Cognitive Connie
Understanding variables & operationalisation
Before a single participant enters a study, researchers face a fundamental challenge: psychology is full of rich, abstract concepts — "stress," "memory," "aggression," "intelligence" — that exist in the mind but cannot be directly measured. Turning these constructs into something a measuring instrument can actually record is the work of operationalisation, and it shapes everything that follows.
Key figures
Gustav Fechner
1801–1887German philosopher and psychologist who pioneered the operationalisation of mental phenomena. His Elemente der Psychophysik (1860) established that sensory experiences — brightness, loudness, weight — could be measured systematically through controlled procedures mapping physical stimulus intensity to subjective perception. Fechner's work laid the groundwork for the idea that psychology's abstract constructs could be made measurable.
John B. Watson
1878–1958Argued forcefully in his 1913 Behaviourist Manifesto that psychology could only be scientific by restricting itself to objectively observable and measurable behaviour. Watson's insistence on operational definitions of both stimuli and responses — rather than unobservable mental states — became the foundation of the behaviourist tradition and permanently shaped how psychological variables are operationalised.
Stanley Stevens
1906–1973American psychologist who formalised the classification of measurement scales (nominal, ordinal, interval, ratio) in a landmark 1946 Science paper. Stevens' typology determines which statistical procedures are valid for a given measure, making it a cornerstone of operationalisation decisions: knowing whether your DV is interval-scaled or merely ordinal changes which analyses you can legitimately apply.
Key concepts
Independent variable (IV)
The variable that the researcher deliberately manipulates or varies across conditions. It is the presumed cause in a cause-and-effect relationship. Good operationalisation of the IV specifies exactly what is changed (e.g. "0 mg, 100 mg, or 200 mg of caffeine in a 250 ml drink") rather than leaving it vague (e.g. "caffeine condition").
Dependent variable (DV)
The variable that is measured to assess whether the IV had an effect. It is the presumed outcome or effect. The DV "depends" on the IV. Operationalising the DV means specifying exactly how it is measured: not "reaction time" but "time in milliseconds from onset of a visual cue to button press on a USB response pad."
Operationalisation
The process of defining a variable in terms of the specific, measurable procedure used to assess or manipulate it. Converts an abstract construct into a concrete, replicable operation. Crucially, different operationalisations of the same construct (e.g. "anxiety" as GAD-7 score vs heart rate vs skin conductance) may yield incompatible findings — not because either study is wrong, but because they measured different facets of the construct.
Confounding variable
A variable that varies systematically with the IV and independently affects the DV, making it impossible to determine whether any observed change in the DV was caused by the IV or the confound. The key word is "systematically" — if it varies randomly rather than consistently with the IV, it is an extraneous variable that adds noise but does not bias the estimate of the IV's effect.
Extraneous variable
Any variable beyond the IV that could affect the DV. Unlike a confound, an extraneous variable varies randomly across conditions rather than systematically with the IV — adding noise and reducing statistical power, but not creating a false directional bias in the results. In practice, distinguishing extraneous from confounding variables requires careful analysis of whether the variable co-varies with condition assignment.
Control variable
A variable that is deliberately held constant across all conditions to prevent it from becoming a confound. For example: same testing room, same time of day, same experimenter, same task materials. Controlling variables improves internal validity — if only the IV differs between conditions, any change in the DV can be attributed to it with greater confidence.
Directional (one-tailed) hypothesis
A hypothesis that specifies both that the IV will affect the DV and the direction of that effect — for example, "sleep deprivation will decrease memory performance." Used when prior theory or evidence strongly predicts the direction. Allows a one-tailed statistical test, which is more powerful for detecting an effect in the predicted direction but cannot detect effects in the opposite direction.
Non-directional (two-tailed) hypothesis
A hypothesis that predicts an effect but does not specify its direction — for example, "sleep deprivation will affect memory performance." Used when the direction of effect is unclear or when the researcher wants to remain open to effects in either direction. Requires a two-tailed statistical test. The null hypothesis always takes the complementary position: no significant effect.
Test your knowledge
Frequently asked questions
What is the difference between a confounding variable and an extraneous variable?+
An extraneous variable is any variable beyond the IV that could affect the DV. If it varies randomly across conditions, it adds noise (reduces power) but does not bias the direction of results — it is extraneous. If it varies systematically with the IV — consistently higher in one condition than another — it becomes a confounding variable that can produce false positive or false negative results. For example, if participants in the experimental condition are tested in the morning and control participants in the afternoon, time of day is a confound rather than merely an extraneous variable.
Can a study be valid even with a weak operationalisation?+
A study can be internally valid (its manipulation genuinely causes the observed change) while having low construct validity — meaning the operationalisation measures something real but not quite the intended construct. For example, a study showing that "hitting a Bobo doll more after watching an aggressive model" might be internally valid (the model did cause more hitting) while having construct validity problems (is hitting an inflatable toy the same construct as real-world aggression?). Strong operationalisation is necessary for conclusions to extend meaningfully beyond the specific measurements used in the study.
Why is a null hypothesis always framed as "no effect"?+
The null hypothesis (H₀) is framed as "no effect" because null hypothesis significance testing (NHST) works by calculating the probability of observing the data if H₀ were true. Researchers then ask: is this probability low enough (below alpha, typically .05) to reject H₀ in favour of the experimental hypothesis? You cannot directly prove the experimental hypothesis — you can only reject its logical opposite. A null result (failing to reject H₀) means the evidence was insufficient to conclude an effect exists; it does not prove no effect exists.
Sources
Last reviewed July 2025- 1.
Campbell D.T. & Stanley J.C. (1966). Experimental and Quasi-Experimental Designs for Research. Rand McNally.
+About this source
Foundational text on research design validity, introducing the concepts of internal and external validity and threats to each.
- 2.
Field A. (2018). Discovering Statistics Using IBM SPSS Statistics (5th ed.). Sage.
+About this source
Widely used statistics and research methods textbook covering variables, hypotheses, and quantitative design.