Introduction
Nutrition knowledge, which includes knowledge of concepts such as dietary guidelines and sources of various nutrients [
1], is an important determinant of diet-related behaviour [
2]. In particular, nutrition knowledge can influence consumers’ ability to identify healthy foods and manage diet-related chronic diseases [
3‐
5]. Nutrition knowledge is influenced by a myriad of factors, including sociodemographic characteristics and socioeconomic status. Research has shown that consumers who are older, female, and have higher income and education perform better on assessments of nutrition knowledge in cross-sectional studies [
6‐
10]. Moreover, nutrition information may be more accessible to consumers with higher literacy, thereby increasing nutrition knowledge [
11,
12]. Other behavioral factors more directly connected to nutrition knowledge also warrant exploration, as research has shown that individuals with specific dietary goals or practices may seek out nutrition information to a greater extent than those without diet-related goals [
13,
14]. Motivation to change diet-related outcomes, including weight status and management of conditions such as type 2 diabetes, could potentially drive knowledge-seeking behavior [
5,
13,
14].
Consumers obtain nutrition knowledge from numerous sources, such as national nutrition policies, dietary guidelines, and food cultures that might influence uptake of or exposure to nutrition information [
15‐
20]. A variety of tools have been used to measure nutrition knowledge across countries [
21], with most studies using unique tools tailored to specific study populations [
1,
2,
8,
9]. The use of disparate tools creates challenges for comparing nutrition knowledge levels and corresponding determinants across studies, geographic contexts, and populations [
21,
22]. This is also a barrier to conducting between-country studies focused on the role and effectiveness of specific nutrition policies in increasing consumer nutrition knowledge. Overall, very few cross-country studies on nutrition knowledge have been conducted [
7‐
9,
23].
The Food Processing Knowledge (FoodProK) score was developed to measure nutrition knowledge based on consumers’ ability to understand and apply the concept of food processing in a functional task [
24]. The focus on processing levels is consistent with increased messaging related to minimizing processed food consumption in dietary guidelines [
15‐
19]. Given that processing is not specific to a given population or context, this measure can serve as an indicator of consumer nutrition knowledge that can be used across studies [
24], lending to the interpretation of cross-country research in this area.
To this end, the current study sought to compare nutrition knowledge levels based on the FoodProK among adults in five countries: Australia, Canada, Mexico, the United Kingdom (UK) and the United States (US). In particular, this study aimed to expand our understanding of the correlates of functional nutrition knowledge to include not only sociodemographic and socioeconomic characteristics, but also body mass index (BMI), and dietary behaviors that potentially influence interest in nutrition information. Correlations between FoodProK scores and self-reported nutrition knowledge and health literacy were also examined to assess how the FoodProK performed in comparison with these measures across countries, as they may be used as proxies for nutrition knowledge.
Results
Sample summary
A total of 22,824 respondents completed the IFPS survey. A subsample of 22,102 respondents from Australia (
n = 3997), Canada (
n = 4170), Mexico (
n = 4044), the UK (
n = 5363), and the US (
n = 4527) were included in the study after removing respondents with missing data on the covariates of interest. Table
2 presents characteristics of respondents included in the analysis, stratified by country. Each country had approximately equal proportions of male and female respondents. In all countries, the greatest proportion of respondents were from the majority ethnic group and reported their income adequacy as “neither easy nor difficult to make ends meet.” The majority of respondents in all countries reported being the primary food shoppers in their households, did not follow specific dietary practices, and were placed in the “adequate” health literacy category.
