Factors influencing women’s choice of weight-loss diet

Abstract

We developed a typology of eight minimally overlapping weight-loss diet methods and used it to survey 151 women dieters on their choice of diet in the previous 12 months, their motivations to diet, and their eating disorder symptomatology. Canonical correlations revealed a potentially problematic “thin, quick, and easy” association of methods and motives, as well as a more healthful “thin, natural, life-style” association. Both featured the pursuit of thinness but not health. In fact, health was rated by dieters as the poorest motivator of dieting. The results highlight the importance to women dieters of short-term aesthetic concerns over long-term health.

Keywords

diet eating behavior eating disorders health behavior overweight women’s health

Introduction

Dieting for weight loss describes any modification to food consumption that is intended to reduce body weight, adiposity, and/or body size (McFarlane et al., 1999). In comparison to our knowledge of why people diet, why they achieve or fail to achieve their weight-loss goals (Green et al., 2009), how and why they change their diet (Chapman and Ogden, 2009), and the physical (e.g. National Institute for Health and Clinical Excellence (NICE), 2006) and mental health (e.g. Stice, 2002) consequences of dieting, we know little about the implications of the choice of diet method. In this study, we surveyed women dieters to discover which diet outcomes are important to them and how this relates to their choice of diet method and their disordered eating symptomatology. Due to the plethora of diets and diet products, we commenced with a systematic categorization of diet methods.

Weight-loss dieting

Success rates for long-term weight-loss dieting are low (Mann et al., 2007). Despite this, a large majority of women have tried diets, many making multiple attempts at weight loss (French et al., 1999). Dieting is a substantial industry, one that is highly profitable, earning billions of dollars each year worldwide (Berzins, 1999), with approximately AUD790 million per annum of earnings in Australia alone (Stark, 2012). Dieting is also a mainstream activity with important implications for physical and psychological wellbeing, and yet diet products are minimally regulated (Rose, 2010) and readily available from a variety of everyday shopping outlets. The media and Internet targeting female audiences also often focus on dieting strategies, diet plans, or diet “tips” (Van den Berg et al., 2007).

Concerns have been raised about the quantitative constraints imposed by diets—in terms of energy restriction, frequency and size of meals, type of energy consumed (from natural foods to supplements), proportion of nutritional components (e.g. carbohydrates, proteins, and fats) (e.g. Shikany et al., 2007)—and the extent to which these constraints deviate from nutritional guidelines for physical health (Freedman et al., 2001). In addition, mental health concerns have been raised with diets that encourage inflexible thinking, unrealistic goals, or perfectionism characteristic of anorexic symptoms (Stice et al., 2010), or the imposition of strict rules about certain foods at certain times characteristic of bulimic symptoms (Westenhoefer, 1991).

Diets that deviate from standard nutritional guidelines are also known to be more difficult to maintain (Herriot et al., 2008). However, consumers of such diets, particularly those with pre-existing body image concerns, may not attribute failure to achieve or maintain weight loss to limitations of the diet itself, but as failure on their part, and thus experience negative self-esteem, negative affect, and poor self-image (McFarlane et al., 1999). Furthermore, there is evidence that such individuals are more likely to start another equally unsuccessful diet, engage in “yo-yo” dieting, and/or binge eating (e.g. Amigo and Fernández, 2007).

Diet motives and methods

Given the prevalence of, and problems inherent with, weight-loss dieting, it is surprising that the method of diet has received so little empirical attention. Instead, diet behaviors have been quantified in terms of generic psychological and/or behavioral factors relevant to dieting, such as intention to diet, diet frequency, duration, or intensity of effort around dieting. In its simplest form, this has involved a dichotomous response to the question of whether one is or is not dieting (e.g. Jenks and Higgs, 2010). As a basic form of gauging problematic dieting, studies have assigned an arbitrary point at which diet frequency becomes “chronic dieting” (e.g. Neumark-Sztainer et al., 2002). Where specific diet methods have been cited, this has typically been in the context of a comparison of low-incidence behavior. For example, research has measured the frequency of artificial sweetener or diet pill use (e.g. Yahia et al., 2011).

Weight-loss dieting has also been equated to cognitive and/or behavioral “restraint” of food. Perhaps the most widely recognized scale in this research area, since its inception in 1975, is the Restraint Scale that was developed to identify chronic dieters (Herman and Mack, 1975). This measure has two subscales labeled “Concern for Dieting” and “Weight Fluctuations.” As the subscale names suggest, their items assess diet-related cognitions (e.g. “Do you give too much time and thought to food”) and outcomes of dieting (e.g. “What is the maximum amount of weight you have ever lost in one month?”), rather than a specific focus on the various dieting behaviors of particular methods.

Of particular relevance to this study, there is also limited research into the link between motives to diet and the diet method chosen. The research that exists has only considered diet choice in very broad terms. This research has revealed two basic motivations for dieting: weight loss to improve appearance and weight loss to improve or prevent health issues, with appearance motivations more likely to be associated with unnatural and potentially unhealthy strategies such as excluding food groups, vomiting, skipping meals, and using laxatives and diuretics (Putterman and Linden, 2004).

