Reading magazine articles about dieting and associated weight control behaviors among adolescents

Study population and study design 

The current study is a secondary data analysis using data drawn from Project EAT (Eating Among Teens), a school-based survey of nutrition and weight concerns among adolescents. The study sample consisted of 4746 middle and high school students. The University of Minnesota Human Subjects Committee granted approval for the current study. Additional detail about study design and methodology has been documented elsewhere [12].

Measures 

Magazine Reading was assessed with a single item question, “How often do you read magazine articles in which dieting or weight loss are discussed?” Response categories were: “never,” “hardly ever,” “sometimes,” or “often.” Frequent reading refers to those indicating reading magazine articles about dieting either “sometimes” or “often.”

Healthy, unhealthy, and extremely unhealthy weight control behaviors were assessed by the question, “Have you done any of the following things in order to lose weight or keep from gaining weight in the past year?” followed by a series of behaviors in random order.

Binge eating was assessed with the one question, “In the past year, have you ever eaten so much food in a short period of time that you would be embarrassed if others saw you (binge eating)?” (“yes”/“no”).

Body satisfaction was assessed using a 10-item modified version of the Body Satisfaction Scale [13] (Cronbach alpha = .92). Self-esteem was assessed using a modified version of Rosenberg’s Self Esteem Questionnaire [14] (Cronbach alpha = .78). Depressive mood was assessed by indicating the extent of experiencing one or more symptoms from a seven-item checklist [15] (Cronbach alpha = .82).

Weight status was derived from height and weight measurements taken by trained research staff using standardized equipment and procedures. Gender- and age-specific cut points for body mass index (BMI) were based on reference data from the Centers for Disease Control and Prevention growth charts [16].

Gender, school level, ethnicity/race, and socioeconomic status (SES) were based on self-report. Ethnicity/race was assessed using a single item of self-report of ethnic identity. The construction of the SES variable has been described previously [12]; the prime determinant of SES was parental level of education.

Analysis 

Chi-square tests of independence and Cochrane-Armitage Tests for Trend were used to compare the sociodemographic characteristics of adolescents who frequently read magazine articles about dieting (Table 1). Logistic and linear regression models were used to test the associations between reading magazines about dieting and weight control behaviors and psychosocial distress, controlling for age, weight status, SES, and race/ethnicity, and stratified by gender (Table 2, Table 3). As each of the psychosocial variables was significantly associated with weight control behaviors, the three psychosocial variables were then added as covariates to the logistic models. In a final step, the interaction between magazinereading and each of the psychosocial variables (e.g., self-esteem, depressive mood, and body satisfaction) were individually entered into logistic regression models to predict unhealthy weight control behaviors. All analyses were conducted using SAS (Release 8.0) [17].

Table 1. Numbers and Percentages of Frequent1 Readers of Dieting/Weight Loss Articles by Sociodemographic Characteristics by Gender

		Girls			Boys
		n	%	p value	n	%	p value
	Total	1016	45.6		323	15.4
	School level
	Middle school	299	40.1		125	16.9
	High school	709	48.8	0.001	196	13.3	.024
	Weight status4
	Underweight	25	38.5		7	6.8
	Average weight	571	44.6		176	13.6
	Moderately overweight	196	46.6		45	12.6
	Very overweight	117	47.4	0.1992	61	20.1	.0022
	Race/ethnicity
	White	511	49.2a3		88	7.4d
	Asian	215	47.4a		88	23.1ef
	Hispanic	52	47.3ac		28	20.3ef
	African-American	156	38.9bc		14	25.1e
	Native American	31	35.2bc	0.004	9	12.9df	.001
	Other	42	42.0ac		92	20.0ef
	SES group
	Low	180	41.4		85	27.2
	Low-middle	189	45.4		59	14.9
	Middle	264	47.5		88	14.7
	Upper-middle	219	46.6		52	9.4
	High	136	46.0	0.1532	27	8.5	< .0012

SES = Socioeconomic status.

1 Report reading sometimes or often. 2 Cochran-Armitage Test for Trend. 3 Proportions with different letters are significantly different (p < .05). Proportions with the same letter are not. 4 Respondents were classified as underweight (BMI <15th percentile), average weight (BMI 15th percentile to <85th percentile), moderately overweight (BMI >85th percentile to <95th percentile), and very overweight (BMI ≥95th) percentile.

