Research Methods

Front 1

Authority

Back 1

anyone we trust to have more knowledge than us and we trust to be truthful to us
-people trust authority figures and experts even when they might be wrong

Front 2

Use of Reason

Back 2

must be careful to avoid faulty logic and not to base logical reasoning on false premises

Front 3

Empricism

Back 3

process of learning things through direct observation or experience and reflection on those experiences

Front 4

Belief Perseverence

Back 4

tendency to continue believing something even when there is contradicting info

Front 5

Confirmation Bias

Back 5

tendency to look for evidence that supports our beliefs

Front 6

Availability Heuristic

Back 6

tendency to overestimate frequency of salient / memorable events

Front 7

Determinism

Back 7

all events have causes
scientists can examine extent to which behaviour is influenced by free will, degree to which some behaviours are more free than others, and what limits might be on free choices

Front 8

Statistical Determinism

Back 8

events can be predicted, but only with probability greater than chance

Front 9

Discoverability

Back 9

by scientific methods, causes can be discovered with confidence

Front 10

Systematic Observations

Back 10

include precise definition of phenomena being measured, reliable and valid measuring tools, accepted research methodologies, and system of logic for drawing conclusions

Front 11

Objectivity

Back 11

eliminate human factors like expectation and bias
objective observation can be verified by more than one observer

Front 12

Introspection

Back 12

participants in experiment perform task and give
detailed description of conscious experience of task
(subjective, unverifiable)

Front 13

Data-Driven

Back 13

researchers expect conclusions about behaviour to be supported by evidence of objective info gathered through systematic procedure

Front 14

Falsification

Back 14

can be disproven

Front 15

Pseudoscience

Back 15

any field of inquiry that appears to use scientific methods and tries to give that impression, but is actually based on inadequate, unscientific methods and yields results that are generally false
-relies on anecdotal evidence, sidesteps disproof, reduces complex phenomena to overly simplistic concepts
Effort Justification
-phrenology or subliminal CDs

Front 16

Description

Back 16

identify regularly occurring sequences of events

Front 17

Prediction

Back 17

behaviour follows laws; regular relationships exist between variables

Front 18

Explanation

Back 18

know what caused behaviour to happen

Front 19

Causality

Back 19

involves covariation, experimental control, time sequence with cause preceding effect, theoretical structure, ruling out alternatives

Front 20

Application

Back 20

various ways of applying learned principles of behaviour

Front 21

Beneficience and Non-malfeasance

Back 21

constantly weigh costs and benefits, produce greatest good

Front 22

Fidelity and Responsibility

Back 22

constantly aware of responsibility to society; uphold highest standards of professional behaviour

Front 23

Integrity

Back 23

be honest

Front 24

Justice

Back 24

everyone treated equally and fairly, reduce bias

Front 25

Respect for People's Rights and Dignity

Back 25

protect welfare and rights of participants

Front 26

IRB

Back 26

determines whether project meets ethical guidelines
-assess risk to subjects
-controversial

Front 27

Informed Consent

Back 27

sufficient info to decide whether to participate
-need to include basic description, length of study, ability to quit anytime, confidentiality and anonymity ensured, researcher contact info available, opportunity to review final results of study, signatures
-need to be careful with infants, children, prisoners, and disabled people

Front 28

Deception

Back 28

needed for participants to act naturally and not botch results

Front 29

Assent

Back 29

researchers give as much info as possible to gauge whether child is willing to participate and rewards given after

Front 30

Debriefing

Back 30

experimenter answers questions and explains purpose of study
can have problem with leakage
amoung of time spent debriefing depends on complexity of study, presence and degree of deception, level of potential distress

Front 31

Dehoaxing

Back 31

revealing purpose of experiment to participants

Front 32

Desensitizing

Back 32

reducing stress / negative feelings from experience

Front 33

Animal Research

Back 33

need to justify study, care and acquire animals properly, and use them for educational and not research purposes
prefer less stressful procedures, proper training, vet inspections, videos or simulations over real animals
controversial, but it has benefited animals and humans

Front 34

Internet Research

Back 34

problems with consent and debriefing

Front 35

Scientific Fraud

Back 35

plagiarism and data falsification
can be caught through peer review process or attempts at replication

