Let's code

LIBNAME mydata "/courses/d1406ae5ba27fe300 " access=readonly; DATA new; set mydata.gapminder; LABEL ALCOHOL="Alcohol Consumption" HIV_RATE="HIV Rate" LIFE_EXPECTANCY="Life Expectancy";

PROC SORT; by COUNTRY;

PROC ANOVA; CLASS HIV_RATE; MODEL ALCCONSUMPTION=HIV_RATE; MEANS HIV_RATE/DUNCAN;

RUN;

LIBNAME mydata "/courses/d1406ae5ba27fe300 " access=readonly; DATA new; set mydata.gapminder; LABEL ALCOHOL="Alcohol Consumption"  HIV_RATE="HIV Rate"  LIFE_EXPECTANCY="Life Expectancy";   /*data management for hiv rate*/ IF HIVRATE >= 1.3 THEN HIV_RATE = "4"; ELSE IF HIVRATE >= 0.4 THEN HIV_RATE = "3"; ELSE IF HIVRATE >= 0.1 THEN HIV_RATE = "2"; ELSE IF HIVRATE > 0 THEN HIV_RATE = "1"; /*data management for alcohol consumption*/ IF ALCCONSUMPTION >= 9.99 THEN ALCOHOL = "4"; ELSE IF ALCCONSUMPTION >= 5.92 THEN ALCOHOL = "3"; ELSE IF ALCCONSUMPTION >= 2.56 THEN ALCOHOL = "2"; ELSE IF ALCCONSUMPTION > 0 THEN ALCOHOL = "1"; /*data management for life expentancy*/ IF LIFEEXPECTANCY >= 76.64 THEN LIFE_EXPECTANCY = "4"; ELSE IF LIFEEXPECTANCY >= 73.13 THEN LIFE_EXPECTANCY = "3"; ELSE IF LIFEEXPECTANCY >= 64.22 THEN LIFE_EXPECTANCY = "2"; ELSE IF LIFEEXPECTANCY >= 0 THEN LIFE_EXPECTANCY = "1";

PROC SORT; by COUNTRY;

PROC FREQ; TABLES HIV_RATE ALCOHOL LIFE_EXPECTANCY; RUN;

/* univariate graphs using new variable from data management */

PROC GCHART;  VBAR HIV_RATE/Discrete type=PCT Width=10; RUN;

PROC GCHART;  VBAR ALCOHOL/Discrete type=PCT Width=10; RUN;

PROC GCHART;  VBAR LIFE_EXPECTANCY/Discrete type=PCT Width=10; RUN;

/* bi variable graps using raw data */ PROC GPLOT; PLOT ALCCONSUMPTION*HIVRATE;

RUN;

The graph is unimodal, the peak is on category 2 which is HIV rate between 0.1 and 2. It is skewed to the left which shows that a few countries have very low HIV rates while most countries have a high rate.

This graph shows that alcohol consumption has a uniform distribution at around 25% percent. This shows that all groups have similar consumption

This graph shows that life expectancy has a uniform distribution at around 25% percent. This shows that all groups have similar life expectancy  

I summarized the data into 3 new variables, HIV_RATE, ALCOHOL and LIFE_EXPECTANCY. For the data management, I looked at the detailed frequency tables for the three variables that I am studying and separated the data into four equal percentages. For all three new variables, 1 is the lowest and 4 the highest.

For HIV rate, most countries fall in category 2 which is a rate between 0.1 and 0.399. 

For Alcohol consumption the frequency is very close in all four categories with a slightly higher percentage for category 3. 25.67% of the countries have alcohol consumption of between 5.92 and 9

Life expectancy also has very close frequencies for all four groups  with category 1 showing the smallest percentage. 24.61% of the countries have life expectancy of less than 64.22 

1) Below is the SAS program

/* Start program */

LIBNAME mydata "/courses/d1406ae5ba27fe300 " access=readonly;

DATA new; set mydata.gapminder;

LABEL ALCOHOL="Alcohol Consumption"  HIV_RATE="HIV Rate"  LIFE_EXPECTANCY="Life Expectancy";   /* Because the variables contains random data as each country has its own value which dont necessarily match others, we group them into categories depending on the values. This is to avoid situation where we get single summaries */

