MEASUREMENT AND METHOLOGY Flashcards
1
Q
MEAUREMENT AND METHODOLOGY
A
- assessment behaviours, attitudesm mental constructs personality, and mental health
2
Q
Intelligence
A
- mental construct specifically defined
- intelligence is NOT IQ
- IQ is score of inteligence
- unlikely that IQ captures all intelligence
3
Q
Alfred Binet
A
- developed IQ and first IQ test (Binet Scale) = mental age/chronological age x 100
- Intelligence stops developing after age 16
4
Q
Mean IQ of Americans
A
- is 100
- SD = 15 or 16
5
Q
Stanford Binet Intelligence Scale
A
- revised version of Binet’s intelligence test
- best known predictor of future academic achievement
6
Q
Lewis Terman
A
- from standford university and revised the Binet Intelligence Scale
- studies of children with higher IQ are better adjusted (gifted children)
7
Q
Weschler Adult Intelligence Scale (WAIS)
A
- most commin IQ test for dults
- organized by subtests
8
Q
Wescler Intelligece Scale for Children (WAIS-R)
A
- for children age 6-16
9
Q
Wechsler Preschool and Primary Scale of Intelligence (WPPSI)
A
- for children age 4-6
10
Q
Goodenough Draw a Man Test
A
- notable for cros cultural application an simple directions to draw the very best picture
- are marked via detial and accuracy, not talent
11
Q
IQ correlates mostly with who?
A
- biological parents and socioeconomic status of parents (income or job type)
12
Q
John Horn and Cattell
A
- found fluid intelligence decines with old age and cyrstaliized intelignece does not
13
Q
Robert Zajonc
A
- relationship between birth order and intelligence
- firstborns more intellignece than 2nd
- greater spaces between children = higher intelligence
14
Q
Charles Spearman
A
- general factor in human intellignece called “g”
15
Q
Achievement test
A
- measure how well you know a particular subject
- measure past learning
16
Q
Aptitude test
A
- meaure your innate ability
- predict future perofmrance
17
Q
Objective test
A
- subects can’t make own answers, are already structure
18
Q
Structure test
A
- more objectively scored than projective tests
- most are self-reported (but still subject to response bias)
19
Q
Q-sort, Q-measure techniques
A
- process of sorting cards into normal distrbution
- each card with a personality statmente and neutral ones and place at the hump and definind or undefining cards are put into the sides
20
Q
Mineesota Multiphasic personality Inventory (MMPI)
A
- OG for mental illnes but now a peronality meaure
- T/F questions
- items to discriminate between different disorders
- high validity becau of items and 3 validity scales (lying, carelessness, faking)
21
Q
California Personality Inventory (CPI)
A
- use for more “normal” and less clinical groups
22
Q
Myer-Briggs Type Indicator (MBTI)
A
- derived from Jung’s personality theory (archetypes)
- 2 answer per question then given 4 letter personality type
1) introvert/extravert
2) sensing/intuition
3) feeling/thinking
4) judgment/perception
23
Q
Julain Rotter
A
- creater internal/external locus of control
24
Q
Projective tests
A
- subjects to create own answer through expression of conflicts, need, impulses
- content is interpreted by administrators
- some may be score objectively
25
Rorschach Inkblot Test
- subject desrcibe what they see in 10 inkblot
- complex scoring
- questionable validity
26
Thematic Apperception Test (TAT)
- cards of interpersonal scences are shown and subject tells story of each card to reveal personality
- used to measure need for achievmenet
- needs, press, persnology are terms with test
27
Reosenweig Picture Frustration (p-F) stud
- cartoons where 1 person is frustrating the other
| - subject ask to describe how frustated person responds
28
Word Association Test
- used in conjucnction with free association tecniques
| - word called out, subject says next word that comes to mind
29
Rotter Incomplete Sentence Blank
- similar to word association
| - subject finish sentences
30
Draw a Person Test
- asks subject to draw a person of each sex and tell a story about them
31
Beck Depression Inventory (BD)
- not to diagnose depression but for severity of depressive symptoms and track course of symptoms
32
Empirical-keying or criterion keying approach
- constructing assessment instruments invovles selection of items that discriminate between various groups
- response determines if person is in particular group or not
- e.g. Strong-Campbell Interest INventory
33
Vocational Test
- what extent individual's interest and strengths match those already found by professionals in particualr job field
34
Lie detector test
- measure arousal of sympathetic nervous system which becomes stimulated by lying (and anxiety)
35
walter Mischel
- critical of personality trait-theory and personality tests in general
- felt situations decided actions (not traits)
36
Anne Anastasi
- researched intelligence in relation to performance
37
F-scale or F-ratio
- measurement of facism or authoritarian personality
38
Bayley Sclaes of Infant Developent
- not intelligene tests
- mesure senory/motor development of infants to identify mentally retarded children
- poor predictors of later intelligence
39
Research design
- how research attempts to examine hypothesis
- differnt quetions = different approachs
- some are called more scientific than others
40
Scientific approach to psychology involves (3)
1) testable hypothesis
2) reporductible experiment can be replicated
3) operationalized defintiion of concept under study
41
Field study
- takes place in naturalistic setting
- less control over enviornment
- generates more hpyothesis than can be proved
42
Experimental design
- takes place in controlled setting
| - draw causal conslusions from experiment
43
IV
- manipulates IV by applying it in experimental or treatment condition by withhodling if from the control condition
44
DV
- does not control the dependent variable but examineshow IV effects the DV
45
Confounding variable
- attempts to minimize or eliminate confounds
| - variables in environment that might also affect DV and blue effect of IV on DV
46
Sample or subgroup
- drawn from population bc is impossible to include all members
- sample must be representative of population and unabiased
47
Random sampling
- applied to achieve representative population and unbias sample
- every member of population has an equal chance of getting chosen
48
Conviencance sampling
- used when random sampling is not possible e.g. a group of psych students
49
Stratified sampling
- to make results more generalizable then convienanc sampling
- aims to match emographic characteritcs of sample to demographics of population
