Stats Flashcards
When is a MANOVA used
With one or more independent variables and two or more dependent variables
Between subjects versus Within subjects designs
And MIXED
Between-subjects (or between-groups) study design: different people test each condition, so that each person is only exposed to a single user interface.
Within-subjects (or repeated-measures) study design: the same person tests all the conditions (i.e., all the user interfaces).
MIXED study design utilizes both BETWEEN GROUPS subjects AND WITHIN GROUPS variables
SEM Standard Error of Measurement in creating Confidence Intervals
68% confidence interval is the score given + or - the SEM
95% confidence interval is the given score + or - TWICE (x 1.96) the SEM
99% confidence is the given score + or - 2 1/2 times SEM (x 2.58)
A test developer is calculating a test’s _____ when dividing the number of true positives ID by the test by the number of true positives PLUS false negatives
Sensitivity
The WAIS-IV’s Global Ability Index (GAI) is based on subtests for which of the following Indexes?
Verbal Comprehension and Perceptual Reasoning
Taylor-Russell Tables provide an estimate of the incremental validity of a test.
The greatest validity comes with a selection ratio of __ and a base rate of __.
Need BASE RATE, SELECTION RATIO, and VALIDITY COEFFICIENT
What happens to the SEM when sample size increases?
- SEM decreases as sample size increases.
Test Unfairness
Occurs when for 2 groups (ie men /women):
Criterion-related Validity is SAME
Criterion scores are SAME/SIMILAR, but
Predictor Scores are significantly DIFFERENT
Rancho Los Amigos Level of Cognitive Functioning Scale
A patient who receives a rating of Level V (Confused – Inappropriate Response) is confused but non-agitated, responds to simple commands, gives random or fragmented responses to complex stimuli, has impaired memory and selective attention, and is unable to retain new information.
Co-efficient of stability
Measures reliability of test across 2 occasions with separate examinees
Co-efficiency of Internal Consistency
Measures the reliability of a factor that fluctuates in intensity or severity over time
Test Bias: What does Slope Bias describe?
When a predictor’s validity co-efficient differs for different groups
Wonderlic Personelle Test - R evaluates what?
A brief test of mental ability for adults.
Solomon 4 Group Design
Evaluates the effects of pre-testing
SAT Reasoning Test’s purpose
Predict the college success of high school seniors
What type of test is the EPPP
A measure of basic psychology knowledge
Multiple Regression Analyses advantages over ANOVA
- Eliminates need for 2-stage analysis involving Global Significance Test followed by “fine grain” significance tests
- Not limited to categorical variables
- Enables researchers to determine if adding additional independent variables affects dependent variable beyond the effects found by the initial independent variables
Rational Economic model of decision making views decision makers as attempting to make
Optimal decisions
3 Components of a Needs Assessment
Organization Analysis, Task Analysis, and Person Analysis
Example of a point-biserial correlation
pregnant or not with multiple item stress test
What are you finding out about a predictor when you divide the number of (true positives) by the number of (true positives plus false negatives)
A predictor’s SENSITIVITY
A psychologist that is testing people would likely find that in comparison with people with a low need for achievement, people with a high need for achievement would prefer which level of task difficulty?
MODERATE difficulty
Spearman- Brown formula is used for what?
Reliability coeffecient- To estimate the effects of increasing or decreasing the length of a test on it’s reliability coefficient RANKS
What kind of distribution has a mean that is less than the median and a median that is less than the mode?
Negatively Skewed
Don’t forget left (lower) and right (higher)
Canonical Correlation
IE: when using scores on 8 subscales of a test to predict the scores on 5 subscales of another test
Why would you chose a MANOVA over 2 ANOVAs?
Control for experimentwise error rate; see which factor(s) are truly important; may reveal differences that aren’t apparent when doing them separately.
Moderator vs Mediator variable
A moderator variable effects the direction or strength of the relationship
A mediator variable explains the relationship - HOW or WHY
Understand confidence intervals and relationship to SE of measurement
68% confidence interval would be the raw score plus and minus (+ and -) the SE of the measurement
In an ANOVA that is comparing the effect of 4 individual variables on 1 dependent variable. What is a significant F value?
