Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
129 Cards in this Set
- Front
- Back
Response bias:
demand characteristics |
aspects of the study that people consider to determine the purpose
- behavior may be affected by knowledge whether accurate or not - changes participant reactivity |
|
Response bias:
social desirability bias |
behave in ways they think is socially acceptable
|
|
Response bias:
leniency bias |
unrealistically favorable rating to known person
(ex) professor/class ratings |
|
Response bias:
central tendency bias |
clusters responses around middle of scale
|
|
Response bias:
acquiescence response set |
"yea sayers"-- tendency to agree
|
|
ceiling effect
|
when scale does not have enough higher ratings to capture variability
|
|
floor effect
|
when scale does not have enough lower ratings to capture variability
|
|
volunteerism bias
|
volunteers differ from non-volunteers:
-more highly educated, higher social class, higher intelligence, higher need for approval, more social, more likely to be females |
|
ways to enhance response rate:
|
-include clear instructions about goal of project
- guarantee anonymity - small gift/lottery - follow up at 2-3 week intervals - response rates > 50% are good |
|
QXR reliability:
Test-retest |
administer, let time pass, readminister to the same group of people
-used to test the consistency of measure |
|
QXR reliability:
split-half reliability |
split the survey in half and correlate
useful with different versions of questions assessing the same construct -treats two halves of a measure as alternate forms |
|
Increasing QXR reliability
|
-increase the number of items
-standardize admission process -items that are clearly and accurately written |
|
QXR validity:
content validity |
do questions cover range of behaviors considered to be part of dimension that you're addressing (measuring what you want to measure)
- measures all facets of a given social construct (how adequately does instrument sample from important behaviors) |
|
QXR validity:
construct validity |
does it measure what it claims to measure
|
|
correlational research
|
tells us about the degree & direction of relationship
-no attempt to manipulate variables -no causal relationships |
|
experimental research
|
tells us if causal relationship exists between one variable and another
-manipulates IV to test influence of DV -compares DV under control/experimental -equates Ps in different conditions -Controls unwanted effects of extraneous variables on DV |
|
correlation
|
measure of association
-describing length of a linear relationship -gives direction & magnitude of relationship -DOESN'T tell: causation |
|
direction of correlational relationships
|
positive: as x increase, y increases
negative: as x increases, y decreases none: x and y are unrelated |
|
magnitude of correlational relationships
|
strength of association - Pearson's r (-1 to +1)
- sign tells: direction - number tells: magnitude Rule: <.3 = weak, .3-.6 = moderate, .6 = strong |
|
third-variable problem (Spurious correlations)
|
X and Y caused by some other variable
|
|
Directionality problem
|
does X cause Y or Y cause X
|
|
independent variable
|
manipulate independent variable to test influence on dependent variable
- you manipulate this! |
|
dependent variable
|
value depends on the independent variable
- DV usually compared using experimental or control groups |
|
random sampling
|
Generalizability
-each member has an equal chance of being chosen to be in study |
|
random assignment
|
Internal validity
- equal chance of Ps to be assigned to control or experimental condition (or other conditions) |
|
internal validity
|
when results can be confidently attributed to the effect of an independent variable
|
|
confound
|
when IV varies with another variable
-extraneous variable that exerts differential effect -threat to internal validity 1. age 2. cohort (generational effects) 3. time of assessment |
|
external validity
|
generalizability
-how applicable are results to other populations |
|
selection bias (external validity)
|
error due to systematic differences in characteristics of those who take part in a study and those who don't
-results when non-comparable criteria are used to enroll participants in an investigation |
|
remedies for volunteer bias
|
make interesting appeal, non-threatening, state importance of research, state why target population, offer small reward, avoid research that is stressful, use public or private commitment to volunteering, have someone known make appeal
|
|
reliability
|
ability to produce similar results on repeated administrations
-precision and error: true score and measurement error |
|
interrater reliability
|
multiple observers: how closely to observers agree when coding same observations
-% agreement, or Cohen's kappa where Pr(a)= observed agreement & Pr(e) = probability of agreement based on chance (takes into account disagreement based on chance) |
|
internal consistency
|
degree of relatedness of individual scale items
-split half reliability (correlation b/w two halves) |
|
validity (general)
|
testing: extent to which scale measures what is intended
