According to Thomas Kuhn, the role of measurement in science is to isolate anomalies and make them unavoidable in theory appraisal. (17) One aspect of measurement error is inadequate study design, usually indicated by flaws in reasoning. Here are a few of frequent flaws that can be found in governmental defenses of marijuana policy based on rejected and outdated findings.

Affirming the Consequent.

The National Household Survey (NHS)(18) and the Monitoring the Future Project (19) (MFP) have produced a 20 year data base of on drug use in the United States, and these are the most widely cited indicators of the success or failure of national policy as the numbers trend up or down. (20)

The surveys do not distinguish between use and abuse of drugs, the NHS has problems with underrepresented population segments, and the MFP has a problem with overreporting by young students. All of these provide serious problems for using the data as a policy diagnostic. (21)

During the early 1990's a downward trend in high school marijuana use reversed. As marijuana use was falling during the 1980's perception of great risk associated with using marijuana increased. As marijuana use began to increase recently, risk perception dropped. (See Figure 1.) [Exhibit 1. Deleted from text file] Federal policy is based on the assumption that use is inversely related to risk perception, and a graph of those two indicators is offered as the scientific justification for present policy, which is to increase risk perception. (22)

No null hypothesis is evaluated; there is no interest in trying to prove the theory false.

People who perceive great risk in using marijuana regularly can be recoded as believers, they believe that marijuana is harmful. There is a complementary percentage of the sample who are non-believers, who do not believe that regular marijuana use is harmful. According to the present policy paradigm, there should be little or no difference between the number of regular users and the number of non-believers.

One way to test this hypothesis is to compare the ratio of regular users to non-believers for the data presented in the NHS and the MFP and compare the ratios among different drugs. A generous standard is to define "little or no difference" as a ratio above .7, and the government paradigm would lead one to expect that little difference in the ratios between drugs, and an improving ratio as age and/or education increase.

Table 1 compares the ratios of users to non-believers for marijuana, alcohol, and tobacco. Regular use is defined as past month use for all drugs, and perceived risk was evaluated as to regular marijuana use, use of 5 alcoholic drinks or more in sequence within the last two weeks, and the use of a pack of cigarettes a day.

There is little or no difference between non-believers and regular users of alcohol and tobacco, and the difference between the two groups drops as age and education increase. (Table not included.) This does not provide sufficient data to reject the hypothesis that there should be little or no difference between the groups, and does not contradict the theory that use is inversely related to perceived harmfulness. (The ratios in Table 1 for alcohol exceed 1.0, and suggest that many past month users agree 5+ drinks is risky and have milder consumption levels.)

There is a big difference between the ratio for marijuana and the ratios for alcohol and tobacco. With one interesting exception, the ratios for marijuana fall significantly below .7, and the ratios drop consistently as age and education increases. This provides sufficient reason to reject the hypothesis that there is no difference between non-believers and regular users.

Many people don't believe marijuana is harmful, and they don't use it anyway. The prominent exception is that during the 1980's, and only with high school students, the ratio was above .7 (though still below alcohol and tobacco).

Table 1. Ratios of Users to Non-Believers

All Ages

1985 1991 1992 1993

Marijuana

used in past month 9.3 4.8 4.4 4.3

non-believers 25.2 22.2 22.3 23.3

ratio .369 .216 .216 .185

Alcohol

used in past month 58.3 50.9 47.8 49.6

non-believers 40.1 41.0 38.2 39.1

ratio 1.45 1.24 1.25 1.27

Cigarettes

used in past month 31.4 27.0 26.2 24.2

non-believers 43.2 36.5 35.9 34.5

ratio .727 .740 .730 .701

Source: National Household Survey

Table 2. Ratios of Marijuana Users to Non-Believers, 12th Graders.

1975 1976 1977 1978 1979 1980 1981 1982 1983 1984

30 day-Mj 27.1 32.2 35.4 37.1 36.5 33.7 31.6 28.5 27.0 25.2

No Grt Risk 56.7 61.4 63.6 65.1 58.0 59.6 42.4 39.6 37.2 33.1

Ratio .478 .524 .557 .570 .629 .565 .745 .720 .726 .761

1985 1986 1987 1988 1989 1990 1991 1992 1993 1994

30 day-Mj 25.7 23.4 21.0 18.0 16.7 14.0 13.8 11.9 15.5 19.0

No Grt Risk 29.6 28.7 26.5 23.0 22.5 22.2 21.4 23.5 27.5 35.0

Ratio .862 .813 .792 .783 .742 .631 .645 .506 .564 .543

Source: The Monitoring the Future Study, the University of Michigan

Table 3. Ratios of Non-Believers to Regular Marijuana Users

Education Age 18-25 Age 26-34 Age 35+ All Ages

1993 All Levels .165 .095 .039 .078

1992 < High School .242 .180 .058 .127

High School .177 .126 .034 .087

Some College .130 .130 .022 .074

College Grad .073 .079 .030 .040

All Levels .161 .120 .033 .080

1993 < High School .253 .193 .040 .123

High School .192 .122 .051 .098

Some College .165 .081 .055 .090

College Grad .094 .041 .021 .034

All Levels .183 .097 .041 .083

Source: National Household Survey, SAMSHA Advance Report #5, 3/94; #7, 7/74.

2. During the same years, the ratio would plummet as age and education increased.

Marijuana policy had high school students convinced marijuana was dangerous for a while, but then lost its credibility. This supports present policy. However marijuana policy loses credibility as people grow and exercise independent reasoning skills. See Table 3.

.