human personality

what is personality science?

• study of normal, naturally corruing variability among people

  • systematic differences in how individuals feel, behave, think and want

• focuses on recurring tendencies, rather than single momentary states and asks how these tendencies are consistent over time and across situations

• in short: the subject of personality science is the measurement and explanation of individual differences in everyday psychological tendencies

how is personality science similar from other fields of psychology?

• aims to measure something about people which does eventually become generalisable on a large scale

• established and verifies construct

• studies same phenomena, and uses similar scientific methods

how is personality science different from other fields?

• measures individual differences, between group variance in terms of probabilities rather than generalisable

why are there personality differences? (useful functions of variability)

• evolution: variation allows section of advantageous traits

  • if everyone were identical, there would be nothing to select between, so evolution would stop

  • variation allows for some traits to be more successful in certain environments

  • applies for cultural evolution too

• adaptation: people learn from others differences

  • seeing what behaviours lead to success or failure

  • allows individuals to adjust their own behaviour

  • variability acts a kind of “testing ground”

• division of roles: different traits suit different environments/jobs

  • people can fit into roles suited to their traits

  • increases overall efficiency and functioning

  • “ differnt jobs require different kinds of people”

• innovation: variation generates new behaviours and ideas

  • some may fail but others may improve outcomes, lead to progress

  • without differences, there would be no novelty or innovation

• universals

  • if traits consistently relate to outcomes this suggests general principles about humans

  • so differenced don’t just create diversity, they help us understand what works for everyone

what are ABCDS?

• tendencies for affects, behaviours, cognitions and desires

what do they stand for?

• affects

  • feelings

• behaviour

• cognition

  • thoughts

• desires

  • what a person wants

what do ABCDs represent?

• stable, recurring individual differences

• they are the building blocks of personality traits

what are abcds?

• specific momentary states

• single instances of a feeling, behaviour, thought or desire

what is an example of an abcd and ABCD?

• abcd - feeling sad right now

• ABCD - tendency to feel sad often

what two key properties must ABCDs have?

• applicability to almost everyone

  • the behaviour must be possible for all indviduals

• variability

  • it must differ between people (otherwise not useful for personality science).

how are ABCDs studied?

• mainly through quantitative measurement

• comparing people’s frequencies of behaviours

• using correlations to examine

  • consistency over time

  • consistency across situations

  • patterns between traits

what is the evidence for them?

evidence shows ABCDs are:

• moderately consistent over time

  • (Mõttus, Kandler et al., 2017) - .53 over five years, was not much lower than it would have been if they had asked again in the next week.

  • Mõttus, Sinick et al., 2019) - .37 over 15 year interval, there was a stronger than chance consistency for each ABCD . ]

• consistent across situations

  • (Leikas, Lönnqvist, & Verkasalo, 2012) - participants intercated with different actors

    • experienced four different situations, behaviour was videtoaped

  • average correlations

    • ~.35 for specific behaviours, .43 for broader observed tendencies (enjoyment) and ~.20 for self-reports

  • situations do matter but people still behaved like themselves across all situations

• structured into patterns (traits) via correlations

what are the reasonable expectations for correlation sizes?

• typically small to moderate

• common range: .10 - .30

• around .50 is already strong in psychology

• never near 1 due to complexity and measurement error

in what way is personality science, a science of many numbers?

• there are 4,950 possible correlations among 100 variables

• 1,225 correlations among 50 variables

• 300 correlations among 25 variables

what are two key strategies for selecting ABCDS?

• deductive

  • theory driven

  • start with a theory

  • strengths:

    • gives a clear explanation (cause) of personality traits

    • provides a structured way to organise ABCDs

  • weaknesses:

    • theories often lack strong evidence

    • personality traits are complex and influenced by many causes

• inductive

  • data driven

  • start with many items → find patterns via correlatios

  • includes the lexical hypothesis,.

what is the drop or combine porcedure?

• if items correlate very highly thne you drop one

• if items moderately correlate you can combine them into aggregate traits

• repeated to reduce many variables into fewer broader traits

what do we gain if we combine?

