Whether individuals are aware that they are influenced by an environmental cue or not could be decisive for the cue’s effects. On the one hand, individuals who are aware that they are being influenced may attempt to correct for the influence (Wegener and Petty, 1995; see Mussweiler and Damisch, 2008). According to psychological reactance theory, this is because individuals try to restore their freedom of choice when they feel it to be threatened (Brehm, 1966; Miron and Brehm, 2006). Attempting to restore one’s freedom of choice can result in what has been termed “contrast effects”—that is, diminished or reversed effects of the threatening external stimuli such as environmental cues (e.g., Lombardi et al., 1987; Strack et al., 1993; see further Zhu et al., 2015). Individuals’ freedom of choice can even be threatened by internal self-imposed threats, arising from choosing specific alternatives and having to give up on others (Iyengar and Lepper, 2000; Miron and Brehm, 2006; Steindl et al., 2015). Thus, even self-imposed dieting-compatible behaviors could evoke reactance over time (Ungar et al., 2015).

However, there is evidence that contrast effects do not occur even though individuals are aware of being influenced. For example, when healthy instead of unhealthy snacks were placed near the cash register at train station snack shops, customers more often chose healthy snacks. This effect was maintained when a sign that read “We help you make healthier choices” was placed near the cash register (Kroese et al., 2016; see further Payne et al., 2016; Bruns et al., 2018). Moreover, individuals have been shown to successfully use environmental cues as reminders to implement their intentions. For instance, coffee shop customers who accepted a $1 coupon toward a future purchase along with a flyer that depicted a stuffed alien and mentioned the alien would be on the cash register as a reminder to redeem the coupon were more likely to use the coupon 2 days later when a stuffed alien was placed at the cash register, compared with customers who received a flyer that did not announce the stuffed alien (Rogers and Milkman, 2016).

While the role of awareness regarding the effects of environmental cues has as yet not been clearly determined, previous research has established the important role of goals (see Papies, 2017). The prominently discussed mechanism of how environmental cues influence behavior is that of activating goals (Förster et al., 2007; Papies et al., 2008; Papies, 2016a,b, 2017; Weingarten et al., 2016). This is possible as goals are embedded in cognitive structures that interlink the goals with mental representations of distinct environmental cues and behaviors. Such cognitive structures emerge when individuals repeatedly perform a behavior with similar goals in the same context (Papies, 2016b). For example, an individual’s repeated dieting attempts with the goal of losing weight may be linked to diet products; subsequently, seeing such diet products can activate the individual’s dieting goal (Anschutz et al., 2008b).

Goals are assumed to drive goal pursuit as they represent end states associated with a reward or positive affect (see Custers and Aarts, 2010). Correspondingly, activated goals have been shown to have stronger effects on behavior when the positive affect related to the goals was higher—in other words, when goals were highly valued by the individuals holding the goals (see Bargh, 2016; see Papies, 2016b; Weingarten et al., 2016). Also in line with this, environmental dieting cues have been shown to have a particularly pronounced effect on individuals who hold a chronic dieting goal, the restrained eaters (Herman and Mack, 1975; Herman and Polivy, 1980; Weingarten et al., 2016; Polivy and Herman, 2017; see Buckland et al., 2018; Masterson et al., 2019). For example, only restrained eaters tasted fewer meat samples in a butcher shop when exposed to a poster on the entrance door that announced a weekly recipe as being low-calorie and “good for a slim figure” (Papies and Hamstra, 2010). Similarly, the portrayal of the exaggeratedly thin Giacometti figures (Stämpfli and Brunner, 2016) was found to specifically influence restrained eaters (Stämpfli et al., 2017). A meta-analysis of the effects of environmental dieting cues on food intake supported the important role of dieting goals. Overall, dieting cues were found to reduce food intake only to a minor extent. In contrast, for individuals with a strong dieting goal, including both restrained eaters and dieters, a small-to-moderate effect was found (Buckland et al., 2018). While restrained eaters are defined as having a chronic dieting goal or “concern for dieting” (see section “Measures”; Dinkel et al., 2005), dieters can be characterized as actively engaging in a weight management attempt at the moment and thus having a current dieting goal.

