitative-empirical papers, and say so when a study is interpretive.

• detect_design → recommend → fit with as_handle=true → audit_result to enumerate the checks the design owes.
• Panel / 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 and romano_wolf for the many-outcome family-wise correction reviewers expect.

Match the toolchain to the reviewer pool, and report the effect size the venue wants. A run end-to-end (synthetic data, real returns) is in the JF execution walkthrough.

Checklist

• Design matches the theoretical form (process → qualitative; variance → quantitative)
• Qualitative: theoretical sampling logic stated; access/immersion specified
• Qualitative: multiple triangulated sources; trustworthiness safeguards named; reflexivity addressed
• Process work: temporal leverage built in (real-time/retrospective; turning points)
• Quantitative: identification explicit and subordinated to the organizational mechanism
• Quantitative: measurement validity and (if needed) multilevel structure handled
• Obvious alternative explanations are designed against, not just discussed later

Anti-patterns

• Convenience sampling dressed up as theoretical sampling
• Qualitative work with no transparency about coding, sources, or fieldwork depth
• Treating a fancy estimator as the contribution when the question needs none
• A quantitative paper that reads as applied econometrics with organizational variables bolted on
• A design that can show that something happens but never how/why organizing produces it
• Omitting reflexivity in interpretive work at a journal that expects it

Output format

【Design】qualitative (ethnographic/process/historical) / quantitative (type)
【Why it fits】link to the theoretical question and process/mechanism
【Sampling/identification】logic + key threat addressed
【Data sources】list + triangulation / measurement plan
【Temporal leverage】how the design captures process (if applicable)
【Rigor safeguards】trustworthiness + reflexivity, or identification checks
【Next skill】orgstud-data-analysis