doodlestudies00HOW RESEARCH WORKS
Not all studies can answer the same question.
Some studies watch what happens to people. Others change one thing and compare. Reviews pull many studies together. Knowing which kind you are reading tells you how far a headline can go.
Check a study01Principles
02THE SHORT ANSWER
Definitions here follow the NIHR glossary
and Cochrane's guide to its reviews.
OBSERVING
A cohort study follows a group over time and compares people exposed to something with people who were not. Investigators observe, they do not intervene.
Check a studyTESTING
In a randomised controlled trial, the options being compared are randomly allocated, which helps keep the groups alike in everything else.
Why randomiseSUMMING UP
A systematic review uses explicit methods to find, select and appraise all the research on one question, to reduce bias.
See the numbers03A QUICK STUDY CHECK
Pick a type of study to see what it does. The bar shows how well that design is suited to showing that one thing causes another.
Based on the NIHR research glossary and the Cochrane Library. A rough guide to study designs, not a verdict on any single study.
04TWO NUMBERS WORTH KNOWING
One number from the history of fair tests,
one on why single studies are rarely the last word.
6 treatments
Scurvy treatments compared side by side in 1747
James Lind, a Scottish naval surgeon unsure which of many proposed scurvy treatments to use, compared six of them in a prospective controlled trial.
FAIR TESTS · JAMES LIND LIBRARY
1m+ papers
Research papers published each year
Sense about Science says around 28,000 scholarly and scientific journals use peer review, publishing over 1 million research papers a year.
JOURNALS · SENSE ABOUT SCIENCE
With so many papers, systematic reviews matter. Cochrane says its reviews are updated as new studies appear, because new results can change a review's conclusions.
05FOUR QUESTIONS TO ASK
These questions help you place a study before deciding how much weight to give it. They follow the definitions used by UK health researchers.
If investigators only observed, the study can show a link, not prove a cause.
Look for randomisation. Without it, differences such as age can explain the result.
Blinding participants, carers and assessors lowers the risk of bias.
A systematic review of all the studies on a question usually tells you more than one new paper.
06WHY RANDOMISE
A confounder is a factor linked to both the thing being tested and the outcome. NIHR's example: if the people in one group are younger, a lower risk of death in that group might be due to age rather than the treatment.
Randomisation is used to minimise those imbalances. Ideally the people enrolling participants do not know the random sequence, so they cannot steer who goes where. Placebos, made to look like the real thing, help keep people blind to their group.
07Questions
A correlation means two things tend to go together. Observational studies can find such links, but a confounder, such as age, may explain both. Testing cause usually needs a fair comparison such as a randomised trial.
NIHR describes it as an observational study that follows a defined group over time, comparing people exposed or not exposed to a factor of interest. It can look forward (prospective) or use past records (retrospective).
The comparison group in a trial. Its members do not receive the new treatment under study, so researchers can see how the new option compares with no treatment or a known one.
Not quite. A systematic review finds and appraises all the relevant studies. A meta-analysis is the statistical step of combining their results, which not every review includes.
Cochrane says the results of new studies can change a review's conclusions, so its reviews are updated as new evidence becomes available.
08Sources
ABOUT THIS NAME
doodlestudies.com is part of the DoodleSale portfolio. It isn't in an auction right now, but you can ask about leasing it.