Overall, based on the presented research, we expected that individuals with restrained eating tendencies, and thus a chronic dieting goal per definition, might be able to successfully apply environmental dieting cues for themselves to facilitate the pursuit of their longer-term goal of losing weight.

In addition to the question of whether individuals can deliberately apply environmental cues for themselves to facilitate the pursuit of a personal goal, the present study addressed a second question that, to date, has received little attention: whether or not environmental cues in general (van Kleef and van Trijp, 2018) and dieting cues in particular can be applied effectively over a longer period of time (Papies, 2016b; Buckland et al., 2018). This is essential regarding the dieting goal, as losing weight requires time.

The scant evidence on environmental dieting cues which are applied over a longer period of time has revealed ambiguous findings. On the one hand, two studies on snacking frequency over 2 weeks indicate that an environmental dieting cue (a photograph of a slim model) can reduce the influence of unhealthy eating habits (Ohtomo, 2017). On the other hand, a similar environmental dieting cue (a picture of a slim model) was found not to help individuals lose weight during two “weight loss programs” that utilized eating diaries over periods of 1 week (Klesse et al., 2012; see for a further example, Stice et al., 2001). In the first weight loss program, the cover of an eating diary depicted either a thin model or a measuring tape; in the second, the cover of the eating diary depicted either a thin model or the same model adjusted to appear normal in size. At the end of a week, the participants exposed to a thin model had gained some weight or lost less weight compared with the participants exposed to the normal-size model or the measuring tape. The results of these studies were explained by the discouraging effect of the thin models which made the participants perceive their weight loss goals as being less attainable (Klesse et al., 2012).

A 2 (cue: visually dieting-related vs. visually neutral cue) × 2 (awareness: being aware vs. not being aware of the cue’s facilitating influence) between-subjects design was applied, with losing weight over time as the dependent variable. This design resulted in a condition in which participants applied the visually neutral cue which was not presented to them as being a dieting facilitator, the control condition Rothko not aware (see section “Cues”), and three experimental conditions related to dieting: one in which participants applied a visually dieting-related cue which was not presented to them as being a dieting facilitator, the experimental condition Giacometti not aware, and two in which participants were told that their cue had been shown to facilitate dieting in the past and in which the participants either applied a visually neutral cue (experimental condition Rothko aware) or a visually dieting-related cue (experimental condition Giacometti aware).

Briefings for the study were held at the university campus in November during the 2 weeks before the beginning of the study. The briefing sessions were held separately for each condition, with a maximum of five participants in each briefing session.

At the briefing, the participants were first asked to sign a written informed consent form which explained that they would be weighed with a body composition monitor scale at the beginning and end of the study and that they would have to record their weight daily during the 26 weeks of the dieting study. They were also informed that the study team would not include any physicians or nutrition counselors and that it would be the participants’ responsibility to evaluate if it would be advisable to contact a physician due to their participation in the study. They were advised that they should only participate if they felt psychologically and physically able to do so and if there were no medical reasons for them not to participate. In addition, participants were informed that they had the right to end their participation at any time without giving a reason and that weight loss below the normal weight range was not encouraged. Finally, they were asked not to discuss the study with other participants.

After the participants gave their written informed consent, also at the briefing, they answered the restrained eating questions, received their dieting diaries and adhesive stickers, and were verbally instructed on how to use the stickers and how to record their data. Depending on the condition, they were either informed or not that their cues were dieting facilitators.