Table 2
Sample Characteristics (n = 22,102), International Food Policy Study, 2018
Age Group |
18–29 years | 21.2 (845) | 19.2 (800) | 29.8 (1204) | 19.1 (1026) | 20.7 (934) |
30–44 years | 26.5 (1060) | 24.7 (1029) | 32.3 (1305) | 24.4 (1307) | 25.2 (1141) |
45–59 years | 24.7 (988) | 25.9 (1078) | 28.5 (1155) | 26.2 (1407) | 25.7 (1165) |
60+ years | 27.6 (1104) | 30.2 (1263) | 9.4 (380) | 30.3 (1623) | 28.4 (1287) |
Sex at Birth |
Male | 49.0 (1959) | 49.6 (2069) | 47.6 (1925) | 48.4 (2609) | 48.4 (2192) |
Female | 51.0 (2038) | 50.4 (2101) | 52.4 (2119) | 51.3 (2754) | 51.6 (2336) |
Ethnicity |
Majority | 76.0 (3039) | 79.6 (3320) | 78.7 (3183) | 89.1 (4776) | 75.9 (3438) |
Minority | 24.0 (958) | 20.4 (850) | 21.3 (861) | 10.9 (587) | 24.1 (1089) |
Education Level |
Low | 42.0 (1682) | 41.3 (1723) | 19.5 (789) | 48.6 (2605) | 58.4 (2645) |
Medium | 32.6 (1302) | 33.7 (1407) | 13.2 (535) | 23.1 (1240) | 9.9 (445) |
High | 25.4 (1013) | 25.0 (1040) | 67.3 (2720) | 28.3 (1518) | 31.7 (1437) |
Income Adequacy |
Very difficult to make ends meet | 8.8 (353) | 8.5 (353) | 12.1 (490) | 6.9 (367) | 9.6 (435) |
Difficult to make ends meet | 19.2 (768) | 19.7 (822) | 31.8 (1286) | 18.4 (985) | 20.0 (905) |
Neither easy nor difficult to make ends meet | 37.6 (1502) | 36.8 (1534) | 38.7 (1564) | 36.4 (1955) | 33.9 (1535) |
Easy to make ends meet | 23.5 (939) | 22.4 (935) | 13.9 (564) | 24.5 (1314) | 21.8 (987) |
Very easy to make ends meet | 10.9 (435) | 12.6 (525) | 3.5 (141) | 13.8 (742) | 14.7 (665) |
Body Mass Index |
< 18.5 | 3.1 (123) | 3.3 (136) | 2.1 (85) | 3.0 (162) | 3.5 (157) |
18.5–24.9 | 35.9 (1437) | 33.6 (1400) | 39.8 (1608) | 34.7 (1861) | 30.8 (1395) |
25.0–29.9 | 26.4 (1054) | 28.7 (1197) | 29.9 (1207) | 26.8 (1437) | 27.8 (1259) |
≥ 30.0 | 21.1 (842) | 24.4 (1019) | 15.5 (629) | 16.8 (903) | 27.3 (1235) |
Missing | 13.5 (541) | 10.0 (418) | 12.7 (515) | 18.7 (1000) | 10.6 (481) |
Food Shopping Role |
Primary shopper | 71.7 (2864) | 72.0 (3000) | 74.9 (3029) | 74.2 (3981) | 73.3 (3319) |
Not primary shopper | 7.1 (284) | 6.0 (249) | 5.1 (205) | 4.7 (253) | 6.6 (299) |
Shared equally with others | 21.2 (849) | 22.0 (921) | 20.0 (810) | 21.1 (1129) | 20.1 (909) |
Dietary Practices |
No specific dietary practices | 87.0 (3477) | 90.3 (3765) | 88.1 (3564) | 87.0 (4665) | 88.6 (4012) |
One or more dietary practices (i.e., vegetarian, vegan, pescatarian, religious practices) | 13.0 (520) | 9.7 (405) | 11.9 (480) | 13.0 (698) | 11.4 (515) |
Dietary Efforts Scorea | 2.8 (2.2) | 2.6 (2.1) | 2.6 (2.3) | 3.1 (2.1) | 3.0 (2.3) |
Health Literacy |
High likelihood of limited literacy (score 0–1) | 26.7 (1040) | 19.4 (810) | 30.5 (1234) | 31.8 (1707) | 25.4 (1150) |
Possibility of limited literacy
(score 2–3) | 24.7 (964) | 23.2 (966) | 31.2 (1261) | 20.5 (1097) | 20.2 (913) |
Adequate literacy (score 4–6) | 48.6 (1897) | 57.4 (2394) | 38.3 (1549) | 47.7 (2559) | 54.4 (2464) |