Research aims

In Study 1, an analysis of explicit and implicit attributes of diet products and strategies was used to develop a typology of “diet method,” and an analysis of the associated claims/outcomes of the diet was used to develop a typology of outcome-relevant beliefs about dieting. Using these typologies, an online survey was conducted in Study 2 to examine women’s motives for dieting in relation to their choice of diet method and their eating disorder symptomatology.

Study 1

The absence of a standardized measure to assess diet methods was noted by Putterman and Linden (2004). A thorough examination conducted by the authors of psychological, health, and medical journals, using the databases PsycTESTS, PsycINFO, PsycARTICLES, and MEDLINE Complete, failed to uncover any instrument that purports to categorize and measure diet methodology. Hence, the first step was to survey the diversity of methods, followed by classification of these methods via qualitative content analysis (Hsieh and Shannon, 2005). In the initial stage of this process, a comprehensive list of diet products was created. A total of 84 diet products and 139 marketing claims were included in these analyses. They were identified from descriptions of diet strategies, diet products, and diet claims/outcomes from promotional material or personal anecdotes; from interviews with laypeople and health professionals; and from online, television, and magazine sources. The name of the product, a brief description of how dieters follow the method, and claims concerning the positive and (where available) negative outcomes of the method were collated from the following sources:

Retail outlets (Melbourne, Australia);

Television advertisements (Melbourne, Australia);

The World Wide Web, using the search terms “diet,” “weight loss diet,” and “diet plan” which were entered into the search engines Google and Bing, and the first 50 sponsored and unsponsored links were investigated for their relevance;

Seven laypeople aged in their 20s and 30s were interviewed about their history of diet product use;

Three health professionals working in the obesity field were consulted.

Diet method typology

Descriptions of how to follow diet strategies and products were examined for common themes using conventional qualitative content analysis. Initially, diets were separated into two categories: those that provide instruction only, and those that provide a physical product such as prepared meals or powders. Then, themes in the form of “methods” were examined within these groups. Table 1 presents the diet methods extracted, including a brief description of their requisite dieting behavior and/or rules, and examples of commercial products/strategies which qualify.

Table 1.

Typology of diet methods and diet outcomes.

Diet method	Description	Example product
Meal Plan	Provides the dieter with a daily meal plan to follow. These diets do not provide pre-prepared food, but they give a clear plan of what the dieter should eat at each meal	CSIRO total wellbeing diet, GutBusters
Calorie Counting	Requires the dieter to count calories and restrict the number of calories they consume each day	Weight Watchers, Calorie King Australia
Low GI	Requires the dieter to eat foods that have a low GI rating (the rate at which the food raises blood sugar level)	Ultimate EnerGI Health Club, The Low GI Diet: Lose Weight With Smart Carbs
High Protein/Low Carb	Encourages the dieter to consume high-protein foods (e.g. meat, nuts, and eggs) at the expense of high-carbohydrate foods (e.g. grains and pastas)	The Atkins Diet, Zone Diet
Meal Replacement	Provides the dieter with specifically formulated weight-loss products (e.g. “shakes,” soups, mousse, or “bars”) that replace most or all of their normal daily meals	Celebrity Slim, Optislim, Tony Ferguson Weightloss Program
Pre-Prepared Meal	Provides the dieter with pre-prepared meals	Lite n’ Easy; Jenny Craig
Supplement	Chemical supplements are added to the dieter’s normal diet and involve the consumption of pills, powders, or liquids that claim to burn fat, enhance metabolism, suppress appetite, or cleanse/detoxify the body	Blackmores Weight-Loss Accelerate, Xantrax tablets
Special Food	Requires the dieter to eat “special” foods or drinks throughout the day; these foods or drinks either replace normal meals or are to be consumed with every meal	Cabbage Soup Diet, Wu-Yi Tea, The Lemon Detox Diet
Diet outcome	Variable label	Example claim
Thinness (preoccupied with thinness)	Thinness	How to burn stomach fat
Health (concerned with health)	Health	The Weight-Loss Diet Meal Plans model healthy eating patterns that can be followed
Permanent weight loss	Permanent Weight Loss	Keep Weight Off For The Long Haul
Achieving quick weight loss	Quick Weight Loss	Fast track your weight loss and fitness with The Biggest Loser Express Kit
Making lifelong changes	Lifelong Changes	You can … practice a lifelong diet to maintain a healthy weight
Requiring little effort	Low Effort	Discover how to lose weight … easily with our simple weight-loss diet
Not interfering with things I enjoy	Easily Integrated	The Flexible Way to Lose Weight & Still Live Your Life
Knowing how the diet works	How Diet Works	Learn more with Bodytrim
Cost of diet product	Cost	Tony Ferguson Offers An Affordable Weightloss Solution That Works
Avoiding unnatural diet product	Natural	Natural Diet Products to help you lose weight safely

CSIRO: Commonwealth Scientific and Industrial Research Organisation; GI: glycemic index.

