1.1 historical review of psychometry

1.1 historical review of psychometry

Psychometry during 19th century

• origin can be tracked to mid 1800s → psychophysics established a discipline by Gustav Theodor Fechner

• first quantitative approach to study psychological perceptions

• The Fechner law

psychophysics studies

• the way physical stimuli is experienced by the subject

• e.g., loudness of a sound or brightness of a light

The Fechner law formula

The Fechner law

• it relates the relationship between change in a physical stimulus & how that changes is perceived by the subject

Historical review 19s in England

• Francis Galton

Francis Galton

• concerned with the study of individual differences → individual differences do not only exist, but they can also be objectively measured

• collected data from a large number of subjects in order to understand individual differences

• applied statistical methods to the collected data

• understood the relationship between mental abilities & other variables

• conceived the concept of correlation (linear regression)

Francis Galton measured

• hearing (highest audible tone)

• sight (depth of perception)

• physical variables (head width, height, forearm length/strength)

Psychometric → historical review (FRANCE)

• Alfred Binet

Alfred Binet developed

• first IQ test → the Binet-Simon test

• first one recognizing that tests must be adapted to individuals

The goal of IQ test → the Binet-Simon test

• identifying children with learning disabilities

Binet-Simon test

• comprehensive for its time: including verbal comprehension, reasoning, memory & abstract thinking

• adapted to different age levels & individuals

Standford-Binet intelligence scale

• upgraded version of Binet-Simon test

• still applied nowadays

Psychometric → historical review during the 20th century

• WW1 → new recruits needed to be evaluated to determine if they were fit for a leadership position

• Robert Yerkes → developed Alpha & Beta tests for the Army (1917)

Alpha & Beta tests

• Developed by Robert Yerkes

• Both tests measured intelligence

• OBS: Beta test was applied to illiterate soldiers as it was non-vernal

Lewis Terman

• Adapts the Binet-Simon & popularizes IQ score

• IQ = mental age/ chronological age * 100 → first attempt at norming

Norming is the

• process of determining the typical performance of a group of individuals in a given test

• it allows us to compare individuals

Raw scores as true Scores were used until

• 20th century

Why was using Raw scores as true Scores were until 20th century a mistake?

• because every measurement is subject to a certain amount of error

Classical Test Theory (CTT)

• developed from 1920-1960

• Observed score = true score + error

Louis Thurstone made a major contributions to

• Factor analysis → which had been proposed by Charles Spearman in 1904

Lee Cronbach proposed

• reliability measure → still utilized today

Item Response Theory

• 70s century

• focus shifted from test development to item development (Rasch & Lord)

• The item became the unit of analysis

Psychometric networks characterize behaviour as

• complex interrelated systems

Three different ways of understanding a psychological construct

• Reflective model

• Formative model

• Network model

Reflective model

• Construct causes (→) the observed variable

---

• The underlying construct CAUSES the observed variables: Depression → sadness, sleep problems, low energy, loss of interest

  • So the logic is → I have depression → therefore I experience these symptoms.

• The symptoms are considered reflections/indicators of the underlying depression.

→ 💡 Remember: Reflective = construct causes symptoms.

Formative model

• Observed symptoms/variable causes (→) the construct

---

• Different observed variables combine to form the construct :Sadness + sleep problems + low energy + loss of interest → Depression

  • The observed characteristics create/form what we call the construct.

💡 Remember: Formative = variables FORM the construct.

Network Model

• Construct is a cluster of interconnected observed variables

---

• we don't necessarily need one underlying construct causing everything, instead the symptoms directly influence each other: Can't sleep → tired → stop socializing → feel lonely → feel sad → sleep worse

  • "depression" can be understood as the whole network of interconnected symptoms, rather than one hidden thing causing all of them.

💡 Remember: Network = symptoms affect each other.

Psycometrics is a

• scientific discipline concerned with the measurement of psychological variables

• measures latent variables that cannot be observed & have to be inferred from other observed variables

Why did psychometrics developed so much historically?

• People needed ways to measure differences between people & put them into categories or make decisions about them.

• Because of this need, psychologists developed better psychological tests & measurement methods.

CTT (test-centered) & IRT (item-centered) are :

• the two prevalent theoretical approaches to test development

Evaluation procedures resulting in a profile of the evaluated peron

• Psychological assessment

The process through which we conclude about latent variables using the information from observed variables is called

• Inference

When the differences that we observe are attributed to the differences in a latent variable, then we are talking about a

• Reflective model

• Latent variable → THE OVAL CICLE (something we cannot directly observe)

• Observed variables → THE SQUARES (things we can actually measure/observe)