When it is greater than 1.0. OR the when the numerator is greater in a ratio of mean square between (numerator) over mean square within (denom)
High scores on scales 4 and 9 on the MMPI would likely look like what type of behavior?
Antisocial behavior and impulsivity
4= psychopathic deviant; 9=hypomania
The Theory of Work adjustment indicates that what is the biggest predictor of a worker being forced out of a position is what?
“Satisfactoriness”
When will there be more false positives than false negatives?
When a populations base rate is low, and the accuracy of a test is high (over 50%; ie:95%). Because it will be more likely to misdiagnose those NOT having the condition than misdiagnosing those having it that do not.
What does G. Herek (1992) attribute to violence against LGBTQ folks?
Heterosexism
What are the Raven Colored Progressive Matrices?
A item Intelligence test for 5 -11 year olds; older adults (65+), or those with moderate to severe learning difficulties
What does Item Response Theory reflect?
Status of latent trait or ability- good for Criterion-Keyed. Tailored, and Diagnostic/Employment screening; but NOT for TEACHER-MADE tests (too big of sample needed?)
ANCOVA
Statistically remove error variance from an extraneous variable
What approach was used to define the Big 5 Personality Traits?
Lexical - WORDS- based on definitions in the dictionary originally; personality comes through via language
A discrimination score of “0” means what on an academic achievement test?
That there was no difference in answering between low achieving students and high achieving students
Which has the lowest reliability out of a 3-point, 7-point, True False, or Free Recall test?
True False
Confidence Intervals need what numerical component?
Standard Error of Measurement
What is a Wide Range Achievement test?
It is a brief achievement test that measures reading comprehension, spelling, and arithmatic
Using an ANOVA: when?
When there are 2 or more independent variables
What is Structural Behavior Modeling used for?
Evaluate the predictive influences of multiple latent factors
Leptokurtik means what about a distribution?
A higher than normal peak
Difference between parametric and not parametric tests?
Important to know if sample not random
Research Design 1: Counterbalanced
A WITHIN SUBJECTS design; Used to control “carryover effects”; using different levels of the independent variable with different participants or groups, in a different order.
Research Design 2: Randomized Block Design
Used to control the effects of a confounding variable by making sure the individuals or groups are equal on that variable blocked into groups) before then randomly assigning individuals to treatment groups
Research Design 3: Solomon 4 Group Design
Special factorial design to evaluate the effects of pretesting on post-test scores. Int/Ext VALIDITY
Design 4: Non-equivalent Groups Design
A quasi-experimental research design in which participants CANNOT BE RANDOMLY ASSIGNED to different treatment groups.
If a tests reliability coefficient is zero, what else do we know?
That the Standard Error of Measurement for the test is equal to the test’s standard deviation
Types of Test Construction
- Factor Analytic method (correlations with some items and less with others)
- Theoretical method (reflect traits identified by experts as reflective of a specific personality theory
- Criterion (empirical criterion keying- items that have been shown to distinguish between people who have the trait and people who don’t)
BARS evaluation
- Behaviorally Anchored Rating Scales
- A criticism is that it requires raters to guess at what kinds of behaviors they would expect rather than actual observed behaviors
What does research say about effects when comparing day, swing, and night shifts
Night shift has the most with detrimental effects to health, and may or may not be similar to swing shift with regard to impact on social lives
What does research say about Single-session Psychological Debriefing?
It does not prevent PTSD but may actually exacerbate symptoms
Interrupted Time Series research design
Using a SINGLE Group DV measured at regular intervals before and after treatment
What type of scoring would be most useful when scores will determine how much participants learned during a 2-week workshop
Percentage scores
% of each domain mastered
Compliance
When a person changes their behavior due to social influence in order to obtain reinforcement or avoid punishment
ICC- What is an Item Characteristic Curve?
The relationship between a latent ability and the performance on a test item
Fleider’s Contingency Model of Leadership (LPC)
A low LPC focuses on completing tasks
What scale on the MMPI identifies if someone may be malingering, trying to look bad, or is being careless when answering
F Scale
MANCOVA
To control for a confounding correlating variable, multiple dependent variables
Meta analysis utilizes an “effect size” to show what?
