Search
Create
Log in
Sign up
Log in
Sign up
PSYC 3006 exam 1
STUDY
Flashcards
Learn
Write
Spell
Test
PLAY
Match
Gravity
Terms in this set (91)
statistical validity
accuracy of the p-value on which a statistical decision is made
construct validity
degree to which the theory/theories behind the research study provide the best explanation for observed results
external validity
extent to which the results of a particular study generalize to other people, places, conditions
generalization
process of inferring something about a population based on sample findings
ecological validity
appropriate generalization from lab to real life situations
internal validity
extent to which we're confident the IV changed the DV
confounding variables
threaten internal validity (i.e. maturation, history, testing, instrumentation, regression to the mean, selection, attrition, diffusion of treatment, sequence effects)
subject effects
changes in participant's behavior due to being in the study
validity
extent to which procedure measures what it's supposed to
control
procedure used to counteract potential threats to the validity of the research
exact replication
repeating the study exactly
systematic replication
testing a theoretical/procedural modification of the original procedure that will produce desired results only if the original findings were accurate
conceptual replication
generating and testing difference research hypotheses from the same problem statement
single-blind procedure
experimenter/assistant is blind to conditions
double-blind procedure
both researcher and participant are blind to assignment of conditions
automation
using technology to test
objective measures
based on empirically observable data/events which 2 can easily agree on
percent agreement
agreement between multiple observers (# agree/# observations x 100)
kappa
indicates true/chance agreement
participant selection
when well done, increases external validity
general population
group of all organisms, events, things we're interested in (i.e. all children)
target population
subset of population in which research is primarily interested in, population to which we want to generalize findings (i.e. all elementary school kids)
accessible population
subset of target population available to research (i.e. local elementary school kids)
representative sample
adequately reflects population characteristics
sample
subset of accessible population on which measures are taken
random sampling
means every population member has an equal chance of being selected, selections don't affect each other
stratified random sampling
draw separate random samples from each of several subpopulations
ad hoc sampling
drawing participants randomly from an accessible population
principle of initial equivalence
holds that all groups must be statistically equal at the start of experiment
free random assignment
involves assigning participants to conditions so one doesn't effect the other
matched random assignment
good when working with small number of participants, use potential confounding variables for matching, the similar pairs are separated
maturation
DV changes occur solely due to participants getting older/more experienced
history
DV changes due to variation in events outside study's IV effects
testing
DV changes due to more practice with the testing
attrition
participants drop out differentially across groups
sequence effects
in repeated measure/within subjects design, order of conditions may cause DV changes
hawthorne effect
changes in behavior due to getting special attention
demand characteristics
experimentation leads to participant behaving differently
5 ways of subject assignment
1. random assignment
2. matched random assignment
3. elimination (removing persons beyond a certain level of a variable)
4. constancy (allowing persons in a study having only a specific level of a variable)
5. build a variable into study's design and examine the effect of levels on DV
true experiment's 5 characteristics
1. RH stating IV's effects on DV
2. 2 or more IV levels
3. assign participants in an unbiased way
4. strong procedures for testing causal relationship
5. uses specific controls to lower threats to internal validity
standard error
average amount off we expect a value to differ from the distribution of means
95% confidence interval
interval within which theres a 95% chance of finding true population value
comparison distribution
distribution to which you compare your sample when NH is true
central limit theorem
as number of samples increases, the mean of the distribution of means approximates the population mean
underestimate
variance of sample will usually _____ population variance
assumptions of independent means t tests
population distributions are normally distributed, variances are equal across populations
independent means t-test by regression
answers question - does the regression coefficient differ from zero or are we any better off in predicting the scores from their group means than if we'd used only the grand mean?