methodology: methodological soundness/appropriateness external: generalizable results |
|
face validity
|
how well it "appears to" measure something
|
|
criterion validity (concurrent and predictive)
|
relation to other tests measuring similar phenomena administered to same people
|
|
naturalistic observation
|
3rd party observations
|
|
participant observer research
|
observer becomes member of group to be studied
-useful for secretive or isolated groups Problems:objectivity, reactivity, privacy & consent |
|
observational research
|
-careful record keeping (distinguishes naturalistic observation from casual (non-systematic) observation)
*Coding scheme is important - uses variety of measures, privacy of participants safe-guarded |
|
observer bias
|
may know goals of study (use blind observer)
-interpretation vs. recording |
|
Data recording methods (Observational Research)
*Field Notes |
used for direct observation
-systematic, selective, recording devices |
|
Data recording methods (Observational Research)
*Content analysis |
for textual and photographic materials
|
|
Data recording methods (Observational Research)
*Narrative record |
chronological description of all behaviors that occur in given setting- qualitative analysis
-Pros: more info, good starting pt -Cons: time consuming/demanding, not economical if you have hypothesis |
|
Data recording methods (Observational Research)
*Time Sampling (Interval recording) |
observing target behavior during selected times:
- useful for frequent behaviors - (ex) observe for 15 s, record for 15 s |
|
Data recording methods (Observational Research)
*Event Sampling |
target behavior serves as unit of analysis
- wait for behavior to begin recording, useful for infrequent behaviors -duration record: length of time behavior occurs - frequency-count record: number of times behavior occurs (short duration behaviors) |
|
Archival research
|
using existing data/records
- records could be biased - no possibility of reactivity - good external validity Cons: records might be hard to get, not generated for research (lack details, accuracy, reliability) |
|
theory
|
formalized set of concepts that organizes observations and inferences and predicts and explains phenomena
-unites seemingly disparate facts - generative, lead to empirically testable data, (ex) evolution, operant condtioning |
|
general APA principles
|
-beneficence/nonmalficence
- fidelty/responsibility - integrity - justice - respect for people's rights/dignity (confidentiality) |
|
ways of acquiring knowledge
|
tenacity (persistence)
authority (consult expert sources) intuition (gut feeling) |
|
science
|
A system of knowledge and procedures for gaining knowledge through careful and unbiased observation and analysis
|
|
rationalism
|
application of reasoning
|
|
deductive
|
general to specific:
deriving specific hypothesis from general idea |
|
inductive reasoning
|
specific to general: specific facts to general conclusion
-data leads to theory |
|
empiricism
|
gaining knowledge by observation:
-causal vs. systematic |
|
characteristics of Science
|
lawfulness of events: nature is predictable
asking empirical ?s: solvable applying control: objectivity self-correcting: replication, public verification, peer review |
|
falsifiability
|
must be able to be proven wrong for it to be testable hypothesis
|
|
pseudoscience
|
doctrine or belief system that pretends to be a science
-non falsifiable hypothesis, unwillingness to look closely at phenomenom, burden of proof lies with person making claim, failure to change/update theory |
|
heuristics
|
judgements made under uncertain conditions;
short cuts that lead to correct answer, but lead to biases in certain specifiable conditions |
|
heuristics:
availability |
likelihood estimated on basis of how easily it can be brought to mind
|
|
heuristics:
representativeness |
category memberships based on similarity (Linda)
|
|
heuristics:
anchoring and adjustment |
numerical estimates
|
|
Bar graphs
|
IV is categorical and DV is continuous
*use to make categorical comparisons |
|
Line graph
|
IV continuous displayed on x axis, (DV on Y) conveys change from point to point
*use to convey trends |
|
scatter plot
|
conveys overall impression of relationship between two continuous variables
-meaningful clusters of dots imply correlations |
|
pie chart
|
conveys proportions or percentages
|
|
error bars
|
used on bar graphs
- standard error of measure (SEM); measures how far your sample mean is likely to be from the true population mean SEM= standard deviation/square root(n) |
|
frequency distributions
|
set of mutually exclusive categories in which actual values are classified and count of values in each category
- "histogram" in graph form (y axis is frequency count #, x axis is categories) |
|
between-subjects design
|
different groups of subjects are randomly assigned to the levels of independent variable, data averaged for analysis
|
|
within-subjects design
|
single group of subjects is exposed to all levels of the independent variable, data are averaged for analysis
|
|
error variance
|
the variability among scores NOT caused by the independent variable
