#passiondrivenstatistics

For the Anova Procedure, I decided to test whether there is a relationship between attending a drug/alcohol treatment program and the age at full remission of alcohol abuse. 

Explanatory (Categorical)--->  Ever went to drug/alcohol treatment program 

Response (Quantitative)----> Age at full remission of alcohol abuse 

First, I subsetted my data with the following code to make sure I only get results from 18-24 year olds, and only from individuals who have had at least one alcoholic drink in their lives: 

If AGE in (18:24); If S2AQ1=1;

This is the code I used to run my ANOVA: 

PROC ANOVA; CLASS S3DQ2a6; MODEL S2BQ3F=S3DQ2a6; MEANS S3DQ2a6;

When examining the association between age of full alcohol remission (quantitative response) and whether or not an individual went to treatment (categorical explanatory), an Analysis of Variance (ANOVA) revealed that among young adults, those who attended treatment achieved full remission of alcohol abuse (Mean=19.26, s.d. ±2.40) around the same age as those young adults who did not attend treatment (Mean=19.15, s.d. ±2.25).

The Association between Motivation, Coercion, and Alcohol Abuse Treatment in Young Adults

1. Background

            Alcohol and substance abuse/addiction is an increasing problem among adolescents and young adults in the United States, and the onset of problematic drinking usually occurs between the ages of 18 and 24 (Bukstein, Cornelius, Trunzo, Kelly, & Wood, 2005). Alcohol abuse and binge drinking among young adults in college and non-college populations can lead to sexual assault, injury, and in some cases, death. (Stone, Becker, Huber, & Catalano, 2012) According to a cross-sectional survey of college students from a sample of 119 public and private colleges, binge drinkers are over eight times more likely to have gotten hurt or gotten in trouble with the campus police or local police. (Stone et al., 2012) While there is no scientific cure for substance abuse/addiction, effective treatment can help an individual to obtain sobriety.

            Three fundamental clinical issues are associated with treatment outcomes for populations with drug and alcohol addiction: coercion, motivation, and therapeutic alliance (Wolfe, Kay-Lambkin, Bowman, & Childs, 2013). The best predictors of who enters treatment and how successful they are in treatment are environmental factors such as family factors, motivation, and access to treatment (Bukstein et al., 2005). Motivated individuals are more willing to engage in therapy and to take an active role in the treatment process (Wolfe et al., 2013).

            Motivation is categorized into two primary types: internal and external motivation. Internal motivation stems from an intrinsic desire to change upon recognition of one's substance use as problematic. (Goodman, Peterson-Badali, & Henderson, 2011) An example of internal motivation might be the desire to change upon noticing unpleasant results (e.g. getting fired from one's job due to inappropriate behavior caused by excessive drinking). Contrarily, external motivation is a result of pressure from family and friends (Wolfe et al., 2013). 

            The present study will examine young adults from the National Epidemiologic Survey of Alcohol and Related Conditions (NESARC). The goals of the analysis will include 1) establishing the association between alcohol abuse and treatment seeking patterns in young adults, and 2) determining whether motivation (internal/external) or coercion has a higher success rate in terms of treatment outcome (e.g. will the individual continue to abuse alcohol after leaving treatment, or will they stay sober?)

2. Method

            2.1 Sample

            The sample from the first wave of the National Epidemiologic Survey on Alcohol and Related Conditions (NESARC) represents the civilian, non-institutionalized adult population of the United States, and includes persons living in households, military personnel living off base, and persons residing in the following group quarters: boarding or rooming houses, non-transient hotels and motels, shelters, facilities for housing workers, college quarters, and group homes. The NESARC included over sampling of Blacks, Hispanics and young adults aged 18 to 24 years. The sample included 43,093 participants. The study was done by computer-assisted interviews. 