Diet outcomes

Diet claims, diet product marketing materials, and discussions with laypeople and health professionals resulted in the identification of 139 diet claims (both explicit and implicit) concerning diet outcomes. These included reference to characteristics of the weight-loss claims made by the diet, through to the claims unrelated to the dieter’s weight loss. From these claims, 10 diet outcomes were extracted and these are presented in Table 1, as well as the abbreviated label name assigned to them, and an example of a corresponding claim relating to each outcome.

Study 2

The diet methods and diet outcomes extracted in Study 1 were used to evaluate diet decision-making in a sample of 151 women who were dieting. They completed an online survey that measured their frequency of use of each diet method identified in Study 1, as well as symptoms of eating pathology. To measure outcome evaluations relevant to particular diet methods (cf. the theory of reasoned action and theory of planned behavior, for example, Miller, 2005), participants also provided a weighting of the importance to them of each diet outcome identified in Study 1. This approach has been successful in quantifying food choice outcomes in non-dieting contexts (e.g. Shepherd and Towler, 1992) and was extended in this study to investigate relationships between the importance that dieters place on particular diet outcomes, their choice of diet method, and disordered eating symptomatology.

In terms of achieving sustainable weight loss, the consensus view is that a desire for rapid weight loss is counterproductive (NICE, 2006), that successful dieting is best motivated by long-term health concerns and not a short-term drive for thinness (Putterman and Linden, 2004). Therefore, it was hypothesized that a diet method–outcome analysis would reveal a split between health/long-term versus appearance/short-term cognitions and further reveal which diet methods are associated with these profiles.

There is ample evidence that dieting increases the risk of developing an eating disorder (e.g. Bohon et al., 2009); however, the nature of this relationship remains unclear (e.g. Stice, 2002). The focus in this study on diet methods aimed to further our understanding of the relationship between dieting and eating pathology by identifying which diet methods, outcomes, and method–outcome combinations are associated with eating pathologies, specifically, preoccupation with eating and high levels of eating restraint. It was hypothesized that diet methods motivated by a drive for thinness and short-term outcomes would be associated with disordered eating symptomatology.

Method

Participants

A sample of 151 women who described themselves as currently on a weight-loss diet completed the survey. Most resided in Australia (93.4%) with the remaining participants from other developed countries that use English as their predominant language. Their average age was 30.9 years (standard deviation (SD) = 11.7 years), with 39 percent below 24 years, 54 percent between 25 and 54 years, and 6 percent over 55 years. In comparison to recent Australian census data, young adults between 18 and 25 years were overrepresented in our sample (cf. 39% in our sample vs 14% in the general population; Australia Age Structure, 2014). Application of World Health Organization (WHO, 2006) body mass index (BMI) classifications reveals that 5.3 percent of our participants were underweight (<18.5 kg/m²), 46.4 percent were within the normal weight range (18.5–24.9 kg/m²), and 48.3 percent were overweight (25.0–29.9 kg/m²) or obese (>30.0 kg/m²). The mean BMI (26.6, SD = 6.9) and proportion of obese women (22.3%) in our primarily Australian sample are comparable to Australian norms (BMI M = 26.9, proportion obese = 22.2%; Cameron et al., 2003), but substantially lower than that of North American women according to the recent National Health and Nutrition Examination Survey (BMI M = 28.7, proportion obese = 35.8%; Flegal et al., 2012).

Participants were recruited using a brief written invitation which contained the URL link to the online questionnaire. This invitation was distributed via email, announcements on Australian university online notice boards for various undergraduate units, and advertised on social networking and online classified advertisement sites. A snowball technique was also employed. Participation was voluntary and anonymous, and no inducement was offered.

Materials

The questionnaire included the following measures.

Demographic and weight information

Participants were asked to indicate their age, country, height, current weight, and “ideal” weight. Participants were also asked to report their history of weight-loss success using a 4-point Likert scale: “I lost more weight than I wanted to lose” (1); “I lost as much weight as I wanted to lose” (2); “I lost less weight than I wanted to lose” (3); and “I gained weight” (4).

Diet methods

Participants used Likert scales anchored between “never” (0) and “always” (10) to indicate their frequency of use of each diet method (identified in Study 1) over the previous 12 months. An expandable textbox also allowed dieters to describe their dieting methods in their own words.

Diet outcomes

Participants used Likert scales anchored between “never” (0) and “always” (4) to indicate their level of agreement with statements about the importance of each diet outcome.

Eating pathology

The Eating Disorder Examination-Questionnaire (EDE-Q) is a self-report questionnaire (Fairburn and Beglin, 1994) that assesses attitudinal, emotional, and behavioral symptoms of disordered eating over the past 4 weeks. Participants used Likert scales anchored between “no days” (0) and “every day” (6) to indicate their frequency of experiencing symptoms related to two subscales of the EDE-Q: Eating Restraint (Cronbach’s α in this study = .73) and Eating Concern (Cronbach’s α in this study = .78).

