Development and validation of a brief general and sports nutrition knowledge questionnaire and assessment of athletes’ nutrition knowledge

Nutrition knowledge (NK) is a modifiable determinant of dietary behaviour [1, 2] and dietary practices are known to influence athletic performance [3]. Therefore, there has been much interest in the assessment of athletes’ general nutrition knowledge (GNK) and sports nutrition knowledge (SNK) [4, 5]. There are limitations with some of the NK questionnaires that have previously been employed, including testing outdated recommendations, lack of comprehensiveness, lack of cultural appropriateness, and insufficient validation [6]. The issue of inadequate validation is common in evaluation measures used in nutrition education research [7]. Researchers may overlook comprehensive methodologies because the time taken to develop and validate questionnaires can be prohibitive [8].

The Nutrition for Sport Knowledge Questionnaire (NSKQ) [9] was developed in 2017 and overcame many of the aforementioned limitations. The NSKQ was based on current sports nutrition recommendations [10, 11] and is comprehensive, containing 89 questions across six nutrition sub-sections. Moreover, it was developed with a panel of international sports dietitians and validated using a robust methodology that included both Classical Test Theory (CTT) [8] and Rasch analysis [12].

Other recently developed SNK questionnaires also represent an improvement on previous tools and are of similar length to the NSKQ [1315]. When collecting data to assess the validity of the NSKQ [9], and (later) when using the NKSQ to evaluate Australian Footballers’ NK 1 response and completion rates were low. The NSKQ takes athletes on average 25 min to complete. Some researchers have reported that the ideal questionnaire completion time to optimise response rates is 13 min or less [16]. Galesic et al. [17] reported that, in general, online-questionnaires that were perceived to be long were started and completed by fewer respondents, with less time spent on questions at the end of questionnaires. In contrast, a meta-analysis of factors influencing response rates of online-questionnaires found that questionnaire length had a very small effect (r = 0.001) on response rates [18], but the paper did not report on completion rates.

To our knowledge, there is no published data specifically on the effect of questionnaire length on response and completion rates in athletic cohorts. Both elite and non-elite sportspeople are often pressed for time, balancing busy training schedules with other commitments [19]. Therefore, it is feasible that long questionnaires would be daunting to athletes and the professionals working with them, and in part, explain the low response and completion rates of the NSKQ.

The aforementioned factors demonstrate a potential need for shorter measurement tools to assess NK amongst athletes. Appropriate validation of such tools is critical if they are to be used by dietitians and coaches with the intention of influencing athlete behaviour and ultimately performance. Rasch analysis is a method which allows researchers to produce short measurement tools that include both difficult and simple items, and is therefore a suitable method for use in developing a brief NK questionnaire [20]. The Rasch model presumes individuals with higher levels of knowledge are more likely to obtain a high NK score and that easy items are more likely to be answered correctly by all respondents [20]. This is advantageous because it means the focus of validation is on the questionnaire as a whole, rather than on individual items [21].

The aims of this study were to:

  1. 1.

    re-assess data used to validate the NSKQ to develop an abridged version of the questionnaire (the A-NSKQ) and

  2. 2.

    compare and contrast response rates, completion rates and NK scores between NSKQ and A-NSKQ in a cohort of non-elite AF and netball athletes.

It was hypothesised that the A-NSKQ would achieve higher completion rates than the NSKQ, and produce NK scores that are comparable to a related cohort (non-elite AF players).