            2.2 Measures

            Survey data from 18-24 year old respondents will be included, as the study will focus exclusively on young adult behavior. The study measures three categorical variables: internal motivation, external motivation, and coercion. Internal motivation was assessed through problems that directly affected the drinker ("Ever continue to drink even though causing health problem," "More than once want to stop or cut down on drinking" "Ever had job or school troubles because of drinking"). External motivation was assessed based on the influence of people in the drinker's social circle ("Ever continue to drink despite causing trouble with family/friends", "Did not seek help because afraid of what boss, friends family, or others might think.") [WCM3] Coercion was assessed by a history of arrest or legal problems tied to drug use/drinking ("Ever get arrested/have legal problems because of alcohol/drug use".) These variables are compared against the primary response variable ("Ever Went to Drug/Alcohol Rehabilitation Program").

3. Implications

            The effects of alcohol abuse are disastrous, especially on college campuses. Hardly a week goes by without a newspaper headline announcing an alcohol-related death or accident. It is important for colleges to become better equipped to deal with alcohol abuse. For example, the study may find that young adults are less inclined to go to treatment and stay sober if they are worried about what their friends think. If that is the case, colleges should take that information and find ways of providing confidential treatment on campus to enable students to get help without worry about external judgment.

References

Bukstein, O. G., Cornelius, J., Trunzo, A. C., Kelly, T. M., & Wood, D. S. (2005). Clinical predictors of treatment in a population of adolescents with alcohol use disorders. Addictive Behaviors, 30(9), 1663-1673. doi: 10.1016/j.addbeh.2005.07.013

Goodman, I., Peterson-Badali, M., & Henderson, J. (2011). Understanding motivation for substance use treatment: The role of social pressure during the transition to adulthood. Addictive Behaviors, 36(6), 660-668. doi: 10.1016/j.addbeh.2011.01.011

Stone, A. L., Becker, L. G., Huber, A. M., & Catalano, R. F. (2012). Review of risk and protective factors of substance use and problem use in emerging adulthood. Addictive Behaviors, 37(7), 747-775. doi: http://dx.doi.org/10.1016/j.addbeh.2012.02.014

hi there! aside from posting the screen captures here in Tumblr (see below), I've uploaded a pdf version in Dropbox, which  is still in vector graphics and has a higher resolution.

Title

 The Association between income per person and life expectancy rate among countries worldwide with  varying urbanization rates.

Vijaya Phanindra Sarma, Passion Driven Statistics Course, Wesleyan University

  Introduction

It is a known fact that average income of a person of a country is a good indicator to study the standard of living of citizens of that country. A country’s Income per Person(1) or average wealth of a citizen tends to be closely related and linked to other socio economic indicators in general like better healthcare, longer life expectancy rates, high standard of living, economic freedom and ability to spend beyond needs and necessities. In general citizens of developed countries with high income have longer life expectancy rates with lower infant mortality rates. Furthermore the rapid urbanization as a result of rural migration caused by the lure of opportunities that cities can offer might be having both positive impact in the form of increased economic growth, better healthcare, educational and other facilities. It might be the case that high income per person of a country contributed to high life expectancy rates and also high urbanization rates in a country to higher income per person levels.

  Research Questions

Is longer life expectancy rate associated with high income per person of a country

Is the association between high income per person and life expectancy rate is same across countries with high urbanization rates

Methodology

 Sample

GNI (Gross National Income) Per Capita, Life Expectancy is taken from the World Bank development indicators for all the available countries for years 2001-2011. 

Urban Rate is taken from the International Human Development Indicators for years 2000-2010

As there is missing data for some countries, average value of 10 years of data is considered.

Measures

The original data is was from the Gapminder(2) code book.

These data is of different years between different selected variables.

Data is missing for many countries, below tables gives snap shot of missing data.

  So Based on the code book definition, generated an alternative dataset for same variables and from the same source

Main Source of search is  http://www.google.co.in/publicdata/directory

Income Per Person Quantitative variable is binned into four categories as per world Development indicators classification(3)

       Income below 1025 "LOW INCOME"        Income between 1025 and 4035 "LOWER MIDDLE INCOME"        Income between 4035 & 12475 "UPPER MIDDLE INCOME"        Income more than 12475 "HIGH INCOME";

Mean and Standard Deviation of Income Per Person income_2001-11 variable. PROC Univariate Procedure

Income per Person Groups and Urbanization Rate (Categorized quantitative Variables). PROC FREQ procedure.