Results

Evaluating the typology of diet methods and diet outcomes

As expected, participant age was positively associated with BMI, r = .44, p < .01 (see Table 2). Increased age was also associated, albeit very weakly, with less use of calorie counting as a diet method and reduced importance of thinness and increased importance of making lifelong changes and knowing how diets work, r = −.18, −.24, .21, .16, p < .05, respectively. BMI was positively associated with a slight increase in use of pre-prepared meals, r = .21, p < .01, increased importance of health, ease of integration, naturalness, and cost, r = .28, .18, .17, and .21, p < .05, respectively, and reduced importance of low effort, r = −.17, p < .05. Given the size of the correlation coefficients, it is evident that neither age nor BMI is a strong predictor of choice of diet method or importance of diet outcome.

Table 2.

Descriptive statistics and bivariate correlations (Pearson’s r values).

Diet method	1	2	3	4	5	6	7	8	9	Item-total r
1. Calorie Counting	–									.18
2. Meal Plan	.08	–								.12
3. Low GI	.14	.02	–							.14
4. Meal Replacement	.06	.22*	.09	–						.24
5. Supplement	.20	.13	.10	.41**	–					.27
6. High Protein/Low Carb	.13	.28*	.29**	.14	.12	–				.37
7. Special Food	.08	.13	−.04	.24*	.19	.23*	–			.12
8. Pre-Prepared Meal	.08	.13	.05	.19	.12	−.02	.06	–		.15
9. Age	−.18*	−.05	.02	−.07	−.13	−.14	−.16	.11	–	–
10. BMI	−.11	.13	.10	.08	−.05	−.13	−.12	.21*	.44**	–
M (range 0–10)	2.97	1.99	1.86	1.46	0.99	1.32	0.84	0.77	30.9
SD	3.52	2.94	2.92	2.70	2.30	2.46	2.07	2.10	11.7
% used method in 12 months	55.00	41.70	35.80	31.80	25.20	29.10	20.50	15.20	–
Diet outcome	1	2	3	4	5	6	7	8	9	10	11
1. Thinness	–
2. Health	.11	–
3. Lifelong Changes	−.23*	.16	–
4. Permanent Weight Loss	−.28**	−.15	.37**	–
5. How Diet Works	−.45**	.06	.40**	.39**	–
6. Low Effort	−.02	.02	−.08	.01	.04	–
7. Easily Integrated	−.08	−.01	.03	.07	.11	.54**	–
8. Quick Weight Loss	.45**	.17	−.36**	−.43**	−.49**	.03	.08	–
9. Natural	−.19	.22*	.25*	.11	.40**	.20	.26*	−.16	–
10. Cost	−.00	.10	.01	.01	.06	.39**	.28*	.12	.17	–
11. Age	−.24**	−.03	.21*	.12	.16*	−.12	.15	−.12	.10	.10	–
12. BMI	−.10	.28**	.15	.02	.00	−.17*	.18*	.05	.17*	.21*	.44**
M (range 0–4)	2.61	2.46	3.09	2.90	2.79	2.69	2.64	2.57	2.57	2.50	30.9
SD	1.11	1.05	0.93	1.02	1.21	1.06	1.05	1.19	1.32	1.21	11.7

GI: glycemic index; BMI: body mass index; SD: standard deviation.

Weak or non-significant bivariate correlations, item-total correlations, and Cronbach’s α (.27) confirmed little consistent overlap between use of the different diet methods (with the exception of a moderate correlation between use of meal replacements and supplements, r = .41, p < .001; see Table 2). Use of multiple methods in the past 12 months ranged from none to all eight diet methods. Of the 13.2 percent of dieters who did not report using any of the methods listed, most described using a very general (“I’ve just been on a ‘general diet’ where I try to eat very, very healthily for a period of time”) or eclectic approach (“My own no-chocolate diet!”).

All 10 diet outcomes co-occurred substantially and were highly endorsed by the majority of dieters (for responses other than “never,” endorsement ranged from 88.7% to 99.3% of the scale; see Table 1). A notable exception was Health, which was least endorsed by dieters and only correlated weakly with one other outcome, Natural (r = .22, p < .01). Table 2 presents the results of bivariate correlations between the outcomes. As expected, Thinness-related outcomes were associated with Quick Weight Loss and negatively associated with How Diet Works, Permanent Weight Loss, and Lifelong Changes. That is, dieters who prioritize thinness also prioritize quick results over understanding how their diet will achieve these results or the need to make lifelong changes for permanent weight loss. Furthermore, Permanent Weight Loss, Lifelong Changes, and How Diet Works all positively correlated with each other, and all three of these aforementioned diet outcomes negatively correlated with Thinness and Quick Weight Loss.

