To undertake this study was purposely created a questionnaire composed of different parts intended to access different information about the sample at study: Part I, Sociodemographic data; Part II, Food knowledge; Part III, Eating habits; and, Part IV, Information about a healthy diet. In order to measure the participants’ food knowledge in Part II were included questions intended to evaluate the students’ degree of knowledge about a healthy diet, as well as other statements about which the students were asked to state their accordance measured on a 5-point Likert scale varying from 1 (strongly disagree) to 5 (strongly agree) (Likert, 1932). In Part III of the questionnaire were included different questions about the participants’ eating patterns and also about their resting habits. To evaluate the level of information about a healthy diet in Part IV were included different statements intended to characterize where the participants obtained the information about this subject, measured on a 5-point Likert scale.

It was undertaken a descriptive cross-sectional study on a non-probabilistic sample of 670 Portuguese university students belonging to one of the 5 schools of the Polytechnic Institute of Viseu (IPV) and the Catholic University of Viseu. The questionnaires were applied only to adults (aged 18 or over) and after verbal informed consent. All ethical issues were followed when designing and applying the questionnaire, which was approved by the Ethical Committee with reference nº 13/2017.

Height and weight were obtained by self-response and used to calculate the Body Mass Index (BMI), dividing weight (kg) by height squared (m.). Then, the BMI scores were classified into classes according to the standards of the International Classification: underweight (BMI < 18.50 kg/m.), normal weight (18.50 ≤ BMI ≤ 24.99 kg/m.), overweight (25.00 ≤ BMI ≤ 29.99 kg/m.) and obese (BMI ≥ 30.00 kg/m.) (WHO, 2006).

Basic descriptive statistics tools were used for the exploratory analysis of the data and in all tests the level of significance considered was 5%. All data analysis was made using the software IBM SPSS (version 25). To assess the relations between some of the categorical variables it was used the Chi square test. The Cramer’s V coefficient was used to evaluate the strength of the significant relations found between some of the variables under study. This coefficient varies from 0 to 1, and for V ≈ 0.1 the association is considered weak, for V ≈ 0.3 the association is moderate and for V ≈ 0.5 or over, the association is strong (Witten & Witte, 2009).

In order to minimize the possibility of interpretative divergences and to assign a clear and precise meaning to the variables, creating the possibility of making them measurable, the conceptual and empirical operationalization of the variables was performed.

To create the variable food knowledge, the first step was to attribute a quotation to each question included in Part II of the questionnaire (Table Nº 1). Then, it was calculated a subtotal of each question and these subtotals were used to calculate the total for each participant, thus giving the score for the variable food knowledge. This variable corresponded to values varying from 9 to 45, and which could be interpreted as follows: [9 - 15[ = poor food knowledge; [15 - 30[ = satisfactory food knowledge; [30 - 40[ = good food knowledge; [40 - 45] = very good food knowledge.

Table 1 Table 1 Operationalization of the variable participants’ food knowledge

The process to calculate the variable eating habits was the same that was used for the variable food knowledge. In this case it was given the quotation 1 when the answer was adequate to healthy eating habits and the quotation 0 when it was not (Table Nº 2). The variable eating habits ranged from 0 to 24, interpreted as follows: [0 - 5[ = poor food eating habits; [5 - 12[ = satisfactory eating habits; [12 - 20[ = good eating habits; [20 - 24] = very good eating habits.

Table 2

(1) Quotation takes value = 1 when the answer was adequate to healthy eating habits; and 0, when it was not

The sample was composed by 670 university students, from which 63.4% were women and 36.6% were men. The age of the participants ranged from 17 to 70 years old, being on average 21.85 ± 5.51 years. As it can be observed in Table Nº 4, 46.0% of the participants were in the age group of 19-21 years. As for the level of education, most of the participants (92.3%) attended a license degree, followed by the participants who attended others degrees (6.7%) and only a few percentage of them (1.0%) attended a master degree. Regarding the area of studies, 42.8% of the participants’ studies were from areas that were not specified in the questionnaire, followed by the students from areas related to health (34.6%). The majority of the participants, 73.9%, had a normal weight and only 3.4% were obese. The results further showed that only a minority of the participants, 5.2%, practiced a high competition sport, 77.8% had not health problems and 9.4% of them had food allergies.

