Skip to main content

Implicit Attitudes Towards Weight, One’s Own Body and its Relation to Food in Women with Overweight and Obesity



Theoretical models emphasize the importance of implicit self-related weight attitudes for the maintenance of body dissatisfaction. Even though body dissatisfaction is increased in obesity, only general implicit weight-related attitudes have been investigated so far. Therefore, the present study assessed self-related and general implicit weight attitudes and their relation to food.


Women with overweight and obesity (OW; n = 71) and women with normal weight (NW; n = 44) completed three implicit tasks to (1) assess attitudes towards persons with normal weight and overweight in general, (2) attitudes towards one’s own body, and (3) the association between one’s own body and food.


While both groups showed an implicit preference towards persons with normal weight relative to persons with overweight, only women with OW showed a significantly stronger negative implicit attitude towards their own body and a stronger association between food and one’s own body. Additionally, self-related and not general implicit weight attitudes correlated significantly with body dissatisfaction and eating pathology.


The results highlight the importance of self-related implicit attitudes and their relation to body dissatisfaction and eating pathology in women with overweight and obesity. Targeting these self-related implicit weight attitudes might help to improve obesity treatments.


Current studies underline the severity of body dissatisfaction among individuals with overweight and obesity as elevated levels compared to normal weight controls have consistently been found (e.g. Weinberger et al., 2016). Besides body dissatisfaction's negative effects on emotional well-being (Schwartz & Brownell, 2004), experimental and longitudinal studies yield robust evidence that it is a risk factor for disordered eating, thereby contributing to the etiology and maintenance of overweight and obesity (Jansen et al., 2008a, 2008b; Neumark-Sztainer et al., 2006). Given this evidence and the limited long-term effectiveness of weight-loss programs (Dombrowski et al., 2014), body image interventions have been integrated in evidence-based weight loss treatments and some initial promising results have been reported (Jansen et al., 2008a, 2008b; Olson et al., 2018). This notwithstanding, a more fundamental understanding of body dissatisfaction and its relation to eating pathology in overweight and obesity could provide novel treatment options by a better alignment of current gold standard treatments to the identified maintaining mechanisms.

Cognitive-behavioral theories (Vitousek & Hollon, 1990; Williamson et al., 2004) postulate that body dissatisfaction results from and is maintained by the activation of dysfunctional body-related schemata, which foster biased information processing involving memory, attention, and interpretation. Accordingly, understanding and changing these dysfunctional cognitive biases is crucial for improving body dissatisfaction.

In the context of overweight and obesity, the most frequently addressed cognitive bias has been the tendency to attribute negative characteristics to individuals with overweight and obesity [weight bias] which is also present in the devaluated subgroup itself (Watts & Cranney, 2009). However, theoretical models also emphasize the importance of weight-related self-schemata over general weight-related schemata, as the latter are shared by most people (Vitousek & Hollon, 1990). Hence, even though individuals with overweight and obesity might share the societal view on obesity, these attitudes may not necessarily be self-relevant and internalized by them. This assumption is supported by studies confirming the existence of discrepancies in persons with overweight and obesity between their attitudes towards obesity in general and attitudes towards themselves on negative and positive traits (see, e.g. Carels & Oehlhof, 2013). In line with this, numerous studies confirm the negative effects of weight bias internalization in persons with overweight and obesity as it is associated with psychological distress, severe eating pathology, and poorer weight loss outcome (Carels et al., 2010; Mensinger et al., 2016). Notably, though, most studies assessing weight bias internalization and its associations with other variables (like e.g. eating behavior) exclusively applied explicit self-report measures (Pearl & Puhl, 2018). However, implicit measures possess incremental explanatory and predictive validity over and above explicit measures as they tap more spontaneous impulses that are less governed by reflective cognition (Friese et al., 2008; Greenwald et al., 2003; Rudolph et al., 2010). In line with this, meta-analytic evidence suggests a moderate correlation between explicit and implicit measures in general, which is influenced by moderator variables like, e.g., conceptual, and procedural correspondence (Hofmann et al., 2005). Concerning weight-related biases, heterogeneous results have been found ranging from weak negative to moderate positive correlations (for an overview see Watts & Cranney, 2009). Hence, implicit measurement may complement explicit measurements and help minimize the influence of social desirability and demand characteristics (Greenwald & Lai, 2020).

To our knowledge, no study to date has assessed implicit negative attitudes to oneself being overweight in individuals with overweight and obesity. Moreover, only two studies with normal weight participants (Parling et al., 2012; Ritzert et al., 2016) assessed implicit associations between the self being thin or fat using self-statements (e.g. “me being thin” or “me being fat”) and evaluative categories (“good/bad”, “disgusting/attractive”, “fearful/attractive”, respectively). Both studies found a positive bias towards oneself being thin, while no devaluation of oneself being fat was evident. However, given the high prevalence of weight-based discrimination as well as the elevated levels of body dissatisfaction in individuals with overweight and obesity (Puhl & Heuer, 2009; Weinberger et al., 2016), individuals with overweight and obesity might differ from normal weight participants’ implicit self-weight attitudes. So far, two studies investigating women with the diagnosis of a binge eating disorder (BED) compared to individuals with overweight and obesity without an eating disorder confirmed the importance of assessing self-relevant implicit attitudes over general weight biases [implicit self-esteem (Brauhardt et al., 2014); implicit self-discrimination (Rudolph & Hilbert, 2014)] as these biases were predictive of eating pathology over and above weight status and/or experiences of weight stigma and general implicit weight bias.

Thus, the aim of the present study was to assess implicit attitudes towards one's own body, beyond the assessment of a general implicit weight bias, in women with overweight and obesity (OW) compared to women with normal weight (NW). Furthermore, given the relationship between self-stigmatizing attitudes and eating pathology using explicit measures (Brauhardt et al., 2014; Pearl & Puhl, 2018; Rudolph & Hilbert, 2014), we were interested in the implicit association of one’s own body and high-calorie food in overweight and obesity. We hypothesized a comparable strength of general weight bias across groups irrespective of participants’ own weightFootnote 1 (measured by the Weight IAT). By contrast, females with OW were hypothesized to show a greater negative implicit attitude towards their own body (as indexed by the Body-SC-IAT), as well as stronger implicit associations between their own body and high-calorie food (Body-Food-SPF). Furthermore, we hypothesized that implicit devaluation of one’s own body (Body-SC-IAT) is related with explicitly assessed body dissatisfaction and eating pathology, while this was not expected for the general weight bias (Weight IAT).

