Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical...
-
Upload
dustin-roberts -
Category
Documents
-
view
214 -
download
1
Transcript of Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical...
![Page 1: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/1.jpg)
Model Regress Linear 3Factor Excel 2013V0F 1
byMilo Schield
Member: International Statistical Institute
US Rep: International Statistical Literacy Project
Director, W. M. Keck Statistical Literacy Project
Slides at: www.StatLit.org/pdf
/Model-Regress-Linear-3Factor-Excel2013-6up.pdf/Model-Regress-Linear-3Factor-Excel2013-1up.pdf
Model using RegressLinear 3Factor in Excel 2013
![Page 2: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/2.jpg)
Model Regress Linear 3Factor Excel 2013V0F 2
Goal: Summarize association before/after control for Gender
Required output: Create and upload your worksheet*:
1. Generate two charts (slides 4 and 15). Slide 4: Show trend-line, equation and R2.
Slide 15: Show trend-lines. Show regression model.
2. Generate/show averages (slide 3) .
3. Generate/show output from regression (slide 9). Data: www.StatLit.org/xls/Pulse-Regress-Worksheet.xlsxNote: Male is already in column D in this worksheet. Demo output: www.StatLit.org/pdf/Pulse-Regress-Output.pdf
Subjects are college students. Male: 1 for men; 0 for women.
![Page 3: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/3.jpg)
Model Regress Linear 3Factor Excel 2013V0F 3
Analyze Data:Enter Formula into K3:L4
Actual male-female differences:• Average weight: 158.3 - 123.8 = 34.5 pounds• Average height: 70.75 – 65.40 = 5.35 inches
Question: How much of the male-female weight difference (34.5#) is due to gender (male vs. female) and how much is due to the difference in heights?
Analyzing a whole into parts is called “decomposition”.
![Page 4: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/4.jpg)
Model Regress Linear 3Factor Excel 2013V0F 4
Chart #1
.
y = 5.0918x - 204.74R2 = 0.616
90
115
140
165
190
215
60 65 70 75
We
igh
t
Height
Weight vs Height
![Page 5: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/5.jpg)
Model Regress Linear 3Factor Excel 2013V0F 5
Decompose Male-Female Weight Difference: 1st try
Actual male-female differences (slide 3):• Average weight: 158.3 - 123.8 = 34.5 pounds• Average height: 70.75 – 65.40 = 5.35 inches
Model Weight on Height (slide 4)• Expected Weight = -204.74 + 5.09 * Height
Decomposition of male-female weight difference: • Due to Height difference: 5.09*5.35 = 27.23#• Due to Sex (Gender) difference: 34.5# – 27.2# = 7.3#Inadequate!!! Sex and height are confounded in slope.
![Page 6: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/6.jpg)
Model Regress Linear 3Factor Excel 2013V0F 6
Model Weight by Height & Sex:Four Step Process
Step 1. From Data Toolbar, select Data Analysis (in the Analysis section). Select Regression
Step 2. Regress Weight on Height and Gender
Step 3. Generate Y values given X for models
Step 4. Generate two trend lines on XY plot
![Page 7: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/7.jpg)
Model Regress Linear 3Factor Excel 2013V0F 7
1) Data Toolbar, select Data Analysis. Select Regression
.
![Page 8: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/8.jpg)
Model Regress Linear 3Factor Excel 2013V0F 8
2a) Regress Weight on Height and Sex
.
![Page 9: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/9.jpg)
Model Regress Linear 3Factor Excel 2013V0F
Weight = -117.6 + (3.69*Height) + (14.7*Male).
9
2b) Results: Regress Weight on Height and Sex (Male?)
Formatting and formula are optional
![Page 10: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/10.jpg)
Model Regress Linear 3Factor Excel 2013V0F 10
3) Expected Weights at selected Heights for Men and Women
Create formula in L33 predicting weight:
Pull L33 down
![Page 11: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/11.jpg)
Model Regress Linear 3Factor Excel 2013V0F 11
4a) Start with new chart:Select Data; Select “Add”
.
![Page 12: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/12.jpg)
Model Regress Linear 3Factor Excel 2013V0F 12
4b) Add Two New Series
.
![Page 13: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/13.jpg)
Model Regress Linear 3Factor Excel 2013V0F 13
4c) After Adding Two New Series,Press “OK”
.
![Page 14: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/14.jpg)
Model Regress Linear 3Factor Excel 2013V0F 14
4d) Select Data Point. Format Data Series. Select ‘Solid Line’
.
90
115
140
165
190
215
60 65 70 75
We
igh
t
Height
Weight vs Height
![Page 15: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/15.jpg)
Model Regress Linear 3Factor Excel 2013V0F 15
4e) Add Regression Equation. Final Result
.
90
115
140
165
190
215
60 65 70 75
We
igh
t
Height
Weight vs Height
Weight = -117.6 + 3.69*Height + 14.7*Male
Male
Female
![Page 16: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/16.jpg)
Model Regress Linear 3Factor Excel 2013V0F 16
Decompose Male-Female Weight Difference: 2nd try
Multivariate ‘regression’ model (slide 9 or 15):Weight = -117 + (3.7*Height) + (14.7*Male)
Difference in average heights: 5.35” (slide 3)Difference in average weights: 34.5# (slide 3)
•14.7 pounds due to gender difference – after controlling for height.
•19.8 pounds due to height difference – after controlling for gender: 3.7 #/inch * 5.35 inches
Moral: How you take things into account matters!
![Page 17: Model Regress Linear 3Factor Excel 2013 V0F 1 by Milo Schield Member: International Statistical Institute US Rep: International Statistical Literacy Project.](https://reader035.fdocuments.in/reader035/viewer/2022062805/5697bfa11a28abf838c95db6/html5/thumbnails/17.jpg)
Model Regress Linear 3Factor Excel 2013V0F 17
Decompose Male-Female Weight Difference: Summary
Decompose 34.5# male-female weight difference.
1st try: Regress weight on height (R2 = .62)•27.2 pounds due to height difference• 7.3 pounds due to gender differenceProblem: Gender, height and weight are confounded
2nd try: Regress weight on height and sex (R2 = .66)•19.8# due to height – after controlling for gender•14.7# due to gender – after controlling for height
Moral: How you take things into account matters!