Table 4 Table 4 Sociodemographical characterization of the sample Source: Own elaboration, based on the research data

For the whole sample the mean score for the participants’ food knowledge was 35.94 ± 4.40 (scale from 9 to 45), meaning that –in general– the students had a good food knowledge. Most of the participants, 77.6%, identified the correct food wheel for the Portuguese population (Figure Nº 1) and revealed a very good degree of knowledge about different nutritional aspects, with most values over 4, on average, on a scale from 1 (poor) to 5 (very good) (Table Nº 5). Moreover, 74.0% of the participants believed that it is extremely important to follow a healthy diet. The results further showed that 26.5% of the participants answered that in the course they attended are frequently included programmatic contents related to a healthy diet (never or rarely: 25.1%; sporadically: 10.8%; sometimes: 19.9; always or almost always: 17.8%).

Figure 1 Figure 1 Question about the identification of the correct Portuguese Food Wheel Source: own elaboration, based on FAO (2019)

Table 5 Table 5 Mean scores obtained for the participants’ degree of knowledge about some nutritional aspects (scale from 1 = poor to 5 = very good) Source: Own elaboration, based on the research data

When seen by gender (Table Nº 6), for both men and women the participants’ food knowledge was good. In view of these results there were not found significant differences between genders, meaning that gender did not influence the participants’ food knowledge (. = 0.598). Nevertheless, in previous studies it was found that women tended to have a better food and nutrition knowledge than men (Kresić, Kendel, Pavicić, Cvijanović & Ivezić, 2009; Yahia, Wang, Rapley & Dey, 2016).

Table 6 Table 6 Participants’ food knowledge according to some of the variables under study Source: Own elaboration, based on the research data

It was also analysed the participants’ food knowledge according to the age group (Table Nº 6) and a higher percentage of the participants aged 22 or over, 22.1%, had a very good food knowledge, when compared to those from the other age groups (age . 18y: 9.5% and 19y ≤ age £ 21y: 12.2%). The results of the Chi square test showed that there were significant differences among age groups, meaning that age group influenced the participants’ food knowledge. However, the association between the variables was weak (V = 0.119). These results may be due to the fact that the older participants were in more advanced years of the courses they attended and therefore had already acquired more knowledge.

As for the participants’ food knowledge according to the area of studies (Table Nº 6), it was found that, regardless the area of studies, most of the participants had good food knowledge (health: 71.1%; food: 55.8%; sport: 85.6%; other: 74.5%). However, it is important to note that a higher percentage of the participants from health and food areas had a very good food knowledge, 22.8% and 37.2% respectively, when compared to those from sport (9.6%) and other areas (7.7%). Independently of the area of studies, none of the participants had a poor food knowledge. The results of the Chi square test showed that there were no significant differences between the areas of studies (. = 0.463). In the study performed by Jones et al. (2015), it was found that students who had taken a college-level nutrition course tended to have better food nutrition knowledge than those who had not.

Regarding the influence of BMI class on the participants’ food knowledge, as it can be observed in Table Nº 6, there were no significant differences among the different BMI classes, meaning that this variable did not influence the participants’ food knowledge. The results further show that independently the BMI class, most of the participants had a good food knowledge.

There were also investigated other possible associations, namely the presence of a health problem and the influence of practicing a high competition sport. However, there were no significant differences in none of those cases (Health problems: . = 0.978; Practice of high competition sport: . = 0.938), meaning that the presence of a health problem and the practice of a high competition sport did not influence the participants’ food knowledge.

For the sample at study, the mean score for the participants’ eating habits was equal to 11.42 ± 2.27 (scale from 0 to 24), which means that in general the participants’ eating habits were satisfactory. These results are not in line with previous scientific research where it was found that university students’ eating habits are usually characterized by unhealthy food choices (Deliens, Clarys, De Bourdeaudhuij & Deforche, 2014). A considerable percentage of the participants, 40.8%, ate 5 meals per day and only 0.6% ate less than 3 meals per day (3 meals/day: 7.6%; 4 meals/day: 25.9%; 5 meals/day: 40.8%; 6 or more meals/day: 25.1%). Furthermore, 6.6% of the participants indicated that usually do not eat breakfast, 17.8% do not take a morning lunch, 1.0% skip lunch, 5.7% do not eat the afternoon lunch, 1.7% skip dinner and 40.7% of the students do not eat the evening supper. In another study developed in Spain it was found that 98% of the university students in the study had breakfast and almost 75% of them had at least four meals per day (García-Meseguer, Burriel, García & Serrano-Urrea, 2014). The results also showed that, in general, the participants’ breakfast included milk (54.0% of positive answers), bread (45.8% of positive answers), cereals (32.7% of positive answers) and coffee (23.9% of positive answers). It is important to highlight that only few students answered that at breakfast they consumed chocolate (2.4%), chips (0.1%), Coca-Cola (2.1%), snacks (4.6%), juices (5.5%) or cakes (12.1%). When the participants were asked about what they ate at lunch, 30.1% of them answered that they ate soup, 28.4% consumed vegetables, 28.2% ate pasta, 21.6% fish, 16.0% potatoes, 2.5% yoghurt, 15.5% soft juices, 4.3% juices, 23.6% salad, 10.4% chips, 58.2% meat, 38.5% rice, 27.8% fruit, 9.4% a sweet desert and 44.9% drank water. Several studies found that the intake of fruits and vegetables among university students is generally low (Peltzer & Pengpid, 2015).