Relationships between importance of diet outcome and choice of diet method

Canonical correlations via Statistical Package for the Social Sciences (SPSS^™) multivariate analysis of variance (MANOVA) were conducted to examine relationships between the rated importance of each diet outcome and the frequency of use of each diet method. With all canonical correlations included, relationships between the two sets of variables were significant, F(80, 852) = 2.29, p < .001, but not once the first and second canonical correlations were removed, F(48, 668) = 1.19, p = .18. Therefore, only significant relationships in the first two pairs of canonical variates were interpreted. The first canonical correlation was r = .68, representing 46.7 percent overlapping variance for the first pair of canonical variates. The second canonical correlation was r = .42, representing 18.1 percent overlapping variance for the second pair of canonical variates (canonical loadings and weights are shown in Table 3). The first canonical variates pair extracted 18.9 percent of the variance from the diet outcomes set and 19.8 percent of the variance from the diet methods set. The second canonical variates pair extracted 10.9 percent of the variance from the diet outcomes set and 12.8 percent variance from the diet methods set. That is, together, the two canonical variates accounted for 29.8 percent of diet outcomes and 32.6 percent of diet methods.

Table 3.

Canonical loadings and standardized canonical weights for canonical analysis with diet outcomes and diet methods.

	Variate 1		Variate 2
	Loadings	Weights	Loadings	Weights
Diet outcomes set
Thinness	.66	.47	.45	.82
Health	.27	.18	.26	.09
Lifelong Changes	−.50	−.39	.33	.31
Permanent Weight Loss	−.54	−.19	−.08	−.31
How Diet Works	−.42	.02	.29	.35
Low Effort	−.39	−.35	.45	.28
Easily Integrated	−.50	−.37	.35	.19
Quick Weight Loss	.50	.09	−.18	−.43
Natural	−.11	.19	.36	−.11
Cost	.05	.23	.36	−.20
Diet method Set
Calorie Counting	.67	.55	.36	.41
Meal Plan	.52	.37	−.41	−.44
Low GI	.15	.01	.60	.54
Meal Replacement	.26	−.06	−.14	.14
High Protein/Low Carb	.48	.18	.26	.21
Supplement	.48	.27	−.42	−.55
Special Food	.56	.41	−.11	−.02
Pre-Prepared Meal	−.05	−.18	−.29	−.25

GI: glycemic index.

In terms of the first canonical variate, inspection of the loadings in Table 3 (using r > .3 as the criterion for inclusion) reveals high positive loadings on Thinness and Quick Weight Loss and high negative loadings on Permanent Weight Loss, Lifelong Changes, Easily Integrated, How Diet Works, and Low Effort. This “thin, quick, and easy” profile, one that exists at the expense of concerns over longer-term factors with dieting, mirrors the most salient and common aspects of diet marketing claims. It was also associated with high positive loadings on Calorie Counting, Special Food, Meal Plan, Supplement, and High Protein/Low Carb. In terms of the second canonical variate, there were high positive loadings on Thinness, Low Effort, Natural, Cost, Easily Integrated, and Lifelong Changes suggestive of a “thin life-style” profile that was associated with high positive loadings on Low Glycemic Index (GI) and Calorie Counting, and high negative loadings on Supplement and Meal Plan. It is interesting to note that health outcomes were not relevant to either of the two variates.

Relevance to weight-loss success

In terms of diet effectiveness, two separate multiple regressions conducted via SPSS^™ revealed that neither weight-loss methods (individually or in combination) nor weight-loss motives (again, individually or in combination) predicted significant variance in weight-loss success, R² = .07 and .10, p > .05, respectively.

Relevance to eating pathology

To examine the relevance of these results to disordered eating, two multiple regressions were conducted via SPSS^™ with frequency of use of the diet methods as predictors and each of the two eating pathology measures from the EDE-Q as dependent variables. These regressions revealed that only Calorie Counting (β = .36) and Supplements (β = .24) were individual predictors (t = 4.88, 2.96, p < .001 and .01, respectively) of Restraint (R² = .28, p < .001) and that only Calorie Counting (β = .34) and Special Food (β = .19) were individual predictors (t = 4.77, 2.66, p < .001 and .01, respectively) of Eating Concerns (R² = .34, p < .001).

Discussion

A typology of diet method and diet outcome

A typology of eight diet methods—to our knowledge, a first of its kind—was developed from analysis of data from product marketing information and interviews with health professionals and laypeople/consumers. The eight diet methods exhibited little overlap in relation to their use over the previous 12 months confirming that the typology contains reasonably non-redundant methods that do not overlap systematically in terms of their use by dieters.

Perhaps not surprisingly, reported use of diet methods generally reflected the availability and popularity of the diet products/advice in society, with calorie counting, meal plans, low GI diets, and meal replacement strategies featuring prominently, and supplements and use of special foods being least common. The only surprise was the low frequency of use of pre-prepared meals. Although these are prominent in marketing and in grocery store food aisles, they were least popular among our sample of dieters, perhaps reflective of their relatively high cost.