-common to all experimental designs, is handled differently in each design |
|
sources of error variance
|
-individual differences among subjects
-environmental conditions aren't constant across levels of IV -fluctuations in physical/mental state of an individual subject |
|
handling error variance
|
-hold extraneous variables constant and treat Ps as similarly as possible
-match subjects -increase effectiveness of IV (strong manipulation) -randomize error variance across groups (random assignment) -inferential statistics |
|
between subject designs:
*Single-Factor Randomized Groups Design |
1. the randomized two group design-- simplest to conduct, but amount of info yielded may be limited
2. the randomized multiple group design; (multiple control group design)-additional levels of IV can be added 3. parametric design; different levels of IV represent quantitative difference 4. nonparametric design: if diff levels of IV represent qualitative differeces |
|
placebo effect
|
a measurable, observable, or felt improvement in health or behavior not attributable to a medication or treatment that has been administered
|
|
matched-group designs
|
used when you know that some variable correlates with your DV
-P's matched on that variable before being studied |
|
matched-pairs design
|
the matched-groups equivalent to the randomized 2 group design to randomized two-group design
|
|
matched-multi group design
|
equivalent to the randomized multigroup design (with matching)-- difficult when more than 3-4 groups
|
|
parametric design
|
when you manipulate IV quantitatively
|
|
nonparametric design
|
when you manipulate IV qualitatively
|
|
factorial designs
|
adding a second independent variable to a single-factor design
-two components to assess: main effect of each IV & interaction b/w IVs |
|
main effect
|
-manipulation one of the IV's produces a change in the DV
the separate effect of each IV analogous to separate experiments involving those variables (factorial design) |
|
interaction
|
-the effect of one IV on the DV is dependent on the other IV
when the effect of one IV changes over levels of the 2nd (factorial design) |
|
factorial nomenclature
|
number of numbers- number of factors
value of number- number of levels |
|
higher-order factorial design
|
more than two IVs; complexity increases as factors increase
-# of main effects and interactions increase -# subjects req'd increases -volume of materials and amt. time needed to complete experiment increases |
|
mixed design
|
-includes between subjects & within-subjects factor in the same design
-allows you to evaluate effects of variables that can't be manipulated effectively w/in subjects -complex mixed designs would include >2 factors, w/ any combo of between subjects & w/in subject factors |
|
covariate
|
a correlational value (self esteem, GPA) in an experimental design
-subtract out the influence of the covariate to reduce error variance *makes design more sensitive to effects of the IV |
|
quasi-independent variable
|
correlational variable that looks like an experimental variable (gender);
|
|
Quasi-experimental design:
*time series design |
make several observations of behavior before and after introducing your IV
O1 O2 O3 treatment O4 O5 O6 |
|
Quasi-experimental design
|
including a quasi-independent variable in an experimental design
-resulting design looks like a factorial experimental design *must NOT be interpreted as causing changes in the DV because it is a pre-existing variable |
|
Quasi-experimental design:
*Interrupted Time Series Design |
make several observations before & after some naturally occurring event
(ex) car accidents before and after new 55 mph speed limit |
|
Quasi-experimental design:
*Equivalent Time Samples Design |
repeatedly introduce the treatment condition, alternated with periods of observation w/out the treatment
-treatment, O1, no treatment, O2, treatment, O3, no treatment, O4 |
|
Quasi-experimental design:
*Nonequivalent Control Group Design |
uses existing groups, includes a time series component and a control group that isn't exposed to IV
-O1O2treatmentO3O4 -O1 O2 O3 O4 |
|
Pretest-Posttest design
|
pretest administered before exposure to experimental treatment, true experimental design
- used to assess impact of some change on performance - pretest sensitization? |
|
Developmental Designs:
*The cross-sectional design |
Different cohorts assessed at the same time--thus they are of different age
Ps from different age groups are run through a study at the same time, create "cohort" groups based on age, allows collection of developmental data quickly *tells about differences, not changes |
|
Drawbacks to Cross-sectional designs
|
-Cohort effect
-may not be appropriate studies using widely ranging age groups |
|
cohort effect
|
subjects of a given age are affected by factors unique to their generation
|
|
Pros and Cons of cross-sectional design
|
Pros: inexpensive
Cons: no direct measure of change; only age differences - difficulty in establishing the equivalence of measures - are results limited to the particular time of assessment? - confounds age differences w/ cohort differences |