Materials and Methods


The study was approved by the ethical committee of the local university (614/2015BO2). The data are part of a baseline assessment of an ongoing RCT started in 2016 on the efficacy of body image interventions in overweight and obesity (results of the RCT will be presented elsewhere). General inclusion criteria for the study were (a) female gender, (b) 18 ≤ age (years) ≤ 69, (c) German language skills on a native level, and (d) the absence of a lifetime diagnosis of an eating disorder. Females with OW had to have a Body Mass Index (BMI) between 25 and 45,Footnote 2 while participants with NW had to be within the normal range (18.5 ≤ BMI < 25). Exclusion criteria included (a) presence of severe physical illness, (b) current presence of substance abuse or addiction, manic episode, suicidal ideation, psychosis or irregular intake of antidepressants, (c) current pregnancy or lactation, and (d) the current participation in weight loss programs or interventions related to eating and weight.Footnote 3 The absence of a lifetime diagnosis of a clinical or subclinical eating disorder was assessed using the diagnostic items of the German version of the Eating Disorder Examination interview (EDE; Hilbert et al., 2004), while all other mental disorders were screened by the Structured Clinical Interview (SCID; Wittchen et al., 1997).

An a-priori calculated power analysis with GPower (Faul et al., 2007) based on an effect of f = 0.035 with α = 0.05 and a power of 0.90 resulted in a total sample size of N = 88 (44 per group). Seventy-one women with OW and n = 44 age-matched women with NW participated in the study.Footnote 4 Groups did not differ on sociodemographic variables. However, women with OW scored higher on all psychopathological variables, including significantly more diagnoses of mental disorders (see Table 1).

Table 1 Descriptive characteristics on self-reported variables for women with overweight and obesity (OW) and women with normal weight (NW)


Implicit Tasks

  1. (1)

    Weight IAT. The Implicit Association Test (IAT; Greenwald et al., 1998) is a widely used task designed to assesses the relative strength of associations between two concept and attribute categories. In the present Weight IAT, “overweight” and “normal weight” served as concept categories represented by six silhouettes of female bodies with either normal weight or overweight (278 × 500px). Six positive (friendly, attractive, motivated, happy, hard-working, intelligent) and six negative adjectives (lazy, inactive, messy, unhealthy, unattractive, unpopular) represented the attribute categories “good” and “bad” and were matched according to number of characters and syllables in German. This IAT consisted of seven blocks with either 24 or 48 trials plus two or four practice trials (see Table 2). In the combined blocks, one concept and one attribute category shared the same response button (“left CTRL”; “right CTRL”). This was “overweight” and “bad” and “normal weight” and “good” for the compatible trials, while response-mapping was reversed for the incompatible blocks (see Fig. 1a). Consequently, the difference in task performance between compatible and incompatible blocks is computed as the IAT score (see section “Presentation and Scoring” for more details).

  2. (2)

    Body-SC-IAT. To assess implicit attitudes towards one’s own body, a single-category (SC)-IAT (Karpinski & Steinman, 2006) was used, which facilitates the interpretation of the score as the relative evaluation of this concept category. The concept category (“one’s own body”) was represented by six photographs of participants’ own body, which were taken in six standardized positions (front, back, left, right, 45°, 225°) in standardized underwear on a separate day. Six positive and six negative emoticons were used as stimuli for the attribute category. The Body-SC-IAT consisted of three blocks, with two relevant mixed blocks (Table 3 for details). In the mixed blocks, 60% of all trials showed stimuli of the combined categories (30% attribute; 30% concept) that required classification using the same button, whereas the remaining 40% were stimuli of the other attribute category (Fig. 1b). Again, the IAT score was computed as the difference in task performance between the compatible and incompatible blocks (see Section “Presentation and Scoring” for more details).

  3. (3)

    Body-Food SPF. The Sorting-Paired Features Task (SPF; Bar-Anan et al., 2009) was used to assess the relative strength between food and one’s own body. In each trial, a concept-attribute pair was presented, comprising a photograph as concept stimulus and a word displayed beneath in blue font as an attribute stimulus. “Body” and “vase” served as concept categories being represented by photographs of participant’s own body or pictures of a vase. “High-calorie food” and “tree species” served as attribute (food: chips, pizza, tart, cheese, cookies; tree species: oak, fir tree, spruce, beech, birch). The latter served as control condition, as no salient associations with the other categories were expected. Participants were asked to classify the presented picture-word pair by pressing the button corresponding with one of the four possible combinations (i.e., body/food, vase/food, body/tree, and vase/tree). Prior to the experimental trials, participants familiarized themselves with the four response keys (“D”,”C”,”K”,”M”) and pictures [body and vase]. Then, five relevant SPF blocks were presented, each including four practice trials and 60 test trials (Fig. 1c).

Table 2 Blocks of the Weight IAT
Fig. 1

Schematic illustration of the implicit tasks. a Weight IAT: In this example, a compatible trial is presented. Participants had to classify the presented stimulus by pressing the left button, if the stimulus was a negative trait or an overweight stimulus, and by pressing the right button, if the stimulus was a positive trait or a normal weight stimulus. b Body-SC-IAT: In the present example, the stimulus set of a participant with overweight in incompatible trials is displayed. The participant had to press the left button if the stimulus displayed was her own body or a negative emoticon and the right button, if the stimulus displayed was a positive emoticon. c Body-Food SPF: In the Body-Food SPF, response-mapping stayed the same across all blocks. Participants had to classify the presented picture-word pair according to the four possible response keys: body-food, body-tree, vase-food, vase-tree. Own body pictures were presented in color during the experimental tasks (approval for using the body pictures in the publication was given by the depicted person)

Table 3 Blocks of the Body-Single-Category (SC) -IAT

Presentation and Scoring

All implicit measures were presented on a 22″ screen and a standard QWERTZ computer keyboard was used to collect responses. Following previous research addressing correlates of individual differences in IAT measures (see e.g. Egloff & Schmukle, 2002), we kept task-sequence, compatibility order, and trial lists constant across participants. Precompiled pseudo-random trial lists were used that balanced task requirements including their sequence-order. All verbal stimuli were presented in German, the native language of our research participants.

