skills/full-company-analysis/references/modules/M004-hypothesis-driven-research.md

<!-- Generated loss-aware reference mirror from God_Level_Public_Company_Financial_Analyst_Job_Guide_V6_99_ALL_SUB70_FIXED.docx. Canonical source remains the bundled DOCX. -->

<!-- Module: 004 | Title: Hypothesis-Driven Research -->

## PART I - ANALYST OPERATING SYSTEM | MODULE 004

# Hypothesis-Driven Research

> Mission. Turn vague curiosity into explicit hypotheses, falsification tests, and prioritized research questions.

## Decision output

Objective: Turn vague curiosity into explicit hypotheses, falsification tests, and prioritized research questions. The completed work product must be reproducible from evidence, show the downstream financial or decision effect when material, state the strongest contrary case, and define a dated update rule.

## Explicit operating procedure

1. Convert the decision question into explicit competing hypotheses, not one favored story. Each hypothesis must predict observable evidence.

1. List the evidence that would be expected if each hypothesis were true and the evidence that would be surprising. Prioritize tests with the highest ability to discriminate between explanations.

1. Assign a starting confidence range or qualitative prior based on base rates and known evidence. Do not manufacture numerical precision when no defensible probability exists.

1. Sequence research so cheap, high-information tests occur before expensive channel work or complex modeling. Stop a line of inquiry once it can no longer change the decision.

1. Record disconfirming evidence with the same prominence as confirming evidence. A hypothesis that cannot be falsified is not a usable research hypothesis.

1. At conclusion, state which hypothesis best fits the evidence, what remains unresolved, and the next observation that could reverse the ranking.

## Required evidence and model bridge

- Primary-source set: source log, assumption register, decision journal, model-change log. Preserve exact document/version, date, period, and source location for every material factual input used in hypothesis-driven research.

- For each key concept - falsifiable claims, base rates, alternative explanations, Bayesian updating, variant perception, precommitment - state whether it is a reported fact, analyst calculation, management claim, external estimate, or judgment. Quantitative concepts must retain raw components and units; qualitative concepts must retain the specific evidence and counterevidence.

- Map only economically relevant findings into the model or decision record. Process-control modules such as hypothesis-driven research may have no direct valuation line; in that case document the downstream error or governance risk the control prevents.

## Metrics and calculation controls

| Metric / concept | Construction | Required validation |
| --- | --- | --- |
| hypothesis specificity | % of research hypotheses stated with a measurable variable, expected direction/magnitude, time horizon, and explicit falsification condition. | hypothesis specificity: Audit a sample back to dated evidence and decision records; verify the stated threshold/score is reproducible by an independent reviewer and tied to a defined decision consequence. |
| expected value of information | Expected decision-value improvement from resolving a question: sum of probability of each possible answer × change in decision value, less research cost. | expected value of information: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. |
| probability update magnitude | Absolute change in assigned scenario or thesis probability after new evidence, measured in percentage points and linked to the triggering evidence. | probability update magnitude: Document scenario definitions and probabilities; verify probabilities sum appropriately, inputs are independently sourced, and sensitivity is recomputed rather than manually overridden. |



## Worked application

> Case: replace "pricing power is strong" with a measurable price, volume, and timing claim.

- Reconstruct the relevant reported fact from primary evidence before interpreting the case. For hypothesis-driven research, show the raw components rather than only the resulting ratio or narrative.

- Build the causal chain through falsifiable claims, base rates, alternative explanations, Bayesian updating, then identify which link is directly observed and which link remains an assumption.

- Calculate hypothesis specificity, expected value of information, probability update magnitude from sourced components under the reported/base interpretation and at least one skeptical alternative interpretation.

- Translate the difference between cases into the variable that matters for hypothesis-driven research: evidence quality, revenue, operating profit/NOPAT, free cash flow, invested capital, financing/dilution, risk, or valuation. Mark non-applicable links instead of inventing them.

- Expert consistency test: rank questions by how much evidence can move value, probability, or downside.

- Precommit the specific future filing, KPI, customer/supplier observation, regulator action, or market input that would materially invalidate the hypothesis-driven research conclusion.

## Failure tests

- FAIL if falsifiable claims cannot be defined and reproduced from the source pack.

- FAIL if the research question is not falsifiable, has no competing explanation, or evidence collection is selected after the analyst already knows which answer it supports.

- FAIL if the hypothesis-driven research conclusion depends on an unstated assumption, unreconciled definition, or evidence that cannot be traced to its source/version.

- FAIL if evidence materially inconsistent with the hypothesis-driven research conclusion is omitted, reclassified, or dismissed without a documented definition, materiality, causal, timing, and source-quality analysis.

## Completion test

A senior reviewer must be able to reproduce the hypothesis-driven research conclusion, vary the most sensitive assumption independently, trace the change through the model, understand the strongest opposing case, and identify the next evidence that would force an update. If any link is missing, the module remains open.