Personality & Social Psych - Weeks 1-3

Lectures 1-9

Psychometrics

Field of psych and stats concerned with measurement theory and test applications of the quantitative assessment of latent psychological constructs

Test

Systematic applications of principles that attempt to measure important psychological attributes (e.g., personality, mental disorders, etc.)

Latent

Unobserved/hidden

Constructs

A label for a domain of covarying behaviours (e.g., introversion, cognitive ability, psychopathy)

Importance of Psychological Tests

• Use people’s responses to infer latent psychological attributes

• Used to inform important decisions (e.g., screening, psychopathology, education, etc.)

• Obtain info efficiently

Psychological Measurement

• Process of assigning numbers to objects so that specific properties of the objects are represented by properties of numbers

• Requires:

  • precision and accuracy of tests

  • that psychological constructs are real

Sampling

Selecting to observe some part of a target population to estimate characteristics of interest within that population

Representativeness of Samples

Biased samples can produce over- or under- estimates of population values

→ response rate/non-respondents/dropouts/volunteers

Psychological Test Scores

• Scoring: process of generating a numerical value from the responses to a psychological test, which is designed to quantitatively represent the relevant psychological construct

• Objective Scoring: e.g., standarised questionnaires

• Subjective Scoring (assessors’ judgement): e.g., projective tests

Concerns with Interpretation of Test Scores

• Cultural Bias: test may have been developed in one culture but items have different meaning in another culture

• Procedural Bias: perhaps one applicant completed online and another in person

• Faking: respondent knows certain response will look better to employer

• Lack of Insight: respondent may genuinely believe they are organised even when they evidently are not

Tests Can Hurt

• Psychological phenomena are complex but tests reduce them to a single number

• People are labeled (e.g., label of “gifted” may give child more opportunities but also more pressure)

Transformation and Standardisation

• Standardisation scores is the process of transforming test scores into universal indexes called standardised scores

• Transformation: a mapping of X to itself; a change

Why Standardise?

• Allows comparisons

• Compare results obtained from different scales that measure same attribute

• Compare results obtained from different versions of same measurement

Examples of Transformations

• Z-scores: change a set of scores so that set has mean of 0 and standard deviation of 1

  • z = x - μ / σ

  • X is raw score, μ is pop. mean, σ is pop. SD; z score of 1 means “1 SD above the mean”)

  • percentiles can be attached to z-scores

• T-score: change a set of scores so that set has a mean of 50 and SD of 10

  • T = (10 x z) + 50

  • T = 60 means “1 SD above the mean”

• Area transformations (quartiles, deciles, percentiles)

Normative Data (“Norms”)

• Normative Data: a collection of scores derived from a (large) representative sample of a population

• Purpose:

  • records of population attributes (policy-making, interventions, etc.)

  • compare an individual’s attributes to pop. levels (norm-based interpretations)

  • can be broken down by sub-group

Validity

• General Definition: the degree to which a claim is correct or true

  • applies to constructs, test, test-items, individuals, samples, norms, etc.

  • EX: “Does a test measure what it claims to measure?”

• Psychometric Definition: the appropriateness, usefulness, or meaningfulness of test (measurement) scores and their interpretations

•