The word, "salud" or "for health" is sometimes spoken before the drinking of a beverage, and many beverages are indeed ingested for health reasons, including milk, tea, and even alcohol. However, health impacts from such beverages likely do not happen immediately, and may in fact take years or even decades to manifest. Figuring out whether these beverage favorites are affecting health is a difficult task, but Scientific AmeriKen is not afraid! Using the power of correlation to analyze the data of the world, Scientific AmeriKen shall attain this elusive information. Furthermore, Scientific AmeriKen will also delve into the impacts of worldwide consumption of Scientific AmeriKen itself!

Lots has been said regarding many of the beverages of this study. For example, tea is extremely healthy, alcohol fights heart attacks, milk for good bones and so on. It is therefore the hypothesis of this experiment that these health legends will be supported by our analysis. Given this hypothesis we anticipate milk and tea to have favorable health profiles, while coffee and alcohol will have mixed health profiles. We expect Scientific AmeriKen consumption to provide the greatest health associations of all.

Data was collected from around the web, including per capita consumption by country of alcohol, milk, tea, coffee, food, as well as Scientific AmeriKen's page views by country. This data was correlated against death per 100,000 individuals by country caused by cancer (all causes), asthma, cardiovascular disease (CVD), stroke, alcohol, road accidents, diabetes, endocrine disorders and dementia. Datasets were correlated using excel to attain the coefficient of correlation (R2) and the correlation coefficient (R). The R value was then used with the degrees of freedom (number of matches in the two compared datasets - 2) to calculate the p-value, using Graphpad online software (Click here). Initial work used nearly all 200+ countries, however, industrialization was a more important factor, thus Scientific AmeriKen narrowed analysis to only those countries that appear on the human development index list (Click here, about 45 countries). Additionally, data for potential data confounders was included (life expectancy and per capita income, and HDI). All data can be found here, and sources for the data can be found here.

Initially data was plotted on a graph to attain the R2 value, and then the square root of this value was used to determine the p-value. This effort is exemplified in the figure below, that shows the positive, no and negative association of alcohol with cancer, stroke and diabetes, respectively. A positive correlation could be restated as, "Countries with higher consumption of X also have higher death by Y" and negative, "countries with higher consumption of X also have lower death by Y". No association would mean no relationship between the beverage and the cause of death.

Next, upon discovery of Excel's "Correl" function, which provides R values by just highlighting columns, the process became much quicker and the R values are presented for each relationship below.

In the above figure the boxes are shaded by whether its a positive or negative correlation, and secondly by whether the p value is significant. P value is a determination of the percentage of likelihood that the given relationship is due simply because of chance. It was arbitrarily chosen that a p value below 0.05 may actually be a real phenomenon, while a p value below 0.0001 may REALLY actually be a real phenomenon. Alcohol was the only beverage to fall into this latter category with being positively associated with cancer, while at the same time being negatively associated with diabetes and asthma. Although not as strong as expected, alcohol consumption was also associated with alcohol deaths. Milk was generally beneficial, while strangely coffee and tea did not seem to be linked with any outcomes. Scientific AmeriKen usage was linked with death by dementia.

The power of linear regression analysis allows the ability to pursue questions such as, "do people in countries that drink a lot of tea live longer?", Unfortunately, this analysis can also quickly generate a lot of data, leaving the investigator to suffer hair loss and ask a second question, "what does it all mean???". There were findings of note, in particular the strong relationship between per capita consumption of alcohol and death by cancer from all causes. This finding is also supported by the association that countries that have a lot of cancer death also have substantial death by alcohol (figure below). Surprisingly, tea was not especially healthy in this analysis, nor was coffee, which one may suspect to be either good or bad, but in actuality, is seemingly inconsequential. Interestingly, dementia was positively associated with milk and coffee consumption as well as Scientific AmeriKen usage. One would have to suspect that since the onset of dementia is usually later in life, that these two beverages and Scientific AmeriKen webzine may actually be prolonging life so that dementia can become a possibility. Scientific AmeriKen happily accepts this explanation, although the data is suggestive that scientific AmeriKen usage is negatively associated with life expectancy, the p-value is luckily above 0.05 (0.06) and thus this ridiculous notion can be cast aside.

Above is the relationship between different causes of death. Intriguingly, some diseases are tightly associated with one another, which may signify a similar disease process is involved. For example, stroke and cardiovascular disease are highly correlated in this analysis, and in fact, it is thought that conditions like chronic inflammation may cause both diseases. Similarly, asthma, diabetes and endocrine disorders may share mechanisms. Yet, some associations may have simpler explanations, such as the negative relationship between cancer and asthma, which cancer typically occurs later in life while asthma earlier, would make sense. Then again, other correlations such as the strong relationship between countries with high stroke death also having high road accidents. Scientific AmeriKen can only speculate that the headaches caused by bad traffic is taking its toll. All in all, Scientific AmeriKen has reached only one firm conclusion and that is that this study has created more questions than it has answered.

Click here to discuss at Science Chat


Return to the cover page Many experiments - see what we've done! Learn about the creator of this site Selling cups and stuff for years now - only sold a couple! Question? Comment? Suggestion? Reach us today! More money means more science!  Science is our one and only goal - support today and make the world better! Free subscription - sign up today! To the Homepage!