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;
62 Cards in this Set
- Front
- Back
Pearson r is used to:
|
Assess the relationship between two variables that are measured with continuous data(age, exam score)
|
|
Partial correlation describes the:
|
Index of the relationship with the influence of a third variable removed.
|
|
ANOVA is used to:
|
Test hypotheses based on population values.
|
|
A T-test is used to:
|
Analyze data from a study that has one independent variable (with only two levels) and one dependent variable.
|
|
In a negatively skewed distribution:
|
Most scores are very high, the mean is lower than the median, and the median is lower than the mode.
|
|
Counterbalancing is used in studies to:
|
Control for order effects by presenting multiple treatments to subjects in varying orders for different subgroups.
|
|
Maturation refers to:
|
any internal change (biological or psychological) that occurs in subjects while an experiment is in progress - and has a systematic effect on the DV - MAY be a threat to internal validity
|
|
Instrumentation:
|
Can be a threat to internal validity, when the measuring process changes between the pre and post tests (e.g. raters become better at rating with practice)
|
|
Path analysis is useful for:
|
As a causal modeling technique, it can be used for examining unidirectional causal relationships among a set of measured traits
|
|
To use trend analysis, you need:
|
a quantitative independent variable
|
|
Autocorrelation refers to:
|
Observations obtained close together in time from the same subjects tend to be highly correlated
|
|
Best control for practice effects in an experiment:
|
Counterbalancing (
|
|
Difference between external & internal validity:
|
External: concerned with generalizing study to real world
Internal: plausability regarding causal relationships between study variables |
|
Eigenvalue
|
Stat that indicates the degree to which a particular factor (in a factor analysis) is accounting for varaiability (so it really tells us the explanatory power of the factor)
|
|
Standard score
|
Type of norm-referenced score that is interpreted in terms of how many SD units a score falls above or below the mean (such as z-scores & T-scores)
|
|
Multiple baseline design
|
Single-subject design in which a treatment is applied sequentailly across settings, behaviors, or subjects (baseline data is first collected in each setting)
|
|
Advantage of MANOVA over multiple ANOVA's is:
|
reduces the probability that a Type I error will be made
|
|
The standard errors of the mean, measurment, & estimate express error in terms of:
|
Standard deviation
|
|
Construct Validity
|
How weel does the test measure the theoretical construc (e.g. intelligence) that is attempting to measure
|
|
Range of the standard error of measurement:
|
0 to the standard deviation of the test
|
|
Range of the validity coefficient
|
-1 to +1
|
|
Range of reliability coefficient
|
0 to +1
|
|
In terms of reliability & validity, increasing test length will:
|
affect reliability more than validity
|
|
Rejecting a false null hypothesis means:
|
Your study has power
|
|
If you go from a one-tailed test to a two-tailed test, you are..
|
Lees likely to correctly reject the null (as you are losing power)
|
|
History refers to:
|
A threat to internal validity when any EXTERNAL event that affects scores on theDV (besides the experimental treatment)
|
|
Statistical Regression:
|
Tendency for extreme scores to revert to the average upon re-testing (batting average)
Can be a threat to internal validity |
|
Differential Mortality
|
Threat to internal validity, when people who drop out of one of the groups, differ in significant ways from people who remain in the study
|
|
Controlling threats to internal validity:
|
1)Random assignment - most powerful method
2)Matching - identifying subjects who are similar & pairing off (bestwhen sampl size is small) 3) Blocking - analyzing the statistical effects of an extraeous variable 4)Holding constant - only including subjects who are homogenous 5)ANCOVA - statisical strategy for removing the effects of extraneous variables |
|
THREATS to external validity
|
1)Int. btw Selection & Tx -
2)Int. btw. Hx & Tx -can't generalize to other settings time/period 3)Int. btw. Testing & Tx - can't geberalize unless had a pretest 4)Demand Chracteristics - cues in setting allowed subjects to guess hypothesis 5)Hawthorne Effect 6)Order Effects - problem with repeated measures design |
