EPPP--Statistics and Research Design
Terms
undefined, object
copy deck
- Formula for the standard error of the mean
-
Standard Deviation
----------------------
square root of sample size - To increase the standard error of the mean, what do you need to do to the sample size and standard deviation?
- Decrease the number of individuals in the sample size and increase the standard deviation.
- Define Selection
- Pre-existing subject differences that threaten the internal validity of a research study...these could account for differences between the groups on the dependent variable.
- What is Path Analysis
- A way to test assumptions about the causal relationships among multiple variables without actually manipulating variables and doing a true experiment.
- F Ratio
-
MSB--mean square between (between group variance)
-----------------------------
MSW--within-group variance - What is a factorial design?
- Several independent variables are assessed at the same time. It allows for the interaction of several variables to be assessed.
- What does LISAREL mean?
- Linear Structural Relationships. A structural equation modeling technique--desiged to assess the degree to which a causal model involviing multiple variables is supported by observed data.
- What does LISREL involve?
- Latent and observed variables. Unidirectional and bidirectional causal relationships.
- Regression to the mean
- The tendency of extreme values or characteristics to be less extreme.
- Pearson r
- The relationship between X and Y variables can be depicted on a scatterplot as a straight line.
- Name several threats to internal validity
-
Maturation
Testing
Instrumentation
Selection
Statistical Regression - Maturation
- Any internal change that occurs in the subjects while the experiement is in progress. I.e., getting hungry, tired, etc.
- History
- Any external event that affects scores or status on the dependent variable.
- Instrumentation
- A change in DV may be observed from pretest to post-test because the nature of the measuring instrument has changed. I.e., the scale broke.
-
Rosenthal Effect
(Pygmalion Effect) - Behavior of subjects changes as a result of experimenter expectancies, i.e., teacher expectations.
- Hawthorne Effect
- Tendency of subjects to behave differently due to the mere fact that they are participating in research, i.e, changing the work environment.
- Cross-Sequential Study
- Involves studying groups of subjects that are divided on the basis of age. Involves studying them for a period of time. Controls for cohort effects, less costly than longitudinal studies.
- Multiple Baseline Study
- A single-subject study that involves the sequential application of a treatment across different baselines. Does not entail withdrawal of treatment.
- One-group time-series
- Administering multiple pretests and post-tests to one group of subjects before and after a treatment is administered. Controls for maturation, testing, and regression.
- What is power?
- The probability of rejecting a false null hypothesis, or detecting a true effect of the independent variable.
- How would you increase power?
- Increasing reliability, or decreasing error variance reduceds the power that you will "miss" a true effect.
- Central Limit Theorem
- Predicts that as sample size increases, the shape of the sampling distribution of means approaches normality, regardless of the shape of the population distribution.
- Scheffe Test
- Post-hoc test that is considered to be the most conservative. Most likely to result in a type II error.
- Type II error
- Incorrect retention of the null hypothesis.
- Canonical Correlation
- Used to identify the relationship between multiple predictors and multiple criterion variables.
- Validity
- Looks at accuracy...does the measure what it purports to measure.
- Trend Analysis
- Quantitative independent variable, non-linear. Repeated measures design.
- Formula for Z-Score
-
(X) Raw Score - (M) Mean
---------------------------
Standard Deviation - Percentile Ranks (PR)
- Flat or rectangular distribution; non-linear
- Type I Error
- Rejecting a true null hypothesis. Thinking you have something when you don't.
- Alpha Level
- The probability of making a type I error. Level of significance.
- Beta Level
- Probability of making a type II error.
- Power
- The probability of rejecting a false null hypothesis. Can it detect a true difference between population means.
- Name four ways to increase power:
-
1. Increase the sample size
2. Increase alpha
3. Use a one-tailed test
4. Increase the difference of the population means under study. - Mann-Whitney U
- A nonparametric test that's equivalent to an independent samples t-test.
- Demand characteristics
- Cues in the experimental setting that allow subjects to deduce the experimental hypothesis.
- Quasi-experimental research
- Subjects are not randomly assigned to treatment groups. This threatens internal validity.
- Percentile ranks
- Always have a rectangular or flat, distribution. The same number of scores fall within any given range.
- Z-Scores
- Any distribution can be converted into z-scores. The z-score distribution will maintain the shape of the original distribution. The mean always = 0 and the SD = 1.
- Chi-Square
- Research that involves categorizing observations. Do obtained frequencies differ?