Research Flashcards
IV
independent variable or experimental variable
DV
dependent variable
what happens with the IV in an experiment
the IV gets manipulated by the researcher
what happens with the IV in an experiment
Manipulated by the researcher
Examples of an independent variable
counseling, exercise, mindfulness
Examples of an dependent variable
weight, IQ, drinks, time, money spent,
quasi-experimental
research that does not use random sampling
systematic sampling/kth sampling
1st participant is randomly sampled, then the next is picked every 10th person
non-probablity sample
subjects are selected by the researcher
convenience sampling
the intact existing group is used with no random sampling
quota sample
your sample has the same type of characteristics that existing in population being studied
hypothesis
hunch/idea
RA fisher
father of statistics and experimental design
each experiment has what two hypothesis
null hypothesis, experimental hypothesis
null hypothesis
no difference/significance between the control and experimental groups
test of significance
a statistical test that assesses whether a result obtained from an experiment is important enough or not
level of sigificance
confidence level of research
probability of difference between groups alpha level
social sciences significance
p or probabilty is less than .05
if p is <.05
probability that differences are less than 5% chance
the lower the p (.01 or .001) the?
the chance factors or the more convincing the experiment
greater statistical power, the more?
confidence in the validity of the research
two types of resarch errors
Type I and Type II
Type I error
reject null when true
false positive
Type II Error
accept null when false
false negative
the probability of making a type I error is
equal to the level of significance
If your p value is .05 then?
.05 is the probablity of making a type I error
If your p value is less than .05
accept null hypothesis
extraneous variables are
errors
internal validity
makes the conclusions of a causal relationship credible and trustworthy
High internal validity demonstrates causal link
external validity
can the expreimental findings be generalized to other people/groups
instrumentation threat
threat to validity in the instrumentation or measurment methods
maturation
threat to validity effect of time rather than IV on text subject results
statistical regression
low scores move up and high scores move down toward the mean
external threat to validity
findings will not generalize to real world
t-test/student’s t-test
tests a hypothesis between two normally distrubuted samples and must have 30 participants
correlated t-test
same group measured on 2 occasions (pre/post test)
What does an Anova do
compares more than 2 groups / analysis of variance
results of Anova is what
F value
more than one DV requires
MANOVA
what is an ANCOVA
analysis of covariance/an adjustment to the groups
hawthorn effect/reactive effect
an individual being observed does better when being observed
correlational research asks
does a relationship between an IV and DV exist
Pearson product-moment correlation/correlation coefficient
correlation of choice in a counseling study
range of a correlation coefficient
-1.00 to 0 to +1.00
correlation coefficient
a statistical relationship between two variables
what is a perfect correlation
-1 or +1
what is a negative or inverse correlation
one variable goes up when another goes down
what is a zero correlation
0.00
no relationship
gaussian curve is also a
bell shaped curve and is normally distributed
measures of central tendancy
mean, median, mode
mean
average or most useful measure
median
middle or 50th percentile
mode
most often occuring amount/top or high point of graph
in a normal distrbution the mean, median and mode are what
the same value
a curve with 2 points
bimodal curve
multimodal curve
two modes
distrbution tail to right
positively skewed
distribution tail to left
negatively skewed
y-axis
ordinate - goes up and down like the letter y
x-axis
absissa
histogram
bar graph
range
measure of variablity or difference between the highest and lowest
SD or Standard deviation is
the square root of the variance
95%
number of scores/cases that fall between 2 +/- standard diviations
99.70%
number of scores/cases that fall between 3 +/- standard diviations
z-score
standard deviation
t-score
mean is 50 and score is 10 points above or below the mean
stanines
standard 9 scores and divide the distrbution into 9 equal intervals hwere the mean is 5 with a standard diviation of 2
descriptive statistics include:
range,
variance,
standard deviation
not experimental
nominal scale
Independent of each other - identify and classify, are qualitiative - eye color, blood type
ordinal scale
describe variables that are rank ordered - high med low
interval scale
describe numbers that are scaled at equal distances - Fahrenheit
ratio scale
each number is measured from zero - height, weight
survey
questionnaire to a sample population
