1. Introduction
The relationships among knowledge, attitudes and behaviors regarding nutrition are not fully understood. However, nutrition knowledge is considered necessary to incorporate healthier food habits, and recent work indicates that it may play a small but pivotal role in this process [
1]. There is no evidence sustaining strong associations among the above-mentioned factors. Many experts claim that this is because of the systematic inadequacy of the validation methods in the field [
2,
3,
4,
5,
6]. Therefore, it is necessary to develop validated nutritional questionnaires to correctly assess the possible connection between behavior and nutrition knowledge.
Among the general population, attitude and education are known important factors for sensible nutritional choices [
7,
8]. Professional athletes who already recognize that such choices can strongly affect their performance. For them, the key factors are nutrition knowledge, followed by attitude and sources of nutritional information. Moreover, nowadays, athletes reach the professional level at a younger age than before, and this makes them more vulnerable to poor diet choices [
2]. In the context of elite sports, sport-specific questionnaires are the preferred tool to assess nutrition knowledge, since the questions are pertinent to the demands of the type of sport. Additionally, in this context, the correct assessment of nutrition knowledge is particularly important since this is a key modifiable variable that can affect athlete performance [
9].
The latest systematic review [
4] showed that most questionnaires on nutrition knowledge for athletes and coaches tested outdated recommendations, missed comprehensiveness, failed to cultural suitability and needed an adequate validation. Several recent works improved the quality of methodologies, although they were restricted to specific sports and had a very limited sample size [
10,
11,
12,
13]. Furthermore, most of these studies superficially rely on the subjectivity of experts’ panels to assess the validity of the content [
10,
14], and do not cover multilingual scenarios. Some of them compromised the quality of responses by distributing the questionnaires online to athletes, without any supervision [
14,
15,
16]. This might have allowed for possible sharing and discussion of the responses among participants. Finally, many validated questionnaires include lots of items and require time to be completed. Thus, there is still work to be done for improving assessment tests on nutrition knowledge.
The aims of this study were: (i) to develop a compact reliable knowledge questionnaire on nutrition knowledge for young and adult athletes (NUKYA) for assessing the main bases of nutrition knowledge specifically in sports team; and (ii) to psychometrically validate the questionnaire.
4. Discussion
Validated tools for assessing nutrition knowledge among athletes are needed, with the final aim of investigating the relationship between nutrition knowledge and dietary habits in this population [
15]. This study reports the development and validation of a reliable and compact sports nutrition questionnaire in both English and Spanish. The questionnaire was validated for its content, construct, reliability and feasibility. To our knowledge, this is the first study to develop and validate a short sports nutrition questionnaire in both English and Spanish.
The final questionnaire includes 24 questions that comprise 59 items. It is divided into four sections: macronutrients (30 items), micronutrients (19 items), hydration (7 items) and food intake periodicity (3 items). The questionnaire can be answered in 12 min on average and is centered on issues that are considered priorities in the education of elite athletes. It does not include aspects such as supplements intake, weight control and alcohol consumption, because they are subjected to nutritionist advice, discouraged or forbidden for athletes (e.g., alcohol consumption). Questions are based on recent guidelines and have been thoroughly revised and adjusted according the commentaries received from experts. On the contrary of other similar validated questionnaires on sports nutrition knowledge [
15], NUKYA is considerably shorter and requires less time to be completed. Moreover, it has been validated for a range of ages (13 to more than 25 years) that includes adolescents, paying special attention to the wording and terms used. We assessed the reliability and validity of our questionnaire using a robust methodology and a large sample of individuals with predictably different levels of sport nutrition knowledge. Moreover, to assure the quality of the responses, we always supervised the administration of the questionnaire. Nothing suggests that the questionnaire cannot be successfully administered to adults older than those participating in the study.
The NUKYA questionnaire was developed through an iterative process of consultation with different groups of experts. First, the questionnaire was validated at the design stage, and then during content validation, using the CVI and obtaining excellent content validity indices. As Trakman et al. [
19] pointed out, quantitative methods such as CVI are effective but unusual when validating the content of sports nutrition knowledge questionnaires. Experts from the panel that validated the content and respondents of the pilot study agreed that the questionnaire covered the concepts it intends to measure, so that face validity seemed to be fulfilled. The feasibility also seems adequate according to the low administration time and the really few demands for clarification from participants.