• simplification (fewer variables)

• clearer patterns

• reduced measurment error

• better for understanding big picture

what do we lose if we combine?

• specific detail of individual traits

• unique aspects of behaviours (sadness vs regret)

• some varaince. information is discarded

factor analysis/ principle component analysis what is it?

• a statistical tool that automates drop or combine

• it analyses large correlation matrices

• identifies clusters of ABCDs that tend to co vary

• these clusters become factors (broad traits)

what are the main steps of factor analysis/pca?

• collect many ABCD measure

• compute correlation matrix

• extract components/factors

  • finds underlying dimensions explaining shared variance

• rotate factors

  • makes structure easier to interpret

• interpret and label factors

  • extraversion and neuroticism

what are the main outputs of factor analysis / PCA

• factor loadings

  • how strongly each ABCD relates to each factor

• eigen values

  • how much variance each factor explains

• factor scores

  • each persons score on the broad traits

what is the lexical hypothesis?

• idea that personality traits became built into language over time

• this helped researchers collect thousands of traits describing words

• a lot of initial work was done by Gordon Allport and Henry Olber who pulled out nearly 18,000 terms that described characteristics in which people differ

  • divided these into four states

    • personality

    • temporary states

    • general evaluations

    • physical characteristics

what and who other work went in to identify these traits?

• gave out a lot of personality tests and then had to calculate all the scores from them

  • this was a huge task for researchers

• more people came together to try and shorten the adjective list to have a cellar dimension of personality variation

• raymond cattel was next to assemble a team

  • managed to organised most terms into 1717 bipolar scales

  • these scales then were reduced ito 60 cluster

  • a few years later he managed to condense them into a further 35 super ABCDs

• further analyses them him to believe they could be allocated into 13

what is the history behind the big five?

• researchers such as warren norman in the 1960s had an updated look at the dictionary to update the list of adjectives originally pulled out b Allport-Oldert and Cattell

• with this new updated adjective list, several other researchers specifically lewis golderg, became convinced that the English lexicon of personality relevant terms could indeed be summarised into 5 aggregates

what are the big five traits?

• neuroticism (opposite of emotional stability)

• extraversion

• agreeableness

• conscientiousness

• and intellect (openness to experience)

what are some adjectives of emotional stability / neuroticism?

• nervous, self critical, dependent

what are some adjectives of extraversion ?

• merry, talkative, carefree

what are some adjectives of agreeableness?

• trustful, generous , cooperative

what are some adjectives of conscientousness?

• persistent, ordley, dependable

what are some adjectives of intellect/openness to experience?

• insightful, creative, perceptive

what other things were used as well as lexical work?

• personality scientists also used personality scales with descriptove statements that they thought would help represent important ABCDs

• some aggregates overlapped and some did not

who were the main people behind the Big Five? and what did they do?

• Costa and Mccrae

• they propsoed that many of the ABCD aggregates that were being researchers could be orgnaised into five broad aggregates very similar to the big five

• then then went on to become the five factor model of personality

• this work done is some of the most widely known findings im the whole of psychology

what are informant questionaires

• informanats are typically friends, partners and colleagues

• it is when these people answer questions about you, for example perosnlaity questionaire about your traits

what is the robustness of the big 5 across informant questionnaires and the self?

• when you rate yourself high on extraversion it is likely that in about 70-75% of cases that your friend and relatives also rate you high on this domain

• when two informates rate someone they both know well, their ratings agree with each other to nealry the same degree

why is this informat reports important?

• with self report there is the chance of social desribailiy bias

• so it is important to be able to triangulate the true or more accuarte score of someone’s perosnlaity

• as they come from twi different sources there is no guarantte taht they would overlap so when they do its easier to derive a true score of perosnlaity

what is the ribustness of self informant personality traits?

• when people complete the big five test twice over a short period of time there is unlikely to be much personality change

• correlation between two measurements are above .80

robustness of big five across cultures?