In each trial, stimuli stayed on screen until the correct response was given and visual error feedback was provided. After the target stimuli, there was a 400 ms blank inter-trial pause. Mapping reminders stayed on screen for the entire task.

All implicit tasks were scored using so-called D scores which transform latencies and errors into a common metric thereby helping to optimize the analyses of response-time data (Greenwald et al., 2003). We used the D2 score, for which latencies in erroneous trials are recorded until the correct response is given. The general logic underlying IAT-type implicit measures is that it should be easier to classify stimuli when two highly associated categories share the same response key. For the Weight IAT and the Body-SC-IAT, this would be expected for compatible trials (faster RTs/highly accurate) compared to incompatible trials (slower RTs/error prone). Thus, the difference in mean response times between incompatible and compatible task conditions are calculated and standardized by the integrative standard deviation of all trials. In the Weight IAT, more positive scores therefore indicate that normal weight is more positively associated (and overweight is more negatively associated) in relative terms. In the Body-SC-IAT, more positive scores indicate a positive evaluation of one’s own body in relative terms.

For the Body-Food SPF, a total score was computed, which reflects the aggregated overall strength of all four associations in this task. Positive scores point towards the existence of associative structures. To disentangle this score, relative differences for each of the four associations were computed, which reflected the difference in task performance for one particular type of stimulus pair (e.g., body-food) and the mean of the other three pairs. Thus, the interpretation of the contrast scores depends on the choice of the comparison category. Positive scores indicate a higher relative strength of the respective association relative to the other three associations.


(1) As in previous studies (Weinberger et al., 2016), the validated German version of the Body Shape Questionnaire (BSQ; Pook et al., 2002) was used to assess body dissatisfaction, with higher scores indicating more body image concerns during the last 4 weeks. Cronbach’s α was α = 0.97 for the present sample. (2) The validated German version of the Beck’s Depression Inventory-II (BDI-II; Hautzinger et al., 2006) was used to self-report depressive symptoms during the last 2 weeks with higher scores indicating higher symptom severity (Cronbach’s α = 0.91 for the present sample). (3) The German version of the Eating Disorder Examination Questionnaire was used (EDE-Q; Hilbert et al., 2004) which has also been validated in obese samples (Aardoom et al., 2012) with higher scores indicating higher eating pathology severity. Cronbach’s α was α = 0.94 in the present sample. (4) The shortened form of the Fat Phobia Scale (FPS; Stein et al., 2014) was used to assess weight bias attitudes consisting of 14 pairs of adjectives on a semantic differential (e.g. lazy—motivated). For each item pair, participants had to indicate which adjective best describes the target person (either normal weight or overweight) on a 5-point Likert scale. Higher scores indicate more negative attitudes towards the target person. Cronbach’s α was α = 0.86 for ratings of the normal weight and α = 0.89 for ratings of the overweight woman. In addition, we included the adjectives used in the Weight IAT as more direct comparison option (Cronbach’s α = 0.89 for the normal weight woman and α = 0.89 for the overweight woman). (5) To control for potential confounding variables, the words used in the SPF task (high-calorie food; tree species) were rated concerning their valence and familiarity on a 5-point Likert-scale including palatability ratings for food words (SPF-Rating).


Participants were recruited via announcements in the local press, flyers, and the university’s mailing list. Eligible participants were invited to a face-to-face diagnostic assessment conducted by trained psychologists using the diagnostic items of the EDE and the SCID. During the diagnostic assessment, height and height were measured and BMI was calculated. Written informed consent was obtained. Furthermore, participants had to provide sociodemographic information and to fill in the questionnaires (BSQ; EDE-Q, BDI-II) using Unipark, an online survey platform. The standardized pictures of the participants were taken on a separate day. The whole body from the ankle to the neck (face omitted) was visible; eye-catching characteristics (e.g., tattoos, plasters) were removed from the pictures using Corel PaintShop Pro X7 (OW group: n = 4; NW group: n = 5). There was no significant difference in the proportion of modified pictures between both groups (χ2(1) = 1.236, p = 0.266).

As the results presented here are part of a baseline assessment of an RCT to investigate the effects of body image interventions, two other appointments on different days were scheduled with the participants prior to the implicit measures to assess body image by means of a thought-sampling procedure during a mirror confrontation, as well as to measure body-related attentional biases (results will be presented elsewhere). On a separate day, the three implicit tasks including the remaining questionnaires (FPS; SPF-Rating) were scheduled. First, the Body-SC-IAT was conducted, followed by the Weight IAT and the SPF. After the tasks, participants with NW were reimbursed and another appointment was scheduled with women with OW participating in the RCT.

Data Analysis

Statistical analyses were conducted using IBM SPSS (Version 26). Mean substitution was used to account for missing questionnaire data (BSQ/BDI-II/EDE-Q: n = 2 for each group; FPS: n = 2 for the group with OW; SPF: n = 1 for the group with OW). In total, 71 women with OW were available for the Weight IAT and the Body-SC-IAT, whereas for the SPF only n = 57 women were available because recruitment started later due to a delay in piloting the experimental paradigm.

Using box plot analyses outlined by SPSS, two outliers for the Weight IAT (n = 1 of each group), one for the Body-SC-IAT (n = 1 of the group with OW) and one for the SPF-Rating (n = 1 group with NW) were identified and excluded from the analyses (see Table 4).

Table 4 Descriptive characteristics of the dependent variables for women with overweight and obesity (OW) and women with normal weight (NW)

Hypotheses were tested by means of univariate ANOVAs with Group (OW vs. NW) as between subjects’ factor. If assumptions of sphericity were not met (Mauchly’s Sphericity Test: p < 0.05), degrees of freedom for dependent variables were adjusted conservatively using Greenhouse–Geisser correction. Levene’s test was used to ensure homogeneity of variance across samples, which was especially relevant because of the different sample sizes. If homogeneity was violated, a nonparametric test (Mann–Whitney U-Test) was conducted. Effect sizes of the group differences and interactions are reported by partial eta squared (ηp2), whereby values > 0.01 are considered small, > 0.06 are moderate, and > 0.14 are large (Cohen, 1988).