IF HIVRATE >= 10 THEN HIV_RATE = "1. Very High"; ELSE IF HIVRATE >= 2 THEN HIV_RATE = "2. High"; ELSE IF HIVRATE >= 1 THEN HIV_RATE = "3. Low"; ELSE IF HIVRATE > 0 THEN HIV_RATE = "4. Very Low";

IF ALCCONSUMPTION >= 15 THEN ALCOHOL = "1. Very High"; ELSE IF ALCCONSUMPTION >= 10 THEN ALCOHOL = "2. High"; ELSE IF ALCCONSUMPTION >= 5 THEN ALCOHOL = "3. Low"; ELSE IF ALCCONSUMPTION > 0 THEN ALCOHOL = "4. Very Low";

IF LIFEEXPECTANCY >= 80 THEN LIFE_EXPECTANCY = "1. 80+      "; ELSE IF LIFEEXPECTANCY >= 70 THEN LIFE_EXPECTANCY = "2. 70-79.9"; ELSE IF LIFEEXPECTANCY >= 60 THEN LIFE_EXPECTANCY = "3. 60-69.9"; ELSE IF LIFEEXPECTANCY >= 50 THEN LIFE_EXPECTANCY = "4. 50-59.9"; ELSE IF LIFEEXPECTANCY > 0 THEN LIFE_EXPECTANCY = "5. 0-49.9";

PROC SORT; by COUNTRY;

PROC FREQ; TABLES HIV_RATE ALCOHOL LIFE_EXPECTANCY;

RUN;

/* End program */

I made frequency distributions table for three variables which are HIV Rate, Alcohol consumption and Life Expectancy. The raw data is random and does not have any grouping or pattern. Because of this I decided to group the results into ranges.

From all 213 countries in this study, 67.35% of them have very low HIV rate and 6.12% have very high HIV rate. This shows that HIV rate is very critical in a few countries while majority of the countries have very low HIV rate

For alcohol consumption, only a few countries have very high consumption. Most of the countries are within the ranges of very low to high

For life expectancy, most countries are between 70 and 80. A few countries have a very low life expectancy of less than 50.

I reviewed all code books and found that Gap Minder has information of interest to me. The variable which I am interested in is HIV Rate as my country and region is heavily affected by it. There are a number of variables which I believe contributes to the spread of HIV. For example, alcohol consumption, employment rate, urban rate. There are also some variables which are consequences of HIV such as life expectancy, suicide rate.

From what I have observed in my area and from reading about the HIV rate from other parts of the world, HIV seem to be common among people as young as 20 years old up to 50 years. I have observed through social media that young girls are mostly at risk of early infection because they interact with older, working men. This is because through the working men they can easily afford things such as luxury clothes, phones, laptops, entertainment. They end up putting themselves at risk of HIV infection.

I have decided that I would like to see if there is any association between HIV rate and Alcohol consumption. Is HIV rate higher in countries where there is high alcohol consumption?

I have designed my cookbook to include HIV rate and alcohol consumption which are the variables I would like to study.

I have consulted some literature to see what research has been done on HIV rate and alcohol consumption. The sources that I reviewed are the following:

1.       Drug and Alcohol Consumption and Trade and HIV in the Caribbean: A Review of the Literature

http://www.jscholaronline.org/full-text/JAID/203/Drug-and-alcohol-consumption-and-trade-and-HIV-in-the-caribbean-a-review-of-the-literature.php

 2.       Integrating HIV/AIDS and Alcohol Research

http://pubs.niaaa.nih.gov/publications/arh333/167-178.htm

 3.       Alcohol Use and Sexual Risk Behaviour: A Cross-Cultural Study in Eight Countries

http://www.who.int/substance_abuse/publications/alcohol_sexual_risk_crosscultural.pdf

 4.       REVIEW OF RESEARCH ON ALCOHOL AND HIV IN SUB-SAHARAN AFRICA

http://www.mrc.ac.za/policybriefs/AlcoholSubSaharan.pdf

From the literature reviews, young people who consume alcohol are more likely to engage in sex, mostly at a younger age and with multiple partners. This increases the risk of HIV infection. Women working in hotels, restaurants, bars, night clubs seem to be at risk of engaging in alcohol related sexual behavior. Heavy drinkers are also at risk of HIV infection. People who drink alcohol may delay seeking treatment for HIV