Self-reported Nutrition Knowledge |
Not at all knowledgeable | 5.6 (223) | 4.1 (169) | 2.8 (114) | 9.4 (502) | 5.8 (263) |
A little knowledgeable | 31.5 (1261) | 30.1 (1256) | 30.4 (1228) | 39.4 (2111) | 28.8 (1306) |
Somewhat knowledgeable | 41.4 (1653) | 44.4 (1850) | 53.0 (2141) | 35.7 (1914) | 41.2 (1864) |
Very knowledgeable | 17.4 (696) | 18.2 (762) | 12.2 (495) | 12.6 (674) | 18.7 (844) |
Extremely knowledgeable | 4.1 (164) | 3.2 (133) | 1.6 (66) | 3.0 (161) | 5.5 (250) |
Comparisons across countries and correlates of FoodProK scores
Within each country, the mean scores across food categories were similar, as demonstrated by the narrow range in scores (Table
3). Australia was an exception as it had the widest mean score range across categories (0.9–1.4), including the lowest dairy score and one of the highest mean scores for the fruit category. Within each food category, mean scores were similar across countries, with dairy scoring lowest across the five countries.
Table 3
Food Processing Knowledge Score by Country
Canada | 1.3 (0.6) | 1.3 (0.6) | 1.2 (0.6) | 1.3 (0.6) | 5.1 (1.6) |
Australia | 1.4 (0.6) | 1.3 (0.7) | 0.9 (0.7) | 1.3 (0.7) | 5.0 (1.8) |
United Kingdom | 1.2 (0.6) | 1.2 (0.7) | 1.1 (0.7) | 1.3 (0.7) | 4.8 (1.9) |
Mexico | 1.4 (0.6) | 1.1 (0.6) | 1.0 (0.6) | 1.3 (0.7) | 4.7 (1.6) |
United States | 1.2 (0.7) | 1.1 (0.7) | 1.0 (0.6) | 1.2 (0.7) | 4.6 (1.8) |
Five countries combined | 1.3 (0.6) | 1.2 (0.7) | 1.0 (0.6) | 1.3 (0.7) | 4.8 (1.7) |
Based on the linear regression analysis (Table
4), those classified as having ‘adequate health literacy’ or the ‘possibility of limited health literacy’ had higher FoodProK scores compared to respondents with a ‘high likelihood of limited literacy’ (β:1.28; CI:1.21, 1.35;
p < .001; β:0.76; CI:0.68, 0.84;
p < .001). Self-reported nutrition knowledge was significantly associated with FoodProK score, as respondents who reported they were ‘very knowledgeable’ (β:0.81; CI:0.67, 0.96,
p < .001), ‘somewhat knowledgeable’ (β:0.75; CI:0.61, 0.88;
p < .001), and ‘a little knowledgeable’ (β: 0.65; CI: 0.52, 0.79;
p < .001) scored higher on the FoodProK compared to those who reported that they were ‘not at all knowledgeable.’ Those who reported being ‘a little knowledgeable’ had lower FoodProK scores than those reporting being ‘somewhat knowledgeable (β:-0.09; CI: -0.15, -0.34;
p=0.002) or ‘very knowledgeable’ (β:-0.16; CI: -0.23, -0.08;
p<0.001). Respondents who stated they were ‘a little knowledgeable’ had significantly higher FoodProK scores than those who selected ‘extremely knowledgeable’ (β:0.50; CI:0.34, 0.66;
p < 0.001), and those who reported being ‘extremely knowledgeable’ had significantly lower FoodProK scores than those who reported being ‘somewhat knowledgeable’ (β:-0.59; CI: − 0.75, − 0.44;
p < 0.001) or ‘very knowledgeable’ about nutrition (β:-0.66; CI: − 0.82, − 0.50;
p < 0.001).