The participants were then weighed individually on the body composition monitor scale (Tanita BC-545). They removed shoes and heavy clothing items when they were weighed. In addition, care was taken that participants removed jewelry and belts and had no electronic medical implants in order to ensure valid and safe measurements. In an attempt to prevent any encouragement of unhealthy diets, all participants received a flyer that stated that it should not be their goal to lose weight as fast as possible, but instead to develop a joy in eating in a healthy and well-balanced manner over the long run. To support participants’ motivation to participate in the study, the flyer also stated that weight fluctuations are a normal part of dieting.

The study began in the first week of December and ran for 26 weeks, until the end of May. In June, participants attended debriefing sessions, which took place in the same location as where the briefings had been held. Most of the participants (88.8% of the 107 remaining in the study sample) were weighed in the first week after the study had ended, and some of them were weighed in the second week (5.6%) or third week after the study’s end (more precisely after 17 days, also 5.6%). At the debriefing, the dieting diaries were collected, and participants were measured individually using the body composition monitor scale. Afterwards, all participants were informed of the study’s purpose and were debriefed, both verbally and in writing.

The data of 107 participants were analyzed. Regarding the exposure to the cues, it is important to note that all participants were exposed to the cues in their dieting diaries in which they had been advised to record their weight daily. Regarding the use of the stickers that depicted the cues, 54.2% of the 107 participants stated that they applied the stickers (indicating a value of 5 or higher on a 7-point scale), and 40.2% stated that they had not applied the stickers (indicating a value of 3 or lower). The first group saw the stickers about four times per day on average according to their own statements. An analysis of variance (ANOVA) with cue, awareness, and their interaction revealed that the sticker usage did differ depending on the participants being in an awareness condition or not, F(1,103) = 4.07, p = 0.046, η²_p = 0.04 (adjusted R² = 0.032). Participants to whom the cues were presented as dieting facilitators (i.e., participants in the awareness conditions) agreed more strongly that they used the stickers (M = 4.72, SD = 2.17) than participants to whom the cues were not presented as dieting facilitators (i.e., participants in the no awareness conditions, M = 3.76, SD = 2.46). There was no main effect of cue, that is, participants who applied the visually dieting-related cue did not agree more strongly that they used the stickers than participants who applied the visually neutral cue, F(1,103) = 1.09, p = 0.299, η²_p = 0.01, or an interaction of cue and awareness, F(1,103) = 0.83, p = 0.364, η²_p = 0.01. The participants (N = 107) most often affixed the stickers in their kitchen (most commonly on or in the fridge), on items that they brought with them “on the go,” and in the eating area.

An ANOVA with cue, awareness, and their interaction revealed no main effect of cue, F(1,103) = 0.29, p = 0.594, η²_p = 0.00, or awareness, F(1,103) = 2.28, p = 0.135, η²_p = 0.02, on participants’ percentage weight change. Cue and awareness also did not interact, F(1,103) = 0.15, p = 0.703, η²_p = 0.00 (adjusted R² = −0.004).

However, when restrained eating was included as a covariate, an analysis of covariance (ANCOVA), which also included cue and awareness, and all interactions of cue, awareness, and restrained eating, with the percentage weight change as the dependent variable, revealed a three-way interaction of cue, awareness, and restrained eating, F(1,99) = 6.85, p = 0.010, η²_p = 0.07; η² = S⁢SE⁢f⁢f⁢e⁢c⁢tS⁢SC⁢o⁢r⁢r⁢e⁢c⁢t⁢e⁢d⁢T⁢o⁢t⁢a⁢l = 0.06 (adjusted R² = 0.061).

Analyzing the interaction of cue and awareness at values of restrained eating with a moderation analysis, using the PROCESS macro for SPSS v3.4 (model 3 was used; Hayes, 2018), revealed the regions where cue and awareness interacted significantly (restrained eating values greater than 4.49 and less than 2.05, see Table 1; model summary: R² = 0.123, p = 0.065).

