Research Toolkit

Lessons learnt hard, written down. One technique among several — developed for mechanism-based empirical work in the organizational social sciences, and offered take-it-or-leave-it.

Featured

Companion interactive apps to Yeah, Yeah, But… and Therefore, a manuscript in progress. Manuscript site →

Numerical NarrativesUsing descriptive statistics and visualizations to move readers from possible to plausible to probable.
Read the primer →Open the app →
Differences-in-InferencesThe three-model architecture — prior, novel, reconciled — as a working tool.Open → Contribution ArchitectureBuild and refine an Elster diagram.Open →
Six core frameworks arranged around a hub: Coleman's Boat, Elster's Diagram, Differences-in-Inferences, Aspirational Abstract, Possible-Plausible-Probable, Empirical Etiquette

These frameworks carry the whole toolkit. Everything on the three pages below is one of them, applied to a phase of the work.

Developing

The Three-Story Approach

Three theoretically motivated, mechanism-specific accounts before you may claim evidence for one. The world is equifinal; one story cannot be confirmed and two collapses into a straw man.

Developing

Coleman’s Boat

Macro condition down to individual situation, across to action, and back up to the outcome. The trip is the explanation; the top arrow is only the correlation.

Developing

Elster’s Diagram

A hypothesis earns belief by generating implications that are observed — and rivals lose it by generating implications that are not. Cogs and wheels, not covering laws.

Framing

The Aspirational Abstract

Write it first and aspirationally, as the paper you wish you had. Its job is not to describe the project but to decide whether the project deserves your time — and to let weak ideas die cheaply.

Delivering

Possible → Plausible → Probable

Belief updates in increments. Possible: the phenomenon plus evidence for more than one account. Plausible: a first-glance figure where the accounts diverge, before controls. Probable: multivariate models, research designs, and simulations.

Delivering

Empirical Etiquette

Selection, attrition, naturalness, scaling. Justify the context as a strategic asset, and name the threats to inference before a reviewer does.

Delivering

Differences-in-Inferences

Three models — the prior result, the novel result, the reconciliation. The interaction is where the mechanism lives, and the reconciliation is the contribution.

Research Design & Methods Workshop

The ladder below follows the sequence of my Research Design & Methods Workshop, which I have run at INSEAD, USI, SKEMA, the Stockholm School of Entrepreneurship, UC3M, INCAE, Michigan and Minnesota. Each rung produces an artifact that a participant carries into the next session, and the tools on this page and my manuscript in progress, Yeah, Yeah, But… and Therefore, follow the same sequence.

Doctoral students and early-career researchers can work up the ladder one rung at a time with their own project. Instructors are welcome to run it as a workshop with their own students.

THE ARC WHAT YOU BUILD WHERE IT GOES POSSIBLE PLAUSIBLE PROBABLE THEREFORE 1. The question that denies a prior A phenomenon a reader cannot wave away, and a question no one can answer with “of course.” Introduction 2. Three stories Three live, mechanism-specific accounts. Each with a subject and a verb. Each given its best case. Theory 3. Elster diagram & design brief Three mechanisms, each with a signature that distinguishes it. The brief ends in a threats list. Theory 4. The three-model table The arc written as columns: prior result, novel result, reconciliation. Results 5. Research design & threats The ideal experiment named, then the most credible real route to a counterfactual. Methods 6. The credible comparison Ideal design beside actual, the gap annotated threat by threat. Own it before a reviewer finds it. Methods 7. Contrastive figure & code The naïve estimate and the design-based one, plus code a reader could rebuild your world with. Results 8. Aspirational abstract & introduction Six sentences, six jobs — then six paragraphs making the same promise at a larger scale. Abstract & Intro
The workshop ladder. Each artifact is the input to the next (maize, down the spine) and welds into a section of the manuscript (grey, to the right). Eight artifacts are not eight tasks — they are one paper. The belief shift the whole chain produces is the contribution.
01 · Developing IdeasEstablishing a phenomenon, specifying ignorance, and articulating the rivals your evidence must rule out.Open → 02 · Framing ArgumentsYeah, yeah, yeah… BUT. The aspirational abstract and the introduction as an opening statement.Open → 03 · Delivering InsightsThe numerical narrative, empirical etiquette, and the three-model reconciliation.Open →