Psychological tests are measurement instruments that must demonstrate validity, reliability, and accuracy. Validity refers to how well a test measures what it intends to measure, which can be assessed through face validity, concurrent validity, or predictive validity. Reliability means a test produces consistent results, demonstrated through test-retest or split-half reliability. Accuracy depends on standardizing results based on a representative group's average performance to create comparable standard scores. Common types of tests include intelligence, personality, aptitude, and interest inventories.
1. 1
Psychological TestingPsychological Testing
MeasurementMeasurement
•• A psychological test is a kind of measuringA psychological test is a kind of measuring
instrument, like a ruler or scale.instrument, like a ruler or scale.
•• As such, it must meet the requirements ofAs such, it must meet the requirements of
any good instrument:any good instrument:
–– ValidityValidity
–– ReliabilityReliability
–– AccuracyAccuracy
ValidityValidity
•• Validity is the extent to which a testValidity is the extent to which a test
measures what it is intended to measure.measures what it is intended to measure.
•• Many ways of assessing validity exist.Many ways of assessing validity exist.
Some of them are:Some of them are:
–– Face validityFace validity
–– Concurrent validityConcurrent validity
–– Predictive validityPredictive validity
2. 2
Face ValidityFace Validity
•• A measure hasA measure has face validityface validity if the items onif the items on
the testthe test appearappear (on the “face of it”) to(on the “face of it”) to
measure what the test is supposed tomeasure what the test is supposed to
measure.measure.
•• A weak form of validity. Some tests seemA weak form of validity. Some tests seem
to have little face validity and yet are validto have little face validity and yet are valid
measures by other criteria.measures by other criteria.
Concurrent ValidityConcurrent Validity
•• A test hasA test has concurrent validityconcurrent validity if its resultsif its results
agree with those of an established testagree with those of an established test
(one whose validity is accepted).(one whose validity is accepted).
•• ProcedureProcedure
–– Administer the new test.Administer the new test.
–– Administer the established test.Administer the established test.
–– Compute PearsonCompute Pearson rr
•• A high PearsonA high Pearson rr indicates that the test isindicates that the test is
valid.valid.
Predictive ValidityPredictive Validity
•• A test hasA test has predictive validitypredictive validity if its resultsif its results
make predictions that agree withmake predictions that agree with
observation.observation.
•• ProcedureProcedure
–– Administer test to a group of individuals.Administer test to a group of individuals.
–– Collect data on performance related toCollect data on performance related to
variable measured by the test.variable measured by the test.
–– Compute PearsonCompute Pearson rr..
•• A high PearsonA high Pearson rr indicates good predictiveindicates good predictive
validity.validity.
3. 3
ReliabilityReliability
•• An instrument isAn instrument is reliablereliable to the extent thatto the extent that
repeated measurements of the same thingrepeated measurements of the same thing
under the same conditions agree.under the same conditions agree.
•• Two kinds of reliability can be established.Two kinds of reliability can be established.
–– TestTest--retest reliabilityretest reliability
–– SplitSplit--half reliabilityhalf reliability
TestTest--Retest ReliabilityRetest Reliability
•• ProcedureProcedure
–– Administer the testAdminister the test
–– Wait awhile, then administer it again to theWait awhile, then administer it again to the
same individuals.same individuals.
–– Compute the correlation (PearsonCompute the correlation (Pearson rr) between) between
the two results.the two results.
•• Test is considered reliable if theTest is considered reliable if the
correlation is high (e.g., above +.90)correlation is high (e.g., above +.90)
Problems With TestProblems With Test--RetestRetest
ReliabilityReliability
•• MayMay overestimateoverestimate a test’s reliability if peoplea test’s reliability if people
remember how they responded to items on firstremember how they responded to items on first
administration and simply answer that wayadministration and simply answer that way
again. This can be combated by using anagain. This can be combated by using an
alternate formalternate form on second administration.on second administration.
•• MayMay underestimateunderestimate a test’s reliability if thea test’s reliability if the
characteristic being measured changescharacteristic being measured changes
substantially between administrations of thesubstantially between administrations of the
test.test.
4. 4
SplitSplit--Half ReliabilityHalf Reliability
•• ProcedureProcedure
–– Administer the test once.Administer the test once.
–– Randomly split items into two groups with halfRandomly split items into two groups with half
the items in each group (split halves).the items in each group (split halves).