Front 36

Basic Research

Back 36

designed to understand fundamental psychological phenomena
describing, predicting, and explaining fundamental principles of behaviour and mental processes
(ex, dichotic listening)

Front 37

Applied Research

Back 37

designed to shed light on solutions to real-world problems
(ex, cell phone and driving studies)

Front 38

Lab Research

Back 38

conditions of study can be specified more precisely and participants can be selected and placed in different conditions of study more systematically
-easier to control for ethics

Front 39

Field Research

Back 39

environment more closely matches situations encountered in daily life
-important because it cannot be replicated in lab, can help confirm findings of lab studies and can make discoveries that make immediate difference in participants lives

Front 40

Mundane Realism

Back 40

how closely study mirrors real-life experiences

Front 41

Experimental Realism

Back 41

extent to which research study has impact on subjects
forces them to take matters seriously and involves them in procedures

Front 42

Quantitative Research

Back 42

data collected and presented in form of numbers
-includes statistical analysis

Front 43

Qualitative Research

Back 43

includes studies collecting interview info, detailed case studies or observational studies
analytical narratives that summarize project's main outcome
-includes content analyses and interviews

Front 44

Empirical Questions

Back 44

must be answerable with data and terms must be precisely defined
-may evolve out of everyday observations of behaviour, need to solve practical problem, attempts to support or refute theory, or unanswered questions from completed study
1. Observe behaviour
2. Apply problem
3. Test theory
4. Generate new question from past research

Front 45

Operational Definitions

Back 45

definition of concept or variable in terms of precisely described operations, measures, or procedures

Front 46

Converging Operations

Back 46

understanding of some behavioural phenomenon is increased when series of investigations converge on common conclusion

Front 47

Serendipity

Back 47

act of discovering something while looking for something else

Front 48

Theory

Back 48

set of logically consistent statements about some phenomenon that best summarizes existing empirical knowledge
organizes knowledge in form of precise statements of relationships among variables
provides tentative explanation for phenomenon and serves as basis for making predictions about behaviour

Front 49

Constructs

Back 49

hypothetical factor not observed directly; existence is inferred from certain behaviours and assumed to follow from certain circumstances

Front 50

Induction

Back 50

logical process of reasoning from specific events to general

Front 51

Deduction

Back 51

reasoning from set of general statements toward prediction of event

Front 52

Productivity

Back 52

good theories advance knowledge by generating great deal of research

Front 53

Falsification

Back 53

science proeeds by setting up theories and attempting to disprove them

Front 54

Parsimony

Back 54

theories include minimum number of constructs and assumptions necessary to explain phenomenon adequately and predict future research outcomes

Front 55

Replication

Back 55

study that duplicates some or all of procedures of prior study (ex, training researchers)

Front 56

Extension

Back 56

resembles prior study and may replicate part of it but adds at least one new feature

Front 57

Measurement Error

Back 57

any factor that introduces inaccuracie into measurement

Front 58

Reliability

Back 58

repeatability over time, populations, locations

Front 59

Content Validity

Back 59

measures what it is supposed to measure

Front 60

Criterion Validity

Back 60

refers to whether test can successfully predict some future behaviour or whether it is meaningfully related to some other measure of beahviour

Front 61

Construct Validity

Back 61

refers to whether test adequately measures some construct

Front 62

Convergent Validity

Back 62

if scores on test designed to measure some construct are correlated with scores on another test theoretically related to construct, convergent validity would be high

Front 63

Discriminant Validity

Back 63

if scores on test designed to measure some construct are uncorrelated with scores on another test that should be theoretically unrelated to construct, discriminant validity would be high

Front 64

Scales

Back 64

Nominal - categorical data, numbers not meaningful
Ordinal - ordering of numbers is meaningful
Interval - each unit increase reflects same change in underlying measure without true zero
Ratio - same as interval with true zero

Front 65

Descriptive Statistics

Back 65

summarize data collected from sample of particpants
Central tendency - mean, mode, median
Variability - range, standard deviation, variance, interquartile range (for outliers)
Graphs - histogram, frequency distribution, stem and leaf display