To assess the relationship between the implicit tasks (D2 scores) and the self-reported data (body dissatisfaction, eating pathology) as well as BMI, Pearson product-moment correlation coefficients were computed, if the assumption of normality was met (Shapiro–Wilk test: p > 0.05). If not, Spearman’s rank correlation coefficient was used instead.


Weight IAT

The univariate ANOVA (Group: OW vs. NW) for the implicit weight bias revealed no significant effect of Group [F(1,111) = 0.118, p = 0.732, ηp2 = 0.001]. In accordance with our first hypothesis, both groups displayed an equally strong implicit weight bias preferring persons with normal weight to persons with overweight (see Fig. 2).

Fig. 2

D2 Scores with standard errors of the implicit tasks in women with overweight and obesity (OW) and women with normal weight (NW). *p < 0.05 **p < 0.01

Correlations: There was no significant correlation between implicit weight bias and explicit attitudes towards persons with normal weight (original FPS: r = − 0.082; p = 0.386; prolonged FPS: r = − 0.094; p = 0.321), but there was a statistical trend between implicit weight bias and explicit attitudes towards persons with overweight (original FPS: r = 0.182; p = 0.054; prolonged FPS: r = 0.171, p = 0.070).


The univariate ANOVA (Group: OW vs. NW) for the Body-SC-IAT revealed a significant effect of Group (F(1,112) = 3.812, p = 0.027, ηp2 = 0.033). In line with our second hypothesis, women with OW showed a negative implicit attitude towards their own body, while women with NW displayed a neutral attitude (shown in Fig. 2).

Body-Food SPF

SPF-Rating: There were no significant group differences concerning ratings of valence, familiarity and palatability of the words used in the SPF [all Fs(1,99 resp. 98) ≤ 2.617, ps ≥ 0.109, ηp2s ≤ 0.026; see Table 1 for more details].

SPF: In line with our second hypothesis, the univariate ANOVA (Group: OW vs. NW) for the total score of the SPF was significant (F(1,99) = 8.104, p = 0.003, ηp2 = 0.076). That is, the aggregated total score in the task was stronger in women with OW than in women with NW. To disentangle which of the four associations drives the total score, more specific scores were computed by contrasting one response category with the three remaining categories. Only the contrast for the association between one’s own body and food showed a significant main effect of Group [F(1,99) = 4.182, p = 0.022, ηp2 = 0.041], while this was not found for the contrast scores of the other associations [body-tree: F(1,99) = 2.224, p = 0.069, ηp2 = 0.022; vase-food: F(1,99) = 0.452, p = 0.252, ηp2 = 0.005; vase-tree: F(1,99) = 1.038, p = 0.156, ηp2 = 0.010].Footnote 5


Implicit measures: There were no substantial correlations between the different implicit tasks, neither in the whole sample nor in the subgroups (all rs ≤|0.108|; all ps ≥ 0.215).

Explicit measures: In line with our last hypothesis, the scores of the Body-SC-IAT showed small but significant correlations with self-reported eating pathology (r = − 0.276; p = 0.002), body dissatisfaction (r = − 0.283; p = 0.001) and BMI (r = − 0.178; p = 0.029), while there were no significant correlations between the scores of the Weight IAT and these variables (all rs ≤|0.082|; all ps ≥ 0.193), except for a trend towards significance concerning BMI (r = − .152; p = .053). The SPF total score correlated modestly but significantly with self-reported body dissatisfaction (r = 0.263; p = 0.004),  eating pathology (r =  0.272; p = 0.003), and BMI (r = 0.195; p = 0.025).


Recent studies highlight the need to target body dissatisfaction in persons with overweight and obesity as body dissatisfaction is associated with a variety of unhealthy behaviors and conditions such as reduced emotional well-being, disordered eating, and less successful long-term weight loss (maintenance) (Carraça et al., 2011; Gall et al., 2016; Olson et al., 2018). According to theoretical models (Vitousek & Hollon, 1990; Williamson et al., 2004), body dissatisfaction results from and is maintained through underlying weight-and shape-related (self-)schemata which are automatically activated and foster biased information processing. As the study of these implicit biases has largely been neglected so far, the present study investigated general and self-related implicit weight attitudes and their relation to eating pathology and high-calorie food in women with overweight and obesity compared to normal weight controls.

In line with our first hypothesis and previous studies (Puhl & Heuer, 2009; Watts & Cranney, 2009), there was no difference concerning the level of general weight bias between women with overweight and obesity and women with normal weight. Both groups showed and implicit and explicit preference towards persons with normal weight relative to persons with overweight. Thus, our results confirm that persons with overweight and obesity indeed internalize the disadvantaged societal attitudes towards their own in-group as shown in previous research using explicit self-report measures (Macho et al., 2021; Pullmer et al., 2021). Interestingly, there was an inverse correlation between BMI and implicit weight bias score, which means that persons with higher BMI showed this preference to a lesser extent. This is in line with other studies showing that the level of weight bias decreases with higher BMI classes (Marini et al., 2013; Schwartz et al., 2006). Importantly, there was no significant correlation between the implicit weight bias and either self-reported body dissatisfaction or eating pathology.