Table 4
Sociodemographic, behavioural, and knowledge-related correlates of the Food Processing Knowledge Score, International Food Policy Study, 2018 (n = 22,102)
Country |
Australia vs. Canada | 0.07 | −0.01, 0.14 | 0.08 |
Australia vs. Mexico | 0.22 | 0.13, 0.30 | *0.001 |
Australia vs. United Kingdom | 0.09 | 0.01, 0.16 | *0.02 |
Australia vs. United States | 0.40 | 0.32, 0.48 | * < .001 |
Canada vs. Mexico | 0.15 | 0.06, 0.23 | * < .001 |
Canada vs. United Kingdom | 0.02 | −0.05, 0.09 | 0.61 |
Canada vs. United States | 0.33 | 0.25, 0.41 | * < .001 |
Mexico vs. United Kingdom | −0.13 | −0.21, − 0.05 | *0.002 |
Mexico vs. United States | 0.18 | 0.10, 0.27 | * < .001 |
United Kingdom vs. United States | 0.31 | 0.23, 0.39 | * < .001 |
Age group |
30–44 years vs. 60+ years | −0.17 | −0.24, − 0.09 | * < 0.001 |
45–59 years vs. 60+ years | − 0.10 | −0.17, − 0.04 | *0.002 |
60+ years vs. 18–29 years | 0.13 | 0.04, 0.21 | *0.002 |
Sex |
Female vs. Male | 0.26 | 0.21, 0.32 | * < .001 |
Ethnicity |
Majority vs. Minority | 0.19 | 0.11, 0.26 | * < .001 |
Education Level |
Medium vs. Low | 0.02 | −0.05, 0.08 | 0.58 |
High vs. Medium | 0.01 | −0.05, 0.07 | 0.80 |
High vs. Low | 0.03 | −0.03, 0.08 | 0.40 |
Income Adequacy | −0.02 | −0.04, 0.00 | 0.12 |
Body Mass Index |
< 18.5 vs. 18.5–24.9 | −0.19 | −0.34, − 0.04 | *0.01 |
25.0–29.9 vs. < 18.5 | 0.18 | 0.03, 0.34 | *0.02 |
≥ 30.0 vs < 18.5 | 0.21 | 0.05, 0.36 | *0.008 |
Missing vs. 18.5–24.9 | −0.32 | − 0.41, − 0.23 | * < .001 |
Missing vs. 25.0–29.9 | −0.33 | 0.42, − 0.24 | * < .001 |
Missing vs. ≥30.0 | −0.31 | −0.41, − 0.21 | * < .001 |
Food Shopping Role |
Primary shopper vs. Not primary shopper | 0.00 | −0.12, 0.11 | 0.93 |
Primary shopper vs. Share equally with others | −0.06 | −0.12, 0.00 | 0.05 |
Share equally with others vs. Not primary shopper | 0.06 | −0.06, 0.18 | 0.36 |
Dietary Practices |
One or more dietary practices (i.e., vegetarian, vegan, pescatarian, religious practices) vs. No specific dietary practices | −0.31 | −0.39, − 0.23 | * < .001 |
Dietary Efforts Score | −0.13 | − 0.14, − 0.11 | * < .001 |
Health Literacy |
Possibility of limited literacy (score 2–3) vs. High likelihood of limited literacy (0–1) | 0.76 | 0.68, 0.84 | * < .001 |
Adequate literacy (score 4–6) vs. Possibility of limited literacy (score 2–3) | 0.52 | 0.46, 0.58 | * < .001 |
Adequate literacy (score 4–6) vs. High likelihood of limited literacy (0–1) | 1.28 | 1.21, 1.35 | * < .001 |
Self-reported Nutrition Knowledge |
A little knowledgeable vs. Not at all knowledgeable | 0.65 | 0.52, 0.79 | * < .001 |
A little knowledgeable vs. Somewhat knowledgeable | −0.09 | −0.15, − 0.34 | *0.002 |
A little knowledgeable vs. Very knowledgeable | −0.16 | −0.23, − 0.08 | * < .001 |
A little knowledgeable vs. Extremely knowledgeable | 0.50 | 0.34, 0.66 | * < .001 |