|
Developmental Designs:
*The Longitudinal Design |
a Single group of ps is measured several times over some period of time (months or years)
-avoids the generation effect that may hurt cross-sectional study |
|
Pros and Cons of longitudinal design
|
Pros: provides a direct measure of age changes
Cons: costly, subject loss or attrition results in non-representative samples, and non-equivalent samples across time -measures become obsolescent & questionably equivalent, -repeated testing effects; and results may be limited to cohort assessed *cross-generational problem *confounds age with time of testing |
|
cross-generational (cohort) problem
|
results from a longitudinal study on one generation may not generalize to anoter
-subject mortality -multiple observation effects (repeated IQ test) -practical difficulties |
|
Developmental Designs:
*Time lag designs |
(ex) SAT
-same ages studied at different times -cannot tell about age related changes, but can tell about cohort effects |
|
Developmental Designs:
*The Cohort-Sequential Design |
combines a cross-sectional and longitudinal component in the same design
*allows you to test for, but not eliminate, generation effects |
|
Single subject design (why it is used)
|
to evaluate macro-level effects on neighborhoods, communities, and larger systems
-also small systems like families and individuals -used more in applied vs. basic research |
|
single subject designs
|
involve repeated, systematic measurement of a DV before, during, and after the manipulation of an IV
(usually DV is characteristic of human being and IV involves application of some intervention) |
|
limitations of single subject design
|
-ethical: withholding an effective treatment
-practical issues -averaging: obscuring results -generalizability -inter-subject variability |
|
single-subject baseline design
|
1. repeated observation
2. consistent observational technique: same measurement and criteria used 3. distinction b/w intervention and non-intervention 4. decision rules: how do you determine if intervention was successful? 5. experimental control-use individuals as their own controls |
|
Pros of single-subject approach
|
-rich set of data
-score not masked by group averages -makes IDing and controlling sources of error variance easy -focus on individual behavior may reveal subtle effects of an IV lost w/ group approach -causal relationships can be established w/ few subjects |
|
Cons of single-subject approach
|
-time consuming/tedious
-limited generalizability -observer bias |
|
Classes/Phases for single-subject research
|
1. no intervention (baseline, withdrawal)
2. intervention 3. reversal |
|
baseline & withdrawal
|
baseline: if it occurs prior to any intervention
withdrawal: occurs after an intervention |
|
intervention
|
systematic behaviors aimed at assisting a client in dealing with concerns
|
|
reversal
|
rarely used phase type; occurs when an intervention technique is applied to increase a behavior it had previously ben used to decrease
|
|
No intervention phases
intervention phases |
"A"
"B" or "C" for additional factors |
|
two questions in single subject evaluation
|
1. has there been a change in area of concern?
2. is there a functional relationship b/w intervention and observed change? |
|
change in level
|
change between phases
|
|
change in slope
|
change between phases (either no accompanying change in level, or with one)
|
|
carryover
|
a delayed change; occurs later in the phase
|
|
overlap
|
scores fluctuate; some due to measurement, some to activity of extraneous variables
|
|
B (intervention only) design
|
type of single-subject design that allows for measurement of change over the course of an intervention
-no evidence of causation -no baseline -DV |
|
AB (baseline and intervention) design
|
consists of no intervention baseline phase (A) and an intervention phase (B)
|
|
ABA (Basic withdrawal) Design
|
allows for more reliable establishment of a relationship b/w intervention and outcome than in the AB design
|
|
multiple baseline designs
|
allows for evaluation across clients, situations, or problems (client systems)
|
|
issues to consider w/ single-subject design
|
choosing a stability criterion
dealing with uncontrolled variability determining generality of findings (inter-subject replication) generality requires replication- subjects & settings |
|
drifting baselines
|
baseline that doesn't stabilize but continues to show systematic variations (drift)
-may drift up or downwards -can be taken into account and effects of variables can be determined |
|
unrecoverable baselines
|
behavior cannot be returned to original baseline after at treatment (carryover effects) or partially recovered
|
|
unequal baselines between subjects
|
baselines for different subjects level off at different values
|
|
inappropriate baseline levels
|
baseline levels that are too high or low may mask effects of a treatment
|
|
smoothing
|
average set of 2 or more observations (when data is unstable)
|
|
detecting change (single-subject)
|
-change in mean level
-immediate change in level -change in trend -latency of change |