Front 66

Inferential Statistics

Back 66

draw conclusions about data that can be applied to wider population

Front 67

Systematic Variance

Back 67

result of some identifiable factor, either variable of interest or factor not controlled for

Front 68

Error Variance

Back 68

nonsystemic variability due to individual differences between participants and any random, unpredictable effects

Front 69

Hypothesis Testing

Back 69

null hypothesis and alternative hypothesis

Front 70

Alpha Level

Back 70

probability of obtaining particular results if null hypothesis is really true

Front 71

Type I Error

Back 71

rejecting null hypothesis when it is true (chance of occurrence equal to alpha)

Front 72

Type II Error

Back 72

fail to reject null hypothesis when it is false when sample size is small or measurements aren't reliable

Front 73

Effect Size

Back 73

estimate of magnitude of difference among sets of scores while taking into account amount of variability in scores

Front 74

Meta-analysis

Back 74

uses effect size analyses to combine results from several experiments that use same variables even though they are likely to have different operational definitions

Front 75

Confidence Interval

Back 75

range of values expected to include population value with certain degree of confidence
(margin of error)

Front 76

Power

Back 76

chance of rejecting null hypothesis when it is false
-affected by alpha level, effect size, and sample size

Front 77

Experiment

Back 77

systemic research study in which investigator directly varies factor(s), holds all other factors constant and observes result of variation

Front 78

Mill's Inductive Logic

Back 78

Method of agreement: if X, then Y
if X is regularly followed by Y, X is sufficient for Y to occur and could be cause of Y

Method of difference: if not X, then not Y
if Y does not occur when X does not, X is necessary for Y to occur

Front 79

Manipulated IVs

Back 79

Situational - manipulated environment
Task - participants complete intellectual activity
Instructional - given instructions on how to do task

Front 80

Confounds

Back 80

uncontrolled extraneous variable
Distributed Study Time example

Front 81

Dependent Variables

Back 81

behaviours measured in experiment
Ceiling effects - all scores high
Floor effects - all scores low

Front 82

Subject Variables

Back 82

not manipulated; pre-existing attributes of participants

Front 83

Geraldo Rivera courtship study

Back 83

confederates asked strangers if they wanted to have sex in a van

Front 84

Statistical Conclusion Validity

Back 84

extent to which researcher uses statistics properly and draws appropriate conclusions from statistical analysis

Front 85

Construct Validity

Back 85

quality of operational definitions for variables

Front 86

External Validity

Back 86

generalize beyond experiment to other populations, environments and times

Front 87

Internal Validity

Back 87

confound free

Front 88

Threats to Internal Validity

Back 88

Pre-post studies - History, Maturation, Regression to the mean, Testing and Instrumentation
Participants problems - Subject selection (groups are not equivalent), Attrition (subjects do not complete study and participants left may be different than ones who pulled out early)

Front 89

Between-subjects designs

Back 89

each level of IV contains diff group of subjects
-necessary with subject variable or deception
-more subjects needed, more costly for time and resources, differences in nonequivalent groups

Front 90

Within-subject / repeated-measures designs

Back 90

each participant exposed to every level of IV
-fewer participants needed, eliminates individual differences
-however, can have sequence effects

Front 91

Random Assignment

Back 91

each subject has equal chance of being placed in any group in study

Front 92

Block Randomization

Back 92

each condition of study has participant randomly assigned to it before any condition is repeated second time

Front 93

Matching

Back 93

participants grouped together on trait then distributed randomly to diff groups in experiment
-use when small sample might cause random assignment to fail
-can only use if matching variable correlates with dependent variable and if measuring matching variable is feasible
-can be problem when 2 groups sampled from diff populations that differ on factor being used as matching variable (can enhance regression affect)

Front 94

Sequence / Order Effect

Back 94

first part of study influences performance later
Progressive Effect: performance changes steadily from trial to trial
Carryover Effect: when particular sequence of conditions might produce effect different than another sequence

Front 95

Counterbalancing

Back 95

alter order of experimental conditions
use more than one sequence, works better for progressive
Complete, Partial (Latin square), Reverse

