3.5.1, 3.2.1, 3.2.2 - Biological correlation investigation

A question about age and drug attitudes

A biological-topic correlation can investigate attitudes to drug use without asking anyone to take a drug or disclose personal use.

Our example question is: Is adult age associated with approval of nonmedical recreational drug use? The two co-variables are age and attitude score. A correlation measures both for each person and examines their relationship. Neither is assigned by the researcher. Consequently, there is no manipulated independent variable or experimental control group in this plan.

A non-directional alternative hypothesis predicts a relationship between age in completed years and the defined approval score in the sampled adults. The null hypothesis predicts no population association between those measurements. A directional alternative would specify a positive or negative relationship, but it needs an evidence-based prediction decided before inspecting results. Without a sufficient basis for one direction, our plan uses a two-tailed test.

Positive correlation would mean that higher ages tend to accompany higher approval scores; negative correlation would mean higher ages tend to accompany lower scores. Little or no monotonic association would mean that higher age does not consistently accompany higher or lower scores. These patterns describe relationships, not causes. A correlation of age with attitudes could reflect cohort experiences, education or social context. Even if age necessarily comes before the current attitude measurement, the design does not separate those alternatives.

In an experiment, by contrast, a researcher might assign adults randomly to receive neutral information or an ethically suitable comparison information sheet, then measure attitude scores. The information condition would be an IV, the attitude score a DV, and the comparison condition would help judge the intervention’s effect. The experimental alternative hypothesis would predict a difference in approval scores between the information conditions, rather than an association with age. Random allocation balances possible participant differences in expectation; it differs from random sampling, which concerns selection from a population. That experiment answers a different question from the age correlation and is not the required correlational practical.

Operationalising and protecting the investigation

The measurements and safeguards must be decided before recruitment so that the question is both answerable and ethical.

Record age in completed years. For a simple attitude tool, use the same statement for everyone: “Nonmedical recreational drug use is acceptable.” Ask adults to rate agreement from 1, strongly disagree, to 9, strongly agree, with 5 labelled neither agree nor disagree. Explain “nonmedical” as use outside medical treatment and tell participants not to provide information about their own drug use. Higher scores operationalise greater approval of this statement. A single broad item is easy to administer, but different interpretations of “drug” limit validity; it does not measure addiction, actual use or the safety of any substance.

This score is ordinal: responses have an order, but we cannot assume that the psychological difference between 1 and 2 equals that between 7 and 8. Interval measurement requires meaningful equal numerical intervals; age in years supplies equal-interval numerical information, but the ordinal attitude score still supports an analysis based on ranks. Nominal measurement classifies cases into categories without a numerical order, such as which local club recruited each adult. Club names cannot supply the ranked attitude scores needed for this analysis. Pilot the wording with appropriate volunteers to check clarity and comfort. A future refined tool could specify a safe hypothetical context or use several balanced items, but its scoring must be declared and justified before data collection.

Recruit consenting adults aged 18 or over from a clearly defined accessible population, using an opportunity sample with a useful spread of ages. Record exactly who was eligible, how invitations were made and how many participated. A small local volunteer sample is feasible but may overrepresent particular social backgrounds. Do not describe it as randomly selected. A probability sample from an appropriate sampling frame could improve representativeness; practical access and nonresponse still matter.

Seek supervisor approval before conducting the practical. Provide information about the topic, voluntary participation, the right to skip an item and withdrawal arrangements. Avoid pressure from teachers, employers or relatives. Collect no names, diagnoses or personal-use histories; use participant codes and a secure separate withdrawal mechanism if codes need to be linked temporarily. State the date after which genuinely anonymous aggregate data cannot be removed. Exact ages can help identify someone in a small group, so do not publish individual records or identifiable combinations. Use respectful debriefing and explain that the study is about attitudes rather than judging or labelling people.

The topic should not encourage drug use, include distressing material or expose illegal behaviour. If suitable recruitment and safeguards cannot be arranged, analyse supplied data instead and accurately report that limitation. Analysis practice must not be presented as a completed personal data-collection exercise.

Collecting paired data consistently

Each participant must contribute one age and one attitude score that remain correctly paired throughout analysis.

Give the same wording, instructions and response scale privately. Allow a person to decline without explanation. Record each complete pair under a code, check that scores lie within 1–9 and ages meet eligibility, and record missing responses honestly. A missing attitude cannot be replaced with a guessed “neutral” score. Decide how incomplete pairs will be handled and report exclusions; do not selectively discard a valid pair because it weakens the expected relationship.

Standardisation improves reliability because differences are less likely to arise from different instructions. Private responding may reduce pressure to give socially approved answers, although social desirability can remain. The item’s stability could be assessed with a suitable repeat measurement, but a later response may also reflect genuine attitude change. Reliability does not guarantee validity: consistent answers to an ambiguous statement could still fail to measure the intended attitude well.

For real data collection, retain a dated record of the procedure, sampling, consent arrangements and raw paired data securely. That record allows the abstract and discussion to reflect what actually happened, rather than a planned method retrospectively presented as completed.

Describing and displaying the relationship

A scatter diagram makes the pattern of paired scores visible before inferential testing.

The following synthetic demonstration data were constructed for learning; they are not responses from real participants.