ethnographic research
holistic and inductive , overall dynamics
inductive reasoning
genernalize based on specific observations
deductive reasoning
general principles inform a hypothesis
halo effect
Thorndyke - rate on one characteristic but really influenced by another
horn effect
a attribute you find negative will influcence your decision
rosenthal effect
experiement expectations might influence change
double blind
both the researcher and the groups do not know who is in what group
norms
normal, typical average person who to is being studied
N
number of subjects in a study
N=1
single subject design - case study
AB design
A = baseline meansurement, B=apply intervention then see if something changed
ABC design
two treatment interventions
ABA design
when a treatment returns to baseline measurement
ABAB design
Baseline, treatment, baseline, treatment, ending on a treatment phase
counterbalancing
the way stimulus is presented can bias a study. Counterbalancing changes the way interventions are presented to groups
percentile rank
percentage is different than percentile: percentage is scoring 50% of test questions, pecentile is scoring better than 50% of test takers
raw score
unaltered scores
metaanaylsis
combine multiple research studies
parametric
interval/ratio
Fall on a continuum
t-test, anova, chi-square test
nonparametric
nominal or ordinal
do not lie on a continuum
Mann Whitney U test, Kruskal Wallis test, Wilcoxon’s signed ranks test
Empirical rule in a normal distribution:
All scores will fall within 3 standard deviations from the mean as 68% at 1 SD, 95% at 2 SD, and 99% 3 SD
standard error of measurement
how much measured test scores are spread around a “true” score
directly related to a test’s reliability
the smaller the sample size
the greater increase of error or decrease in reliability
interrater reliability
the extent to which independent evaluators produce similar ratings in judging the same abilities or characteristics in the same target person or object
correlation coefficient
measures the linear correlation between two sets of data
Pearson product-moment correlation coefficient
parametric test to measure the linear relationship between two normally distributed variables
Random error
Chance errors
mainly affects precision
systematic error
Consistent or proportional difference between observable and true values
affects the accuracy of a measurement
types of systematic error
Response bias
social desirability bias
Experimenter drift
Kurtosis
Stratified Random Sampling
Helps you pick a sample that reflects the groups in your participant population(e.g., matching census percentages)
Cluster Sampling
Selecting subgroups without randomly selecting individuals within those groups. E.g., randomly selecting classrooms
Purposeful Sampling
Selecting people based on who is likely most knowledgeable about the topic, or because they represent needed characteristics. Example: selecting CEOs for a qualitative study about the impact of power on personality
directional hypothesis
does the IV increase the DV
non-directional hypothesis
How does the IV affect the DV
does it increase or decrease?
t-test
test between 2 groups
Chi square ( χ 2 ) tests
used when t-tests cannot be performed because the data do
not resemble a normal distribution
Mann Whitney U test
non-parametric test between 2 groups
t-test equivariant
Kolmogorov Smirnov Z procedure
test between 2 groups for less than 25 participants
Used instead of the Mann-Whitney
Kruskal Wallis test
Used instead of the Mann Whitney when there are more than two IV groups
Equivalent to anova
Wilcoxon’s signed ranks test
Paired group tests
Nonparametric equivalent to dependent t-test for one group only (pre and post-test)
Friedman’s rank test:
Used instead of the Wilcoxon for paired group tests of more than two groups
Analysis of
variance (ANOVA)
used instead of a t-test for multiple groups
Analysis of
covariance (ANCOVA)
Testing two groups and controlling for a possible confounding variable
Multiple analysis of covariance
(MANCOVA)
Similar to ANCOVA but with multiple dependent variables
Between group t-tests =
“independent,” “unpaired”
(e.g., CBT vs. placebo)
Within group t-tests =
“dependent,” “paired”
(e.g., pre and post-test)
Coefficient of Determination
To determine the amount of shared variance between two variables
(i.e., effect size), we square
The Factorial ANOVA statistical test is used to determine:
if two or more sets of groups are significantly different from each other on your variable of interest. Your variable of interest should be continuous, be normally distributed, and have a similar spread across your groups. In addition, you should have enough data (more than 5 values in each group).
Dependent t-test
paired t-test
Compare means of 2 groups
Spearman’s rank correlation coefficient
nonparametric test that measures the strength and direction of association between two ranked variables.
equivalent to Pearson’s coefficient