|
Ways to increase External Validity
|
1)Random Selection-ensuress equal representativeness
2)Naturalistic Research - controls for Hawthorne & demand characteristics 3)Single/Double Blind Research-subjetcs/researchers don't know which treatment has been assigned 4)Counterbalancing - controls for order effects |
|
True vs. Quasi Experimental Research
|
True = investigator randomly assigns subjects to different groups
Quasi = random assignment is not possible |
|
Developmental Research
|
1)Longitudial Design = same people studied over time
2)Cross-Sectional = different groups, divided by age are studied at the same time 3)Cross-Sequential = representative samples of different ages groups are assessed on two or more occasions |
|
Single-Subject Designs
|
1)AB Design - single baseline foloowed by single tx
2)Reversal (Withdrawal) Design - baseline, treatment, return to baseline, (ABA, or ABAB) 3)Multiple Baseline - IV is sequentially administered across 2 or more subjects, behaviors, or settings |
|
Z-Score
|
A measure of how many SD's a given score is from the mean
|
|
T-Score
|
based on 10 point intervals, where T=50, as the mean, and every 10 points above or below is one SD
|
|
Stanine Scores
|
distribution of 9 equal intervals 1-9, mean = 5 and SD = 2
|
|
Standard Error of the Mean
|
expected error of a given sample mean. It is obtained by dividing the SD by the square root of N
|
|
Type I Error:
|
conclude there is a difference, when in fact there is not
|
|
Type II Error
|
concluding there is no difference, when in fact there is
|
|
Factors that affect Power
|
Sample size = increases
Alpha - increases One-tailed - increases Magnitude of Population Difference - increases |
|
T-Test
|
(Parametric test used to test hypotheses about two means)
One Sample = comparing one sample to a population mean Independent Samples = comparing 2 means that are unrelated Correlated Samples = compares means that are matched or related to each other |
|
Post Hoc Tests
|
1)Schaffe = most conservative (least Type 1 error)
2)ukey is best for pairwise comparisons |
|
ANOVA
|
One-way = 1 independed variable & more than 2 independent groups
Factorial (2-way) = used when there are more than one independent variable |
|
MANOVA
|
used when a study has 2 or more dpendent variables and one or more independent variable
|
|
Mann-Whitney
|
non-parametric test used forrank-ordered data when there are more than 2 independent groups
|
|
Factors Affecting the Pearson r
|
1)Linearity - assumes the reationship is linear
2)Homoscedasticity - assumes dispersion of scores is equal 3)Range of Scores - assumes a wide range of scores |
|
Misc. Correlation Coefficients
|
Point-Biserial =relates 1 continous variable & 1 dichotomous
Contingency = correlation btw. 2 nominal variables, that each have at least 3 categories |
|
Spearman's Rho
|
Used to correlate 2 variables that are ordinally ranked
|
|
Eta
|
used to measure continous data when the relationship is non-linear
|
|
Canonical Correlation
|
used to relate 2 or more predictors to 2 or more criterion variables
|
|
Discriminant Function Analysis
|
scores on 2 or more variables are combined to determine whether they can be used to predict which criterion group a person belongs in (i.e. high or low achievement group)
|
|
Suppressor Variable
|
an extraneos variabel that results in a spuriously LOWER correlation (it suppresses the relationship)
|
|
Trend Analysis
|
statitical technique for looking at the pattern of change between two quantitative variables (include linear, quadratic, cubic, & quartic)
|
|
Rosenthal Effect
|
when the experimenter's expectancies influence subject's responses on a dependent variable in the direction predicted
|
|
Solomon 4 group design is used to
|
Evaluate the effects of testing
|
|
ANCOVA is used to
|
reduce error variance due to an extraneous variable
|
|
In regards to SD & Validity, the Standard Error of Estimate has:
|
a direct relationship with SD of the criterion & an indirect relationship with validity
|
|
Item Response Theory is based on the premis that:
|
when test content is of varying difficulty, uniform scales of measurement can be applied to persons of different ability level
|
|
In weighting variables in a multiple regression equation, we use:
|
line of best fit
|
|
Cluster analysis is used for the purpose of:
|
deriving several subgroups from a cluster of dependent variables
|
|
Advantage of ANOVA over multiple T-Tests is:
|
Reduces Type I error
|