Mean differences between experimental and control groups
Moderator vs Mediator
Moderator- effects the STRENGTH of the relationship
Mediator- are RESPONSIBLE for
What is a NONLINEAR TRANSFORMATION?
When score transformations alters the rank order and relative size of the distance between scores
What will raising a cutoff score due to T/F Positive and Negatives?
Raising the cutoff score will decrease T and F positives and increase T and F negatives
Stepwise Multiple Regression would be used for what?
To reduce the number of predictors needed to account for criterion variability.
What part of the Rorschach test is useful for evaluating intelligence?
Organizational Activity
With regard to Holland’s Self-Directed Search, what is the frequency with which a respondent’s CODE occurred in the standardized population is called what?
Commonness
KUDER - RICHARDSON 20 test
A high correlation (>.8 or .9) is associated with a homogenous test
Latin Square is what kind of statistic??
Administering each level of the independent variable to each participants so that it is given the same number of times to each, IN EACH POSITION
What test is based on the PASS model of intelligence?
A measure of basic cognitive processes; The Cognitive Assessment System is based on the PASS model (Planning, Attention, Simultaneous processing and Sequential processing.
If a researcher changes is alpha level from .01 to .001, what is likely the result with regard to correctly or incorrectly rejecting or retaining a true or false null hypothesis?
Less likely to incorrectly reject a true null hypothesis
If a researcher gets a .85 correlation for participants raw scores and predicted scores, what correlation co-efficient will he get if he converts those scores to non-normalized t-scores?
Also .85
What area of scoring would you expect an child with ASD without language impairment to obtain the lowest scores on a WISC-V?
Coding and Symbol Search (processing speed)
Student’s t-test
Compare the means of scores obtained by 2 groups
2 Way ANOVA
Two independent variables
Chi Square
Analyzes frequency (nominal) data t
What information does the slope of the Item Characteristic Curve tell us?
Discrimination (between low and high achievers), Item difficulty, and probability of getting an item correct by guessing!
A test developer is wanting to evaluate the effects of a training program on sales success of newbies; which type of test validity would be important?
Criterion-related
If designing achievement test where reliability was most important, what type of test design would you be LEAST likely to choose?
T/F
NOT M/C, essay, FIB
When an assumption in an ANOVA is violated, what is the best strategy to use to analyze the data
Krusty-Wallis LOL
Kruskal
When conducting an Item Analysis, one item has a p-value of .85. What does this mean?
That the item is easy.
P-Values with Item Analysis
Lower = more difficult
Higher + easier (more people answered it correctly)
How to discriminate using an Item Analysis?
Calculating D statistic; this is the number of people who got it wrong from the number of people that got it right
What is Bonferroni for?
- Control experiment-wise error rate
- Reduce significance level for each comparison
When factors are OBLIQUE, what does that mean?
That correlation coefficient for any two factors is greater than 0.
OBLIQUE vs ORTHOGONAL
Non-correlated vs Correlated
When using a split-half method for reliability, what should you remember?
Splitting using even or odd numbers will likely result in a reliability co-efficient that over-estimates the reliability
What statistic (correlation coefficient) would you use when correlating associations between 2 variables that have been artificially dichotomized?
Tetrachoric. Used for comparing 2 artificially dichotomized normal distributions.
Eta correlation coefficient
Used for continuous, non-linear variables
When would you not want to use autocorrelation?
With a repeated measures design
What does “Test the Limits” mean in testing, and when would you do this?
allows extended time or attempts in an effort to evaluate maximum abilities in the individual completing the test. Used after administering the test using standard procedures.
Shrinkage
Associated with cross-validation
When KNOWLEDGE of a person’s score on one test effects how that person is rated on a subsequent test, what is this called?
Criterion Contamination
Construct validity
High correlation w/test representing same attributes and low correlation w/test representing opposite attributes
A low selection ratio and moderate base rate are useful for what?
Maximizing incremental validity
