Class Notes 11-18-2004
Multiple Regression overview
EPSY
6301
18 Nov 2004
Review of performing Project #4 part d, calculating homogeneity of within-groups regression coefficients
Review of univariate and bivariate statistics will be based on Tabaschnick book starting with page 53
In regression we are looking for a line of best fit
- mean of x and mean of y is first "moment"
- mean of the variance is the 2nd moment
we are seeking the regression formula to make predictions about other values of x or y
if you find the correlation coefficient, you are almsot home free because you have the correlation times (standard deviaton of y divided by standard deviation of x)
b = Rxy ( Sy / Sx)
formula 3.31 is the old version / long version of what we looked at prevously
Analyze - Regression - Linear Regression
- DV is the Y value (outcome / response)
- colinearity is just when you use multiple regression
look at pearson product moment result, if it is negative the outcome was negative
- this is not the same as the slope but is related to it
r squared indicates the precent of y that is explained by x
regression is a general form of ANOVA
- ANOVA and multiple regression actually combine for the ANCOVA procedure
As IVs increase, you must increase your sample size
- 15 - 20 subjects are needed per IV
Now look at chapter 5 (skip ch 4 on cleaning up data and screening scores)
in correlation you don't want high correlation, you do want moderate correlation (.3 or .4)
see example on p. 122 (Tabaschnik) multiple regression example
formula on p. 127, if you standardize your IVs then you
what is the goal of multiple regression: to predict
- to get the formula, you must get the coefficients
- so they recommend you standardize coefficients, and that is what the BETA coefficients are
Calculatiions for multiple regression are so complicated that we will not be asked to do hand calculations for this
- it requires matrix algebra
- we will just rely on SPSS
the order you enter the variables in regression makes a big difference
Next discussion will be about MANOVA
18 Nov 2004
Review of performing Project #4 part d, calculating homogeneity of within-groups regression coefficients
Review of univariate and bivariate statistics will be based on Tabaschnick book starting with page 53
In regression we are looking for a line of best fit
- mean of x and mean of y is first "moment"
- mean of the variance is the 2nd moment
we are seeking the regression formula to make predictions about other values of x or y
if you find the correlation coefficient, you are almsot home free because you have the correlation times (standard deviaton of y divided by standard deviation of x)
b = Rxy ( Sy / Sx)
formula 3.31 is the old version / long version of what we looked at prevously
Analyze - Regression - Linear Regression
- DV is the Y value (outcome / response)
- colinearity is just when you use multiple regression
look at pearson product moment result, if it is negative the outcome was negative
- this is not the same as the slope but is related to it
r squared indicates the precent of y that is explained by x
regression is a general form of ANOVA
- ANOVA and multiple regression actually combine for the ANCOVA procedure
As IVs increase, you must increase your sample size
- 15 - 20 subjects are needed per IV
Now look at chapter 5 (skip ch 4 on cleaning up data and screening scores)
in correlation you don't want high correlation, you do want moderate correlation (.3 or .4)
see example on p. 122 (Tabaschnik) multiple regression example
formula on p. 127, if you standardize your IVs then you
what is the goal of multiple regression: to predict
- to get the formula, you must get the coefficients
- so they recommend you standardize coefficients, and that is what the BETA coefficients are
Calculatiions for multiple regression are so complicated that we will not be asked to do hand calculations for this
- it requires matrix algebra
- we will just rely on SPSS
the order you enter the variables in regression makes a big difference
Next discussion will be about MANOVA
Posted: Thu - November 18, 2004 at 09:35 PM
Quick Links
Calendar
Sun | Mon | Tue | Wed | Thu | Fri | Sat |
Categories
Archives
XML/RSS Feed
Statistics
Total entries in this blog:
Total entries in this category: 10
Published On: Nov 18, 2004 09:35 PM
Total entries in this category: 10
Published On: Nov 18, 2004 09:35 PM
