7. RESULTADOS
7.5 PASO 5: DEFINICIÓN DE RANGOS ACEPTABLES PARA LOS
characteristic of some forms of bulimia nervosa, and also carries health risks of its own (Fairbum and Cooper 1984), and the number of girls involved in this extreme behaviour seems is worryingly high, although somewhat lower than levels reported in studies of North American adolescents (French, Story, Downes, Resnick and Blum 1995; Serdula et al 1993). It should be noted, however, that questions relating to past purging do not distinguish between regular, habitual vomiters and those who may have experimented with it once or twice. The question of the frequency and patterning of vomiting episodes in the present sample will be further examined in the longitudinal analyses. Although numbers o f girls involved in each unhealthy practice were fairly low compared with involvement in each healthy practice, altogether the majority (66%) of weight controllers reported having used at least one unhealthy method in the past.
Factor analysis of the weight control behaviours supported the proposition that healthy and unhealthy behaviours tend to cluster together. The strongest factor identified all o f the weight control methods identified by a dietician as ‘unhealthy’, together with two of the ‘neutral’ behaviours, whilst the second factor identified all four of the behaviours classified as ‘healthy’. Reducing starchy foods loaded just below the threshold on both the healthy and unhealthy factors. This may reflect conflicting advice about the desirability and health value o f this approach to weight control. The results of the factor analysis support the validity of dividing dieters into ‘healthy’ and ‘unhealthy’ groups according to their weight control style.
7.7.1 Characteristics Associated with Healthy and Unhealthy Weight Control
Whilst results reported in chapter 5 showed that socio-economic status was not significantly associated with dieting intensity, the analyses presented in this chapter indicate that it is associated with the choice of weight control behaviours
Chapter 7. Cross-sectional analyses: healthy and unhealthy dieting behaviours
grt>up
amongst those who control their weight. Girls in the healthy weight controyiad a significantly higher level o f family affluence than those in the unhealthy group. This finding is similar to those reported in adult samples in Britain and the US which suggest women of high SES are more likely than those of lower SES use healthy, but not unhealthy weight control methods (Wardle and Griffith 2000; French, Perry, Leon and Fulkerson 1995^)
There were no significant associations between age and weight control group, and contrary to expectations BMI did not distinguish between the healthy and unhealthy weight control groups. BMI was, however, significantly higher amongst both weight control groups than amongst non weight controllers. Puberty was associated with unhealthy but not healthy weight control, supporting the proposition that early development may be associated with weight concerns and pathological eating behaviours (Duncan 1985; Wichstrom 1995). All three weight control groups differed for body dissatisfaction, the most intense dissatisfaction linked to the more extreme and risky weight control.
The only significant between-group differences for psychological distress were found between the unhealthy weight controllers and the non-weight controllers. In each case the healthy weight controllers were intermediate but not significantly different to the non weight controllers. These differences were not due simply to the more intense dieting of unhealthy weight controllers, as these effects persisted even after dieting intensity was controlled for. Unhealthy weight controllers were also more conscious of and more accepting of social preferences for thinness than healthy weight controllers and they were more likely to favour a low body weight and to set lower thresholds for the point at which a body is too fat or too thin. As in the case of the psychological distress analyses, differences between the healthy and unhealthy groups were independent of the more intense dieting of unhealthy weight controllers. Healthy weight controllers differed significantly from non weight controllers on all but the classification of a body as too fat. These findings seem to suggest that whilst negative attitudes towards body fat characterise all weight controllers, elevated
Chapter 7. Cross-sectional analyses: healthy and unhealthy dieting behaviours
levels o f psychological distress are typical only of those girls whose repertoire of weight control methods includes unhealthy behaviours.
Both healthy and unhealthy weight controllers were more likely than non weight controllers to eat a low fat diet and practice healthy eating habits. Unhealthy weight controllers were also more likely to eat the recommended daily number of portions of fruit and vegetables, and results for the healthy weight controllers approached significance. These results stand in contrast with those from a study of US adolescents (Story et al 1998) which found that only moderate, not extreme, dieters had elevated levels o f health promoting dietary behaviours and physical activity. The difference between these two studies may reflect the more stringent criteria for classifying individuals as unhealthy dieter used in the US study, in which only those reporting vomiting or use o f dieting pills were classified as extreme dieters. The classification system employed in the present study used much less stringent criteria, according to which an individual could be classified as an unhealthy weight controller on the basis of much less extreme methods such as skipping meals. This means that the unhealthy group consists of a wide range o f individuals, including some whose weight control consists primarily of healthy methods.
Unhealthy weight controllers were much more likely than non dieters to show signs of disordered eating. There were no significant elevations in the likelihood of healthy weight controllers showing signs of disordered eating. Significant differences between healthy and unhealthy weight controllers in anorexic symptoms and bulimic binge eating were independent of differences in the dieting intensity o f the two groups.
There is a certain amount of circularity when examining links between unhealthy weight control and disordered eating (and similarly links between healthy weight control and health behaviours), since weight control style is largely defined using parameters that are closely related to the behaviours being studied. The value o f the analyses presented here lies largely in providing the opportunity to
Chapter 7. Cross-sectional analyses: healthy and unhealthy dieting behaviours
distinguish between the effects of weight control as a generic activity and the specific kinds of activities that it involves. It should be noted that whilst the healthy weight control group excluded anyone involved in unhealthy methods, the unhealthy group includes those using both unhealthy and healthy methods, as numbers of participants using only unhealthy methods were too small for any meaningful analysis. This means that the results relating to behaviours must be interpreted with care.
In the case o f the absence o f links between disordered eating and healthy weight control it seems fair to conclude that involvement in healthy weight control has at most minimal direct connections with serious eating problems. Unhealthy weight control, which is defined largely in terms o f milder forms of the behaviours that characterise eating disorders, may be a factor which predisposes individuals to develop further eating problems, or may itself simply be a symptom of a pre-existing tendency towards disordered eating behaviour.
In the case of healthy eating behaviour the results are more difficult to interpret. Both healthy and unhealthy dieting have a positive effect on healthy eating behaviour, but since the majority o f the unhealthy group use healthy methods as well as unhealthy methods it is not possible to conclude that unhealthy weight control methods are linked to healthy eating behaviour. What can be asserted, however, is that the majority o f dieters, including many o f those involved in practices which would not be recommended, report more healthy eating habits, a lower level of dietary fat and higher consumption o f fruit and vegetables than non dieters.