50
Longitudinal design
- studying same objects at differnt points in lifespand and provides better, mored valid result than most other methods
- time consuming
51
Cross-sectional design
- different subjects of different ages are compared
| - faster, easier than longitudinal
52
Cohort-sequential design
- combines longiudeninal and cross-sectional approaches
53
Within-subjects
- test same person at multiple time points and looks at chnges within person
54
Between-subject design
- compares 2 groups of ppl at the same time point
55
Quasi-experimental esign
- compares 2 groups of people but design is used when it is not feasible/ethical to use random assignment e.g. you cannot assign one group to smoke for 20 years
56
Double-blind
- experiments when neither subject nor experimenter knowns whether subject is assigned treatment or control
57
Placebo
- inactive substance disgused as substance in control group
58
Predicitve value
- degree to which IV can predict DV
59
Generalizability
- degree to which results from experiment can be applied to population and real world
60
Acquiescence
- when ppl agree with opppsing statment
| - "careless" responding
61
Cohort effects
- effect that might result when group is born and raised in particular time period
62
Demand characteristic
- subject act in way that they think experimenter wants or expects
63
Experimenter bias
- researchers see what they want to see
- AKA rosenthal effect
- minimized in double blind experiment
64
Rosenthal effect AKA
- AKA experimenter bias
65
Hawthorne effect
- subjects alter their behaviour because they know they are being observed
- applies to workers altering their behaviour for the same reason
66
Nonequivalent control group
- problematic type of control group when an equivalent canot be isolated
67
Placebo effect
- subjects behave different because thy think they have recieved treatment substance or condition
68
Reactance
- attitude change in response to feeeling that options are limited
- e.g. someone is set on a type of ice cream flavour when they find out it is sold out
69
Selective attrition
- when subjects that drop out of an experiment are different from those that reamin
- remaining sample is no longer random (morality effects)
70
Social desirability
- when subjects do and say what they think puts them in a favourable light
71
Illusory correlation
- when a relationship is inferred when there is actually is none
- e.g. people insist a relationship exists between physical and personality characteristics depsite evidence tht no such relationship exits
72
Meta-analysis
- study that mathematically comvines and summarize overall effects of research findings for topic
- calcualte 1 overall effect size
- needed when conflicting results are found
73
Insittuional review board (IRB)
- all studies have to pass ethical standards
- all subjects are provided with risk and benefit of being in study and then sign a consent form
- Milgram experiment (electric shocks) was the catalyst for higher ethical standards in psychological research
74
Statistics
- process of repreenting or analyzing numerical data
75
Descriptive statistics
- organize data from a sample by showing it in a meaningful way
- do not allow conclusions to be drawn beyond by the sample
76
Percentiles
- most common on standardized test
| - shows position in whole group
77
Frequenecy distrubutions
- how data in study looked
| - might show how often different variables appeared
78
Nominal variables
- descriptive bames
| - no order or relationship among variables but grouping
79
Ordinal variables
- implies order
| - nothing else is known (equal spacing not assumed)
80
Interval variables
- implies order and equal spacing
- do NOT include a real zero
- 0 is arbiturary and does not signify absence of temperature
81
Ratio variable
- have order, equal intervals and a real zero e.g. age - incrases by a year, increases and has an absolute 0
82
Graphs
- used to plot zero
83
Frequency polygon
- plotted points connected by lines
| - often used to plot continuous varaibles
84
Histogram
- consists of vertical bars in which the sides of vertical bars touch
- for discrete varisbles that have clear boundaries
- interval variables where there is order
- bars lined up in order
85
Bar graph
- vertical lines in historgram but o not touch
86
Measures of central tendecny
- where on a number line the data set falls in general
| - 3 types of central tendecny (mean, mode, median)
87
Mean
- average
| - affected by high scores (outliers)
88
Standard error of the mean
- calculate how "off" the mean might be in either direction
89
Median
- find the middle of the set
90
Mode
- most frquenctly occuring value
91
Variability
- additional informtion to central tendency by telling you how the scores are spread out overall
92
Range
- basic measure of variability but subtracting lowest to highest value
- is the spread
93
Variance/standard deviation
- how much variation there is among the number in distrubution
1) how much each score differs from mean
2) square each deviation
3) add all deviations = sum of square
4) divided by n = variance
4) take square root of variance = standard deviation
- large = highly disperesed
94
Normal distribution
- AKA bell curve
- larger sample = greater chance of bell curve
- unimodal (one hump) --> mean, median and mode are all equal
95
z-scores
- normal distribution
- how many standard deviations a score is from the mean
- usually -3 to 3
96
t-score
- transformation of z-score
- mean = 50
- SD = 10
= t= 10(Z)+50
97
Standard normal distribution
- same thing as normal distribution
- mean = 0
- SD = 1
98
Correlations
- part of statistics but neither purely descriptive nor purely inferential
- show relationships NOT causlity between variables
99
Positive correlation
- simple and linear
| - one increases, so does other variable
100
Negative correlation
- simple and liner
| - as one variable goes down, the other goes up
101
Cuvilinear correaltion
- not simple and linear
| - looks like a curved line e.g. arousal and performance
102
Zero correlation
- no relationship between the variables
103
Person r correlation coefficient
- way of numerically calculating and expressing correlations
- range from -1 to +1
- -1 = perfect negative correlation
- +1 = perfect positive correlation
- 0 = no relationship
- strength of relationship is how close it is to -1 or +1
104
Spearman r correlation coefficient
- another correlation and used when the data in form of rank
- used to determine the line of linear relationships
105
z scores with what percentile rank?
