Term
|
Definition
-compares two or more than two groups.
-f statistic tells you whether the variations among the means is significant, but does not tell you where the differences are (use post-hoc testing)
-if ANOVA is not statistically significant, you cannot do more testing |
|
|
Term
|
Definition
-compares only 2 groups (means from groups 1&2)
-can be independent (data from two groups) or related (data from one participant or matched pairs participants)
-can run multiple t-tests but you must modify the alpha level and make it more stringent |
|
|
Term
|
Definition
-use with two or more than two groups
-is more powerful, allows observation of complex relationships, less chance of a type I error |
|
|
Term
| Pearson Product-Moment Correlation-Parametric |
|
Definition
-measures continuous variables
-predictor variable/IV
-criterion variable/DV |
|
|
Term
| Simple Regression-Parametric |
|
Definition
| predictive value of a relationship |
|
|
Term
| Multiple Regression-Parametric |
|
Definition
-multivariate
-one criterion variable and several predictor variables |
|
|
Term
|
Definition
| -dependent variable is categorical (normal vs. impaired) and several predictor variables |
|
|
Term
| Coefficient of Determination-Parametric |
|
Definition
-assist with interpretation of correlations
-Pearson R squared give degree of association between two things |
|
|
Term
| Mann-Whitney U-Non Parametric |
|
Definition
-may be used in place of a t-test
-used when you don't have large n |
|
|
Term
| Wilcoxon Matched Pairs Rank Test-Nonparametric |
|
Definition
-2 dependent samples t-test
-takes the place of related samples t-test |
|
|
Term
| Kuskal-Wallis test-Non Parametric |
|
Definition
-one-way ANOVA
-requires independent samples |
|
|
Term
| Friedmans' Two-Way Analysis of Variance-Non Parametric |
|
Definition
| -substitute for a repeated measures ANOVA |
|
|
Term
| Spearman Rank-Order Correlation-Non Parametric |
|
Definition
-may be used in place of PPMC
-used for ordinal or nominal data
-used if there are few subjects |
|
|
Term
| Chi-Square-Non-Parametric |
|
Definition
-nonparametric alternative to independent t-test or ANOVA
-used to analyze relationships
-used when DV involves classifying individuals into distinct groups (normal vs. abnormal) |
|
|
Term
| Four Levels of Measurement |
|
Definition
1-nominal-categories of qualitative variables
2-ordinal-categories that are ranked; have no idea how much difference there is between groups
3-interval-equal intervals throughout the scale, but no 0 point
4-ratio-best type of data, equal intervals and a 0 point |
|
|
Term
| Four Factors to Control in Measurement |
|
Definition
1-test environment
2-instrument calibration
3-instructions
4-observer bias |
|
|
Term
|
Definition
| -ratio of true score variability to observed score variability (hope it is 0) |
|
|
Term
|
Definition
| reliability of an instrument (expect the same results every time you give an exam) |
|
|
Term
| How do you Establish Agreement/Reliability |
|
Definition
-PPMC: check to see if some behavior predicts another
-Point/Point agreement: looking for % agreement
-Frequency Ratios: record # of times and when a behavior occurs (2 observers) |
|
|
Term
|
Definition
|
|
Term
|
Definition
-how recruited
-how many
-sampling error
-confidence level
-level of variability |
|
|
Term
|
Definition
| -everyone has equal chance to be chosen |
|
|
Term
| Stratified Random Sampling |
|
Definition
| -when you want equal halves |
|
|
Term
|
Definition
| -when the sample is meant to represent the population of an area |
|
|
Term
|
Definition
| -pick every nth person out of a random sample |
|
|
Term
|
Definition
| -when you don't have money for a random sample |
|
|
Term
| Consecutive Sampling (non-probability) |
|
Definition
| -recruit everyone in a range |
|
|
Term
| Convenient Sampling (non-probability) |
|
Definition
| -anyone who will participate |
|
|
Term
|
Definition
| -should be described in enough detail for replication (equipment, calibration, reliability measures) |
|
|
Term
|
Definition
Must be:
justifiable
described
valid
reliable
cited |
|
|
Term
|
Definition
-uses mean, median, mode
-score of an individual described the population
-frequency distribution: amount of data in each class |
|
|
Term
|
Definition
|
|
Term
|
Definition
| Two baseline phases, treatment withdrawn & pt. returns to baseline w/out treatment |
|
|
Term
|
Definition
Modification of ABA design
hard to draw conclusions because of lack of initial baseline |
|
|
Term
|
Definition
one of the most powerful designs
allows for two examinations of treatment and is more ethical because it ends on a treatment phase |
|
|
Term
|
Definition
| similar to ABA withdrawal, but target behavior is turned toward baseline by replacing it with an incompatible behavior |
|
|
Term
|
Definition
| like ABAB Withdrawal, this design is preferred to those which end one a baseline phase. Consider that a lot of COMD behaviors are not easily reversible, though |
|
|
Term
|
Definition
based on A-B comparisons, but comparison is made across several subjects
control is achieved by treating one behavior while keeping another at baseline |
|
|
Term
| Alternating Treatments Design |
|
Definition
compare two treatments
must know generalization properties of both behaviors otherwise improvements in one cannot be attributed to the treatment |
|
|
Term
|
Definition
collected at the beginning of each session
must use same protocol for each probe
allows clinician to look at levels of ability
do not provide feedback
used in single-case design |
|
|
Term
|
Definition
during therapy
prompts & cues are given
not info about generalization
shows how client does DURING therapy |
|
|
Term
| Requirements for Baselines |
|
Definition
at least 3 data points
can't trend in the direction of treatment
spacing should mirror therapy sessions
criterion for stability
patterns=stable |
|
|
Term
|
Definition
| elements of treatment are removed one part at a time |
|
|
Term
|
Definition
| only for multiple baselines, remove one treatment while keeping others constant |
|
|
Term
| Partial-Sequential Withdrawal |
|
Definition
| combination, removal of parts of each treatment sequentially for each baseline |
|
|
Term
|
Definition
| generalization of trained behaviors to untrained stimuli |
|
|
Term
|
Definition
| generalization of trained behavior to untrained behavior |
|
|
Term
| Concurrent Stimulus-Response Generalization |
|
Definition
| both stimulus & response generalization |
|
|
Term
|
Definition
| enables a researcher to determine if the findings from their sample represent the population as a whole |
|
|
Term
|
Definition
positive (trend toward higher scores)
negative (trend toward lower scores)
outliers (random individuals on high/low end) |
|
|
Term
| Assumptions of Parametric Tests |
|
Definition
scores are random sample
sample distribution is normal
USE FOR LARGE SAMPLES |
|
|
Term
|
Definition
| examines differences between two groups |
|
|
Term
|
Definition
looking at if variables are related
degree of strength of a relationship
does NOT tell cause & effect |
|
|
Term
| Parts of a Research Paper |
|
Definition
Abstract
Introduction
Methods
Results
Discussion
References
Acknowledgement
Authorship |
|
|
Term
|
Definition
describe findings
explain modification of data
show tabled information
describe statistics and results |
|
|
Term
|
Definition
restate question
relate to previous findings
limitations
theoretical implications
clinical implications
future research implications |
|
|
Term
|
Definition
order reflects contribution
determined prior to study |
|
|
Term
|
Definition
|
|
Term
| Principles of Research Ethics |
|
Definition
respect for persons
benefits outweigh risk to subjects
benefits & burden of research fairly distributed |
|
|
Term
|
Definition
cite sources of ideas
accurate report of data
conflict of interest (funding)
confidentiality/privacy |
|
|
