Research Methodology

Collector research is not merely the accumulation of facts. It is a controlled process for turning observations, records, comparisons and claims into conclusions that can be explained, challenged and revised.

A sound method keeps the question, evidence trail, competing explanations and level of confidence visible. It prevents a plausible clue, repeated story or attractive attribution from becoming stronger in the record than the evidence allows.

Explore research methodology

Ten detailed topics grouped by the stage of investigation.

10 detailed topics

Frame and plan the investigation

Turn a broad curiosity into a bounded research task with defined scope, alternatives, evidence needs and stopping points.

Build and test the evidence

Trace sources, compare competing explanations and determine whether apparently supporting records are genuinely independent.

Record and revise the conclusion

Preserve the reasoning, state confidence proportionately and keep the work capable of being checked and changed.

The central methodological problem

Research often begins with an object that already carries a label, family story, catalogue attribution or market description. The danger is not that these starting points are always wrong, but that they can silently define the answer before the investigation begins.

A stronger method converts the inherited claim into a testable proposition, records the object independently, creates credible alternatives and preserves every source and reasoning step. The result may still be provisional, but it becomes defensible and useful.

A repeatable research sequence

1. Define the exact question

Name the object or population, the proposition being tested, the alternatives that must be distinguished and the decision the answer will support.

2. Record the starting position

Capture the object's present condition, marks, measurements, provenance, existing claims and the researcher's prior beliefs before interpretation begins.

3. Plan the evidence route

Identify the source families, comparison examples, inclusion criteria, chronological boundaries and likely gaps before collecting large volumes of material.

4. Build competing hypotheses

Create plausible alternatives and define what evidence would support, weaken or rule out each one. A method that can only confirm the preferred answer is not a test.

5. Trace and evaluate sources

Follow claims to their earliest available source, preserve citation detail and distinguish direct evidence, interpretation, repetition and market language.

6. Test corroboration and conflict

Ask whether separate sources are truly independent, whether they address the exact proposition and whether contradictory or negative evidence has been sought deliberately.

7. State confidence and limitations

Write the conclusion at the level the evidence supports, explain the unresolved weaknesses and identify what new evidence would change the result.

8. Preserve and revisit the work

Keep notes, sources, superseded conclusions and review triggers so later researchers can reconstruct the path rather than inheriting only the final claim.

Important distinctions

A research question is not a desired answer

A useful question leaves more than one outcome open. Framing the task around proving a preferred attribution invites confirmation rather than investigation.

Evidence is not the same as interpretation

A mark, record or material feature is evidence. The meaning assigned to it is an interpretation that must be tested against alternatives.

Repetition is not corroboration

Many records may repeat one original source. Corroboration requires sufficiently independent evidence that bears on the same proposition.

Incomplete evidence is not neutral evidence

Missing records can limit confidence, but they do not automatically support or disprove a claim. Their significance depends on what should reasonably have survived or been found.

Confidence belongs to individual claims

Identity may be secure while date, originality, provenance or rarity remain uncertain. A single overall confidence label can conceal important differences.

Revision is not failure

Changing a conclusion when evidence improves is evidence of a functioning method. Silently replacing the old answer destroys the research history.

Detailed Topics

Related Topics