Biology · Research Design, Data & Scientific Practice
Research questions, operational definitions and testable predictions
Two imagined plots have different average plant heights. One question asks how large that difference is. Another asks whether changing a specified condition would change height. The same two numbers cannot answer both questions without further information.
Life, Evidence & Biological Scale introduces evidence and biological organisation. Organisms, Environments & Ecological Niches and Population Ecology & Life-History Patterns develop the importance of boundaries and denominators. Their English chapters are available in this working manuscript; their publication remains pending. The examples here require arithmetic, averages and ordinary graph reading. The sample-variance and confidence-interval example explains its notation as it is introduced; it can be revisited after the simpler comparisons.
Back to the Biology mapBy the end, you can
Turn a broad biological interest into a specified question, measurable variables and predictions that can distinguish explanations.
Name what the question is about
A research question specifies what is being investigated and the kind of answer sought. Describing a population, explaining a mechanism, predicting a future observation and estimating the effect of an intervention are different aims. The original question map uses one imagined plant population to show how the required evidence changes with the aim.
For a descriptive question, the target might be the average height of all plants of a defined kind in one bounded plot at a stated time. For a comparison, it might be the difference between two such averages. A causal question asks about the difference that changing a condition would make while specifying the relevant population, outcome and setting. An association observed under existing conditions is a different quantity.
The words healthy, diverse or successful are too broad to serve as a measurement by themselves. A question about height does not automatically answer a question about survival, reproduction or lifetime fitness. Choosing an outcome gives the investigation a clear target while leaving other biological questions open.
A different question needs a different comparison
Read the diagram in words
Description, association, intervention and prediction use the same imagined plants but ask for different quantities. Every case needs defined units and timing. The causal case additionally needs a justified assignment comparison; a descriptive average does not supply that design.
| Question type | Quantity or relation sought | Evidence required | What this alone leaves unresolved |
|---|---|---|---|
| Description | Average assigned height in the defined plot at one time | Included plants and comparable height records | Does not identify the effect of a change |
| Association | Difference between existing condition groups | Defined groups and comparable outcomes | Other differences may account for the association |
| Intervention | Difference attributable to changing the stated condition | An assignment and comparison design addressing alternatives | Requires justified design assumptions |
| Prediction | A later height record under specified conditions | A prediction stated separately from its evaluation records | Predictive success alone need not identify a mechanism |
| Question type | Evidence required | Connects a question to its evidence requirements |
|---|---|---|
| Describe one population | Specified plants, outcome and time | Connects a question to its evidence requirements |
| Compare existing conditions | Specified plants, outcome and time | Connects a question to its evidence requirements |
| Estimate a change caused by assignment | Specified plants, outcome and time | Connects a question to its evidence requirements |
| Predict a later observation | Specified plants, outcome and time | Connects a question to its evidence requirements |
Height in cm at the stated time
Original conceptual map, with a stipulated plant population and observation interval. The relations describe questions and evidence requirements, not causal pathways, physical distances or actual study findings.
Connect an idea to an observable rule
An operational definition states how an idea is represented in the investigation. In the original descriptive case, height means the assigned vertical extent in centimetres at the stated observation time. Count means the number of included plants under the stated boundary rule. These definitions make two records comparable; they do not make height or count a complete account of plant condition.
A record needs its unit and timing. A height of eight centimetres is an amount at a time. Growth of two centimetres over an interval is a change. A rate would also divide by the interval length. Comparing final heights alone may mix differences already present at the beginning with differences that developed later.
Measurement and biological meaning must remain connected but distinct. Two devices can report the same number while measuring different properties, and two investigators can use the same word while applying different inclusion rules. Clear definitions let a reader identify these mismatches before interpreting a pattern.
Ask which observation would separate explanations
A hypothesis offers an explanation that implies consequences under specified conditions. A prediction states an observation expected if those conditions and the explanation hold. Several explanations can predict the same simple pattern. Agreement with that pattern alone therefore need not distinguish them.
Suppose an imagined association between bright locations and taller plants could arise from a response to light, from different inherited backgrounds occupying those locations, or from both. Merely finding another bright location with tall plants leaves these possibilities open. A comparison that records initial state and inherited background, or changes a condition under a suitable design, addresses a more discriminating question.
Scientific work can also begin with exploration: an unexpected pattern suggests a new question. Its timing matters. A prediction written after seeing the outcome has a different evidential role from one specified before those outcomes were known. Describing this honestly allows discovery to inform a later test without presenting the discovery data as an independent confirmation.
Worked example
Match the question to its evidence
The original map's description case asks for the average height in a defined plot at one time. It needs the target boundary, inclusion rule and height records. It does not need to assume that a change was assigned.
The causal case asks how a specified condition would change height over one interval in the chosen population. It needs a suitable comparison and a design that addresses alternative explanations. Two existing plot averages alone leave initial differences and other conditions unresolved.
The prediction case asks about a later record under stated conditions. A successful prediction can support predictive usefulness, but the same outcome may be predicted by more than one mechanism. State what the observation distinguishes before choosing an explanation.
Check your understanding
Why is tallest plant an incomplete operational definition of population success?
Show reasoning
It specifies an extreme height, not survival, reproduction or the whole population's condition. A study may use that outcome, but it must not silently generalise it to every meaning of success.
Two hypotheses both predict taller plants in bright plots. Does finding that pattern establish one of them?
Show reasoning
No. The observation is compatible with both. More discriminating evidence is needed to separate a response to conditions from different inherited composition or another explanation.
Why record whether a prediction was made before outcomes were known?
Show reasoning
That chronology distinguishes a prespecified test from an explanation suggested by the same observations. Both can be useful, but the latter does not supply an independent confirmation of itself.
What to carry forward
Specify the population, variable, comparison and time; then identify the observation that could actually distinguish the proposed explanations.
References
Resources supporting this lesson’s explanations and examples.