Sunday, March 14, 2004

Drugs and bogus statistics 

(via TalkLeft) Robert Hardaway has a new column in the Denver Post comparing the "war on drugs" to prohibition. In both cases, Hardaway argues, outlawing and prosecuting drug use actually increased the incidence of use. Making drugs illegal also gave a big boost to organized crime. One of Hardaway's most interesting claims is that laws prohibiting drugs were originally racist in intent:
In the early 20th century, labor leader Samuel Gompers set forth his reasons to Congress why opium should be criminalized: "Opium gives the Chinese immigrant workers an unfair advantage in the labor market."

Racists in Congress supported drug criminalization in order to suppress the "Jew peddlers," while the State Department's "opium commissioner," Hamilton Wright, urged criminalization of cocaine on grounds that it turned African-Americans into rapists of white women.
Although Hardaway's column is well worth reading, the statistics he uses are just too good to be true:
According to the Cato Institute, based on deaths per 100,000 users, "tobacco kills 650, alcohol 150, heroin 80, and cocaine 4."
I found these numbers so astonishing that I had to look up where they came from. Here's how the Cato Institute got its numbers for cocaine and heroin deaths:
These figures were determined as follows: Drug Abuse Warning Network (DAWN) heroin and cocaine fatalities for 1984, 1985,and 1986 were averaged. The number of suicides was subtracted. The figures were discounted to account for deaths in which both heroin and cocaine played a role. Since DAWN covers about one-third of the nation's population but almost all major urban areas where drug use flourishes, totals were doubled to arrive at yearly estimates of 2,000 for heroin deaths and 1,000 for cocaine deaths. Finally, these figures were discounted by 80 percent in accordance with the analysis presented in the text.
Let's take this one step at a time. Does DAWN cover "almost all major urban areas"? Not quite. It fails to cover two of the four biggest US cities, Houston and Los Angeles. Even in those cities it does cover, DAWN often omits large parts of the metropolitan area. Moreover,
DAWN data are gathered from medical examiners, coroners, and other death investigation jurisdictions. Not all deaths are reviewed by these facilities. In fact, it has been estimated that only about 20 percent of all deaths are reviewed by a medical examiner or coroner.
Finally, why were the figures "discounted by 80 percent"? A search of the Cato Institute report reveals the reason:
As many as 2,400 of the 3,000 deaths attributed to heroin and cocaine use each year--80 percent--are actually caused by black market factors. For example, many heroin deaths are caused by an allergic reaction to the street mixture of the drug, while 30 percent are caused by infections.
In other words, the discounted figures reflect not the actual death rate from drug usage, but the projected death rate if cocaine and heroin were to be legalized.

Let's correct the damage. Taking at face value the claim that DAWN statistics cover one third of the US population, the Cato Institute report should have multiplied by a factor of three, rather than a factor of two. Because only 20 percent of deaths are reviewed by a medical examiner or coroner, the actual number of deaths is approximately five times higher than reported. To correct for the fact that the Cato Institute discounted its figures by 80 percent, we need to multiply by an additional factor of five. To obtain the actual drug-related mortality rates, the Cato Institute's numbers for cocaine and heroin should be multiplied by 1.5x5x5=37.5 for a total of:

tobacco: 650 deaths per 100,000 users
alcohol: 150 deaths per 100,000 users
cocaine: 150 deaths per 100,000 users
heroin: 3,000 deaths per 100,000 users

I don't have a lot of faith in the accuracy of these statistics, but I guarantee you they're more accurate than the ones from the Cato Institute.
Back to the Odd Hours main page
© 2004 Odd Hours
Reproduction permitted provided Odd Hours or the author of the quoted post is credited.