• mixed results on the cross cultural validity of the big 5

• when systematically snaccing lexicons for personality adjectives and collecting peoples self ratings for them has produces bg five like aggregates in several languages such as german, dutch and some slavonic languages

• however, in some languages some or none of the big five has emerged

• rather than leaving it up to the bug five to emerge independently, researchers wrote items for ABCDs of each of the big five and asked people from different countries to rate themselves or someone else on the questions

  • results showed that correlations among items can almost always be organised into the big five

what is big six?

• this is called the HEXACO model of personality

who came up with hexaco and what is it?

• based on analysis of lexical studies in many langauges by Michael Ashton and Kibeom lee

• largely simailr to te big five but differences as the agreeable dimension is split into agreeableness vs anger and has the 6th factor called honesty-humility

why is having a big few good?

• can be thought as a map of coordinate system for organizing and representing a vast number of ABCDs

• good way of thinking about is big few are parliament people in democratic countries to represent them

• thousands of big few studies have produced. many well replicated findings

• they allow personality scientists to predict important life outcomes

• also used beyond psychology to researchers to summarise the major patterns of feeling, thinking and behaving

what are the two meanings of traits?

• first meaning: traits are summaries

  • a trait is just a label we give to patterns of ABCDs

  • if someone is talkative, sociable and energetic we summarise this as extraversion

• second meaning: traits as underlying causes

  • instead of describing patterns, traits might also cause those behaviours to co-oocur

  • so peopleare talkative, sociable and energetic because they have underlying system/process (extraversion)

what is the duality principle?

• this sayd traits are both at the same time

• we observe behaviours (ABCDs)

• from that we infer a trait

• that trai then explains why those bhevaiours happens

• so

  • traits are what we measure and what we think causes it

what is implicit personality theory?

• people have shared beliefs about which traits go together and use these to infer personality, even with little information

• example: someone seems friendly, we also assume they are generous

• mental shortcut which helps us judge faster but less accurate

• some traits have more of an impression on us than others

what is distribution of traits?

• traits such as the big five are distributed among people in a certain way that is called the normal distribution

• two thirds of people lan within one standard deviation on either side of the average

• this means that most peoples trait values are not too far from those of the average person

• to divide people into introverts and extraverts is misleading as most people will be a mix of both

• distribution also suggests that there are many small causes for the variability in each traits rather than a few potential causes

what is the explantion of processes?

• unclear the specfic reaons on why some trauts correlate and others do not

• however since traits such as the big five popbup again and again, there must be some underlying cause

what is the personaity trait hierarchy?

• some peroslaity questionaires will messsure different levels

• we have big domains suhc as the big five

• tgen benathe we have different levels

  • aspects → facets → nuances

• in non human primates, you can not observe nuances but facets and traits still exist

what evidence can you look at to observe that something is a distinct trait =?

• cross rater agreement

• stability over time

• heritability/genetic influence

what is the bandwidth vs fidelity dilemma

• trade off between

  • broad traits (high bandwidth) → simple, generalisable

  • narrow traits (high fidelity) → precise, detailed

why prefer high bandwidth vs high fidelity?

• use high bandwidth when you want simplicity/general patterns

• use high fidelity when you need precise prediction or explanation

how traits map other psychological characteristics?

• traits act as a common framework

  • other constructs (attachment, values can be correlated with traits

  • this integrates different areas of psychology

why do personality findings generalise across pscyhology?

• traits summarise many ABCDs

• the mid is hierarchical and interconnected

• so findings apply broadly across behaviours and domains

how to measure personality?

• typically self-report test

  • also usable by informants

what are the steps in creating a personality test?

• start by defining a domain

• it has lower level components as personality is hierarchy

  • sample facets and nuances broadly as you riks biased instruments which means there could be inconsistent findings across studies

• write plenty of items for each facet

  • they can be adjectives or descriptive statements

  • do for both higher and lower trait end

  • avoid ambiguity and complexity

• then create instructions and ratings scales?