34:14:2
106
Regression
- step above simple correaltions
| - looks are the variance accounted for
107
Statitical regression
- identify relatioship between 2 variables and makes predictions about 1 variable based on another varible
108
Inferential statistics
- allows generalizability from sample to population
109
Statistics
- numbers that descirbe the sample
110
Parameters
- numbers that describe the popultion
111
Significance
- if numbers that desribe the sample are describing real differences or pattern rather than random variable
- if significant, research can generalize the same findings to population
112
Test of significance
- hoping to reject the null hypothesis (hypotheis that no real differences exisit)
- test of significant shows that statiscs were significant and that you can reject the null hypotheis
113
Alpha level
- a significance level used by most of
114
Type 1 error
- incorrectly reject the null hypothesis
| - a false positive
115
Type 2 error
- incorrecty accept the null hypothesis
| - a flase negative
116
T-tests
- compare means of 2 different groups to see if groups are truely different
- analyze continuous data
- can not test for more than 2 groups
117
Chi-square
- when n are classified into categories or cells
- tells us whether the groups are significantly different in size
- look at patterns or distributions (not means)
- analyze categorical or discrete data
- can assess "goodness of fit" - whether pattern is what would be expected
118
Categorical/discrete data
- data that has been countd rather than measured so limited to positive and whole numbers (usually)
119
ANOVA
- popular because of flexibility
| - analyze differences among means of continuous variables but it more flexible because more than 2 groups
120
1 way ANOVA
- tests whether the means on one outcome or DV are significantly diffferent across groups
121
2 way ANOVA
- test the effects of 2 independnet variables or treatment conditions at once
122
Factorial analysis of variance
- used when an experiment involves more than one IV
- can separate the effects of different levels of different beariables
- split each IV into levels which yield combinations
123
Main effects (factor analysis)
- the effect of IV on DV
124
Interaction effects (factor analysis)
- combine the IV
| - effects of DV change depending on the level of IV
125
Analysis of Covariance (ANCOVA)
- tests whether at least 2 roups co-vary
| - can adjust for preexisting differences between groups
126
Linear Regression
- allows use of correlation coefficients in order to predict one variable from another variable
- meaure relationship but don't describe the relationship
- use correlational data to make predictions based on a line fit with the least-squares method
127
Creating tests with statistics
- ensure that measures are on target
128
Standardized test
- used to create norms and tried out on huge groups of people
129
Criterion-reference test
- measure master parituclar subject e.g. final exam
130
Domain-referenced test
- measure less-defined properties e.g. intelligence and need be checked for reliability and validity
131
Reliability
- how stable the measure is
132
Test-re-test reliability
- measured by the same individual taking the same test more than once
- high test re-test = person would et the same core each time
133
Split-half reliability
- comparing an individuals' performance on 2 halves of th same test
- reveals internal consistency of test
134
Item analysis
- increase internal consistency
- analyze how large group responded to each item on the meaure
- weeds out problematic items and replaced with better questions with discriminatory value
135
Validity
- how well the test mesures a construct
136
Internal validity
- the extent to which the different items within a measure "hang together" and test the same thing
137
External validity (4)
- extent to which test measure what is intends to measure
1) concurrent
2) construct
3) content
4) face
138
Concurrent validity
- whether scores on new measure positively correlate with other measures known to test the same construct
- AKA cross validation
139
Construct validity
- whether the test really taps he abstract concept being measured
140
Content validity
- whether the content of the test covers a good sample of the construct being measured (not just part of it)
141
Face validity
- whether the test items simply look like they measure the construc
142
Campbell & Fiske
- created the multitrait-multimethod technique to determine the validity of test
