Research Methods I Flashcards
1
Q
Goal of Science
A
General approach to understanding natural world
2
Q
Systematic Empiricism
A
Learning based on observations of real world
3
Q
Empirical Questions
A
Questions about the way the world is, some not observable
4
Q
Public Knowledge
A
Work published in journal within context of prior work
5
Q
Three fundamental features of science
A
Systematic empiricism, empirical questions, and public knowledge
6
Q
Pseudosciences
A
activities and beliefs claimed scientific via proponents
7
Q
Three sins of pseudosciences
A
Claim to be scientific, lacks one or more of the 3 features of science, and not empirically based, can not be falsified
8
Q
Clinical practice of psychology
A
diagnosis and treatment of psychological disorders and related problems
9
Q
Empirically supported treatments
A
Studied scientifically and results in improvement compared to no treatment, placebo, or alternative
10
Q
Variable
A
Quantity/quality varying across people/situations
11
Q
Quantitative Variable
A
Quantity measured by assigning number to individual
12
Q
Categorical Variable
A
Quality measured by assigning label to individuals
13
Q
Population & Sample
A
Very large group of People.
Small/subset of populationi
14
Q
Random Sampling
A
every member of population has equal chance to be selected for sample. Difficult/impossible due to less definition of population
15
Q
Convenience Sampling
A
Individuals who happen to be nearby, willing to participate, may not represent population
16
Q
Statistical relationship
A
Occurs between two variables when average score differs across levels of other variable
17
Q
Difference between groups
A
One for of statistical relationships, displayed as bar graph typically
18
Q
Correlation between Quantitative Variables
A
second form of statistical relationships, displayed as scatterplots.
19
Q
Relationships (+/-)
A
Positive when high scores on one variable associate with higher scores on other.
Negative when higher scores on one associate with lower scores on other.
20
Q
Strength of Correlation
A
For quantitative variables use Pearson’s r, range from -1 to +1 good for linear relationships.
21
Q
Correlation and Causation
A
Correlation does not imply causation, independent variable is thought to be cause and is manipulated while dependent variable is the effect or result.
22
Q
Difference between correlation and causation
A
directionality problem where X causes Y or Y causes X. Third-variable problem has another variable, Z causing both X and Y.
23
Q
Research Literature
A
published research in field, enormous for psychology, must be either article in professional journals, or scholarly books in psychology and related fields
24
Q
Professional Journals
A
Monthly/quarterly published either Empirical research report (one< study) with method and conclusions. Review articles summarize previous and present new organization of results. Theoretical article is review article devoted to showing new theory
25
Double-blind peer review
manuscript is submitted to editor- established researcher who sends it to other experts.
26
Scholarly Books
Written for practitioners for use by other researchers and practitioners, monograph= single author or small group for coherent presentation.
Edited volumes have small group or one editor review and write chapters on same topic
27
Databases
Academic Search Premier, JSTOR, and ProQuest for general information. Use PscyINFO for APA information, comprehensive
28
Ethics
Branch of philosophy concerned with morality, set of principles and practices providing moral guidance in field
29
Three principles of Tri-Council Policy Statement
Respect for Persons, Concern for Welfare, and Justice
30
Respect for Persons
Respect autonomy with free, informed, and ongoing consent and protect those incapable of such due to age, impairment, illness, and other issues. Concepts are autonomy and protection
31
Concern for Welfare
No unnecessary risk, privacy is considered, confidentiality, and enough information to assess risks and benefits.
32
Justice
Fair and equal treatment, vulnerabilities considered and marginalized groups are not unjustly excluded, compensation given if needed, meet professional obligations
33
Confederate
Helper pretending to be a real participant
34
Autonomy
Right to make own choices free from coercion
35
Informed Consent
Obtain and document agreement to participate in study after being informed of what may affect decision
36
Privacy
Right to decide what information of selves is shared with others
37
Confidentiality
Maintained by researcher, agreement to not disclose personal information without consent or legal authorization
38
Ethical Conflict Inevitable?