example of reporting
"contrary to our hypothesis, the mean cognitive test scores for the 8 year olds (M=, SD=) was not statistically different from the mean for the 4 year olds (M=, SD=), t(6)=.06, one-tailed test"
single sample t-test
compare single mean to population mean when population standard deviation is unknown
dependent means t-test
compare 2 linked means when population SD is not known
independent means t-test
compare 2 independently-sampled means when population SD is not known
independent means t-test by regression
compares 2 ind-sampled means when pop SD is not known, go about comparison by computing slope (B) of a regression line between 2 means, and determining its statistical difference from zero
statistical power
probability that the study will yield a statistically significant result if the RH is true aka indicates confidence in correctly rejecting the null, like to see at least 80%
cohen's d effect sizes
small (0.2-0.5), medium (0.5-0.8), large (0.8 or more)
larger effect size more power
how effect size effects power
larger sample size more power
how sample size effects power
between-groups variance
variability among group means
experimental variance
shows effect of the IV, want this to be high
extraneous variance
affect of uncontrolled extraneous variables on the study results
systematic between-groups variance
sum of extraneous and experimental variance, necessary for determining causal differences
nonsystematic within groups variance
due to random factors, individual differences
error variance
due to factors that affect some participants but not others within a group, due to chance factors and individual differences
manipulation check
assessment of the effectiveness of the experimental manipulation, evaluates if the manipulation really had its intended effects
procedures for controlling extraneous variance
must make sure groups are equal at start, random assignment, select participants similar on certain variable to control that variable, build potential confounding variable into design, matching participants, treat groups same except IV manipulation, blind, control procedures
ex post facto studies
'after the fact' researcher observes current behavior and attempts to relate it to earlier experiences (unreasonable to infer causal relationships)
single group, post-test only studies
researcher manipulates IV with a single group which is then measured (doesn't control for confounding variables of expectation effects, history, maturation, regression to mean)
single group, pre-test, post-test studies
same as single group, post test only but has a pre-test
time-series design
taking multiple measures over course of longitudinal study
pretest-posttest, natural control-group studies
naturally occurring groups are used, one gets treatment, includes pre and post test
experimental design
control groups and randomization mean
randomized, posttest only control-group design
includes random selection and assignment, compares posttest measures of experimental vs control group
groups are statistically equal
random assignment ensures that
randomized, pretest posttest control-group design
all participants tested on DV, experimental group gets treatment, both are re-tested
multilevel, completely randomized, between subjects design
participants are randomly assigned to 3 or more conditions, may/may not include pretests
solomon's 4 group design
combines randomized, pretest-posttest, control-group design and posttest only control group design
unbiased assignment
achieved through free random or matched random assignment
independent-groups design
different participants in each group
correlated-groups design
same/closely matched participants in each group
single-variable designs (univariate)
only one IV
multivariable (factorial) designs
2 or more IVs
within-groups variance
measure of nonsystematic variation within a group
between groups variance
represents how different the group means are
SS (sum of squares)
sum of squared deviations from the mean
probing
statistically testing the differences between group means
planned comparisons/a priori/contrasts
experimenter predicts which groups will differ and in what direction
post hoc comparison/a post erior/incidental
evaluate pattern of means after
error bars
graphic depiction of variability for that particular measure
standard error of the mean
SD of the distribution we'd get if we took every possible sample of a given size from a population and graphed the means of those samples
between groups variance/within groups variance
F ratio
extraneous variance
due to confounds, want this to be low
cohens d
indicate how much of a standard deviation we expect (or found) as a difference between means
YOU MIGHT ALSO LIKE...
Academic Word Lists - AWL Sublists
giflingua
$12.99
STUDY GUIDE
PSY282 chapter 15
122 Terms
rooster1964
Research Methods Exam 3 Terms
74 Terms
Jacey_Levesque6
PSY 215 Exam 2 UK
91 Terms
deve3
OTHER SETS BY THIS CREATOR
Black Fire Exam 2
38 Terms
natobat
Black Fire Exam 1 2018
25 Terms
natobat
abnormal exam 4
69 Terms
natobat
abnormal exam 3
108 Terms
natobat
;