A total of 90 undergraduate students (79 women and 11 men) were recruited on campus during the orientation days of the semester. Their ages ranged between 19 and 38 years (M = 21.56, SD = 3.83). Their body mass index (BMI, based on self-reported weight and height) ranged between 17.2 and 35.9 kg/m2 (M = 21.9, SD = 3.5, n = 88). Eleven percent of the sample were underweight (i.e. definition of BMI < 18.5 kg/m2), 76 % were of normal weight (BMI ≥ 18.5 and < 25 kg/m2), 9 % were overweight (BMI ≥ 25 and < 30 kg/m2), and 3 % were obese (BMI ≥ 30 kg/m2).
The first step in the study was a Nutrition IAT. Subsequently, a questionnaire on the healthiness of the students’ diets and their preferences for certain foods was distributed to the participants and to be returned within the next three days. After the Nutrition IAT, the subjects went to another room, where a second experimenter presented four measured cups of drinks and a liquid taste evaluation questionnaire. The taste task served as a distraction from the health-related aspects of the beverages (high distraction situation). Finally, a behavioural food choice between fruits and snacks was presented to the participants. In this task, participants were shielded from possible distractions (low distraction situation).
The study was conducted according to the ethical guidelines of the German Psychological Society and of the Declaration of Helsinki. The project was approved by the Research Ethics Committee of the University of Education Schwäbisch Gmünd. On the monitor, prior to starting the Nutrition IAT, participants were informed that beginning the task would be taken to indicate consent to participate.
The participants indicated the healthiness of their diets with one item (healthy dietFootnote 1, i.e., “I eat a balanced diet: a lot of fruit, vegetables, and whole grains, limited consumption of fats, sugars, and cholesterol”, ), and responses ranged from not at all = 1 to very much = 4. In addition, participants rated their preferences for six unhealthy and six healthy foods (i.e., “Please indicate on a scale from 1 = not at all to 4 = very much the extent to which you like the following food.”). Food items were taken from the Nutrition IAT described below. The internal consistency for the preference for healthy food was α = .69, and that for the preference for unhealthy food was α = .52.
The present study’s Nutrition IAT integrated a self-concept measure using the categories of self versus other (see ) rather than pleasant versus unpleasant, the categories of attitude IATs (e.g., ). The stimuli for the target concept self versus others were a series of words; the category “self” corresponded to I, self, my, me, and own, and the category “others” corresponded to they, them, your, you, and others (translated from German; see ). The stimuli for the attribute category were photographs of healthy food items (i.e., apple, banana, carrot, coarse whole meal bread, pepper, and salad) and unhealthy food items (i.e., chocolate, chips, crisps, ice cream, gummy bears, and pizza).
The Nutrition IAT was administered on a laptop computer with the programme Inquisit 3.0.4 (Millisecond Software, Seattle, USA). The IAT procedure consisted of seven blocks . In the first block, the participants practiced target concept discrimination by categorising stimuli as self or other (20 trials). In the second block, they did the same for attribute discrimination by sorting photographs of food articles into the categories healthy and unhealthy (20 trials). The first series of critical trials consisted of 20 practice trials (Block 3) and 60 critical trials (Block 4). The participants categorised items into two combined categories, each of which assigned an attribute and a target concept to the same key (e.g., self + unhealthy for the left key, other + healthy for the right key). In Block 5, the participants practiced the switched key assignment for the attribute stimuli (20 trials). Block 6 and Block 7 (the second critical block) were complementary to Block 3 and Block 4; the only difference was that these blocks included the reversed key assignment that the participants had practiced in the previous block (e.g., self + healthy for the left key, other + unhealthy for the right key). The IAT score was calculated using the improved scoring algorithm proposed by Greenwald et al. . The IAT score is based on individualized differences between mean reaction time for self + unhealthy food and mean reaction time for self + healthy food. Positive scores indicate a stronger association between the self and healthy food than between the self and unhealthy food. No participants’ error rates exceeded 20 % (average error rate: 3.46 %), and no mean latencies exceeded two seconds. Applying the scoring algorithm (see ) to two mutually exclusive subsets of the IAT combined-task trials, the reliability (split-half method) of the IAT score was r = .86.
Food choice (low distraction situation): At the end of the session, the participants were offered a selection of fruits (i.e., apples, bananas) and snacks (i.e., small chocolate bars, sachets with gummy bears), in each case approximately hand-sized. They were asked to choose something (one piece) as a small reward for their participation. Their choices were coded as 1 = fruit and 0 = snack. The choice of fruit was regarded as healthy behaviour, whereas the choice of a snack was regarded as unhealthy behaviour. Three female participants did not take either fruit or a snack. Thus, all analyses relating to the food choice were calculated without the data of these three participants.
Liquid intake in the taste comparison task (high distraction situation): Four transparent cups, each filled with 155 g of drink (i.e., mineral water and three types of soft drinks: coke, sparkling apple juice, lemonade), were used for the taste comparison task to cover liquid intake. The participants were instructed to taste each of the four drinks and to rate their tastes to absorb their attention. They were given 9 min. A liquid taste evaluation questionnaire with four adjectives (i.e., fresh, tangy, natural, and sweet) and an overall taste assessment (i.e., “Please indicate on a scale from 1 = not at all to 6 = very much the extent to which you like the drink.”) was used. After the test, the liquid left in each cup was taken to another room and weighed by another experimenter. Thus, for the participants, the taste rating was the task. They were uninformed that the liquid left in each cup was being used to calculate their intake of each beverage.
The associations between the variables were analysed using zero-order correlations. In addition, we applied regression analyses using hierarchical entry procedures (forced entry) to evaluate the extent to which gender and age (variable Block 1), the three self-report measures healthy diet, preference for healthy food, and preference for unhealthy food (variable Block 2), and the Nutrition IAT (variable Block 3) contributed to the variance in each of the behavioural measures. Final equations were defined by a significant (p < .05) increase in R2 change after the last block was entered.