Yes, can be minimized in risks and maximizing the benefits so they are unequal
39
Nuremberg Code
10 principles in 1947 with Nazi physicians who committed unspeakable actions
40
Declaration of Helsinki
World Medical Council, 1964, human research based on written protocol, or detailed research reviewed by committee
41
Belmont report
Recognized principle of seeking justice, and balancing risks and benefits of groups at societal level
42
Tri-Council Policy Statement: Ethical Conduct for Research Involving Humans
TCPS-2 of the social sciences and humanities research council, institutes of health research, and natural sciences and engineering research council are all in Canada in 1988, 2010 had TCPS-2 arise updating all responsibilities
43
REB
Research Ethics Board, committee responsible for reviewing protocols in research for ethical problems. Has two members with expertise in disciplines, one expert in ethics, and community member with no affiliation
44
Full REB review
Default requirement for human research
45
Minimal risk research
Likelihood of harms is no greater than those encountered by participants in aspects of everyday life, includes surveys, naturalistic observation, questionnaire-based studies
46
APA Ethics Code
Ethical Principles of Psychologists and Code of Conduct, much research concerns clinical practice of psychology, use Standard 8 for research and publication.
47
Consent Form
to be red and signed, good practice to tell participants of risks and benefits, demonstrate procedure, ask if any questions arise, and remind of right to withdraw at any time. Not necessary all time if in public place where person have no belief of privacy
48
Deception
misinforming participants on purpose of study, confederates, phony equipment, false feedback, allowed when benefits outweigh harm
49
Debriefing
Informing participants as soon as possible of purpose of study, revealing deception, correcting other misconceptions, minimize harm which occurred, return mood to happy thoughts.
50
Scholarly Integrity
No fabrication of data or plagiarism, direct quotations, citation provided
51
Ethical responsibilites
To be known and accepted, reading and understanding TCPS-2, REB review, minimal harm is disclosed
52
What should be done with risks with respect to the APA's recommendations?
Are too be identified and minimized, may be accustomed to the risks and so reduce or eliminate as many as you can by modifying research design. Prescreening helps minimize participants at higher risk as well. Finally, active steps to maintain confidentiality,
53
Deception (minimized and Identified)
Generally accepted to wait until debriefing to reveal research question as long as procedure, risks, and benefits are disclosed in informed consent process.
54
Informed Consent and Debriefing
Provide as much information as possible, script or talking points created, informed consent covering all rights of persons, debriefing to identify all risks and benefits to person. When approved remain vigilant to potential risks
55
Measurement
Assignment of scores to individuals so scores show some characteristic of individual
56
Psychometrics
Psychological measurement, needs some systematic procedure for assigning of scores to represent characteristic of interest
57
Psychological Constructs
Variable which can not be assessed by observation, lack capacity to behave in certain manner, internal process are involved.
58
Conceptual definition
Describes behaviours and internal processes making up construct and relation to other variables.
59
Operational Definitions
Definition of variable in terms of measurement, self-report measures (report thoughts/feelings), behavioural measures (observed and recorded behaviours), and physiological measures (recording heart rate, BP, GSR, hormones).
60
Converging Operations
Multiple operational definitions of same construct in or across studies
61
Levels of measurement
Nominal level (categorical variables, no ordering), ordinal level (scores to represent rank-order of individuals, no representation of intervals), interval level (numerical scale with same interval and same interpretation for such intervals throughout, no true zero point however), and ratio level (true zero point representing complete absence)
62
Reliability
consistency of measures,
63
Test-retest reliability
reliability across time, correlation done by scatterplot and Pearson's r. Above .7 is good, .8 and above ideal.
64
Internal consistency
Consistency of answers across items on multiple-item measure, split-half correlation with two sets of items and score computed for each set, calculated using Cronbach's alpha. Above .7 is good, .8 and above ideal.
65
Interrater Reliability
Extent to which different observers are consistent in judgments, calculated with Cronbach's alpha or Cohen's kappa when categorical, above .7 is good but .8 and above is ideal.
66
Validity
Extent to which scores represent variable intended to
67
Face Validity
The extent to which measurement appears to measure the construct of interest to the everyday person. Weak evidence of validity, take it at face value, subject to social desirability bias.
68
Content Validity
Measure covers construct of interest, done by checking method against conceptual definition of construct
69
Criterion/Construct Validity and convergent validity
The extent of scores correlating with other variables it should correlate with.
Convergent is when the variables we expect to correlate do in fact correlate. The focus on correlation.
70
Concurrent Validity
When the criterion is correlated with other constructs and is able to predict the results of other constructs. Focus is on prediction here.
71
Predictive validity
When the criterion is measured at some point in the future after construct has been measured, as the scores predicted the future outcome. You focus on predicting future behaviours.
72
Discriminant Validity
The extent of scores on the measure not correlating with measures of variables it should not correlate with, questions on extraversion should not correlate with openness.
You also do not want too high a correlation either as it suggests the concepts are tapping into the same construct, .9 is excessively high as correlation.
