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;
74 Cards in this Set
- Front
- Back
When is one-way ANOVA used?
|
-one DV
-one IV -three or more levels of the IV (Can use ANOVA with 2 levels) |
|
What are the 2 sources of variance in the one-way ANOVA
|
-Treatment variance
-Error variance |
|
What type of size values (large or small) do we want the 2 sources of variance to be
|
-Treatment variance LARGE
-Error variance small |
|
How is ANOVA represented
|
"F"
F= Treatment variance/Error variance the bigger the "F" value, the more likely to reject the null |
|
What does ANOVA stand for
|
Analysis of Variance
|
|
What is the Error variance attributed to
|
-participants
|
|
What can affect the participants
|
environment: causes participants to react diff when assessed on the DV
history: causes diff. responses to treatment |
|
Equation
Degrees of Freedom for ANOVA |
F = MSbetween/MSerror
MS-mean squared MS = SS/N or SS/N-1 SS: Sum of Squares N: sample size for population N-1: sample size for sample |
|
What is the difference between MSb and MSe
|
with the "b", it is the number of groups (k)-1
with the "e", it is the total number of participants (N-k) |
|
What are the seven steps in ANOVA
|
1)State the null
2)One-way ANOVA (equation F=) Assumptions 3)alpha=0.05 4)Criteria for rejecting the null (p<0.05) 5)run the test 6)Accept/reject the null 7)Post Hoc Test 8)conclusion |
|
When do you reject the null for ANOVA
|
-when there are more than 2 groups only tells us that AT LEAST 2 of the groups are different
|
|
What is the Tukey Post Hoc test
|
Pairwise comparisons between means
|
|
What is one of the most used Post Hoc tests
|
Tukey
-can be used to determine which pair of means are different -controls against type 1 errors |
|
After using the Post Hoc test, when do you reject the null
|
if the mean difference (absolute value) is greater than or equal to the Tukey contrast value
|
|
When are Bivariate Correlation used
|
When there are two different variables that have been measured
|
|
What would the investigator be interested in with Bivariate Correlation
|
In determining if there is a relationship between the two variables
|
|
True or False
If a relationship is found between variables, that indicates a cause and effect |
FALSE
i.e. Height and weight in adults...incr, in weight does not mean an incr. in height |
|
When r=1.00 in correlation...
|
Considered a perfect positive correlation or relationship between the variables
Scores high/low for both variables |
|
When r=-1.00 in correlation...
|
Considered a perfect negative correlation or relationship
Score high for one variable, but low for the other |
|
What are used to view relationships between variables
|
Scatterplots
|
|
What is the most common type of correlation
|
moderately positive/negative correlation
|
|
What are the positive descriptors used with correlation
|
r=1.00 - .80 Strong pos.
r=.80 - .40 Moderate pos. r=.40 - 0 Weak pos. |
|
What are the negative descriptors used with correlation
|
r=0 to -.40 Weak neg
r=-.40 to -.80 Moderate neg r=-.80 to - 1.00 strong neg |
|
In what ways can correlations may be used (3)
|
1.Determine the relationship between 2 variables
2.Determine the reliability of a measuring instrument 3.Determine the validity of a test |
|
What is assumed with correlation techniques
|
there is a linear (straight) line relationship between variables
|
|
How is assumption with correlation techniques checked
|
by drawing and looking at the line of best fit
|
|
Correlation Techniques
Person Product Moment Correlation |
This is a parametric technique that is symbolized by a small letter "r"
|
|
Correlation Techniques
Spearman Correlation |
This is a nonparametric technique also symbolized using "r"
|
|
What are the assumptions for Person Product Moment Correlation
|
1)one group randomly selected from the population of interest
2)both measured variables be interval/ratio data type 3)Data on each variable must be normally distributed 4)variables must have a linear relationship to each other 5)variances of the two variable must be equal |
|
What is the Homoscedasticity assumption
|
Assumption asks if the variances of data points are equally distanced from the line of best fit
|
|
How do we determine if the Homoscedasticity assumption is met
|
ellipse shape circles are drawn around the points to decide if assumption is met
|
|
What is the seven step procedure for Pearson Product Moment r
|
1)Write the null
2)get the Pearson Product Moment r formula 3)determine the alpha level 4)Determine the critical value (state when the Ho is rejected) 5)calculations 6)Compare calc. to critical r 7)state conclusion |
|
What is the Coefficient of Determination
|
=r squared