Unfortunately, the typology was unable to capture all variance in dieting. Tellingly, at least 13 percent of self-described dieters did not endorse any particular method. On the basis of comments received from participants in relation to this, we suggest in future the inclusion of a ninth diet method—“unstructured/lifestyle”—which describes general or eclectic approaches to controlling or selecting foods (e.g. “I didn’t use an actual diet product, I just tried to eat healthier—[I] cut back on [my] consumption of junk food and increased [my] fruit and vegetable intake”). Veganism and vegetarianism would qualify for this category, along with others, although it remains to be determined empirically the extent to which such methods actually constitute “weight-loss dieting” in the outcome beliefs of their adherents.

Interestingly, neither the choice of weight-loss diet nor the underlying motives, either individually or in combination, predicted significant variance in reported history of weight loss. Even assuming that diet methods and motives should be judged entirely on the basis of the amount of weight lost, these results confirm two general themes to emerge from the weight-loss literature: (1) that body weight is strongly biologically determined and as such, achieving and maintaining significant amounts of weight loss volitionally is not realistic for most dieters (Friedman, 2004) and (2) that diet method is not an important determinant of weight loss, even when comparing diets that advocate radically different and even contradictory strategies (e.g. low carbohydrate/high protein vs high carbohydrate/low fat; Johnston et al., 2014).

Importance of diet outcomes explains choice of diet method

Prior research has broadly connected thinness motivations (as opposed to health motivations) with potentially harmful strategies such as excluding food groups and using diuretics and with disinhibited eating (Putterman and Linden, 2004). We examined these thinness motivations in relation to dieters’ ratings of the importance of diet outcomes (“outcome evaluations” in the theory of reasoned action) and revealed a clustering of thinness-related outcome beliefs which incorporated a desire for quick weight loss at the expense of long-term aspects of dieting such as permanent weight loss or understanding how a diet actually works.

This cluster and its implications for diet choice were examined in detail via canonical correlation which revealed two subtypes of association. The dominant one (in terms of explained variance) could be described as a “thin, quick, and easy” association, characterized by a de-emphasis of typically healthful aspects of dieting such as the desire for permanent weight loss, lifelong change, integration of the diet into one’s lifestyle, and understanding how the diet works. We took this profile to be inherently unhealthy and, indeed, it was associated with methods such as calorie counting, use of special foods, meal plans, supplements and a high protein/low carb diet that, with the exception of the high protein/low carb and meal plan diets, were associated with two clinically relevant symptoms of disordered eating—dietary restraint and eating concerns. The second association could be described as a “thin, natural, life-style” association characterized again by a focus on thinness, but this time in conjunction with a desire for low effort, naturalness, low cost, dietary integration into one’s lifestyle, and lifelong changes. Although it too included thinness as a focus and was associated with calorie counting as a method, it included more potentially healthy attributes (e.g. naturalness, integration, lifelong) and was associated with less use of supplements and meal plans.

On these bases, we suggest that health professionals should recognize that a focus on appearance is not only prevalent and dominant in dieters, but that it is characterized by short-term thinking and goal-setting; a reluctance to research the diet method/product in question, its claims, and its risks/costs, and so on; and the use of diet methods some of which are potentially hazardous. Consistent with this idea, a recent review of six randomized controlled trials confirmed that use of non-dieting/non-thinness approaches to food choice and eating yielded improvements in body image and self-esteem, along with improvements in physical/metabolic markers of health, things that a growing number of researchers suggest should be the focus of dieter change rather than BMI or body weight per se (Bacon and Aphramor, 2011).

The thinness-related associations may also explain the counterintuitive finding that the “slow and steady” weight-loss strategy advocated by health professionals is not always most effective and that diet adherence and weight-loss maintenance are more likely in dieters who achieve rapid initial weight loss (Purcell et al., 2014). While the desire for rapid weight loss is inherently risky, our results suggest it is nonetheless natural and consistent with dieters’ own preconceptions and goals and thus may be more meaningful and motivating to them even, paradoxically, in the long term.

Disappointingly, health was the least endorsed of all diet outcomes and was generally unrelated to other types of outcomes. In other words, health was neither an important feature of dieters’ motivation to lose weight, nor was it integrated as part of a broader set of presumably “positive” motivations to diet. In terms of the canonical correlations, health as a diet outcome also did not feature in associations between diet outcome evaluations and choice of diet method. Health’s low prominence among dieters and its low relevance in relation to diet methods are difficult to understand or address. The reason is that the risks to physical health of being overweight are already highly promoted to the general public (cf. the “obesity epidemic”). Thus, we are reluctant to attribute our results to lack of public awareness of the risks of being overweight but, rather, speculate that the results highlight the dominance of appearance as a motivator for behavior change and a relative lack of awareness of the risks of being preoccupied with weight and of attempting rapid weight loss. As Putterman and Linden (2004) put it, the drive for a thinner physique is itself the “toxic factor” in dieting, not dietary change per se. The dominance of appearance over health motives may also be explained by their different “psychological distances.” Evidence across diverse contexts suggests that the costs and benefits of temporally proximal events are psychologically construed in more concrete and less abstract terms than distal events (cf. temporal Construal Level Theory; Trope and Liberman, 2010). In the case of the concrete decisions about eating made each day as part of one’s diet, greater influence would be expected from the typically short-term outcomes associated with appearance over the typically long-term outcomes associated with health.