Somewhat knowledgeable vs. Not at all knowledgeable | 0.75 | 0.61, 0.88 | * < .001 |
Very knowledgeable vs. Not at all knowledgeable | 0.81 | 0.67, 0.96 | * < .001 |
Extremely knowledgeable vs. Somewhat knowledgeable | −0.59 | −0.75, −0.44 | * < .001 |
Extremely knowledgeable vs. Very knowledgeable | −0.66 | −0.82, − 0.50 | * < .001 |
Respondents engaging in one or more specific dietary practices such as vegetarianism had significantly lower FoodProK scores (β:-0.31; CI: − 0.39, − 0.23; p < .001) than those with no specific dietary practices. Respondents who reported efforts to consume less sugar, sodium, trans fat, calories, or processed foods had significantly higher FoodProK scores (β: -0.13; CI: − 0.14, − 0.11; p < .001) compared to respondents not making efforts to modify their eating patterns in these areas. Food shopping role was not significantly associated with FoodProK score.
The oldest age group (60+ years) scored significantly higher on the FoodProK than the youngest age group (18–29 years) (β: 0.13; CI: 0.04, 0.21; p = 0.002). Respondents aged 30–44 years (β: -0.17; CI: − 0.24, − 0.09; p < 0.001) and 45–59 years (β: -0.10; CI: − 0.17, − 0.04; p = 0.002) had significantly lower FoodProK scores than those in the 60+ years category. Females scored higher on the FoodProK than males (β: 0.26; CI: 0.21, 0.32; p < 0.001). Education and income adequacy were not significantly associated with FoodProK score.
Respondents with a BMI < 18.5 or ‘missing’ BMI data had lower FoodProK scores than those with a BMI between 18.5–24.9 (β: -0.19; CI: − 0.34, − 0.04; p = 0.01; β: -0.32; CI: − 0.41, − 0.23; p < .001). Moreover, respondents with BMIs between 25 and 29.9 (β: 0.18; CI: 0.03, 0.34; p = 0.02) or ≥ 30 (β: 0.21; CI: 0.05, 0.36; p = 0.008) had significantly higher FoodProK scores than those with BMIs under 18.5, and those with missing BMI data had significantly lower FoodProK scores compared with respondents with BMIs between 25 and 29.9 (β: -0.33; CI: − 0.42, − 0.24; p < 0.001) or ≥ 30 (β: -0.31; CI: − 0.41, − 0.21; p < 0.001).
As shown in Table
4, respondents from Australia, Canada, Mexico, and the UK scored significantly higher on the FoodProK compared to respondents from the US (β: 0.41; CI: 0.33, 0.49;
p < .001; β: 0.33; CI: 0.25, 0.41;
p < .001; β: 0.18; CI: 0.10, 0.27;
p < .001; β: 0.31; CI: 0.23, 0.39;
p < .001, respectively). Several other country contrasts were also significant. Respondents in Australia had significantly higher FoodProK scores than those in the UK (β: 0.09; CI: 0.01, 0.16;
p = 0.02) and Mexico (β: 0.22; CI: 0.13, 0.30;
p = 0.001). Canadian respondents had significantly higher FoodProK scores than those in Mexico (β: 0.15; CI: 0.06, 0.23;
p = < 0.001). Respondents in Mexico had significantly lower FoodProK scores than the UK (β: -0.12; CI: − 0.21, − 0.05;
p = 0.002).