73
Creating an Operational Definition
Existing Measure is typically best as you save time and effort, already tested to be valid and results can be compared with previous studies. Making own measure still look to literature, strive for simplicity weighed against need for multiple items rather than one, done for content validity, and have way of measuring overall score.
74
Convergent Validity
Extent to which measurement method works in same way as existing measures
75
Socially Desirable Responding
People respond in ways they are expect to, do not want to look bad in eyes of researcher
76
Demand Characteristics
Subtle cues revealing how researcher expects participants to be behave, prevented by making as simple and brief so as not to tempt participants into venting frustrations on results or guarantee anonymity
77
Evaluating Measure
Done by evaluating reliability and validity, if data casts doubt on reliability or validity then ask why, could be a number of reasons why.
78
Experiment
Study designed to answer question of causal relationship between independent and dependent variable
79
Conditions
First fundamental feature of experiments where we see different levels of independent variable
80
Extraneous Variable
Second fundamental feature of experiments which is attempted to be controlled, add noise to experiment and make reading it harder
81
Internal validity
Purpose of experiment to show two variables are statistically related, logic where two similar conditions are created and independent variable creates difference(s) between them
82
External Validity
High if the way the study is done supports generalizing results to people and situations beyond those studied, often said to be mundane realism
83
Psychological realism
Same mental process is used in both laboratory and real world
84
Experiments and external validity
Blanket conclusion of experiments being low in validity is true in part since they learn of psychological processes operating in variety of people and situations
85
Construct Validity
Quality of experiment's manipulations, sees conversion from research question to experimental design or operationalization
86
Statistical Validity
Speaks to whether statistics conducted in study support conclusions made, number of conditions or participants determine effect size, power analysis for any real difference
87
Manipulation Check
Separate measure of construct attempting to be measured on the part of the researcher
88
Controlling Extraneous Variables
Can be done by holding them constant, testing everyone in same place, same instructions, and treating same way
89
Noise
Created by extraneous variables and make it difficult to detect effect of independent variable by adding variability to data and becoming confounding variables
90
Confounding Variable
Extraneous variable differing on average across levels of independent variables,
91
Problems of Confounding Variable
Confound means to confuse which is what the variables do, they provide alternative explanation for any observed difference in the dependent variable. Avoided by holding extraneous variables constant.
92
Between-Subjects
Participants are tested in only one condition. simpler, require less time for testing, avoid carryover effects.
93
Random Assignment
Random process to decide which participants are tested in which conditions, important in all experiments yet unlikely to result in equal sample sizes in different conditions.
94
Block Randomization
All conditions occur once in sequence before any are repeated, occur in random order, generated prior to testing participants.
95
Limits of Random Assignment
Not guaranteed to control all extraneous variables across conditions, participants may be older, younger, not concern because confound is detected when experiment is replicated.
96
Treatment
Any intervention meant to change behaviours for the better, psychotherapies, and medical treatments for disorders and learning, conservation, reduce prejudice, and so on.
97
Treatment and Control Condition
Treatment has participant receive treatment, control has no treatment, if treatment works they then receive the treatment, no-treatment has a placebo given so the placebo effect occurs where the expectancy of feeling better causes improvement.
98
Randomized clinical trial
Psychotherapies and medical treatments involve this type of experiment
99
Placebo Control condition vs. no-treatment control
Receive placebo that looks like treatment yet lacks active ingredient/element responsible for effectiveness.
No treatment has no treatment whatsoever
100
Waitlist Control Condition
Participants will receive the treatment but must wait until treatment condition has received it, alternatively leave out control condition altogether and compare to best alternative treatment.
101
Within-subjects
Participants are tested under all conditions, maximum control on extraneous variables, made to be less noisy. Control extraneous variables, reduce noise, and detect relationship between independent and dependent variable.
102
Carryover effect
effect of being tested in one condition on participants behaviour in later conditions. First condition's potential effect carryover to next conditions
103
Practice Effect
type of carryover effect wherein a task is better performed due to the chance to practice it
104
Fatigue Effect
Task is performed worse due to becoming tired or bored, carryover effect.
105
Context Effect
Changes in the survey when participants perceive stimuli/interpret task in a different context. Examples: Option A changes to option D, influential questions and order, anonymity/confidentiality, purpose of research is presented.
106
Counterbalancing
Testing different participants in different orders, random orders typically, controls order of conditions and detection of carryover effects.
107
Simultaneous Within-Subjects Designs
Presenting all information to the participants in a sequence within a randomized order.
