Research Questions, Hypotheses, and Protocols

Lesson 1.1 · 12 min

From curiosity to a research question

A research question narrows a broad interest into something you can actually investigate. "Does this compound do anything interesting" is not a research question. "Under stated conditions, does exposure to this compound change a specific, measurable outcome in a specific model, compared to a specific control" is closer.

A good question names the variable of interest, the outcome you will measure, and the comparison that will let you tell whether anything happened at all. If you cannot write those three things down, the question is not ready to guide any activity.

Vague questions produce vague records, and vague records produce nothing anyone can trust later, including you. Time spent sharpening the question before you begin is the highest-value hour in the entire project.

Level II assumes you already understand basic RUO handling from Level I or equivalent experience. This module builds the thinking layer on top of that handling competence.

Research question:

A specific, answerable statement of what you intend to find out, naming the variable, the outcome, and the comparison.

Lesson 1.2 · 12 min

Developing a hypothesis

A hypothesis is a specific, testable statement about the relationship between a variable and an outcome, written before you collect data. It is not a wish, and it is not the same thing as a prediction of what you personally expect to happen.

A workable hypothesis has a companion: the null hypothesis, the statement that nothing changed. Writing both forces honesty about what would count as evidence against your idea, not just evidence for it.

Hypotheses should be falsifiable. If no conceivable observation could ever contradict your statement, it is not a hypothesis, it is an opinion dressed in scientific language.

Lesson 1.3 · 14 min

Writing a research protocol

A protocol is the written plan for how a question will be investigated: the materials, the procedure steps in order, the measurements, the timing, the number of replicates, and the criteria for stopping or excluding data. It is written before the work begins, not reconstructed afterward from memory.

The test of a protocol is transferability. If a competent stranger could not follow it and end up doing approximately what you did, the protocol is incomplete. Ambiguity in a protocol becomes noise in the results.

Protocols also state what you will not do: what conditions would cause you to stop, what you will do with unexpected observations, and what counts as a deviation worth recording. A protocol without stopping rules invites decisions made in the moment, which are the decisions most vulnerable to bias.

Protocol:

A written, dated plan describing materials, procedure, measurements, and decision rules, prepared before data collection begins.

Deviation:

Any departure from the written protocol during execution, which should be recorded rather than silently absorbed.