HERALD-TRIBUNE ANALYZES '02 TEEN DRUG USE SURVEY

The 2002 Florida Youth Substance Abuse Survey reported that the
state's biggest teen drug problem wasn't in Miami, Tampa or any county
with an urban center.

According to the survey, Charlotte County ranks as the state's worst
when it comes to teen substance abuse.

The Herald-Tribune analyzed the results and methodology of the survey.
The review included interviews with experts in the field of surveys
and teen drug use, as well as with representatives of the company that
created and administered the survey. The goal of the analysis was to
answer some of the questions a skeptic of the survey's findings would
have, such as:

Were there flaws in the way the survey was administered?

Almost 63,000 of Florida's sixth-through 12th-grade students took the
survey, including 1,039 in Charlotte County. Students were told to
answer honestly and that they would remain anonymous.

A spokesman for the survey company said that in every case, a teacher
or teacher's aide administered the survey and law enforcement officers
weren't involved. Chris McCarty, the survey director at the University
of Florida Survey Research Center, who was asked to review the survey,
said the ideal way to conduct it would have been to put the name of
every Florida student into a computer and let the computer randomly
pick the students to be surveyed.

"That means that even if one student in one school in rural Bay County
is picked, you send someone out to do the survey," McCarty said.

But McCarty said that method isn't cost-effective, so survey officials
used a "cluster sample." First a school was picked, and then
individual classrooms within the school were selected to participate.
"It's a fair way to do a survey, but there are consequences," McCarty
said. "For example, there are schools known more as the druggie
schools than others in the district. Did they happen to pick that
school, or not?"

Officials with the Department of Children & Families, the state agency
responsible for the survey, said Charlotte County Vocational-Technical
Center and every middle and high school in the county participated in
the survey.

Charlotte Harbor Center, a second-chance school where kids caught with
drugs are sometimes sent, was not one of the schools involved. # What
if Charlotte's students exaggerated their answers? Three methods were
used to identify phony answers. Surveys were eliminated when students
embellished their drug use by selecting the highest levels of use for
every drug.

Second, three questions related to a fictional drug, Derbisol. Results
were invalidated when students reported on two out of three questions
that they'd used Derbisol.

Finally, when students gave inconsistent answers, such as reporting
that they had smoked marijuana in the last 30 days and that they had
never tried marijuana, their questionnaires were discarded. Did
Charlotte's demographics have an effect on the survey results?

There are two competing stereotypes about which kids do drugs:
well-to-do whites who have the money to buy drugs or minorities in
inner cities.

Actually, neither is true, said Howard Simon, a spokesman for
Partnership for a Drug-Free America.

"When you talk about where kids are using, and what kids are likely to
use, there is no typical drug user," he said. "It's not just the
quote-unquote poor kids who use, it's not just the quote-unquote bad
kids." # Was Charlotte County's student body fairly represented in the
surveyed population?

A potential problem does exist in the demographic makeup of the
Charlotte County sample.

The report states that almost 5 percent of the Charlotte students
surveyed were American Indian. However, less than 1 percent of
Charlotte's eligible survey population was American Indian that year,
according to statistics from the Department of Education. The
Herald-Tribune also found cases in other counties where the American
Indian representation was higher in the state's survey than in the
eligible survey population of sixth-through 12th-graders.

DCF officials had no explanation for the discrepancy.

"To my knowledge, our goal is to get a representative sample," said
Ken DeCerchio, the director of Substance Abuse Programs for DCF. "It
may not always be possible. I don't see it as a problem with the
validity of the survey." An associate research scientist employed by
the survey company said the high percentage of American Indians isn't
a concern because the report is based on countywide drug use, not use
among various ethnic groups.

But McCarty said the demographic irregularity can't be
dismissed.

"That's a severe oversampling of American Indians," he said.
"Culturally, they're different. It doesn't invalidate the survey, but
it's a factor to consider."

Was this an unbiased survey? DCF paid a Massachusetts-based company
called Channing Bete more than $270,000 to administer the survey.

Channing Bete is primarily a publishing company that sells anti-drug,
anti-smoking and related literature.

According to state purchasing records, Channing Bete has sold more
than $400,000 worth of its products to various government agencies in
Florida over the past three years.

Charlotte County Commissioner Tom D'Aprile said the business
relationship is a conflict of interest that taints the survey results.
"I truly believe there's something else behind this survey. Those
stats aren't accurate at all," he said.

Dan Carmody, a Channing Bete spokesman, said there is no conflict of
interest.

Channing Bete wouldn't provide its salespeople with the survey
results, Carmody said. But a salesperson might refer to the public
announcement of a drop in teen drug use by Gov. Jeb Bush and credit
the company's anti-drug brochures for the decline, he said.

"If it's in the public domain, sure, it could be used," Carmody said.
"Any salesman could use info in the public domain."