 /* ANOVA Analysis Incomeperperson vs Life Expectacy Rate */

/* DUNCAN POST HOC TEST AS the CATEGORICAL VARAIBLES Have more than 2 levels */

PROC ANOVA; CLASS incomeperperson_grp;

MODEL life_expectancy_2001_11=incomeperperson_grp;

MEANS incomeperperson_grp /DUNCAN;

/* ANOVA Analysis Incomeperperson vs Urban Rate */

PROC ANOVA; CLASS incomeperperson_grp;

MODEL urbanrate_2000_10=incomeperperson_grp;

MEANS incomeperperson_grp /DUNCAN;

Results

Univariate

Among all the countries the average income per person is 7320 (Std Deviation 10425)

Of  the four income level groups, 60%  (31.72, 29.66) form the lower income and lower middle income range

Of the four Urban Rate groups, 35% the highest are from the upper middle urbanization group.

Bivariate

ANOVA analysis between Income Per Person Groups (Category Explanatory Variable) and Life Expectancy Rate (Quantitative Response Variable) shows that the association is statistically significant (F Value= 92.50 and P <0.0001)

Further Duncan Test on multiple categories of explanatory variable income per person shows that these 4 income level groups are statistically different.

ANOVA analysis between Income Per Person Groups (Category Explanatory Variable) and Urban Rate (Quantitative Variable) shows that the association is statistically significant (F Value= 62.33 and P <0.0001)

Further Duncan Test on multiple categories of explanatory variable income per person shows that these 4 income level groups are statistically different.

Moderation

Urban Rate confirmed positive and significant association between Income Per Person and Life Expectancy Rates.

Among countries with High Income Per Person and are in Highly Urbanized Cities there is high life expectancy rate.

A bar chart below clearly illustrates association between Income Per Person, Urban Rate and Life Expectancy rate.

Furthermore the highest life expectancy rate of age 80 is found in both High and Upper Middle Urban Cities of a country and of people with high income per person.

This group is visibly absent in the other two groups Middle Urban and Lower Urban Cities 

  Discussion

What might the results mean?

• Countries with higher income per person and with large population living in urbanized cities seems be having longer life expectancy rates.

Strengths

• Results are based on the data provided by the World Bank

• Average value of a decade (10 years) considered for analysis which fairly represents data over a considerable period

Limitations

• Further Research is required to determine if there are any other factors that influence Life Expectancy Rate

• A good example of existence of  other factors is for instance a county might focus to provide better social security even though their income per person is very low, for instance Ecuador with income per person at around USD 1515 (which is Lower Middle Income group) has life expectancy rate of 74.79

(1) A country’s GNI (Gross National Income) Per Capita is dollar value of country’s yearly income divided by its total population

(2)www.gapminder.org

(3)http://data.worldbank.org/about/country-classifications

Well life has thrown a few obstacles since I began this project which put me behind schedule but I'm working to catch up! 

What we have here is the syntax for frequency tables that represent my variables. I've decided to rename my labels for clarity. 

libname mydata "/courses/u_coursera.org1/i_1006328/c_5333" access=readonly;

DATA new; set mydata.gapminder_pds;

/*labels used for variables*/

LABEL incomeperperson="GDP per capita in USD$"

breastcancerper100th="Breast Cancer rate 100,000 per year"

co2emissons="Co2 emission in metric tons"

proc sort; by country;

proc freq; tables country breastcancerper100th incomeperperson co2emissions;

run;

Here is a snippet of the Frequency tables:

Among countries (Gapminder), the correlation between urban rate, life expectancy rate (quantitative explanatory variable) and income per person (quantitative response variable) was 0.61797 (p=.0001) and  0.59629 (p=0.001) suggesting that 38% (i.e. 0.61797 squared) and 36% (i.e. 0.59629 squared) of the variance can be explained by income per person

When examining the association between urbanrate (categorical response variable) and incompe per person (categorical explanatory variable), a chi-square test of independence revealed that there is significant different between income groups and urbanization rate