Table Nº 7 shows the participants’ eating habits according to some of the variables, and when seen by the age group it was found that a higher a percentage of the participants aged 22 or over, 36.9%, had good eating habits when compared to those from the other age groups, but also a higher percentage of them revealed poor eating habits, 1.7%, than the participants from the other age groups. The results of the Chi square showed that the differences among age groups where statistically significant (. = 0.021), and therefore age group influenced the participants’ eating habits. However, the association between the variables was weak (V = 0.093).

Table 7 Table 7 Participants’ food knowledge according to some of the variables under study Source: Own elaboration, based on the research data

According to gender, it was observed that the results were similar for both men and women and that both genders revealed satisfactory eating habits (68.1% for women and 69.2% for men). In this case, there were no significant differences between genders. According to other studies, in non-stress periods, female students tended to adopt better eating habits than male students, but that tendency is different when female students are stressed and in those periods they tended to increase the consumption of high-sugar, high-fat foods and unhealthy snacking, while reducing the intake of healthier foods like vegetables (Economos, Hildebrandt & Hyatt, 2008; Papier, Ahmed, Lee, & Wiseman, 2015; Sharma, Harker, Harker, & Reinhard, 2010).

Regarding the participants’ eating habits according to the BMI class, as it can be observed in Table Nº 7, independently of the BMI class, the majority of the participants revealed satisfactory eating habits, with no significant differences among BMI classes. These results are not surprising, because in other studies it was also reported some inconsistent findings between BMI and dietary patterns (Aparicio et al., 2017).

When the eating habits were analyzed according to the area of studies, it was observed that, regardless the area of studies, most of the participants had satisfactory eating habits. However, none of the participants from food areas revealed poor eating habits, when compared to those from other areas (sport: 0.9%; health: 1.7%; other areas: 1.4%). Nevertheless, in none of the cases the participants’ eating habits were very good. The results of the Chi square test showed that there were significant differences among the areas of studies (.= 0.001) and therefore the area of studies influenced the participants’ eating habits, but with a weak association between the variables.

As it can be observed in Table Nº 8, most of the participants who practiced high competition sports, 51.4%, revealed good eating habits, while the majority of the participants who did not practice, 69.9%, revealed satisfactory eating habits. Therefore, not surprisingly were found significant differences among the two groups (. = 0.015), with a weak association between the variables (V = 0.113). In fact, there are specific guidelines for athletes and proper eating habits provide the achievement of a high performance (Galanti et al., 2014; Meyer, 2015; Shi & Gisolfi, 1998).

Table 8 Table 8 Participants’ eating habits according to the practice of high competition sports Source: Own elaboration, based on the research data

It was also analysed if the presence of health problems influenced the participants’ eating habits and no association was found (. = 0.222), being the eating habits of most of the participants who had a health problem satisfactory (74.3%), as well as the eating habits of those who had not (67.2%).

Since nutrition knowledge is one of the factors that influence eating habits (Spronk, Kullen, Burdon & O’Connor, 2014), it was evaluated if a higher food knowledge was associated with better eating habits (Table Nº 9). The results suggested that a better food knowledge corresponded to better eating habits. The results of the Chi square test showed that food knowledge influenced the participants’ eating habits (. = 0.016), with a weak association between the variables (V = 0.096). Nonetheless, for most of the participants, regardless the food knowledge, their eating habits were satisfactory. In another study by Abraham, Noriega & Shin (2018), it was found that despite the participants’ food knowledge, their food choices were not necessarily healthy.

