Statistics Flashcards
(31 cards)
Quasi-experiment
Research manipulates IV but subjects are not assigned randomly
Stratified random sampling
Assuring each subgroup is randomly sampled in proportion to size
Subjects Designs
- Between subjects - each subject is exposed to 1 level of IV
- Matched subjects - match subjects and assign 1 part of the pair to each group. Ensures equality of groups
- Within subjects - each subject is exposed to both levels of IV
Nonequivalent Group Design
Control group is not necessarily similar to experimental group because random assignment not used. Ex- try out a new reading method. New method for Mrs. Brown's class and placebo for Mrs. Johnson
Demand characteristics
Cues that suggest to subjects what the researcher expects of them
Hawthorne Effect
Behave differently if they know they’re being observed
Variance
Sd squared
Z score
Score-Mean/sd
T-Score
Mean is 50 and Sd is 10
Factor Analysis
Account for interrelationships found among various variables by seeing how the groups of variables “hang together”
ANOVA
More than 2 amounts
Chi-Squared
Categorical data
Norm-Referenced Testing
How an individual performs in relation to others
Domain-Referenced Testing
What the test taker knows about a specified content domain
Reliability
Consistency within the test
Test reliability
Look at correlation between…
- Test-retest: A, A
- Alternate-form: A, B
- Split-half: 1/2A, 1/2A
Standard Error Measurement (SEM)
Index of how much we expect a person’s score to vary from the score the person is capable of. Want 0.
Validity
Extent to which the test measures what it’s supposed to measure
Interval vs. Ratio
Equal intervals. Ratio has a true zero point.
Aptitude Test
Predict what someone can accomplish through training
Achievement Test
Test what someone already knows
Adaptive Test
Computerized, gives harder questions if you keep getting questions right
Ratio IQ
Mental Age/Chron Age x 100
Always will decline with age
Deviation IQ
How well a person performed on an IQ test relative to same-age peers