Our results from the Body-SC-IAT, furthermore, support the assumption that self-related weight biases are more relevant than general weight biases. In line with our second hypothesis, women with overweight and obesity displayed a negative implicit attitude towards their own body, while participants with normal weight showed a rather neutral attitude. This finding in participants with normal weight is surprising as previous research generally found a self-attractive bias (Parling et al., 2012; Ritzert et al., 2016). Although our sample was older, levels of explicit body dissatisfaction were comparable between these studies. However, and in contrast to former studies, we used idiosyncratic pictorial compared to generic, non-specific lexical stimuli (like e.g. “Me being fat is attractive” (Ritzert et al., 2016)), which seem to be more ecologically valid when assessing self-related associations (Bluemke & Friese, 2012). The non-significant correlation between the implicit assessment of the general weight bias and the self-related weight bias, furthermore underpins the need to distinguish between these implicit biases. That is, even though a person with overweight and obesity might hold general negative attitudes towards overweight and obesity, it might not necessarily be self-relevant for this individual. Therefore, the differentiation between general and self-related weight biases is crucial and supported by the fact, that in line with our last hypothesis, only the implicit attitudes towards one’s own body were significantly correlated with explicitly assessed body dissatisfaction and severity of eating pathology. This was furthermore supported by our third implicit task as a stronger link between one’s own body and high-calorie food in women with overweight and obesity compared to women with normal weight was found. The importance of self-related weight biases in regard to body dissatisfaction and dysfunctional eating behavior in individuals with overweight and obese has already been confirmed in cross-sectional and experimental studies using explicit measures (Jiang & Vartanian, 2018; O’Brien et al., 2016; Schvey et al., 2011). Our results extend this previous work by demonstrating this link on an implicit level thereby underlining its spontaneous, non-reflective character.

Concerning clinical implications, it has been shown, that participants with high levels of explicit weight bias internalization benefit less from weight-loss or maintenance interventions, and same holds for holistic health interventions (Mensinger et al., 2016; Pearl, et al., 2018a, 2018b). Future research should therefore (a) investigate the correlation between explicit and implicit weight bias internalization and (b) test if implicit weight-bias internalization possesses incremental predictive validity over and above its explicit measurement in order to understand the possible clinical relevance of implicit weight-related biases. As only minor spontaneous changes in explicit and implicit weight bias internalization have been reported after obesity treatments without a direct intervention (Carels et al., 2010; Mensinger et al., 2016; Pearl, et al., 2018a, 2018b), a next step will be to investigate which interventions are most effective in changing these weight-related biases. There is some evidence that enhancing self-esteem and body image as well as evoking empathy produces changes in explicit weight-related attitudes in individuals with overweight and obesity (Robinson et al., 1993; Teachman et al., 2003). In line with this, a recent pilot study confirmed positive effects in terms of reduced self-reported weight bias internalization and more eating self-efficacy after attending a short cognitive-behavioral group program to specifically reduce weight bias internalization (Pearl, et al., 2018a, 2018b). The promising results of these intervention studies should be further investigated in larger samples and during weight loss programs to assess their additive effect on treatment outcome. Thereby, the use of implicit measures prior to and after the intervention might be useful to understand possible changes in less-controlled, schema-driven mechanisms.

Some limitations of this study should be delineated. First, the order of tasks was not counterbalanced, which might have affected the order of overall effect sizes and reliability. However, we chose to start with the Body-SC-IAT, as we were primarily interested in this self-relevant association and did not want it to be confounded by the activation of general weight-related biases. Further, possible sequence effects would have deteriorated correlations. Second, IAT-like implicit measures are prone to some limitations such as compatibility-order effect (Fiedler et al., 2006), which can bias results because contrast scores might be larger when the compatible block precedes the incompatible one. As we were primarily interested in group differences, however, we kept compatibility order constant across participants as done in previous research (Egloff & Schmukle, 2002). Third, all evidence for a relation between implicit attitudes and the other variables of interest is correlational, hence, the direction of causation needs to be clarified in future studies using experimental and longitudinal designs. Forth, due to well-known gender differences concerning body dissatisfaction (Weinberger et al., 2016) and due to the accompanying RCT on the efficacy of mirror exposure, which so far has mostly been studied in female populations (for an overview see Griffen et al., 2018), participation was restricted to women, which might limit generalizability. Given that the marginalization of males in research in the field of eating disorders is problematic (Murray et al., 2017), future studies should test replicability of these findings in a more heterogeneous sample comprising as well male participants. Furthermore, our sample consisted of treatment-seeking participants due to the accompanying RCT. However, this can be seen as a strength of the study, as there is strong evidence for the need to discriminate between treatment-seeking and non-treatment seeking persons with overweight and obesity, as treatment-seeking individuals might be more vulnerable for negative consequences of their weight (Vieira et al., 2012). Fifth, as actual BMI might differ from participants’ perception of oneself being overweight (see e.g. Chang & Christakis, 2003), future studies should address as well self-perception of overweight and obesity. Finally, as this was the first study to use an implicit measure to assess self-related weight bias, a validation of the Body-SC-IAT including normative data is necessary to further test its use in clinical practice and longitudinal research. Importantly, as we did not to statistically control for the frequency of mental disorders as the higher prevalence of mental disorders in the overweight compared to the normal weight sample reflects naturally occurring differences (Baumeister & Ha, 2007; Kasen et al., 2008), future studies should include clinical populations as control group (e.g. depressive sample) in order to disentangle the influence mental disorders might have on these implicit biases.

To conclude, this is the first study demonstrating the importance of self-related over general implicit weight biases as they might be directly linked to body dissatisfaction and dysfunctional eating behavior in individuals with overweight and obesity. Changes in these measures might be informative with respect to treatment success and targeting the schemata constituting implicit attitudes might complement treatment options for obesity.


  1. 1.

    We assumed weight bias to occur irrespective of one’s own weight due to the fact that general implicit weight attitudes are common and widely held among different groups like health professionals, university students, but also in the devalued subgroup itself, and mixed results have been found concerning differences in the magnitude of weight bias between different weight classes (Anselmi et al., 2013; Elran-Barak & Bar-Anan, 2018; Marini et al., 2013; Roddy et al., 2009; Watts & Cranney, 2009).

  2. 2.

    Participation in the OW group was restricted to a BMI ≤ 45 as, according to the German guidelines for treatment of obesity, bariatric surgery is recommended for higher BMI categories.

  3. 3.

    This inclusion criterion was necessary for the accompanying RCT.

  4. 4.

    The presented unbalanced sample size is due to the accompanying RCT conducted with participants with overweight and obesity demanding a larger sample size than n = 44 in the OW group to detect moderate pre-post effects. For the conducted ANOVA, homogeneity of variance was therefore tested and if not present, non-parametric tests were applied (see Statistics for more information).

