Reading...
Play button
Play button
Use LEFT and RIGHT arrow keys to navigate between flashcards;
Use UP and DOWN arrow keys to flip the card;
H to show hint;
A reads text to speech;
37 Cards in this Set
- Front
- Back
|
What is the standard error of the mean?
|
When data is collected from many samples of equal size, the distribution of data means would be "normal" and any deviation in the means would be wrror. The average deviation of multiple means of numerous samples is the std. error of the mean.
|
|
|
What is the central limit theorem?
|
Assuming an infinite # of equal sized samples are drawn from a population, the SD and mean of the sample means distribution would be equal to the population mean and standard deviation divided by the square root of the sample size.
|
|
|
What is the rejection region? It's size depends on what?
|
THe tail ends of a normal curve - it is the region of unlikely values because it is unlikely a researcher would obtain these values by chance. When values fall in this region you REJECT the null hypothesis. It's size depends on alpha.
|
|
|
When you reject the null and it turns out to be a mistake, this is known as ___.
|
Type I error
|
|
|
When the null is accepted and it is a mistake, this is known as ____.
|
Type II error
|
|
|
What is power? ANd how do you increase it?
|
The ability to correctly reject the null hypothesis. Increase sample size, increase treatment effect, decrease error, use parametric statistical tests, and one-tailed.
|
|
|
WHen selecting a statistical test of difference, how would the DV influence which test you use?
|
w/ nominal or ordinal DV = non parametric tests (Chi sq, Mann Whitney, Wilcoxin)
w/ interval or ratio DV = t-test or ANOVA parametric tests W/2+ DV's = multivariate |
|
|
Name two assumptions for parametric tests.
|
Homoscedasticity (similar variability) & normal distribution
|
|
|
How would you ensure that there is no correlation between the groups being tested?
|
Random assignment. See page 30 for a good chart
|
|
|
How do you calculate df for the following chi square tests... single sample chi square and multiple sample chi square.
|
Single sample df - #columns - 1
Multiple samples df = (#rows-1) X (#columns -1) |
|
|
How do you determine df form t-test single sample, matched or correlated samples, or independent samples?
|
t-test single sample df = N - 1; t-test for matched samples df = # pairs - 1; t-test for indep samples df = N-2
|
|
|
How do you calculate the three types of df for one way ANOVA's?
|
df total = N - 1
df b/n groups = #groups - 1 df w/n groups = df total - df b/n |
|
|
Chi square expectancy frequency for a cell formula?
|
[(sum of rows) X (sum of columns)]/N
|
|
|
Chi square expectancy frequency when you are given N and groups?
|
N/total # of cells
|
|
|
Why would you want to run one ANOVA versus several t-tests?
|
To minimize type I error.
|
|
|
How do you calculate the F ratio?
|
Mean square between groups/MS within groups. Mean square is the average variability within and between groups.
|
|
|
Of the post hoc test options for one way ANOVA's, which is the most conservative and which protects the least from Type I error?
|
Scheffe is the most conservative with Tukey coming in next and Fisher's LSD protects least from type I error.
|
|
|
How many F ratios in a 2-way ANOVA?
|
3 - one for each IV and one for the interaction.
|
|
|
When interpreting a 2-way ANOVA, what do you look for first and then second?
|
First look for the interaction effect and if no interaction effect examine the main effects.
|
|
|
When would you want to focus on a trend analysis?
|
When the IV is quantitative, frequently the outcome is nonlinear so we become more interested in the trend of the data.
|
|
|
What is the correlation coefficient?
|
Describes the relationship b/n X and Y in terms of strength (0 - 1.0) and directions (- or +).
|
|
|
What is the coefficient of determination?
|
Square the correlation coefficient. It represents the amount of shared variability in Y by X.
|
|
|
How is a simple linear regression derived?
|
Based on the "line of best fit" through the scatterplot and is calculated with the "least squares criterion."
|
|
|
What are the three assumptions of bivariate tests?
|
1. Linear relation b/n X and Y.
2. Homoscedasticity 3. Unrestricted range of scores on both X and Y. |
|
|
What is a zero order correlation and a partial correlation?
|
Zero order - the relaion of X and Y without extraneous variables affecting them.
Partial correlation - looks at a relationship of 2 variables with the effect of a 3rd variable removed. |
|
|
Semipartial correlation is...
|
relationship between 2 variables w/ influence of 3rd variable on either X or Y taken out.
|
|
|
In multiple R, how many IV's and DV's what qualities do the X and Y variables have?
|
1+ IV's, 1 DV, Y is always interval or ratio, 1+ is interval or ratio
|
|
|
What is the coefficient of multiple determination?
|
R squared - an index of the amount of variability in Y accounted for by the combo of all X's.
|
|
|
What is multiple regression equation?
|
Y = a + b1X1 + b2X2 + b3X3...
|
|
|
What is multicollinearity?
|
correlation between IV's or the predictors.
|
|
|
What is the difference between stepwise and hierarchical regression?
|
In stepwise regression, predictors are added to the equation in the order of strength of correlation with DV in ascending or descending order. Hierarchical regression is when the researcher adds IV's in order most consistent with theory.
|
|
|
When would you use Canonical Correlation?
|
Allows you to evaluate the relationship between 2 SETS of variables. Allos for prediction of 2 criterion scores based on 2 predictor scores.
|
|
|
When would you use a discriminant function analyses?
|
2+ predictors and 1 criterion - allows prediction of membership in group. Use this when criterion is nominal rather than interval or ratio.
|
|
|
What is used to predict a categorical criterion based on categorical predictors?
|
Loglinear analysis
|
|
|
What test is used to test a theoretical pathway for significance and relationship strength?
|
Path analysis. Path coefficients are calculated with a series of multiple regressions.
|
|
|
LISREL...
|
Linear Structural Relations - structural equation modeling. Determines whether the path diagrams are correct.
|
|
|
Name two tests of structure:
|
Cluster analysis - gather data on a variety of DV's and look for natural subgroups.
Factor analysis. |
