Term
|
Definition
| Given a population with a finite mean, μ, and a finite variance, σ2, the sampling distribution of the mean will approach the normal distribution as N, sample size, increases. |
|
|
Term
|
Definition
| The decision to reject the null hypothesis when it is true |
|
|
Term
|
Definition
| Failure to reject the null hypothesis when it is false |
|
|
Term
|
Definition
| The capacity of a test to reject a false null hypothesis |
|
|
Term
|
Definition
| One in which subjects are assigned randomly to the condition |
|
|
Term
|
Definition
| A distribution comprising values of a statistic. Ex- Distribution of the mean height for Psych394 for the past 50 semesters |
|
|
Term
|
Definition
| Normality, Homogeneity of variance, Independent factors |
|
|
Term
|
Definition
| The effect of one IV ignoring the effect(s) of the other IVs. |
|
|
Term
|
Definition
| Each subject receives only one level of the IV |
|
|
Term
|
Definition
| Each subject receives each level of the IV |
|
|
Term
|
Definition
| The extent to which the simple effects are different in direction and/or magnitude |
|
|
Term
|
Definition
| The experiment controls which levels are used (ex- decides on happy and angry) |
|
|
Term
|
Definition
| The experimenter allows levels to vary by chance (ex- picks two emotions out of a hat) |
|
|
Term
|
Definition
| The effect of one IV at a single level of another IV |
|
|
Term
| Equal measure-to-measure correlation |
|
Definition
| An assumption of repeated measures designs that the correlation between any pair of trials is the same as the correlation between any other pair of trials. |
|
|
Term
| Why is unequal n a problem? |
|
Definition
1. Confound creates conceptual problem, because effects are confounded with sample size, n
2. SScells ≠ SSmain effects + SSinteraction
3. Difficult in deriving orthogonal coefficients |
|
|
Term
|
Definition
1. Randomly discard from large cells until equal n
2. Substitute cell mean for missing scores in small cells
3. Use computer program to find a solution (usually least squares) |
|
|
Term
| List 3 determinants of statistical power and indicate how they relate to power |
|
Definition
1. Effect size: as effect size increases, power increases
2. Sample size: As sample size increases, power increases
3. Type I error: If Type I error raises than test has more power |
|
|
Term
| Why is the expected value of the F-ratio 1 under the null hypothesis? |
|
Definition
| Because the variability among the means is not different from the variability among the scores. |
|
|
Term
|
Definition
- sample size
- error variance
- effect size
- alpha
- others beyond this course
|
|
|
Term
|
Definition
| The magnitude of the difference between the means |
|
|
Term
|
Definition
| The variability among the scores due to random error or chance differences |
|
|
Term
|
Definition
| The probability of making a Type I error |
|
|
Term
| Requirements of orthogonal comparisons |
|
Definition
- ∑a=0
- ∑ab=0
- Comparisons must be apriori (planned)
|
|
|
Term
|
Definition
| A variable that varies systematically with the independent variable |
|
|
Term
| Fully crossed interaction |
|
Definition
| Each level of each IV appears at each level of every other IV |
|
|
Term
|
Definition
- The effect of A across S is not systematic
- Expected means squares
|
|
|
Term
| Assumptions for repeated measures |
|
Definition
- No carry-over effects
- Equal measure-to-measure correlation
|
|
|
Term
| Specific carry-over effects |
|
Definition
There is a known pattern of the carryover
solution: return to baseline |
|
|
Term
| General carryover effects |
|
Definition
effects not specific, but simply due to repeated measurement. (ex- participants become bored or fatigued)
solution: practice trials or breaks |
|
|
Term
|
Definition
| A potential side-effect of counterbalancing order. Effect is due to previous exposure to specific level of IV. |
|
|
Term
| Solutions to violations of equal measure-to-measure correlation |
|
Definition
- MANOVA: treat each level of repeated measures IV as a separate DV.
- Multiply Box's epsilon by any within df
|
|
|
Term
|
Definition
| Expresses a relation between 2 or more variables |
|
|
Term
|
Definition
| Where the regression line crosses y |
|
|
Term
|
Definition
- Rectilinearity
- Homoschedasticity
- x is measured without error
|
|
|
Term
|
Definition
|
|
Term
|
Definition
| Equal variances of the ys at each x |
|
|
Term
| Multiple Regression gives you: |
|
Definition
- an intercept
- slopes for each predictor
- r, the magnitude of the relation
- r2, the variance in y accounted for
|
|
|
Term
|
Definition
| Excessive correlation among predictors |
|
|
Term
|
Definition
| Prediction of x1 is enhanced because x2 suppresses variance irrelevant to y |
|
|
Term
| What's wrong with difference scores? |
|
Definition
- Masks main effects
- Interpretation (only tests interactions)
- if B>2, does not work
|
|
|
Term
|
Definition
- Do not try to predict beyond original x range
- Do a scatterplot to look for leverage
- Truncated distributions reduce the size of r
|
|
|
Term
|
Definition
| Correlation between 2 dichotomous variables |
|
|
Term
| Why should you do a bivariate plot? |
|
Definition
- to check for rectilinearity
- to check for homoschedasticity
|
|
|
Term
| What do you learn by testing the significance of each b in multiple regression? |
|
Definition
| You learn whether each b is a unique predictor of y |
|
|
Term
| df for orthogonal comparisons |
|
Definition
|
|
Term
| Point-biserial correlation |
|
Definition
| Correlation between 1 dichotomous and 1 continuous variable |
|
|
Term
|
Definition
| the inflation or r2 when the number of predictors increases |
|
|
Term
| What is the standard error of the estimate? |
|
Definition
| Syx= the standard deviation of the regression of y on x |
|
|
Term
|
Definition
|
|
Term
| df in hierarchical multiple regression |
|
Definition
|
|
Term
|
Definition
| Correlation between ranks |
|
|
Term
| Fisher's r-to-Z' transform |
|
Definition
| A transformation intended to make r more normal prior to doing any arithmetic, because as rho approaches the absolute value of 1, distribution is skewed. |
|
|
Term
|
Definition
| Correlation between 1 dichotomous and 1 continuous variable |
|
|
Term
|
Definition
| Minimizes the (squared differences of) predicted y's from the observed y's |
|
|