  5. 5.

    Note: Additional analyses including BMI as continuous variable revealed analogous results for the implicit measures (see Appendix (Fig. 3) for details).


  1. Aardoom, J. J., DingemansSlofOp’tLandt, A. E. M. C. T., & Van Furth, E. F. (2012). Norms and discriminative validity of the eating disorder examination questionnaire (EDE-Q). Eating Behaviors, 13(4), 305–309.

    Article  PubMed  Google Scholar 

  2. Anselmi, P., Vianello, M., & Robusto, E. (2013). Preferring thin people does not imply derogating fat people. A rasch analysis of the implicit weight attitude. Obesity, 21(2), 261–265.

    Article  PubMed  Google Scholar 

  3. Bar-Anan, Y., Nosek, B. A., & Vianello, M. (2009). The sorting paired features task a measure of association strengths. Experimental Psychology, 56(5), 329–343.

    Article  PubMed  Google Scholar 

  4. Baumeister, H., & Ha, M. (2007). Mental disorders in patients with obesity in comparison with healthy probands. International Journal of Obesity.

    Article  PubMed  Google Scholar 

  5. Blair, G., Cooper, J., Coppock, A., Humphreys, M., & Sonnet, L. (2021). Estimatr: Fast estimators for design-based inference. (p. R package version). Retrived 18, Aug 2021, from

  6. Bluemke, M., & Friese, M. (2012). On the validity of idiographic and generic self-concept implicit association tests: A core-concept model. European Journal of Personality, 528(October 2011), 515–528.

    Article  Google Scholar 

  7. Brauhardt, A., Rudolph, A., & Hilbert, A. (2014). Implicit cognitive processes in binge-eating disorder and obesity. Journal of Behavior Therapy and Experimental Psychiatry, 45(2), 285–290.

    Article  PubMed  Google Scholar 

  8. Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 39(2), 175–191.

    Article  PubMed  Google Scholar 

  9. Carels, R. A., & Oehlhof, J. B. M. W. (2013). Internalized weight bias: Ratings of the self, normal weight, and obese individuals and psychological maladjustment. Journal of Behavioral Medicine.

    Article  PubMed  Google Scholar 

  10. Carels, R. A., Wott, C. B., Young, K. M., Gumble, A., Koball, A., & Oehlhof, M. W. (2010). Implicit, explicit, and internalized weight bias and psychosocial maladjustment among treatment-seeking adults. Eating Behaviors, 11(3), 180–185.

    CAS  Article  PubMed  Google Scholar 

  11. Carraça, E. V., Silva, M. N., Markland, D., Vieira, P. N., Minderico, C. S., Sardinha, L. B., & Teixeira, P. J. (2011). Body image change and improved eating self-regulation in a weight management intervention in women. The International Journal of Behavioral Nutrition and Physical Activity, 8(1), 75.

    Article  PubMed  PubMed Central  Google Scholar 

  12. Chang, V. W., & Christakis, N. A. (2003). Self-perception of weight appropriateness in the United States. American Journal of Preventive Medicine, 24(4), 332–339.

    Article  PubMed  Google Scholar 

  13. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd edition). Erlbaum.

    Google Scholar 

  14. Dombrowski, S., Knittle, K., Avenell, A., Araujo-Soares, V., & Sniehotta, F. (2014). Long term maintenance of weight loss with non-surgical interventions in obese adults: Systematic review and meta-analyses of randomised controlled trials. BMJ, 348(May), g2646.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  15. Egloff, B., & Schmukle, S. C. (2002). Predictive validity of an implicit association test for assessing anxiety. Journal of Personality and Social Psychology, 83(6), 1441–1455.

    Article  PubMed  Google Scholar 

  16. Elran-Barak, R., & Bar-Anan, Y. (2018). Implicit and explicit anti-fat bias: The role of weight-related attitudes and beliefs. Social science and medicine (Vol. 204). Elsevier Ltd.

    Book  Google Scholar 

  17. Fiedler, K., Messner, C., & Bluemke, M. (2006). Unresolved problems with the “I”, the “A”, and the “T”: A logical and psychometric critique of the implicit association test (IAT). European Review of Social Psychology, 17(1), 74–147.

    Article  Google Scholar 

  18. Friese, M., Hofmann, W., & Wa, M. (2008). When impulses take over: Moderated predictive validity of explicit and implicit attitude measures in predicting food choice and consumption behaviour copyright © the British psychological society. British Journal of Social Psychology, 47, 397–419.

    Article  Google Scholar 

  19. Gall, K., Van Zutven, K., Lindstrom, J., Bentley, C., Gratwick-Sarll, K., Harrison, C., Lewis, V., & Mond, J. (2016). Obesity and emotional well-being in adolescents: Roles of body dissatisfaction, loss of control eating, and self-rated health. Obesity, 24(4), 837–842.

    Article  PubMed  Google Scholar 

  20. Greenwald, A. G., & Lai, C. K. (2020). Implicit social cognition. Annual Review of Psychology, 71(25), 1–25.

    Article  Google Scholar 

  21. Greenwald, A. G., McGhee, D. E., & Schwartz, J. L. K. (1998). Measuring individual differences in implicit cognition: The implicit association test. Journal of Personality and Social Psychology, 74(6), 1464–1480.

    CAS  Article  PubMed  Google Scholar 

  22. Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the implicit association test: I. An improved scoring algorithm. Journal of Personality and Social Psychology, 85(2), 197–216.

    Article  PubMed  Google Scholar 

  23. Griffen, T. C., Naumann, E., & Hildebrandt, T. (2018). Mirror exposure therapy for body image disturbances and eating disorders: A review. Clinical Psychology Review, 65(February), 163–174.

    Article  PubMed  Google Scholar 

  24. Hautzinger, M., Keller, F., & Kühner, C. (2006). Das Beck Depressionsinventar II. Deutsche Bearbeitung und Handbuch zum BDI II. Harcourt Test Services.

  25. Hilbert, A., Tuschen-Caffier, B., & Ohms, M. (2004). Eating disorder examination: Deutschsprachige version des strukturierten Essstörungsinterviews. Diagnostica, 50(2), 98–106.