Table 2. Odds Ratios (OR) and 95% Confidence Intervals (CI) for the Associations of Weight Control Behaviorsa and Frequency of Reading Magazine Articles About Dieting and Weight Loss Among Girls, Controlling for BMI, Age, Race, and SES

	Variable	Healthy Weight Control Behaviors (n = 1975)		Unhealthy Weight Control Behaviors (n = 1330)		Extremely Unhealthy Weight Control Behaviors (n = 287)		Binge-Eating (n = 391)
		OR	CI	OR	CI	OR	CI	OR	CI
	Girls magazine reading
	Never (reference)	1.0		1.0		1.0		1.0
	Hardly ever	2.7	1.9–3.7	1.8	1.4–2.4	1.0	0.6–1.7	1.5	1.0–2.2
	Sometimes	4.8	3.4–7.0	3.3	2.5–4.2	2.7	1.8–4.2	2.4	1.7–3.5
	Often	6.2	3.0–12.7	7.3	4.7–11.4	6.0	3.6–10.0	3.9	2.5–6.2
	Boys magazine reading
	Never (reference)	1.0		1.0		1.0		1.0
	Hardly ever	1.8	1.4–2.5	1.8	1.4–2.4	1.8	1.0–3.4	1.5	0.9–2.3
	Sometimes	2.9	1.9–4.5	2.8	2.0–3.8	2.8	1.5–5.1	1.8	1.1–2.9
	Often	3.5	1.3–9.4	2.1	1.0–4.1	4.0	1.4–11.9	1.9	0.7–5.3

BMI = body mass index; SES = socioeconomic status.

a Healthy weight control behaviors included the responses: exercising, eating more fruits and vegetables, eating less high-fat foods, and eating less sweets. Unhealthy weight control behaviors included the responses: fasting, eating very little food, taking diet pills, vomiting, using laxatives, using diuretics, using food substitutes, skipping meals, and smoking more cigarettes. The extremely unhealthy behaviors comprise the subset of the unhealthy weight control behaviors that include taking diet pills, vomiting, using laxatives, and using diuretics.

Table 3. Mean Scores for Self-Esteem, Depression, and Body Satisfaction by Frequency of Magazine Reading Among Girls and Boys Controlling for Age, Race, BMI, and SES

		Self-Esteema		Depressive Moodb		Body Satisfactionc
		Girls	Boys	Girls	Boys	Girls	Boys
	Magazine reading
	Never	18.0	19.0	12.0	10.8	34.2	37.2
	Hardly ever	17.9	18.6	12.4	11.2	33.5	36.4
	Sometimes	16.9	17.9	12.8	11.4	30.8	36.3
	Often	16.1	18.0	13.9	11.6	27.7	39.9
	p value	< .001	< .001	< .001	.003	< .001	.034

BMI = body mass index; SES = socioeconomic status.

a Possible scores ranged from 4 to 24; a high score indicates high level of self-esteem. b Possible scores ranged from 7 to 21; a low score indicates low level of depressive mood. c Possible scores ranged from 10 to 50; a high score indicates high level of body satisfaction.

Demographic characteristics of youth reading magazine articles about dieting 

Among girls, there were no significant trends of frequent reading by BMI or SES. In contrast, boys who were nonwhite, in the higher BMI percentiles, and among the lower SES groups were significantly more likely to report reading articles about dieting/weight loss (Table 1).

Frequency of magazine reading about dieting and weight control behaviors 

Strong associations were found for weight control behaviors and binge eating with increasing frequency of magazine reading (Table 2). When depressive mood, self-esteem, and body satisfaction were added as covariates to the logistic models, all of the odds ratios of engaging in each of the weight control behaviors and binge eating remained significant for the girls, and all but two remained significant for the boys (data not shown).

Frequency of magazine reading and psychosocial indicators 

The adjusted mean scores for self-esteem, depressive mood, and body satisfaction by frequency of magazine reading are displayed in Table 3. Overall, as frequency of reading dieting-related magazine articles increased, adolescents scored poorer for indicators of psychosocial well-being.

Moderating factors 

As frequent reading of dieting-related magazine articles was associated with unhealthy weight control behaviors, we hypothesized that the relationship between magazine reading and unhealthy weight control behaviors might be moderated by the students’ psychosocial distress. For example, we hypothesized that the main association would be heightened by high levels of depressive mood, poor self-esteem, or body dissatisfaction. To test this hypothesis, interaction variables magazine reading and each of the psychosocial variables (e.g., self-esteem, depressive mood, and body satisfaction) were created and each was entered in a logistic regression model stratified by gender. The analysis concluded that none of the interaction effects were statistically significant in further explaining the association of magazine reading on unhealthy weight control behaviors for either boys or girls (data not shown).