108
Which is better within-subject or between-subject
within-subject, with proper counterbalancing, using one does not preclude using the other however.
109
Subject Pool
Established group of people agreeing to be contacted for participating in research studies
110
Standardization
Unintended bias on part of experimenter where multiple extraneous variables are introduced
111
Experimenter Expectancy Effect
Expectations about participants feelings and behaviour may unintentionally give clearer instructions or encouragement or more time allotted
112
Minimizing unintended variation in procedure
Standardize with written protocol, standard instructions, automate procedure, anticipate questions on part of participants, train many experimenters on protocol together
113
Double-Blind Study
Minimize experimenter expectations by making experimenter and participant blind to the condition which participants are in, may not always be possible
114
Record Keeping
Generate written sequence of conditions before beginning study and testing participant in sequence, good to add basic demographics like date, time, testing, name of experimenter, comments, questions.
115
Pilot Test
Small-scale study to ensure procedure works as planned, can be small size but give you confidence that procedure works as planned.
116
Nonexperimental Research
Lacks manipulation of independent variable, random assignment , or both
117
When to use nonexperimental
When manipulating IV is unethical, impossible, research question may be broad and exploratory, noncausal relationship between variables, single variables.
118
Single-variable research
Interest is in a single-variable, answer interesting and important questions, no questions on statistical relationships between variables.
119
Correlational Research
Measure two variables of interest and no attempt to control extraneous variables and assess relationship between them. No manipulation of IV, only relation between variables.
120
Quasi-experimental Research
Manipulate independent variable yet no random assignment to conditions.
121
Qualitative Research
Data are nonnumerical, not analysed with statistical methods, written description.
122
Internal Validity across Methods
Highest in Experimental, medium in quasi-experimental with some manipulation, and correlation is the lowest, fails to manipulate or randomly assign conditions.
123
Why Correlational Research?
Two reasons: do not believe statistical relationship is causal, and thought to be causal but cannot manipulate independent variable since impossible, impractical, or unethical.
124
Naturalistic Observation
Approach to data collection with behavioural observation, type of field research observing people in environment in which it occurs typically. Observations are unobtrusive, ethically accepted if participants are anonymous, behaviour occurs in public where privacy is not expected.
125
Issues of Naturalistic Observation
Sampling: when, who, where, under what conditions are we observing needs to be precise to make data collection manageable for researcher.
Measurement: can be straightforward but behaviour may not be obvious/objective. When it needs a judgment on part of observer it is called coding, a clearly defined set of target behaviours.
126
Archival Data
Data already collected for other purpose, measurements are more or less straightforward, content analysis wherein systematic approaches to measurement use complex archival data, specify keywords, phrases, or ideas and finding occurrences in data.
127
Multiple Dependent Variables
Possibility of carryover effects arise when experimenter is interested in other dependent variables, can be measured in same order for all participants, counterbalance as well possibly.
128
Manipulation Check
Additional measure of IV, confirms independent variable was manipulated successfully, done at end of procedure to ensure effect of manipulation lasted entire procedure.
129
Measure Same Construct
Operationally define and measure same construct or closely related ones, when multiple dependent variables are different measures they may be combined into single measure of construct. Generally more reliable, ensure internal consistency is computed with Cronbach's alpha.
130
Multiple Independent Variables
Interaction between variables where level of one IV depends on level of the other IV.
131
Factorial Designs
Each level of one IV combined with each level of other IV's to produce all possible combinations. Shown in factorial design table, if three conditions of one IV and two of other IV then 3x2 factorial design with six conditions. Unmanageable if more than two or three levels for each IV.
132
Between-subjects Factorial Design
All IV's are manipulated between subjects, tested in one condition. Simple, no carry-over effects, and minimizes time and effort on participants.
133
Within-subjects Factorial Design
All IV's are manipulated within subjects, tested in all conditions, more efficient and control over extraneous variables.
134
Mixed Factorial Design
Test in all conditions while having some participants tested on IV of one condition instead of both.
135
Nonmanipulated IV's
Measured variables which is not manipulated, commonly participant variables, and are between-subjects factors, experimental so long as one IV is manipulated, causal conclusions only drawn on manipulated IV.
136
Graphing Results of Factorial Experiments
Results can be graphed with one IV on x-axis and other with different bars/lines, y-axis for dependent variable.
137
Main Effect
Statistical relationship between IV and dependent variable averaging across levels of other independent variables.
138
Interaction
Effect of IV depends on level of other IV, crossover interaction where IV for one group changes but then other group has a different change due to IV once again. Typically primary focus of research.
