Install Understanding people

Add the book to your home screen or desktop for a separate reading window.

If your browser has no install option, you can bookmark the book and read it here.

Search
Browse the book & library

How we use evidence

Reading list

Listen to this lesson

Synthetic voice
Choose a section
  1. Read Hear what kind of claim is being made
  2. Read Separate findings and applications
  3. Read Match the method to the question
  4. Read Ask whose experience is represented
  5. Read Keep confidence proportionate
  6. Read Use one check

Andrew · Synthetic narration · Spoken lesson

A spoken adaptation. The written page has the full text and links. Audio streams when played; download the MP3 for offline listening.

On this page

When this book describes a research finding, it should explain what researchers did, what they found and what the finding cannot tell us. When it suggests something to try, it should make clear whether that particular suggestion has been tested.

Three kinds of statement

Kind What the reader should expect Example
Research finding A nearby reference and a description of the relevant setting or limitation Researchers combined studies in which people judged truth and lies. On average, people were right about 54 times out of 100.
Interpretation An explicit possibility rather than a verdict about a person Public disagreement might reflect a concern about how the group will react.
Practical suggestion A reason to try it, without claiming that research has proved the suggestion works Ask what changed between the private conversation and the meeting.

The reading guide explains the status of the framework, exercises and fictional examples. Green action panels identify practical suggestions; their styling is not an evidence grade.

Match the method to the question

Research methods answer different questions. Here is what the terms used in the chapters mean:

Method What researchers do What to watch for
Experiment Change something and compare what happens, often by randomly assigning people to different groups. Whether the study was designed well enough to show cause and effect, and whether its setting resembles yours.
Observational study Record what happens without assigning people to different conditions. Two things happening together does not prove that one causes the other.
Interviews Ask people about their experiences and what those experiences mean to them. An account tells us how someone describes an experience; it may not settle every fact about the event.
Ethnography Spend time observing people in their everyday settings, often alongside interviews. Detailed findings from one setting do not automatically describe every group.
Research review Bring together and assess earlier research on a question. How the authors chose the studies and whether they considered conflicting findings.
Systematic review Use stated methods to search for, select and assess relevant studies. Even a thorough search cannot remove weaknesses in the studies it finds.
Meta-analysis Combine numerical results from several studies. An average can hide important differences between studies or groups.

We look for findings supported by several studies whose methods readers can inspect. One memorable experiment, an expert's reputation or an agency's use of a practice is not enough to establish a rule that applies widely.

Ask who was studied

In their 2010 paper The weirdest people in the world?, Joseph Henrich, Steven Heine and Ara Norenzayan examined how often behavioural research relies on people from Western, educated, industrialised, rich and democratic societies. They compared findings across populations and questioned whether results from these groups can safely be applied to everyone. This is a reason to ask who was studied; it is not a reason to assume how someone will behave from their nationality. Read about the paper.

For each research claim, we ask: who took part, what did the researchers measure, what did they compare, how clear was the result, and how close is that situation to yours?

Describe confidence without false precision

We describe the evidence behind each claim separately. For example, Chapter 1 distinguishes a paper examining earlier research from a study reporting two experiments. Chapter 3 explains when evidence about students is being used to suggest questions about volunteers. These descriptions help you judge how much a finding can tell you. They are not a formal scoring system for research quality.

An average result for a group cannot tell you why one person acted as they did. People within the same group can differ considerably. The book applies this caution to findings about culture, personality and the brain.

Keep sources and examples separate

Sources are linked beside research claims and explained in the reference list. Evidence for one part of an approach does not prove that a whole conversation, worksheet or method works. We say when a suggestion goes beyond the people or setting that researchers studied.

The expanded reference library uses selected research and public guidance. Its source-access log distinguishes research records and abstracts from relevant public-guidance sections and records the limits of each source. It does not imply that every full research article was read, that the selection was systematic, or that the practical exercises have been validated. Explanations and fictional examples are original writing; third-party questionnaires, diagnostic checklists, figures and recordings are not reproduced.

The source list covers all twelve chapters and includes methodological papers, experiments, research syntheses, ethnography and public support guidance. Foundational work is read alongside later accounts where available, and contrary findings are included when they change the explanation. Malle’s challenge to a broad actor–observer claim in Chapter 1 and Vignoles and colleagues’ challenge to a simple East–West contrast in Chapter 7 are examples. The selection is not a comprehensive or continuously updated literature review; publication here does not imply that each cited study represents the latest evidence on its topic.