Sensitivity analyses
Sensitivity analyses indicated that the FoodProK scoring method did not change the pattern of scores across countries or associations between scores and other variables. Irrespective of whether the FoodProK was in the original 8-point format, 7-point format dropping only deli meat, or 6-point format dropping the entire meat category, the same correlates were significant in the regression model, with no meaningful differences in the parameter estimates. Further, the correlations between FoodProK, self-reported nutrition knowledge, and health literacy status were comparable regardless of the scoring approach.
FoodProK scores and relationships between knowledge-related variables
Health literacy and the FoodProK score were moderately correlated (rs = 0.37, p < 0.001). There was a very weak, positive correlation between self-reported nutrition knowledge and each of health literacy (rs = 0.09, p < 0.001) and the FoodProK score (rs = 0.09, p < 0.001).
Discussion
The current study is one of the first to examine differences in nutrition knowledge levels across multiple countries. Based on understanding of levels of food processing, adults from Canada and Australia scored highest on the functional nutrition knowledge test, with adults in the US scoring the lowest. Differences across countries are likely due to a range of factors, including national dietary guidelines and nutrition policies that may influence consumers’ access to and uptake of nutrition information based on the reach and effectiveness of these initiatives [
36]. Country-specific dietary patterns or food culture may also play a role in nutrition knowledge among populations, particularly informal channels of nutrition education such as family food practices and cultural beliefs which contribute to consumers’ implicit understanding of a food’s nutritive quality/properties [
37‐
39]. This ‘prior’ knowledge may reinforce messaging from national education campaigns, or on the contrary, conflict with cultural beliefs around healthy eating in some populations [
40‐
42]. Countries with the lowest FoodProK scores – Mexico and the US – also have among the highest levels of consumption of ultra-processed foods across countries [
43‐
48]. Lower scores in these countries may reflect lower levels of knowledge or different social norms in populations in which highly processed foods are ubiquitously available and consumed.
Although some differences in nutrition knowledge scores across countries were statistically significant, the magnitude of differences was modest. The large study sample size resulted in high levels of power; thus, even modest differences were statistically significant in some cases. For example, Canada vs. US and UK vs. US had modest, but significantly different FoodProK scores (β: 0.33,
p < .001 and β:0.31,
p < .001, respectively) which are difficult to interpret. Modest differences may also reflect similar content in national nutrition guidelines and labelling policies with respect to the NFts that appear on pre-packaged products, which were displayed to respondents as part of the FoodProK [
15‐
19]. Future research should focus on the impact of new national nutrition guidelines on nutrition knowledge, including evaluations of awareness, comprehension, use, and reach of such guidelines documents and associated campaigns.
Overall, cross-country studies of nutrition knowledge to enable comparisons of the current findings are lacking. Grunert et al. (2012) found that adults in the UK had significantly higher nutrition knowledge than respondents from four other European countries [
7]. The authors attributed this finding to the “history of health policies and nutrition-related initiatives,” as well as potential cultural differences among UK respondents compared with the other countries (p. 166) [
7]. While specific policies are not described by Grunert et al. (2012) [
7], the UK was one of the first countries among the six included in the study to adopt dietary guidelines, which may have contributed to consumers’ general nutrition knowledge. We are unaware of any other studies that have examined differences between the five countries included in the current study.
Respondents who reported efforts to modify their eating patterns scored higher on the FoodProK. Individuals with specific diet-related goals likely have a greater interest in nutrition or may rely on labels and other sources of nutrition information more frequently [
13,
14]. Moreover, individuals with dietary preferences may possess greater motivation to obtain nutrition information, which may drive them to improve their knowledge to support specific dietary choices [
13,
14]. This study did not find an association between food shopping role and nutrition knowledge, which may reflect the fact that such tasks are gendered and based on the social organization of society rather than nutrition knowledge [
49,
50].
