Alcohol Consumption: Why do you do it?

advertisement
Alcohol Consumption:
Why do you do it?
Anna Kheyfets
Guillermo Morini
Drew Vinson
Econ 120:Statistics
Professor Yoon
5 May 2009
Introduction
 Research

Question
What variables affect
alcohol consumption?
 Relevance



Alcohol might lead to
substance abuse
Risk to human health
≈ 4% disease worldwide
College students want to drink legally
Alcohol
Consumption
Worldwide
Burden of
Disease
from
Alcohol
Variables Tested
 ŷ = Alcohol consumption (L of pure
alcohol/capita) among adults (15 years)








X1 = Minimum Legal Drinking Age
X2 = Government Education Expenditures
(% of GDP)
X3 = GDP per capita ($1000)
X4 = Unemployment Rate (%)
X5 = Global Peace Index (Score)
X6 = Inflation rate (%)
X7 = Homicide Rate (per 100,000)
n = 65 countries
Predictions

Positively correlated with
alcohol consumption





Unemployment Rate
Inflation Rate
GDP per capita
Homicide Rate
Negatively correlated with
alcohol consumption



Minimum Legal Drinking Age
Global Peace Index
Government Education Expenditures
X1 = Minimum Legal Drinking Age
18 or 21 ?
 The



mission of MADD is to
stop drunk driving
support the victims of this violent crime
prevent underage drinking
 MADD
supports the 21 Legal Drinking Age
Since the 21 Minimum Legal Drinking
Age law was enacted in 1984, an
estimated 25,000 lives have been
saved.
Amethyst
 Movement
of US college presidents
calling for the reconsideration of US
drinking age laws
 Current signatory count: 135

Signed by President David Oxtoby,
Pomona College
“Alcohol education that mandates
abstinence as the only legal option has not
resulted in significant constructive
behavioral change among our students”
Pros of 21 Law
 The




21 Law Saves Lives
Has saved 25,000+ since enacted
Saves 1,000 lives a year
Halved the rate of drunk driving deaths since
1980s
Decreases overall underage drinking
Amethyst vs. MADD

The studies of lives saved quoted by MADD,
do not focus on people under 21


Lives saved between 18-20 are diluted by the fatalities
in the 21-24 group
Alcohol-related traffic fatalities declined (1982 -1992)
•
•
•
•
•
•

United Kingdom:
Germany:
Australia:
The Netherlands:
Canada:
United States:
50% decline
37% decline
32% decline
28% decline
28% decline
26% decline
All countries except for US: min drinking age = 18 yrs.
Factors other than drinking age are involved
Validity of Model
Regression Statistics
F- Stat
Multiple R
0.6646
R Square
0.4417
Adjusted R Square
0.3732
H0: β1 = β2 = β3 = β4 = β5 = β6 = β7 = 0
H1: at least one of βs ≠ 0
Standard Error
3.0543
REJECT H0
Observations
65
ANOVA
df
SS
MS
F
Significance F
7
420.7
60.10
6.442
1.334 E-05
Residual
57
531.7
9.328
Total
64
952.4
Regression
Significant Variables
Coefficients
Standard
Error
t Stat
P-value
8.874
3.4654
2.5608
0.0131
Min Legal Drinking Age
-0.0128
0.0780
-0.1636
0.8707
Education Expenditures
0.3139
0.2955
1.0622
0.2926
GDP per capita ($1000)
0.0909
0.0328
2.7683
0.0076
Unemployment Rate (%)
0.1199
0.1510
0.7943
0.4303
Global Peace Index
-2.8045
1.3115
-2.1384
0.0368
Inflation Rate (%)
-0.0611
0.0720
-0.8478
0.4001
Homicide Rate (per 100)
0.0297
0.0506
0.5865
0.5599
Intercept
Regression Model
ŷ = 8.8743 - 0.0128x1 + 0.3139x2+ 0.0909x3+ 0.1199x4- 2.8045x5- 0.0611x6+ 0.0297x7
Var
ŷ
b
Statistic
Alcohol consumption
Unit Value
Effect on ŷ,
per unit increase
L per capita
8.8743 Intercept
x1
-0.0128 Min Legal Drinking Age
x2
0.3139 Education Expenditures
x3
0.0909 GDP per capita
x4
0.1199 Unemployment Rate
x5
-2.8045 Global Peace Index
x6
-0.0611 Inflation rate
x7
0.0297 Homicide Rate
1 year
-4L
1%
-1.3 L
$1,000
+9L
1%
+ 12 L
1 Unit
- 280 L
1%
-6L
per 100,000
people
+3L
Check for Normality
18
Frequency
16
14
12
10
8
6
4
2
0
-8
-6
-4
-2
0
2
Residuals
4
6
8
More
Results Compared with Predictions

Positively correlated with
alcohol consumption





Unemployment Rate : TRUE
Inflation Rate : FALSE
GDP per capita : TRUE
Homicide Rate : TRUE
Negatively correlated with
alcohol consumption



Minimum Legal Drinking Age : TRUE
Global Peace Index: TRUE
Government Education Expenditures : FALSE
Conclusion

Testing this hypothesis,
where X1 = min drinking age
H0: B1 = 0
H1: B1 ≠ 0
 Using a t-test at the 5% significance level,
we rejected the alternative hypothesis.

Therefore, there is no significant relationship
between the minimum legal drinking age and the
consumption of alcohol per capita in a country.
 This data supports the argument of the Amethyst
Initiative.
Conclusion
 Recalling
our concern about the
connection between alcohol and
disease…
 Governments cannot control any of the
significant variables from our model

Unless the government wants to go to war to
curb alcohol consumption
Acknowledgements
 Professor
Yoon for your guidance and
statistics course.
 Audience
 Sources





www.who.org
www.cia.gov
www.amethystinitiative.org/
www.madd.org
www.photius.com/rankings/murder_rate_of_countries
_2000-2004.html
Download