PROBLEMS WITH A SURVEY
 THAT WAS WIDELY
 PUBLICIZED
 SCRIPPS HOWARD NEWS
 SERVICE

 Date: Friday, July 30, 1999

 Section: Source: By LINDA SEEBACH

 Scripps Howard News Service

 Memo: COLUMN

 Embargoed for SUNDAY release
 (Linda Seebach is an editorial writer for the Denver
 Rocky Mountain News.)

 Edition: 

 Let someone start talking about children, and the brains
 of normally skeptical adults turn to mush.

 The latest example is the release of "Top Ten States to
 Raise a Child," an annual non-event brought to you since
 1995 by an advocacy group called the Children's Rights
 Council.

 It drew media attention, of course. Maine is the new
 holder of first place. Last year's winner, Iowa, dropped
 to 10th.

 The rankings are based on 10 criteria: cases of child
 abuse and neglect, children not immunized by age 2, the
 high school dropout rate, percentage of children in
 poverty, the child death rate, the infant mortality rate,
 lack of early prenatal care, juvenile arrests, teen births
 and number of divorces.

 Well, all those things have something to do with
 children's welfare, to be sure, though they probably
 wouldn't be the top 10 on most people's lists of what
 makes a community a good place to raise kids. Parents
 care about the quality of their local schools, but the
 dropout rate is only one indication of that. Furthermore,
 the statewide dropout rate implies nothing about the
 school district they live in or the individual schools their
 children will attend.

 Still, it is difficult to argue against any of these as
 measures of well-being for children and adolescents. It's
 what the group does with those numbers that makes the
 whole exercise preposterous.

 Some of their criteria are percentages, but referring to
 different groups; the percentage of 2-year-olds that have
 not been immunized, the percentage of high school
 students who drop out. Others are ratios with different
 denominators; the child death rate, for instance, is
 calculated on the basis of deaths per 100,000. And for
 some indicators, like the number of juvenile arrests and
 the number of divorces, they divide the number by the
 state's population.

 They convert all these figures to decimals, add them up
 and divide by 10.

 Unfortunately for their purposes, averaging percentages
 of different-sized groups makes arithmetic nonsense. If
 half the children at a school buy lunch in the cafeteria,
 but only 10 percent of the teachers, you are not entitled
 to conclude that on average, 30 percent (50 plus 10,
 divided by 2) of the school buys lunch.

 Actually it makes even less sense than that; if half the
 children buy lunch and 10 percent of the teachers have
 yogurt for breakfast, is the school's nutrition ranking 30
 percent. 

 But that's not all. You may remember the story about the
 museum tourist who told his companions that the sun
 was 5 billion and 3 years old. He knew, he said, because
 he had visited the museum three years ago and it was 5
 billion then.

 Nobody at the Children's Rights Council can possibly
 understand why that's funny, or they would have known
 better than to have computed their meaningless average
 to six decimal places, though the original percentages
 have only two significant digits.

 The report doesn't even apply its own formula
 consistently. Five states are missing divorce statistics.
 Juvenile crime is missing from others.

 Also, state rankings changed this year, the council's
 legislative assistant told AP, because the group changed
 the criteria.

 They discovered that data on single-parent households,
 unwed births and teen birth overlapped, so this year they
 dropped two of the three. And they also dropped data on
 deaths from drugs and alcohol because not all of them
 affected children.

 But not all divorces affect children, either. And if the
 council is genuinely worried about overlapping categories,
 the ones it used this year overlap plenty. As they know,
 because they themselves say elsewhere on their Web site,
 "Children raised in single parent families are at great risk
 for juvenile delinquency, teen pregnancy, poor grades,
 drugs, dropping out of school and other trouble."

 These people are advocates, after all, and they mean
 well. But that doesn't explain why anybody else should
 take their bizarre arithmetic seriously, or why, year after
 year, reporters write about it as if it made sense.