    Article  Google Scholar 

  26. Hofmann, W., Gawronski, B., Gschwendner, T., Le, H., & Schmitt, M. (2005). A meta-analysis on the correlation between the Implicit Association Test and explicit self-report measures. Personality and Social Psychology Bulletin, 31(10), 1369–1385.

    Article  Google Scholar 

  27. Jansen, A., Bollen, D., Tuschen-Caffier, B., Roefs, A., Tanghe, A., & Braet, C. (2008). Mirror exposure reduces body dissatisfaction and anxiety in obese adolescents: A pilot study. Appetite, 51(1), 214–217.

    Article  PubMed  Google Scholar 

  28. Jansen, A., Vanreyten, A., van Balveren, T., Roefs, A., Nederkoorn, C., & Havermans, R. (2008). Negative affect and cue-induced overeating in non-eating disordered obesity. Appetite, 51(3), 556–562.

    Article  PubMed  Google Scholar 

  29. Jiang, M. Y. W., & Vartanian, L. R. (2018). A review of existing measures of attentional biases in body image and eating disorders research. Australian Journal of Psychology, 70(1), 3–17.

    Article  Google Scholar 

  30. Karpinski, A., & Steinman, R. B. (2006). The single category implicit association test as a measure of implicit social cognition. Journal of Personality and Social Psychology, 91(1), 16–32.

    Article  PubMed  Google Scholar 

  31. Kasen, S., Cohen, P., Chen, H., & Must, A. (2008). Obesity and psychopathology in women: A three decade prospective study. International Journal of Obesity, 32, 558–566.

    CAS  Article  PubMed  Google Scholar 

  32. Macho, S., Andrés, A., & Saldaña, C. (2021). Validation of the modified weight bias internalization scale in a Spanish adult population. Clinical Obesity.

    Article  PubMed  Google Scholar 

  33. Marini, M., Sriram, N., Schnabel, K., Maliszewski, N., Devos, T., Wiers, R., Huajian, C., Shiomura, K., Schnall, S., Bar-anan, Y., Vianello, M., Ayala, A., Dorantes, G., Kesebir, S., Pereira, A., Tulbure, B., Ortner, T., Stepanikova, I., Greenwald, A. G., & Nosek, B. A. (2013). Overweight people have low levels of implicit weight bias, but overweight nations have high levels of implicit weight bias. Plos One, 8(12), 1–9.

    CAS  Article  Google Scholar 

  34. Mensinger, J., Calogero, R., & Tylka, T. L. (2016). Internalized weight stigma moderates eating behavior outcomes in women with high BMI participating in a healthy living program. Appetite.

    Article  PubMed  Google Scholar 

  35. Murray, S. B., Nagata, J. M., Griffiths, S., Calzo, J. P., Brown, T. A., Mitchison, D., Blashill, A. J., & Mond, J. M. (2017). The enigma of male eating disorders: A critical review and synthesis. Clinical Psychology Review, 57(April), 1–11.

    Article  PubMed  Google Scholar 

  36. Neumark-Sztainer, D., Wall, M., Guo, J., Story, M., Haines, J., & Eisenberg, M. (2006). Obesity, disordered eating, and eating disorders in a longitudinal study of adolescents: How do dieters fare 5 years later? Journal of the American Dietetic Association, 106(4), 559–568.

    Article  PubMed  Google Scholar 

  37. O’Brien, K. S., Latner, J. D., Puhl, R. M., Vartanian, L. R., Giles, C., Griva, K., & Carter, A. (2016). The relationship between weight stigma and eating behavior is explained by weight bias internalization and psychological distress. Appetite, 102, 70–76.

    Article  PubMed  Google Scholar 

  38. Olson, K. L., Thaxton, T. T., & Emery, C. F. (2018). Targeting body dissatisfaction among women with overweight or obesity: A proof-of-concept pilot study. International Journal of Eating Disorders.

    Article  Google Scholar 

  39. Parling, T., Cernvall, M., Stewart, I., Barnes-Holmes, D., & Ghaderi, A. (2012). Using the implicit relational assessment procedure to compare implicit pro-thin/anti-fat attitudes of patients with anorexia nervosa and non-clinical controls. Eating Disorders, 20(2), 127–143.

    Article  PubMed  Google Scholar 

  40. Pearl, R. L., Hopkins, C. H., Berkowitz, R. I., & Wadden, T. A. (2018). Group cognitive-behavioral treatment for internalized weight stigma: A pilot study. Eating and Weight Disorders, 23(3), 357–362.

    Article  PubMed  Google Scholar 

  41. Pearl, R. L., & Puhl, R. M. (2018). Weight bias internalization and health: A systematic review. Obesity Reviews, 19(8), 1141–1163.

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  42. Pearl, R. L., Wadden, T. A., Chao, A. M., Walsh, O., Alamuddin, N., Berkowitz, R. I., & Tronieri, J. S. (2018). Weight bias internalization and long-term weight loss in patients with obesity. Annals of Behavioral Medicine, 53(8), 782–787.

    Article  PubMed Central  Google Scholar 

  43. Pook, M., Tuschen-Caffier, B., & Stich, N. (2002). Evaluation des Fragebogens zum Figurbewusstsein (FFB, deutsche version des body shape questionnaire). Verhaltenstherapie, 12(2), 116–124.

    Article  Google Scholar 

  44. Puhl, R. M., & Heuer, C. A. (2009). The stigma of obesity: A review and update. Obesity, 17(5), 941–964.

    Article  PubMed  Google Scholar 

  45. Pullmer, R., Kerrigan, S. G., Grilo, C. M., & Lydecker, J. A. (2021). Factors linking perceived discrimination and weight bias internalization to body appreciation and eating pathology: A moderated mediation analysis of self-compassion and psychological distress. Stigma and Health.

    Article  Google Scholar 

  46. R Core Team. (2021). R: A language and environment for statistical computing.

  47. Ritzert, T. R., Anderson, L. M., Reilly, E. E., Gorrell, S., Forsyth, J. P., & Anderson, D. A. (2016). Assessment of weight/shape implicit bias related to attractiveness, fear, and disgust. The Psychological Record, 66, 405–417.