tells us the proportion of shared variance by the two variables The amount of variance explained in performance on one variable by knowing the performance on the second variable |
|
What is a synonym for regression
|
Prediction
|
|
What must be done first between X and Y variables
|
the Correlation
|
|
What are soem examples of Regression
|
-GRE tests predicting individual's success in grad school
-measuring skinfolds to predict body fat % -determining activity level and measuring BMI |
|
What is the prediction/regression formula
|
Y'=bX + a
**same formula for a straight line** where "b" is the slope and "a" is the y-intercept |
|
What is the difference between Y and Y'
|
Y is the actual observed score
Y' is the predicted score |
|
What does the regression line describe
|
formula for the line describes the line of best fit
|
|
How many points are needed to define a line
|
2
|
|
What can be calculated by knowing 2 points in the line
|
The slope (a) and the y-intercept (b)
|
|
True or False
Estimates are made for the line of best fit using all of the data because we do not know the 2 points to begin with |
True
|
|
what does the slope steepness depend on
|
how the changes in the X variable (predictor) are related to changes in the Y variable (criterion)
|
|
What does the regression line describe
|
formula for the line describes the line of best fit
|
|
How many points are needed to define a line
|
2
|
|
What can be calculated by knowing 2 points in the line
|
The slope (a) and the y-intercept (b)
|
|
True or False
Estimates are made for the line of best fit using all of the data because we do not know the 2 points to begin with |
True
|
|
What does the regression line describe
|
formula for the line describes the line of best fit
|
|
what does the slope steepness depend on
|
how the changes in the X variable (predictor) are related to changes in the Y variable (criterion)
|
|
How many points are needed to define a line
|
2
|
|
What can be calculated by knowing 2 points in the line
|
The slope (a) and the y-intercept (b)
|
|
True or False
Estimates are made for the line of best fit using all of the data because we do not know the 2 points to begin with |
True
|
|
what does the slope steepness depend on
|
how the changes in the X variable (predictor) are related to changes in the Y variable (criterion)
|
|
How is the line of best fit drawn
|
1)Place dot where the mean X score meats the mean Y score
2)locate the y-intercept value 3)connect the two points |
|
True or false
Using the scatterplot and prediction line is a precise way of predicting |
False
|
|
True or false
Once prediction formula is developed, it is use on individuals who were involved in its formulation |
False
Individuals who were NOT involved |
|
What are the X values for each individual entered into
|
The prediction formula to determine a predicted Y score
|
|
What is the minumum number of sujects needed per variable
|
40
|
|
Why is there a minumum requirement of suject with regression lines
|
half of the subjects are used to develope the regression prediction formula and the other half are used to test the formula
-need to determine if it is a good prediction formula |
|
When are predicitons considered to have NO errors
|
when r = 1.00 or -1.00
|
|
What is error of estimates
|
residual
the actual Y value minus the predicted Y value |
|
Definition
Variance error of estimate |
The variance of a sample of residuals
|
|
Definition
Standard Error of Estimate (SEE) |
The square root of variance error of the estimate
also-the standard deviation of the residuals |
|
True or False
Want the standard error of estimate to small |
True
The smaller it is, the more precise the prediction |
|
What will result in a smaller SEE (standard error of estimate)
|
small stnd dev for the Y scores and high magnitude for r
|
|
Formula
SEE |
= sy x (sq rt. of 1-r squared)
sy- stnd dev of Y variable |
|
What are the assumptions for Bivariate regression
|
Same as correlation
1)Randomness 2)normal distribution 3)linear relationship btwn IV and DV |
|
What is multiple Regression
|
has more than on IV
|
|
What is the general purpose for multiple regression
|
asses relationship between one variable (DV, symbolized as y) and several others (IVs, predictor, or x)
|
|
What can the IV be for multiple regression
|
-Correlated or not
-Continuous or categorical -Naturally occurring or manipulated |
|
What is the formula with multivariate regression
|
Yp=a + b1X1 + b2X2 +.....
|
|
Definition
Beta Weights |
The standardized regression coefficients (b in the equation)
|
|
What are Beta Weights used to determine
|
the importance of the IV in relation to the DV
|
|
True or false
Beta Weights are really the raw regression coefficient (raw score) converted to a z-score |
True
|