Sociodemographic differences in knowledge were also observed. Consistent with other literature, functional nutrition knowledge was higher with age and among females [
7,
9,
10]. Existing evidence points to behavioural and attitudinal differences between men and women, as well as different age groups, as a possible explanation for these differences. Women and older age groups appear to be more health conscious, and it is hypothesized that increased interest in healthy eating may result in increased nutrition knowledge due to intentional efforts to seek out nutrition information [
7,
51]. Moreover, nutrition and food tend to be predominantly “female domains,“ [
49,
51] suggesting women may be more likely than men to be exposed to nutrition-related health information, increasing their opportunities to gain knowledge.
The association between ethnicity and nutrition knowledge has not been extensively studied. This study found that the ‘majority’ ethnic group in each country had significantly higher FoodProK scores when controlling for other covariates. Some studies have used other measures of ethnicity such as citizenship status, showing lower nutrition knowledge levels among immigrant populations [
23,
52]. This may be explained, in part, by acculturation, as immigrants in varying stages of assimilation may have different exposure to national dietary guidelines. The amount and type of cultural exposures, among other aspects of immigrant or ‘minority’ experiences, could potentially impact knowledge of country-specific guidance on healthy eating [
52‐
54], as well as familiarity with foods in a new cultural context. Additionally, given racism that excludes some individuals from fully participating in economic and other systems, those not identifying as ‘White’ or ‘non-Indigenous’ may have had fewer opportunities to develop and apply nutrition knowledge and related skills, such as label reading [
54‐
56]. Overall, these factors may result in lower capacity to answer the FoodProK questions.
With respect to BMI, there were notably lower FoodProK scores among those with missing BMI data compared to the other categories, and higher FoodProK scores when comparing the highest BMI categories to the lowest < 18.5 group. Generally, the literature is inconclusive with respect to associations between BMI and nutrition knowledge [
57,
58]. Furthermore, this study relied on self-reported height and weight. US-based studies have shown that weight tends to be under-reported [
59‐
62], and while it is unlikely that data are missing at random, it is difficult to discern what might underlie the BMI associations observed in this study.
The findings also shed light on different methods of assessing nutrition knowledge. FoodProK scores were positively associated with a measure of health literacy, the NVS, which provides a functional assessment of respondents’ ability to understand and apply numeric and descriptive information contained in NFts. Given the focus of the NVS on a nutrition label, this measure might be considered to assess nutrition literacy [
21]. In contrast, a commonly used measure of self-rated nutrition knowledge, in which participants rate their perceived level of knowledge on a scale of 1 to 5, was very weakly associated with health literacy, as well as FoodProK scores. Respondents who rated themselves as ‘extremely knowledgeable’ had lower literacy and FoodProK scores, which suggests that many respondents drastically overestimate their nutrition knowledge. This finding reinforces the need to move beyond single-item measures towards functional tests of nutrition knowledge, such as the FoodProK, in order to capture some of the nuance and complexity of nutrition knowledge.
The strength of this study lies in the large sample size and multi-country design, which enabled comparisons of nutrition knowledge using a functional measure. Several limitations should also be considered. First, the sample was recruited using non-probability sampling, which does not enable the generation of nationally representative population estimates. Moreover, there is potential for social desirability bias given the use of self-reported measures [
59,
60]. There are also limitations of the FoodProK score, as content validity testing demonstrated poorer performance in the meat category compared to other categories [
24]. Sensitivity tests revealed the FoodProK score performed similarly irrespective of whether 6-, 7- or 8-point scales were used; however, further validity and reliability testing of this measure is required, including examining its ability to accurately capture nutrition knowledge in diverse populations and contexts. Modest differences in knowledge may be related to the FoodProK test’s limited ability to detect differences in nutrition knowledge. The large study sample further enabled detection of statistically significant differences in small parameter estimates, which may not reflect meaningful differences in nutrition knowledge across subgroups in all cases. In addition, self-administration of the NVS has not been validated. While this limitation is consistent across all countries, future studies should examine potential differences in self vs. interviewer-administered versions.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.