–– Score the split halves separately.Score the split halves separately.
–– Compute PearsonCompute Pearson rr on the resulting pairs ofon the resulting pairs of
scores.scores.
•• Reliability is high to the extent thatReliability is high to the extent that
PearsonPearson rr is high (e.g., above .90).is high (e.g., above .90).
Advantages/Disadvantage of SplitAdvantages/Disadvantage of Split--
Half ReliabilityHalf Reliability
•• AdvantagesAdvantages
–– Requires only one testing session.Requires only one testing session.
–– Eliminates the possibility that the variableEliminates the possibility that the variable
being measured will change betweenbeing measured will change between
measurements.measurements.
•• DisadvantageDisadvantage
–– No guarantee that the two “split halves” areNo guarantee that the two “split halves” are
equivalent. If not, then methodequivalent. If not, then method
underestimates the reliability of the test.underestimates the reliability of the test.
AccuracyAccuracy
•• A measure isA measure is accurateaccurate to the extent thatto the extent that
its results agree with a known standard.its results agree with a known standard.
•• However, in psychological tests there areHowever, in psychological tests there are
no “known” standards available tono “known” standards available to
compare results to.compare results to.
•• Instead, results areInstead, results are standardizedstandardized..
•• In effect, standardization uses group dataIn effect, standardization uses group data
to create a kind of standard based onto create a kind of standard based on
average group performance.average group performance.
5. 5
StandardizationStandardization
•• ProcedureProcedure
–– Administer the test to a largeAdminister the test to a large standardization samplestandardization sample
chosen to be representative of the target population.chosen to be representative of the target population.
–– Score the test to get theScore the test to get the raw scoresraw scores..
–– Compute the mean (average) and standard deviationCompute the mean (average) and standard deviation
of the scores.of the scores.
–– Use these to createUse these to create standard scoresstandard scores having ahaving a
specified mean and standard deviation.specified mean and standard deviation.
•• Results of tests measuring the same variableResults of tests measuring the same variable
and standardized to the same values can beand standardized to the same values can be
compared.compared.
Examples of Psychological TestsExamples of Psychological Tests
•• Intelligence TestsIntelligence Tests
•• Personality TestsPersonality Tests
•• Aptitude TestsAptitude Tests
•• Interest InventoryInterest Inventory
Intelligence TestsIntelligence Tests
•• Several standardized intelligence tests areSeveral standardized intelligence tests are
available.available.
•• All are designed to measure the kinds ofAll are designed to measure the kinds of
intelligence that are required for goodintelligence that are required for good
academic performance.academic performance.
•• We will be examining these in more detailWe will be examining these in more detail
in another lecture.in another lecture.
6. 6
Personality TestsPersonality Tests
•• These measure various aspects ofThese measure various aspects of
personality.personality.
•• Objective TestsObjective Tests
–– Can be objectively scored.Can be objectively scored.
–– Example: MMPIExample: MMPI
•• Projective TestsProjective Tests
–– Allow person to “project” something ofAllow person to “project” something of
themselves into their responses.themselves into their responses.
–– Scoring is often subjective.Scoring is often subjective.
–– Examples: Rorschach Inkblot Test, ThematicExamples: Rorschach Inkblot Test, Thematic
Apperception TestApperception Test
Aptitude TestsAptitude Tests
•• Aptitude testsAptitude tests attempt to predict whetherattempt to predict whether
someone will be able to acquire some newsomeone will be able to acquire some new
skill or ability with relative ease orskill or ability with relative ease or
difficulty.difficulty.
•• All aptitude tests actually measure aAll aptitude tests actually measure a
person’s current abilities that areperson’s current abilities that are
foundational to the skill or ability to befoundational to the skill or ability to be
acquired. That is, all aptitude tests areacquired. That is, all aptitude tests are
achievement testsachievement tests..
Interest InventoryInterest Inventory
•• I like to mention this one as you may find itI like to mention this one as you may find it
useful to help decide on a career.useful to help decide on a career.
•• Test compares your interests to those of peopleTest compares your interests to those of people
who are happy and successful in their fields ofwho are happy and successful in their fields of
work.work.
•• If your interests match those of people in aIf your interests match those of people in a
given field, then you might want to look intogiven field, then you might want to look into
that field.that field.
•• Caution! DoesCaution! Does notnot determine yourdetermine your aptitudeaptitude forfor
that sort of work!that sort of work!