Reprinted from Eating Disorders Review
July/August 1999 Volume 10, Number 4
©1999 Gürze Books
Eating problems in early childhood can predict more severe eating disturbances in adolescence. However, until recently, there were no tests specifically designed to detect factors that might place young girls at risk of developing eating disorders. In addition, the youngest groups studied have generally been middle school students. Almost no studies have been done among girls in elementary school.
The McKnight Risk Factor Survey-III (MRSF-III) is a questionnaire that examines a number of potential risk “domains” among girls as early as the 4th grade. A recent study found the MRSF-III to be a useful instrument to assess potential risks and protective factors for the development of disordered eating in preadolescent and adolescent girls (Int J Eat Disord 25:195, 1999).
Methods
Several versions of the MRSF were pilot- tested before the MRSF-III was administered to 134 elementary school girls (4th and 5th graders), 243 middle school girls (6th to 8th grade) and 274 high school girls (9th to 12th grade) in Arizona and California.
The ethnic background of the students was as follows: White, 52%; Hispanic, 24%; Asian-American, 13%; African-American, 4%; Native American, 2%, and “Other,” 3%.
The Weight Concerns Scale, the Rosenberg Self-Esteem Scale, the Center for Epidemiological Studied Depression scale and the Child Depression Inventory were included in the administration of the MRSF-III to assess the validity of the new test.
Because of concerns that the younger girls might have difficulty reading and understanding some of the items on the MRFS-III, the Arizona investigators read the questionnaire aloud to 99 4th and 5th graders. The girls were invited to ask questions about any question they did not understand.
Results
The test-retest reliabilities of the MRFS scales were high across all age groups, except for a few cases among the elementary students, according to Dr. Catherine Shisslak, of the University of Arizona, Tucson, who directed the study.
For elementary, middle school, and high school students, the overall test-retest reliabilities for weight control behaviors were high. No significant differences were found between the oral and the non-oral tests.
A few limitations
Dr. Shisslak also outlined a few limitations for the questionnaire. She believes that there are two possible reasons why the test was less consistent and reliable among the elementary school group.
First, the young students usually have lower levels of cognitive functioning that vary more over time than do those in middle or high school students. Like all self-report instruments, distortions and inaccurate reporting can occur.
Second, the younger girls may not have understood some of the terms and might have interpreted them differently from the older girls. Dr. Shisslak notes, for example, that elementary school students may not clearly understand what a “binge” is.