Front 96

Latin Square

Back 96

every condition of study occurs equally often in every sequential position
every condition precedes and follow every other condition exactly once

Front 97

Cross-Sectional design

Back 97

between-subjects design, short duration
-can have cohort effects

Front 98

Longitudinal design

Back 98

within-subjects design, cohort effects eliminated
-can have attrition

Front 99

Cohort Sequential design

Back 99

group of subjects selected and retested every few years with additional cohorts selected and retested every few years
combines cross-sectional and longitudinal

Front 100

Experimenter Bias

Back 100

experimenter expectations can influence subject behaviour
-can automate procedure and use double blind

Front 101

Participant Bias

Back 101

demand characteristics, 'good' subjects
-behaviour of participants can be affected by experimenter's race and gender, demeanor, friendliness, attitude
-control with deception, manipulation checks, field research

Front 102

Independent Groups Design

Back 102

between-subjects design, manipulated IV, random assignment

Front 103

Matched Groups Design

Back 103

between-subjects design, manipulated IV, matching

Front 104

Nonequivalent Groups Design

Back 104

between-subjects, IV is subject variable

Front 105

Within-subjects Design

Back 105

manipulated IV

Front 106

SIngle-Factor Multilevel Designs

Back 106

ability to discover nonlinear effects, ability to address multiple questions and rule out alternative explanations
going beyond 2 levels makes all counterbalancing options available

Front 107

Control Group Designs

Back 107

Placebo - subjects think they are being treated when they're not
Waiting list - ensure control and treatment groups are similar
Yoked - each control group subject yoked to experimental group subject

Front 108

Factorial Design

Back 108

more than one IV

Front 109

Notation System

Back 109

number x number
# of numbers indicates how many IVs
number indicates how many levels

Front 110

Factorial Matrix

Back 110

with 2x2:
Factor A; levels A1 A2
Factor B; levels B1 B2
Conditions A1B1, A1B2, A2B1, A2B2

Front 111

Main Effect

Back 111

overall effect of IV

Front 112

Interaction

Back 112

effect of one IV depends on level of another IV
-row and column means of factorial matrix same but something still happened
-if lines on graph are not parallel, interaction probably exists

Front 113

Possible outcomes for 2x2 factorial design

Back 113

1. main effect for factor A only
2. main effect for factor B only
3. main effects for A and B only
4. main effect for A plus an interaction
5. main effect for B plus an interaction
6. main effects for A and B plus an
interaction
7. interaction only, but no main effects
8. no main effects, and no interaction

Front 114

Mixed Factorial Designs

Back 114

one between-subjects factor and one within-subjects factor

Front 115

PxE Factorial Design

Back 115

subject (person) variable and manipulated (environment) variable

Front 116

Mixed PxE Factorial

Back 116

subject variable and manipulated within-subjects variable

Front 117

Aptitude-Treatment Interaction Designs

Back 117

educational research
Aptitude (subject) x Treatment (environmental)

Front 118

Correlation

Back 118

observes variables and relates them, looks for ways in which people differ from each other (variance among organisms)

Front 119

Correlation Coefficient

Back 119

Pearson's r
ranges from -1 to +1
assumes linear relationship

Front 120

Coefficient of Determination

Back 120

Pearson's r^2
proportion of variability in one variable that can be accounted for by variability in other variable

Front 121

Regression Lines

Back 121

straight lines that best summarize correlation
Y = a + bX

Front 122

Criterion Variable Y

Back 122

variable being predicted

Front 123

Predictor Variable X

Back 123

variable used to make prediction

Front 124

Cross-lagged Panel Correlation

Back 124

investigates correlations bw variables at several points in time

Front 125

Advantages of correlational research

Back 125

some IVs can't be manipulated (subject variables)
practical / ethical reasons

Front 126

Correlations in nature-nurture research

Back 126

compare identical twins vs. fraternal twins, reared together vs. reared apart

Front 127

Multivariate Analysis

Back 127

Y = a + bsXs ...

one criterion variable (Y) but several predictor variables (X)
-when influences of several predictor variables combined, prediction can improve

Front 128

Regression Weights (bs)

