Scribd
Upload
210 views4 pages

Lecture 20 - KEY - Multiple Linear Regression Worksheet

The document summarizes key outputs from a multiple linear regression model examining factors associated with crystal methamphetamine use. [1] It finds that age, race, and lifetime use of other drugs explains 42.1% of the variance in methamphetamine use. [2] Lifetime ecstasy use is associated with 0.376 more instances of meth use, while race alone did not significantly predict use. [3] For a black 27-year-old who has used ecstasy and cocaine, the model predicts around 4 instances of meth use in the past 4 months.

Uploaded by

Camilo Parra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
210 views4 pages

Lecture 20 - KEY - Multiple Linear Regression Worksheet

The document summarizes key outputs from a multiple linear regression model examining factors associated with crystal methamphetamine use. [1] It finds that age, race, and lifetime use of other drugs explains 42.1% of the variance in methamphetamine use. [2] Lifetime ecstasy use is associated with 0.376 more instances of meth use, while race alone did not significantly predict use. [3] For a black 27-year-old who has used ecstasy and cocaine, the model predicts around 4 instances of meth use in the past 4 months.

Uploaded by

Camilo Parra
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 4

UGPH-GU20 Fall 2019

Lecture 20: Multiple Linear Regression KEY

1. You’re a researcher interested in use of methamphetamines. The NYCDOHMH provides


you with a dataset of a sample of New Yorkers with a count of the number of times they
have used methamphetamine in the last four months. In the study you also ask about the
following:
 Sex (male = 0, female = 1)
 Age (In years)
 Race (white, black, hispanic, asian, other)
 Lifetime use of the other club drugs (cocaine, GHB, ecstasy, ketamine).
 Depression: measured using the CES-D. The CES-D is a series of 20 likert items scored
from 0-3. The total score is summed with higher scores indicating higher depressive
symptamotology and a cutpoint of 16 is used to dichotomize participants into
depressed/not depressed.

You generate the below output

Model Summary
Model R R Square Adjusted R Std. Error of the
Square Estimate
a
1 .712 .506 .421 18.087
a. Predictors: (Constant), coke1, a26to39, asian, other, black, hisp, ghb1,
a40plus, ecstasy1, ketamine1

a
ANOVA
Model Sum of df Mean F Sig.
Squares Square
b
Regression 19457.167 10 1945.717 5.948 .000
1 Residual 18973.644 58 327.132
Total 38430.812 68
a. Dependent Variable: crystal1
b. Predictors: (Constant), coke1, a26to39, asian, other, black, hisp, ghb1, a40plus,
ecstasy1, ketamine1
a
Coefficients
Model Unstandardized Coefficients Standardized Coefficients t Sig.

B Std. Error Beta


(Constant) 2.879 6.143 .469 .641
a26to39 3.069 5.736 .061 .535 .595
a40plus .603 8.590 .008 .070 .944
other 6.203 9.142 .068 .679 .500
black -4.874 7.608 -.062 -.641 .524
asian 4.499 11.020 .039 .408 .685
hisp -6.091 6.123 -.106 -.995 .324
ecstasy1 .376 .160 .302 2.350 .022
ketamine1 .339 .204 .216 1.662 .102
ghb1 .735 .178 .474 4.138 .000
coke1 -.250 .153 -.190 -1.637 .107
a. Dependent Variable: crystal1

1a) What percent of the variance in the number of times using crystal methamphetamine in the
last four months is explained by age, race, and lifetime use of ecstasy, ketamine, GHB and
cocaine?
Adjust R^2 = .421.

1b) According to model 1, describe the relationship between lifetime use of ecstasy and the
number of times using crystal methamphetamine in the last four months.

Adjusted for race, age, lifetime use of, ketamine, GHB, cocaine, a lifetime ecstasy user has used
methamphetamine 0.376 more times in last four months than a non-lifetime ecstasy user (β= 0.376;
p = .002).

1c) According to model 1, describe the relationship between race and the number of times using
crystal methamphetamine in the last four months.

There is no statistically significant difference in times of methamphetamine use across


race groups.
1d) How many times would a black, 27 year old who has tried ecstasy and cocaine be predicted
to crystal methamphetamine in the last four months?

Step 1: write out the estimated regression equation

Step 2: enter values for the IVs into the regression to predict the mean value of the DV.

Times of Methamphetamine use = 2.879 + 3.069 A26TO39 + 0.603 A40PLUS


+ 6.203 OTHER + (-4.874) BLACK + 4.499 ASIAN + (-6.091) HISPANIC + 0.376 ECSTASY1 +
0.339 KETAMINE1 + 0.735 GHB1 + (-0.250) COKE1
= 2.879 + 3.069 X 1 + 0.603 X 0
+ 6.203 X 0 + (-4.874) X 1 + 4.499 X 0 + (-6.091) X 0 + 0.376 X 1 + 0.339 X 0 + 0.735 X 0 + (-
0.250) X 1

2. Answer questions based on the table below.

3a) Write out the null hypothesis, alternative hypothesis and the estimated regression equation.

H0: β1 = β2 = … = βk =0
Ha: at least one of β slopes is not 0

Outcome = intercept + (-0.02) Male + 0.09 Age + 0.16 Black + 0.08 Hispanic + (-0.21) Asian
+ (-0.23) Others + (-0.01) FamilyIncome + (-0.06) Educ + (-0.16) ParentalDispMarijUse +
0.02 ParentCheckHW + 0.11 ParentLetKnow + (-0.12) DisapMarijUse + 0.11
NumClassmatesMarij + 0.01 FreqReligiousAttend + (-0.002) EverArrested

3b) Is age confounded by another factor? Explain.

Yes. Age was significant in the unadjusted model (simple regression/univariate model),
and became non-significant in the adjusted model (multiple regression /multivariate
model).

3c) Is disapproval of youth marijuana use confounded by another factor? Explain.

Yes. The coefficients in the univariate and multivariate model are dramatically different.

You might also like