Table 9 Table 9 Participants’ eating habits according to food knowledge Source: Own elaboration, based on the research data

A sleep duration from 7 to 9 hours is necessary for young adults’ optimal health (18-25 years), according to the National Sleep Foundation (Hirshkowitz et al., 2015). In the present study, the participants slept on average 7.01 ± 1.95 hours, which is considered adequate. As for the association between the participants’ eating habits and their sleeping hours (Table Nº 10), it was found that, independently of the sleeping hours, the majority of the participants had satisfactory eating habits. Therefore, no significant differences were found, meaning that the participants’ sleeping hours did not influence their eating habits. The results of the study performed by Nedeltcheva et al. (2009) suggested that a recurrent sleep deprivation is associated with an increased consumption of excess calories from snacks, however without a statistically significant change in the intake of energy from meals.

Table 10 Table 10 Participants’ eating habits according to their sleeping hours

Table Nº 11 presents the participants’ frequency of usage of the different sources of information about a healthy diet and the most privileged sources of information were the parents/family and also the television, with 43.2% of the participants, in both cases, indicating that used those sources of information frequently. These results are not surprising, since parents transmit information about dietary attitudes to their offspring and therefore influence their eating patterns (Deliens et al., 2014; Larson & Story, 2009). It is important to note that 41.8% of the participants indicated that they agree with the fact that their parents/family are concerned about preparing balanced meals for them and 39.2% of them are concerned with the preparation of balanced meals for themselves. When the students were asked if they already attended a class performed by a nutritionist, the majority of them, 69.6%, answered that they did not, against 30.4% of the participants who answered yes.

Table 11 Table 11 Participants’ frequency of usage of the different sources of information about a healthy diet Source: Own elaboration, based on the research data

Most of the participants, 68.9%, had a satisfactory level of information about a healthy diet, 29.7% good and only 1.4% of them revealed a poor level of information about this subject (Table Nº 12). When seen by age group, the results were not totally elucidative, since the participants aged 22 or over were the only ones who revealed a poor level of information about a healthy diet (1.7%), but they were also the ones with a higher percentage of students with very good information about this topic (3.5%). Nonetheless, were found significant differences among age groups (.= 0.020), meaning that this variable influenced the participants’ level of information about a healthy diet. However, the association between the variables was weak (V = 0.108).

Table 12 Table 12 Participants’ level of information about a healthy diet Source: Own elaboration, based on the research data

According to gender, the results suggested that women tended to have a higher level of information about a healthy diet than men, with significant differences between genders (. = 0.001). Again the association between the variables was weak (V = 0.166). In previous studies it was also found that women tend to search more information about health issues than men, especially using the Internet (Bidmon & Terlutter, 2015).

As for possible influence of BMI class, as it can be observed in Table Nº 12, it was found that BMI class in fact influenced the participants’ level of information about a healthy diet (. = 0.041), with a weak association between the variables (V = 0.095). Also the area of studies proven to influenced the participants’ level of information about a healthy diet (. = 0.045), but again with a weak association between the variables (V = 0.095).

There were also analysed other possible associations, namely the influence of the practice of high competition sports and the presence of health problems. The practice of high competition sport did not influence the participants’ level of information about a healthy diet (. = 0.073), but the presence of a health problem did (. = 0.027, V = 0.119). The results further showed that a higher percentage of the participants who had a health problem revealed a very good level of information about a healthy diet (4.2%), when compared to those who had not (1.0%).

Since not all the information about a healthy diet is equally reliable (Diviani, van den Putte, Giani & van Weert, 2015), it was analysed if the level of information about a healthy diet actually influence the participants’ food knowledge (Table Nº 13). The results suggested that a higher level of information about a healthy diet corresponded to a better food knowledge. The results of the Chi squared showed that the level of information about a healthy diet influenced the participants’ food knowledge (. = 0.000), with a weak association between the variables (V = 0.190).

Table 13 Table 13 Participants’ food knowledge according to the level of information about a healthy diet Source: Own elaboration, based on the research data

Table Nº 14 refers to the participants’ eating habits according to the level of information about a healthy diet, and as it can be observed a higher percentage of the participants with a very good information about a healthy diet, 63.6%, revealed to have good eating habits, when compared to the other participants. In this case it was also proved that the level of information about a healthy diet influenced the participants’ eating habits, again with a weak association between the variables.

Table 14 Table 14 Participants’ eating habits according to the level of information about a healthy diet Source: Own elaboration, based on the research data