Back 128

reflect relative importance of each predictor

Front 129

Ecological Validity

Back 129

research with relevance for everyday cognitive activities of people trying to adapt to environment

Front 130

Demand Characteristics

Back 130

aspects of study that reveal hypothesis being tested

Front 131

Hawthorne Effect

Back 131

participant behaviour affected by knowledge on is in experiment

Front 132

Evaluation Apprehension

Back 132

participants want to be evaluated positively

Front 133

Independent samples t-test

Back 133

when 2 groups of participants independent of each other (random assignment, subject variable)
assumes data approximate normal distribution, assumes homogeneity of variance
-when tests of homogeneity of variance indicate problem, use nonparametric tests

Front 134

Paired t-test

Back 134

within-subjects / repeated measures / matched groups

Front 135

ANOVA

Back 135

used for single-factor multilevel designs, tests for presence of overall significance that could exist somewhere among various levels of IV
-use post hoc analysis (Tukey's HSD) to determine level where significance lies
F-ratio shows to what extent mean differences could be due to chance or are result of some other factor

Front 136

Factorial Design

Back 136

any study with more than one IV

Front 137

N-factor ANOVA

Back 137

used for factorial designs using interval / ratio data
-more than one F-ratio
MS = SS/df F = MS / error variance
Simple effects analysis: compares each level of on factor with each level of another

Front 138

df

Back 138

df bw groups = # of groups - 1
df for interaction = df of first factor x df for second factor
df for error variance = (# of subjects per group - 1) x total # of groups

Front 139

Spearman's rho

Back 139

ordinal data

Front 140

Restricting range of variable

Back 140

weakens correlation

Front 141

Causality

Back 141

can attribute when A and B occur together, when A precedes B in time, when A causing B makes sense in relation to theory, when other explanations for co-occurence can be eliminated

Front 142

Factor Analysis

Back 142

large # of variables measured and correlated
-correlation matrix (all possible pairs of tests)
-determines factor loadings

Front 143

Basic Research

Back 143

aims to increase core knowledge about human behaviour and mental processes

Front 144

Applied Research

Back 144

aims to solve real-world problem
-problems include ethical dilemmas (informed consent, privacy, debriefing), reduced internal validity, between-subjects designs (nonequivalent groups, regression problems with matching), within-subjects designs (counterbalancing problems, sequence effects, attrition)

Front 145

Quasi-experiment

Back 145

whenever causal conclusions can't be drawn because less control over variables, random assignment not feasible
-single-factor nonequivalent groups with 2+ levels, nonequivalent groups factorial designs, PxE factorial designs, correlational research

Front 146

Nonequivalent Control Group Design

Back 146

random assignment is not used so control and treatment group already different from each other

Front 147

Interrupted Time Series Design

Back 147

take measures for extended period before and after event expected to influence behaviour
O1 O2 O3 O4 O5 T O6 O7 O8 O8 O10
-used to evaluate trends, can measure several dependent variables, some expected to be influenced and others expected not to change
-if outcome pattern of location 2 matches location 1 it supports treatment / program effect

Front 148

Archival Data

Back 148

info that has already been gathered for some reason aside from research project
-can encounter experimental bias

Front 149

Content Analysis

Back 149

any systematic examination of qualitative information in terms of predefined categories

Front 150

Program Evaluation

Back 150

applied research to assess effectiveness and value of policies or specially designed program
-determine if true need exists for program, who would benefit
-assess whether program is being run according to plan and what changes can be made to facilitate operation
-deals with ethical problems like informed consent, confidentiality, perceived injustice in control groups, avoid conflict with stakeholders

Front 151

Needs Analysis

Back 151

set of procedures for predicting whether population of suffiient size exists that would benefit from proposed program, whether program could solve clearly defined problem, and whether members of population would actually use program

Front 152

Formative Evaluation

Back 152

monitor progress of program during implementation
-if implemented as planned, provide clear and continuous on how program is being used (program audit)

Front 153

Summative Evaluation

Back 153

overall assessment of program effectiveness
-decide whether to keep program

Front 154

Cost-effectiveness Analysis

Back 154

monitors actual cost of program and relates cost of effectiveness of program's outcomes