    Article  Google Scholar 

  48. Robinson, B., “Bean,” E., Bacon, L. C., & O’reilly, J. (1993). Fat phobia: Measuring, understanding, and changing anti-fat attitudes. International Journal of Eating Disorders, 14(4), 467–480.;2-J

    CAS  Article  Google Scholar 

  49. Roddy, S., Stewart, I., & Barnes-Holmes, D. (2009). Anti-fat, pro-slim, or both? Using two reaction-time based measures to assess implicit attitudes to the slim and overweight. Journal of Health Psychology, 15(3), 416–425.

    Article  Google Scholar 

  50. Rudolph, A., & Hilbert, A. (2014). A novel measure to assess self-discrimination in binge-eating disorder and obesity (2005). International Journal of Obesity, 39(April), 1–3.

    Article  Google Scholar 

  51. Rudolph, A., Schröder-Abé, M., Riketta, M., & Schütz, A. (2010). Easier when done than said! implict self-esteem predicts observed or spontaneous behavior, but not self-reported or controlled behavior. Zeitschrift für Psychologie/Journal of Psychology.

    Article  Google Scholar 

  52. Schvey, N. A., Puhl, R. M., & Brownell, K. D. (2011). The impact of weight stigma on caloric consumption. Obesity, 19(10), 1957–1962.

    Article  PubMed  Google Scholar 

  53. Schwartz, M. B., & Brownell, K. D. (2004). Obesity and body image. Body Image, 1(1), 43–56.

    Article  PubMed  Google Scholar 

  54. Schwartz, M. B., Vartanian, L. R., Nosek, B. A., & Brownell, K. D. (2006). The influence of one ’ s own body weight on implicit and explicit anti-fat bias. Obesity.

    Article  PubMed  Google Scholar 

  55. Stein, J., Luppa, M., Ruzanska, U., Sikorski, C., König, H. H., & Riedel-Heller, S. G. (2014). Measuring negative attitudes towards overweight and obesity in the german population–psychometric properties and reference values for the german short version of the fat phobia scale (FPS). PloS One, 9(12), e114641.

    Article  Google Scholar 

  56. Teachman, B. A., Gapinski, K. D., Brownell, K. D., Rawlins, M., & Jeyaram, S. (2003). Demonstrations of implicit anti-fat bias: The impact of providing causal information and evoking empathy. Health Psychology, 22(1), 68–78.

    Article  PubMed  Google Scholar 

  57. Vieira, P. N., Palmeira, A. L., Mata, J., Kolotkin, R. L., Silva, M. N., Sardinha, L. B., & Teixeira, P. J. (2012). Usefulness of standard BMI cut-offs for quality of life and psychological well-being in women. Obesity Facts, 5(6), 795–805.

    Article  PubMed  Google Scholar 

  58. Vitousek, K. B., & Hollon, S. D. (1990). The investigation of schematic content and processing in eating disorders. Cognitive Therapy and Research, 14(2), 191–214.

    Article  Google Scholar 

  59. Watts, K., & Cranney, J. (2009). The nature and implications of implicit weight bias. Current Psychiatry Reviews, 5(2), 110–126.

    Article  Google Scholar 

  60. Weinberger, N., Kersting, A., Riedel-Heller, S. G., & Luck-Sikorski, C. (2016). Body image dissatisfaction in individuals with obesity compared to normal-weight individuals: A systematic review and meta-analysis. Obesity Facts, 9, 242.

    Article  Google Scholar 

  61. Williamson, D. A., White, M. A., York-Crowe, E., & Stewart, T. M. (2004). Cognitive-behavioral theories of eating disorders. Behavior Modification, 28(6), 711–738.

    Article  PubMed  Google Scholar 

  62. Wittchen, H.-U., Zaudig, M., & Fydrich, T. (1997). SKID-I/II Strukturiertes Klinisches Interview für DSM-IV. Beltz Verlag.

    Google Scholar 

Download references


Open Access funding enabled and organized by Projekt DEAL. The authors received no funding from an external source.

Author information




All authors have approved the final article. JB, FS, EN, and JS contributed conception and design of the study. FS was responsible for programming of the experimental paradigms. JB was responsible for data collection and JB, FS, and JS performed statistical analyses. JB and JS wrote the first draft of the manuscript. FS and EN contributed to manuscript revision and mentoring.

Corresponding author

Correspondence to Julia Baur.

Ethics declarations

Conflict of Interest

The authors declare no conflict of interest.

Informed Consent

All participants completed the informed consent process before participating in the present research study.

Research Involving Human and Animal Participants

No animal studies were carried out by the authors for this article.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.



Results with BMI as continuous variable.

See Fig. 3

Fig. 3

Relations of Body Mass Index (BMI) with D2 scores of the implicit measures. Linear regressions were conducted in R (R Core Team, 2021) using the robust HC2 estimator from the estimatr package (Blair et al., 2021). BMI was entered as a mean-centered predictor (average BMI = 28). For the Weight IAT, a positive intercept (b0 = 0.779; p < 0.001) indicated that participants generally evaluated normal weight as positive. However, the absence of a significant slope revealed that this effect was not related with BMI (b1 = − 0.011; p = 0.184; and b1 = 0.004; p = 0.459, after removing three bivariate outliers with a BMI of 45). For the Body SC-IAT, a negative intercept (b0 = − 0 .060; p < 0.05) indicated that participants evaluated their own body as slightly negative (at an average BMI of 28). Additionally, a negative slope (b1 = − 0.008; p < 0.05) indicated that it was the more negatively evaluated the higher BMI was. For the Body-Food SPF, a positive intercept (b0 = 0.245; p < 0.001) suggested that body and high-calory food were positively associated. Further, a positive slope (b1 = 0.005; p < 0.05) indicated that this association increased with BMI. For all scatterplots, regression lines with 95% CI are given. *p < 0.05

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Baur, J., Schmitz, F., Naumann, E. et al. Implicit Attitudes Towards Weight, One’s Own Body and its Relation to Food in Women with Overweight and Obesity. Cogn Ther Res (2021).

Download citation


  • Obesity
  • Implicit attitudes
  • Implicit association task
  • Body dissatisfaction
  • Weight bias