Previous research has noted the multidimensional nature of weight-loss motives (Vartanian et al., 2012), and perhaps long-term health priorities were obscured in this study by immediate cosmetic ones. Perhaps future research could specifically target dieting groups with salient and immediate health concerns, such as Type 2 diabetics. In such populations, a concern with health would be more likely to be a prominent motivator for dietary change and could better highlight how health concerns impact dieting decision-making and behavior.

Limitations

These included a focus on women at the exclusion of men, the use of a convenience sample and reliance on self-report measures, self-selection into the study, and use of correlational methods for inferential analysis. Nor did we survey women with clinically diagnosed disordered eating. Thus, it is impossible to infer causality from the results, and it is difficult to extrapolate from the results to the general population. Furthermore, the physical and psychological sequelae to dieting, methods, and motives cannot be determined from our results.

Concluding remarks

We developed a potentially useful typology of relatively non-overlapping diet methods and diet outcomes. We sought to use these typologies to demonstrate a clear delineation between aesthetic and health motivations for dieting and then reveal the extent to which these different motivations predict differences in diet method. However, our efforts were to some extent thwarted in that health was not only found to be a weak motivator of dieting in that it received the lowest importance ratings of all diet outcomes, it was also generally unrelated to the other diet outcomes surveyed. Health also failed to contribute to the cluster of motivations that predicted choice of diet method. Instead, associations between diet outcomes and methods hinged upon the pursuit of thinness. Specifically, we uncovered a potentially problematic “thin, quick, and easy” motivational profile characterized by use of some diet methods that were associated with eating disordered symptomatology, and a “thin, natural, life-style” motivational profile with more healthful attributes. (Note that neither profile included the pursuit of health.) Collectively, our results attest to the complex interrelationships between weight-loss motives and choice of diet method—relationships that have yet to be fully explored within either the medical/disease-prevention models of obesity or the biopsychosocial models of disordered eating (Neumark-Sztainer, 2005)—and the potency of the pursuit of thinness as a behavioral motivator.

Footnotes

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Amigo

Fernández

(2007) Effects of diets and their role in weight control. Psychology, Health & Medicine 12: 321–327.

Australia Age Structure (2014) http://www.indexmundi.com/australia/age_structure.html (accessed 6 October 2014).

Bacon

Aphramor

(2011) Weight science: Evaluating the evidence for a paradigm shift. Nutrition Journal 10. Available at: http://www.nutritionj.com/content/10/1/9

Berzins

(1999) Protecting the consumer through truth-in-dieting laws. Journal of Social Issues 55: 371–382.

Bohon

Stice

Burton

(2009) Maintenance factors for persistence of bulimic pathology: A prospective natural history study. International Journal of Eating Disorders 42: 173–178. DOI: 10.1002/eat.20600.

Cameron

Welborn

Zimmet

. (2003) Overweight and obesity in Australia: the 1999–2000 Australian Diabetes, Obesity and Lifestyle Study (AusDiab). Medical Journal of Australia 178: 427–432.

Chapman

Ogden

(2009) How do people change their diet? An exploration into mechanisms of dietary change. Journal of Health Psychology 14: 1229–1242.

Fairburn

Beglin

(1994) Assessment of eating disorders: Interview or self-report questionnaire? International Journal of Eating Disorders 16: 363–370.

Flegal

Carroll

Kit

. (2012) Prevalence of obesity and trends in the distribution of body mass index among US adults, 1999-2010. JAMA: Journal of the American Medical Association 307: 491–497.

10.

Freedman

King

Kennedy

(2001) Popular diets: A scientific review. Obesity Research 9: 1S–40S.

11.

French

Jeffery

Murray

(1999) Is dieting good for you? Prevalence, duration and associated weight and behaviour changes for specific weight loss strategies over four years in US adults. International Journal of Obesity and Related Metabolic Disorders: Journal of the International Association for the Study of Obesity 23: 320–327. DOI: 10.1038/sj.ijo. 0800822.

12.

Friedman

(2004) Modern science versus the stigma of obesity. Nature Medicine 10: 563–569.

13.

Green

Larkin

Sullivan

(2009) Oh stuff it! The experience and explanation of diet failure: An exploration using interpretative phenomenological analysis. Journal of Health Psychology 14: 997–1008.

14.

Herman

Mack

(1975) Restrained and unrestrained eating. Journal of Personality 43: 647–660.

15.