Focusing on the upper restrained eating region where cue and awareness interacted significantly—that is, on “highly restrained eaters” with restrained eating values greater than 4.49 (n = 27, 25.23% of the observations)—an ANOVA with cue and awareness as independent variables revealed a significant interaction of cue and awareness on participants’ percentage weight change, F(1,23) = 5.19, p = 0.032, η²_p = 0.18 (adjusted R² = 0.146). Subsequent custom hypothesis tests (contrasts) revealed that, compared with the highly restrained eaters (i.e., with a restrained eating value >4.49) in the control condition Rothko not aware, to whom the visually neutral cue was not presented as a dieting facilitator (M = 0.86%, SD = 2.87), the highly restrained eaters in the experimental condition Rothko aware, who applied the visually neutral cue but were told that it was a dieting facilitator, lost weight (M = −7.72%, SD = 2.47), p = 0.012. Highly restrained eaters who applied the visually dieting-related Giacometti cue and were either not told, in the experimental condition Giacometti not aware (M = −3.06%, SD = 5.33), p = 0.142, or told that it was a dieting facilitator, in the experimental condition Giacometti aware (M = −2.63%, SD = 6.13), p = 0.199, did not significantly lose weight during the study compared with the highly restrained eaters in the control condition. Further, contrasts did not reveal any other statistically significant weight change differences among the participants of the experimental conditions.

Focusing on the lower restrained eating region where cue and awareness interacted significantly (restrained eating values <2.05, n = 9, 8.41% of the observations), an ANOVA did not indicate a significant weight change difference.

To enlarge the limited sample size of the highly restrained eaters (n = 27) which resulted from the moderation analysis, the ANOVA and the subsequent custom hypothesis tests (contrasts) were repeated for the restrained eaters derived from the median split of restrained eating (Mdn = 3.50). The median split allowed us to take a closer look at a larger number of restrained eaters (n = 49). The ANOVA for the restrained eaters (restrained eating values >3.50) with cue and awareness as the independent variables revealed a significant interaction of cue and awareness on percentage weight change, F(1,45) = 5.17, p = 0.028, η²_p = 0.10 (adjusted R² = 0.154). The custom hypothesis tests (contrasts) revealed that, compared with the restrained eaters in the control condition Rothko not aware, who applied the visually neutral cue and were not told that it was a dieting facilitator (M = 0.37%, SD = 3.47), restrained eaters in the experimental condition Rothko aware, who applied the visually neutral cue but were told that it was a dieting facilitator (M = −6.02%, SD = 4.28), lost weight, p = 0.002. Restrained eaters in the experimental condition Giacometti aware, who applied the visually dieting-related cue and were told that it was a dieting facilitator (M = −3.40%, SD = 4.99), p = 0.024, also lost weight compared with the restrained eaters in the control condition. Restrained eaters in the experimental condition Giacometti not aware, who applied the visually dieting-related cue and were not told that it was a dieting facilitator (M = −2.82%, SD = 4.22), tended to lose weight compared with the restrained eaters in the control condition, p = 0.055 (effect at p < 0.10); see Figure 1. Further, simple contrasts did not reveal any other statistically significant weight change differences between the participants of the experimental conditions.

The ANOVA for participants low in restrained eating (restrained eating values ≤3.50, n = 58) indicated an interaction of cue and awareness at the level p < 0.10, F(1,54) = 3.18, p = 0.080, η²_p = 0.06 (adjusted R² = 0.048).

Restrained eaters in the present study deliberately applied cues that were presented to them as being dieting facilitators with some success, and there were no indications of contrast effects. A plausible explanation is that contrast effects may more likely occur if individuals are aware of influences that are incongruent with their goals. In the present study, in which all the participants wanted to lose weight, it can be assumed that the environmental cues that were presented as dieting facilitators were perceived as goal-congruent.

The following evidence supports the assumption that contrast effects can occur for goal-incongruent influences but do not occur for goal-congruent influences of which individuals are aware. When individuals actually trained for a task, they performed better when they were told that the training had a positive influence than when they were told that the training had a negative influence (Nicolao et al., 2016). This suggests that individuals could counteract the actual positive influence from the training when they believed that the training had a negative influence. Importantly, this effect only occurred when the training actually prepared individuals for the subsequent task. This rules out a mere placebo effect (Nicolao et al., 2016)—that is, an effect which cannot be attributed to an actual treatment but to individuals’ belief of being treated (Fontaine et al., 2016).