The questionnaire had satisfactory construct validity, which was ascertained with different methods: (i) a comparison of groups who were expected to obtain different scores; (ii) an analysis of item difficulty and discrimination; (iii) a sectional analysis; and (iv) a Rasch analysis, a model developed according the item–response theory. In the comparison of groups, nutrition students obtained the best scores, followed by university students with no formal studies in health and, finally, by athletes and high school students. Differences between groups were statistically significant in some cases and coherent with the formal education received and age of individuals. In most cases, in the sectional analysis, we observed the same patterns. A few alterations are detected, such as hydration and periodicity of food intake, and they are easily explained. The analysis of item difficulty and discrimination detected six items that could be considered too easy or too difficult. However, these items were retained because they all showed a good discrimination index and were considered important to assess if certain concepts were present in the knowledge structure of the subjects. Additionally, Rasch analysis indicated that participants who performed well throughout were more likely to answer particular questions correct. The person–item map showed a good superposition between item difficulty range and the person measure distribution, so we can state that the test meaningfully measures the ability of all respondents.
Regarding reliability, internal consistency was examined through the Cronbach’s alpha and the split-half methods, both resulting in a good item consistency. Cronbach’s alpha was recalculated by extracting each item once, and values were equally high. Thus, such evidence indicates that items included in the questionnaire are measuring the same underlying concept. Moreover, questionnaire stability was assessed by comparing test and retest scores, and no significant differences were detected between them, either globally or group by group. Finally, the correlation between test and retest was high, and we also obtained significant correlations for each group. These data indicate that the questionnaire has an excellent repeatability. We detected a non-significant but constant increase in retest scores compared to test scores, across the four subject groups. This effect is well described elsewhere [
28] and can be attenuated selecting a long test–retest interval. In our case, the interval was 15 days, whereas an interval of 2 or 13 days is generally recommended. Other factors that can influence this retest effect are test form, test modality and participants’ age.
In their systematic review, Trakman et al. [
15] developed and used two ratings for sports nutrition knowledge questionnaires. One rating was used to assess comprehensiveness and included 11 items (general knowledge, carbohydrates, protein, fat, micronutrients, pre-exercise, post-exercise, during exercise, recovery, fluid, supplements and alcohol). The second rating was used to assess validity and reliability evidence (face validity, content validity, item discrimination, internal reliability, construct validity and external validity). These ratings were based on the guidelines developed by Parmenter and Wardle [
17]. In both cases, one point was awarded for each item covered, with a maximum score of 11 and 6, respectively. Applying these ratings to our questionnaire, the NUKYA questionnaire obtained nine points in comprehensiveness and six in validity and reliability scores, which are in the upper range of those published in the systematic review. We were also able to give a positive response to the complementary question of whether a pilot study was conducted.
The NUKYA questionnaire has also several limitations. The most important one is that it does not include questions on supplements intake, weight management and alcohol consumption. On the one hand, this can be a drawback in certain cases. On the other hand, we obtain a quick assessment of the nutritional knowledge of aspects related to the maintenance of a balanced diet and its adaptation to sport practice. Another limitation is that we did not assess the validity of the questionnaire among different nationalities, although our sample included a small percentage of subjects from foreign countries. Additionally, some items of the questionnaire have high (n = 4) or extremely low (n = 2) difficulty indices. Likewise, Wald test showed that some items (n = 5) could be removed without influencing the model. However, as the test performance is not compromised, these items were retained because they assessed concepts that were considered important in the nutrition education of athletes. The NUKYA questionnaire has been validated for team sport athletes and for Spanish population, although some of the subject groups included individuals from other nationalities. Consequently, the questionnaire would have to be re-validated before being used in individual sport athletes or in other populations. Finally, we were not able to differentiate and compare the responses obtained from athletes in the English and Spanish versions of the questionnaire. This was because the online form allowed athletes to choose the language they felt more comfortable with, and the answers were recorded in the same database without any differentiation.