13344 Chap.6 Case-Control Studies, Bias and the Odds Ratio
Case-Control Studies, Bias and the Odds Ratio
Starting from the outcome
The chapter explains how case-control studies select people with and without an outcome and compare their past exposure, when the design suits a question, and how cases and controls should be defined and matched.
Odds ratios and bias
It works through an odds ratio from a two by two table, explains why relative risk is not appropriate with case-control data, and sets out the selection and information biases the teaching names, including recall bias.
Using it in the essay
Case-control designs are efficient and often the first study of an association, but they are vulnerable to recall bias and confounding.
The chapter ends with the limits to state when you weigh the design against a cohort in an essay.
What this chapter covers
- 01
Working backwards from outcome to exposure
- 02
Incident and prevalent cases
- 03
Choosing and matching controls
- 04
The odds ratio and the two by two table
- 05
Selection bias
- 06
Recall and recording bias
Worked example · free
Odds ratio for childhood bullying and depression
- 1Share exposed: 80 ÷ 200 = 40% of cases and 80 ÷ 400 = 20% of controls.
- 1Odds of exposure: cases 80 ÷ 120 ≈ 0.67; controls 80 ÷ 320 = 0.25.
- 1Odds ratio = 0.67 ÷ 0.25 ≈ 2.67, equal to (80 × 320) ÷ (120 × 80).
- 1Interpretation: above 1, so exposure is more common among cases.
Key terms
- Recall Bias
- Recall Bias is systematic difference in how cases and controls remember past exposure.
- Incident Case
- An Incident Case is a newly diagnosed case, usually preferred when studying causes.
- Selection Bias
- Selection Bias arises when compared groups are not comparable with each other or with the population.
- Matching
- Matching is choosing controls similar to cases on factors such as age, sex and socioeconomic status.
- Information Bias
- Information Bias is systematic error in how data are obtained, including recall bias and interviewer or recording bias.
Case-Control Studies, Bias and the Odds Ratio FAQ
Why are case-control studies good for rare conditions?
Because they start from people who already have the outcome, they avoid following huge numbers of people for years in the hope that a few become cases. The module's example is childhood maltreatment and adult personality disorder, where the outcome is relatively rare.
How many controls should each case have?
The module says that ideally each case is matched to about four controls, which gives the most statistical power, though plenty of published studies use fewer. More than one control group, such as healthy and clinical controls, can also help.
What is the difference between bias and random error?
Random error affects everyone and narrows as samples grow. Bias pushes results away from the truth in one consistent direction, such as cases recalling exposures more fully than controls, and a larger sample does not remove it.
Why is relative risk not used with case-control data?
Relative risk needs the number of people who develop the outcome among all exposed and unexposed people, which is incidence. A case-control study fixes the numbers of cases and controls, so it cannot measure incidence. The odds ratio is used instead and approximates relative risk when the outcome is rare.
Where should controls come from?
From the same population that produced the cases, so that each control would have become a case had they developed the condition. GP lists, hospital patients, neighbourhood residents and random dialling are all possible sources, each with a risk the teaching names.
Should cases be incident or prevalent?
For questions about causes, incident cases, newly diagnosed, are usually preferable. People who have had a condition for a long time are also shaped by factors that affect how long it lasts, so a sample of prevalent cases mixes the causes of onset with the causes of survival.
Assessment move
Practise one odds ratio table a week with your own numbers, and for every published case-control study you read, write down how cases were found, where controls came from and how exposure was measured. Write one sentence for each type of bias in the chapter explaining how it would push an odds ratio up or down in a study you design.
Working through Case-Control Studies, Bias and the Odds Ratio in 13344? Sia is AskSia’s AI Psychology tutor — ask any 13344 Case-Control Studies, Bias and the Odds Ratio question and get a clear, step-by-step explanation grounded in how 13344 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.