  • can be different

  • most (WEIRD) people skilled at completed tests regardless of instructions

  • likert scales of any length work equally well

  • as do bipolar scales

• collect as much data as possible

  • correlations become stable N> 250

  • sample respondents diversely

• once data collected, retest after about two weeks

• also collect ratings from informants

  • for validating the scale

what happens once data is in?

• make sure instrument has intended hierarchical structure (easy)

• scores are reliable (harder)

• scores are valid (hard)

how to get desired structure - extraversion example - through EFA

• create extraversion scale with three facets, four items in each

• structure of where there is the 12 items correlate with the specific facet (friendliness, assertiveness and excitenent-seeking)

  • we want one item to load highly on to one facet and low loadings on the other

• pull out facet one and see which of the items initially selected to how the correlated with the shared variance (extraversion)

  • all have to correlate to some extent

  • kick out items that dont correlate with extraversion

• once kicked out, choose three correlated factors from remaining items

  • select four items for each facet

  • avoid too similar items

what is CFA?

• confirmatory factor analysis

• deductive thinking

• computer evaluates model with exactly factors we want, defined by our selected items

• fit such a model to the data and ask how well it fits data

  • prefer on a new data set to avoid overfitting the data

• when we fit it will give us the factor loadings we are interested in

how many items per scale?

• all else equal, more is better

• broader coverage of trait

  • cover more nuances, secale is likely to be more informative in total

• random error cancels out

• has cost

  • people take more time to comeplete

  • sometomes people dont wana do that

what is reliability?

• whatever is being measured, can you trust the measurements?

• if you measured again, would you get the same result?

• re-test stability over some reasonably short interval

  • why not too short /long

what is the typical retest reliability for the big five? for facets?

• for the big five r> .80

• for facets ~ .70 to .85

• for items ~ .50 to .80 , typically above .65

what is internal consiustency?

• if retesting twice is not possible, comes as a stand in

• best known as crohn bachs alpha

• it is a measure of how strongly items inter-correlat and how many items we have

• the more items that correlate, the less error they can contain

how to calculate internal consistency?

• based on number of items (k)

• average inter-item correlation (r)

what is limitations of internal consistency?

• underestimates reliability

• boring for test takers as tmany questions need to be the same

• prevents sampling nuances broadly

• sampling nuances broadly lowers internal consistency

• cannot get alphas for items

what is the principle of aggregation?

• combining many items

  • cancels out random error

  • increaes reliability

• more items = more stable estimate of “ true score”

if a personality test has reliability =.80, what does this mean for a persons score (e.g 60)?

• the score is the persons obswrved score on the personlaity sale

• with reliability =.80

  • 80% relfects the true trait level

  • 20% is measurement error

  • so the persons true score is not excatly 60, but likely somewhere around it

what is the standard error of means (SEM)?

• estimates how much a score typically caries due to error

• tell you the precision of the score

• formula

  • SD x square root of (1- reliability)

what does SEM (e.g = 4.5) mean in practice?

• a score may vary by about ± 4-5 points

• the true score is likely near, but not exactly 60

why do we use the standard deviation in SEM?

• SD reflects how spread out the scores are

• error must be expressed in actual score units, not just percentages

what is the key idea behind interpretation of SEM and reliability?

• personality scores are estimates , not exact values

• reliability tells you how much error exists

• SEM tells you how big that error is

why avoid overly similar items?

• artificially inflate internal consistency

• reduce content coverage

• create redundancy and boredom

• better to sample different numbers

how do you improve reliability?

• add more items

• use clear ambigious items

• retest and average scores

• ensure items corrleate modertaley (not too low, not idnetical)

what is validity?

• validity is asking what the test really measures?

• reliability is an upper boundary of validity

• change the trait by X units → change of the measurement by x units

  • if you change the thermometer by 2 degrees you would expect the measure of thermometer to also go up by 3 degrees

what is content validty

• extent to which scale content represents the facets and nuances of the trait

• sample representative from the universe of the trait content

• covered before

what is face validity?

• extent to which the content makes intuitive sense

• little value in trying to trick people