Previous research on dieting has indicated that placebo effects can affect weight loss (Panayotov, 2019; see Fontaine et al., 2016). For instance, hotel room attendants who were told that their work-related physical activity exceeded the recommended amount lost weight within 4 weeks, independent of their actual physical activity levels (Crum and Langer, 2007). In the present study, the weight loss effect of the visually neutral cue on participants who believed it to be a dieting facilitator could be considered a placebo effect. However, as we assume that the environmental dieting cues used in this study influenced behavior by activating the dieting goal (see Papies, 2016b), individuals could be seen as treated if the cues actually activated their dieting goal, also when the cue per se was visually neutral.

As a limitation of the present study, stronger demand effects (Sawyer, 1975) for participants in the awareness conditions, and especially in the condition Rothko aware in which participants applied a visually neutral cue and were told that it was a dieting facilitator, cannot be ruled out. In terms of limitations of the present study, it should also be noted that the participants differed in whether they applied the stickers and how often they had seen them, according to their self-reported information. Exposing study participants to a cue more reliably, such as via smartphone alerts, should improve future studies. Further, regarding limitations of the present study, it has to be considered that we did not control for other weight management practices—that would have exceeded the weight management practices imposed by the study’s design—in which participants engaged in during the study. Therefore, it is possible that there were supporting influences regarding losing weight, such as concurrent dieting support from a physician. Another limitation is that the study was not preregistered. As a last limitation that should be mentioned, the effects of intervention studies can diminish after the intervention ends (for example, Inauen et al., 2017). After the present study had ended, the weight development of the participants was not tracked.

In the present study, we found that restrained eaters with their highly valued chronic dieting goal (Papies, 2016b; Weingarten et al., 2016; Polivy and Herman, 2017) could deliberately apply environmental dieting cues to facilitate the pursuit of their personal goal of losing weight. This is consistent with previous evidence showing that activated goals have a stronger effect on behavior when the goals are highly valued by the individuals (Weingarten et al., 2016). Interestingly, although all of our participants could be assumed to have had highly valued dieting goals, given that they chose to participate in a 6-month dieting study, the dieting cues only had an effect on the restrained eaters. As restrained eating was defined to be chronic dieting (Polivy and Herman, 2017), this may suggest that an environmental dieting cue that is implemented over a longer period of time must activate a persistent or chronic goal in order to elicit longer-term effects. Thus, previous evidence (Buckland et al., 2018) and the results of the present study suggest that, while environmental dieting cues can generally have immediate effects on individuals who highly value their dieting goal—that is, also on individuals who may value a goal only shorter-termed such as dieters—effects over a longer period of time require a highly valued persistent dieting goal.

In addition to persisting goals, perceived goal attainability could have facilitated the effects of the environmental cues in the present study that were presented to the participants as dieting facilitators. Accepted as such, the cues could have allowed participants’ goal of losing weight to appear more attainable. Previous research has indicated that perceived goal attainability plays a role in the effectiveness of environmental dieting cues (Klesse et al., 2012). Perceived goal attainability could be decisive for enabling the positive affect related to a goal or the related dieting cue to drive goal pursuit (Anschutz et al., 2008a; Custers and Aarts, 2010). A conceivable way to enhance perceived goal attainability could be social support, for instance, via smartphone chat messages (Inauen et al., 2017).

