Social & Critical Psychology · Simplified
What the Instrument Actually Sees
Every measure of prejudice is named after something larger than the thing it observes.
Nothing here tests you
There is no task on this page that measures anything about the person using it, and you never take any of the instruments described. Vantry, the Uplands and the Lowlands are invented. No group in them stands for any real group anywhere, no figure is a prevalence estimate, a norm or a published effect, and no stereotype content appears at all: the activity is about what an instrument observes, and it would run identically for any pair of invented categories.
The setting
Vantry
An invented country of six regions. Two of them, the Uplands and the Lowlands, have a long and unequal history that everybody in Vantry can describe and disagrees about. Four research teams have each measured what they all call prejudice towards Uplanders, and each has published a figure.
Step 1 of 3
First, a judgement
Four instruments, all described in the literature as measuring prejudice. What is the main thing that differs between them?
Step 2 of 3
One instrument at a time
Read what each team found, then answer one question about it before the rest is shown.
Key terms
- What an instrument observes
- The thing it directly records, before any inference. It is almost always narrower and more specific than the name of the construct.
- Level
- Whether a measure is about a person, a relationship between people, or a distribution across an organisation. Measures at different levels are about different objects, not about the same object at different resolutions.
- Reification
- Where an instrument stops being evidence about a construct and becomes the construct, so that the question of whether it measures prejudice stops being asked.
What does this instrument directly observe, before anything is inferred from it?
Step 3 of 3
Four instruments, four objects
| Instrument | Level | What it directly observes | What it cannot show |
|---|
What this shows
Four instruments, and not four routes to one thing
Key idea: Each of these instruments directly observes something narrow and specific: what people will write on a scale, a difference in sorting speed, which files somebody picked on one occasion, a distribution of appointments over twelve years. Each is then reported as a measurement of prejudice, and the distance between those two descriptions is where the argument lives. The last one is different in kind rather than in quality. It is not a worse measure of individual attitudes; it is not a measure of individual attitudes at all, and the point runs both ways. No individual acting on a prejudice is demonstrated by it, and no study of individual attitudes, however large and however well done, can confirm or refute it, because the two are claims about different objects. That is why asking which instrument is the best measure of prejudice is the wrong question. The useful one is what you want to know, and then which object that is a question about.
Four instruments is a simplification: a fifth, in which colleagues rate how somebody appears to them, sits at a level between the individual and the institutional and is in the longer version. Nothing here is an argument that measuring prejudice is impossible or that the instruments are worthless; each observes something real, and the alternatives to expressed attitudes were developed for good reasons. The criticism is of the inferential step and not of the measurement. Two further cautions. The list of things that could explain each finding is illustrative rather than exhaustive, and a good argument for a further explanation should be credited. And every figure on this page was invented for teaching: the simulated respondents, the percentages and the twelve-year stability are all written, and none of them is evidence about anywhere.
The longer version adds the informant-report instrument, asks you to predict a fictional respondent's behaviour from their score before seeing it, and closes with a set of published-style claims to judge against what the instruments can support. It is at Measuring Prejudice: Instrument Laboratory in the main collection.