Herriot

Thomas

Hart

. (2008) A qualitative investigation of individuals’ experiences and expectations before and after completing a trial of commercial weight loss programmes. Journal of Human Nutrition and Dietetics 21: 72–80. DOI: 10.1111/j.1365-277X.2007.00837.x.

16.

Hsieh

H-F

Shannon

(2005) Three approaches to qualitative content analysis. Qualitative Health Research 15: 1277–1288. DOI: 10.1177/1049732305276687.

17.

Jenks

Higgs

(2010) Reactivity to smoking- and food-related cues in currently dieting and non-dieting young women smokers. Journal of Psychopharmacology 25: 520–529. DOI: 10.1177/0269881109359093.

18.

Johnston

Kanters

Bandayrel

. (2014) Comparison of weight loss among named diet programs in overweight and obese adults. JAMA: Journal of the American Medical Association 312: 923–933.

19.

McFarlane

Polivy

McCabe

(1999) Help, not harm: Psychological foundation for a nondieting approach toward health. Journal of Social Issues 55: 261–276.

20.

Mann

Tomiyama

Westling

. (2007) Medicare’s search for effective obesity treatments. American Psychologist 62: 220–233.

21.

Miller

(2005) Communications Theories: Perspectives, Processes, and Contexts. New York: McGraw-Hill.

22.

National Institute for Health and Clinical Excellence (NICE) (2006) Obesity guidance on the prevention, identification, assessment and management of overweight and obesity in adults and children. NICE clinical guideline 43. Available at: http://guidance.nice.org.uk/CG43/NICEGuidance/pdf/English

23.

Neumark-Sztainer

(2005) Can we simultaneously work toward the prevention of obesity and eating disorders in children and adolescents? International Journal of Eating Disorders 38: 220–227. DOI: 10.1002/eat.20181.

24.

Neumark-Sztainer

Croll

Story

. (2002) Ethnic/racial differences in weight-related concerns and behaviors among adolescent girls and boys: Findings from project EAT. Journal of Psychosomatic Research 53: 963–974. DOI: 10.1016/S0022-3999(02)00486-5.

25.

Purcell

Sumithran

Prendergast

. (2014) The effect of rate of weight loss on long-term weight management: a randomised controlled trial. The Lancet Diabetes & Endocrinology 12: 954–962. DOI: https://dx-doi-org.web.bisu.edu.cn/10.1016/S2213-8587(14)70200-1

26.

Putterman

Linden

(2004) Appearance versus health: Does the reason for dieting affect dieting behavior? Journal of Behavioral Medicine 27: 185–204.

27.

Rose

(2010) Calls to regulate “obesity industry.” Australian Associated Press, 29 March. Available at: http://news.smh.com.au/breaking-news-national/calls-to-regulate-obesity-industry-20100330-r8od.html

28.

Shepherd

Towler

(1992) Nutrition knowledge, attitudes and fat intake: Application of the theory of reasoned action. Journal of Human Nutrition and Dietetics 5: 387–397.

29.

Shikany

Barash

Redden

. (2007) Divergence in popular diets relative to diets consumed by Americans, and implications for diet selection. Medscape General Medicine 9. Available at: http://www.ncbi.nlm.nih.gov/pubmed/18092015?dopt=abstractandotool=iaudeaulib

30.

Stark

(2012) Lifetime of dieting is the new weight loss plan. Sydney Morning Herald, 30 December. Available at: http://www.smh.com.au

31.

Stice

(2002) Risk and maintenance factors for eating pathology: A meta-analytic review. Psychological Bulletin 128: 825–848. DOI: 10.1037//0033-2909.128.5.825.

32.

Stice

Sysko

Roberto

. (2010) Are dietary restraint scales valid measures of dietary restriction? Additional objective behavioral and biological data suggest not. Appetite 54: 331–339. DOI: 10.1016/j.appet.2009.12.009.

33.

Trope

Liberman

(2010) Construal-level theory of psychological distance. Psychological Review 117: 440–463.

34.

Van den Berg

Neumark-Sztainer

Hannan

. (2007) Is dieting advice from magazines helpful or harmful? Five-year associations with weight-control behaviors and psychological outcomes in adolescents. Pediatrics 119: e30–e37. DOI: 10.1542/peds.2006-0978.

35.

Vartanian

Wharton

Green

(2012) Appearance vs. health motives for exercise and for weight loss. Psychology of Sport and Exercise 13: 251–256. DOI: 10.1016/j.psychsport.2011.12.005.

36.

Westenhoefer

(1991) Dietary restraint and disinhibition: Is restraint a homogeneous construct? Appetite 16: 45–55.

37.

World Health Organisation [WHO] (2006). Global Database on Body Mass Index. Retrieved from http://apps.who.int/bmi/index.jsp?introPage=intro_3.html

38.

Yahia

El-Ghazale

Achkar

. (2011) Dieting practices and body image perception among Lebanese university students. Asia Pacific Journal of Clinical Nutrition 20: 21–28.