Psychometrics Flashcards
(27 cards)
why have psychometrics?
to measure and quantify constructs
allow a falsifiable theoretical hypothesis
discrete
finite range of values
dichotomous (more than 2 values)
or polyotomous
continuous
no theoretical limit
must be measured using scales
gives estimates
nominal scale
numbers used as labels
info carried by numbers to identify e.g. male/female = 1/2
mathematical operations
ordinal
ranking system - hierarchy
info carried by numbers - identity and rank order - but no info about distance between observations
eg percentile scores
non-parametric tests
limitation - distribution differences between rank scores
interval
equal intervals between time units - meaningful difference between consecutive numbers - reflects equal distance
no true 0 - cannot calculate
info carried by numbers
eg celcius temperatire
ratio
possesses a 0
eg periods of time
problems - used for Stanford-Binet intelligence test - to see mental and chronological age - measured using different scales, so cannot use one as a proportion of the other
norm-referenced
performance compared with a reference group - used to assess developmental norms (ordinal scale) and within-group norms (normal distribution)
good for exams
BUT norm-referenced testing may be the cause of the decline in educational standards - only compared within same year, not previous year, so decline across years goes unnoticed
may put minorities at a disadvantage
criterion references
whether someone has achieved the required level of performance, advantages of predicting behaviour outside the testing situation e.g. testing for job competence
domain references
indicated how well a person does on a particular type of task - how likely they are in everyday life to show a type of behaviour
test development
1- plan - constructs, population, objectives
2- generate item pool
3- submit item to review
4- pilot
5- evaluate results )quantitative/ qualitative)
6- cross-validation - additional trial administrations
7- standardise test length, sequence, administration
8- administer test
Pretty Girls Suck Penis Each Creating Sticky Arms
selected responses
closed ended, e.g.. MCQs
forced choice questions
+ - easy and objective scoring, improves reliability, testing time is good (many tests at one time), simplifying quantitative analysis
_ - guessing, may be lucky, incorrect answers to objective measures can be due to haste and carelessness, difficult to balance options, certain constructs difficult to measure this way, less flexible then constructed response items
constructed response
open ended and varies limitless, sometimes used to assess personality using projective techniques e.g. test takers respond to ambiguous stimuli like pictures or words
+ - rich data - unique characteristics, more ecologically valid
_ - time consuming, more complex, reliability, test length makes less reliable e.g. need more time for administration so number of times can be included less, content sampling errors and produce less consistent scores
quantitative
statistical, item fairness
qualitative
uses judgements of viewers to determine accuracy and fairness
appropriateness, clarity and grammatical correctness, basic rules, bias
item fairness
see if an item is bias by seeing if test scores have the same meaning for members of different groups
item difficulty
how difficult items are and if they are equally difficult to everyone
calculate relative difficulty of population
item discrimination
degree to which the test differentiates between test takers of different abilities/behaviours/groups
occupational psychology - personnel selection
note what staff does and make a list of cog. skills needed
administer tests that correlate with aspect of the job performance
validate the test
problems with faking and bias
value of ‘g’
educational psychology
assessing dyslexia - give ability tests. if dyslexic, will perform better on non-language assessed parts
assessing ASD - social, communicative and behavioural difficulties - autistic diagnostic observation schedule
scores not accurate or valid enough to ale fine discriminations, can stigmatise children
clinical psychology
Minnesota Miltiphasic Personality Inventory (pathology) - empiricle scales, meaningful
depression - rating scale - Hamilton rating scale, questionnaires
neuropsychological assessment
detects deficits in cognitive abilities, assessing general intellectual functioning, memory, language, reading/writing ect
organic amnesia tests - autobiographical memory interview for retrograde amnesia
executive function e.g. strop test
Burgess and Shallice - sentence completion test
forensic psychology
legal system,
US psychometric tests to see if individuals should be tried in court
need to assess people to see if they have a disorder
could fake - can’t just do a questionnaire
construct validity
people think intelligence is different at different ages, across and between cultures, and psychologists assign dif. meanings to intelligence and other personality traits for example