There is little evidence about the effects of environmental dieting cues over longer periods of time (Papies, 2016b; Buckland et al., 2018). Indeed, some health intervention studies have examined the effects of environmental cues on food choices, sales, or consumption in places such as restaurants and cafeterias over time spans of up to several months (see Allan et al., 2017; Cadario and Chandon, 2018; see Carter et al., 2018). Such public health intervention studies are of significant practical value. However, they do not allow researchers to examine the effect of repeatedly presented environmental cues on particular individuals—with the individuals’ characteristics such as restrained eating which can be decisive for the cues’ effects. This is because, in public eating settings, the clientele does not remain constant. While it may remain relatively constant in on-site eating settings (e.g., in universities or worksite cafeterias), this will not be the case in off-site eating settings such as restaurants, cafes, or cinemas. Thus, mainly short-term effects of environmental cues on different individuals are captured when health intervention studies are evaluated. In line with this, a meta-analysis of public health intervention studies in on-site eating settings, off-site eating settings, and grocery stores found that the effects of healthy eating interventions were largely independent of study duration. Increasing the study duration from 1 to 15 weeks reduced the effect size by only 13% (Cadario and Chandon, 2018).

Thus, further studies are needed to examine the effects of repeatedly presented environmental dieting cues on specific individuals—ideally, with individuals who apply the cues deliberately for themselves. Such studies should address various temporal matters (Scholz, 2019). For example, they could shed light on the aspect of habituation. In general, to apply a distinct cue repeatedly could lead individuals to habituate to the cue, which would involve a decrease in their behavioral response (Rankin et al., 2009; Silvestrini and Gendolla, 2011; Papies, 2016b). Importantly, future studies should have large enough sample sizes. Further, as in the present study it was not controlled for all weight management practices that participants engaged in during the study, which could have supported losing weight, studies which control for more potentially intervening influences are needed to confirm the present study’s findings. In formal terms, future studies should also have preregistered research protocols that indicate, among other factors, the research hypotheses and criteria for data exclusions.

Today’s obesity epidemic with its increased number of obese adults (World Health Organization [WHO], 2020b) indicates the relevance of the present study. In our obesogenic environment (Hill et al., 2003; Papies et al., 2014), to deliberately use environmental dieting cues in tempting situations—for example, at the fridge—to activate our dieting or healthy eating goals could help. An advantage of this approach is that it can be pursued with minimal effort. In contrast, interventions which attempt to modify (instead of activating elements of) the existing cognitive structures that underlie individuals’ health-related behavior, such as making specific plans for how to behave in response to tempting situations, mostly involve a form of training which is normally time-consuming (Haasova et al., 2016; Papies, 2016b).

However, the effects of environmental dieting cues over a longer period of time may depend not only on the presence of highly valued goals which persist over time, as suggested by the present research, but are dependent on individuals’ possibility to pursue their health goals (Papies, 2016b). This possibility is facilitated by environments in which healthy food options are available, accessible, and salient (Papies, 2016b, 2017; see also Kwasnicka et al., 2019). This is in line with nudging (Thaler and Sunstein, 2008; see also Hansen and Jespersen, 2013). Nudging is used as an “umbrella term” for influences of the environment on individuals’ behavior—of which the influenced individuals are usually not aware (Chartrand, 2005)—and has been more specifically defined as enhancing the accessibility or salience of behavioral options to change behavior (Papies, 2017). Placing healthy products near the cash register (Kroese et al., 2016) is an example of a nudging intervention, as is listing healthy food options first on menus and/or highlighting them as healthy (Policastro et al., 2017; as an existing example see the Nutri-Score, Giner and Brooks, 2019). According to the World Health Organization [WHO] (2018), overweight and obesity can be reduced by supportive environments that make the choice of healthier foods the easiest choice (“the choice that is the most accessible, available, and affordable”). Activating dieting goals using environmental cues, ideally with individuals who apply the cues deliberately for themselves according to the results of the present study, in combination with a supportive environment in terms of improved availability, accessibility, salience, and affordability of healthy food options, could thus represent a promising strategy to address the problem of obesity. With regard to the limitations of the present study, however, the present study should be seen as a starting point for future improved long-term studies, which, for example, could have a larger sample size and control for more potentially intervening influences.