Skip to content

Guided reading and advanced methods

Continues the paper-reading workshop: telling a finding apart from a clinical recommendation, seeing design and bias, reading effect sizes correctly.

Who it is for
Psychology students
Limits of this section
Figures are shown only together with their source, outcome measure, follow-up period, comparison and uncertainty. Odds ratios are not converted into probabilities.

Related content already on the site

Stations

Separating findings from clinical recommendations

Learning aim
Distinguish a research finding from the additional judgements needed for a recommendation.
Explanation
Recommendations depend not only on an effect estimate but also certainty, benefits and harms, feasibility, resources and values.
Example (fictional or published study)
An average reduction in symptoms in one study does not mean the same option should be chosen for every person.
Exercise
Find the word ‘effective’ in an abstract’s conclusion and list the comparator, outcome and uncertainty information needed.
Content approval status
Beyaz approval · 18 October 2026

Design and risk of bias

Learning aim
Identify the question a design can answer and major pathways for bias.
Explanation
Randomisation can balance some confounders; blinding, missing data, measurement and selective reporting may still affect results.
Example (fictional or published study)
If outcomes for people who discontinue are omitted, apparent benefit may be exaggerated.
Exercise
Identify how groups were formed, how missing data were handled and whether prespecified outcomes were reported.
Content approval status
Beyaz approval · 18 October 2026

Absolute and relative effects

Learning aim
Explain why a relative effect does not show absolute benefit without baseline risk.
Explanation
Risk and odds ratios are relative measures; risk difference gives an absolute change. A confidence interval shows effects compatible with the data but does not capture every uncertainty.
Example (fictional or published study)
The same relative reduction yields different absolute differences when baseline risk is 2/100 versus 20/100.
Exercise
Find the relative measure in an abstract and use control-group risk to express the absolute difference per 100 people.
Content approval status
Beyaz approval · 18 October 2026

Generalising from animal research to humans

Learning aim
Separate a mechanistic finding from the evidence needed for a clinical outcome in humans.
Explanation
Animal models can inform mechanisms; differences in species, dose, setting and outcomes limit direct clinical generalisation.
Example (fictional or published study)
A procedure that changes behaviour in rodents is not evidence that it treats depression in humans.
Exercise
Find a news summary of an animal study and list the omitted steps involving species, dose, comparator and human studies.
Content approval status
Beyaz approval · 18 October 2026

General information, not diagnosis or personal advice.

← Student Library