Drawing instead of answering to evaluate the effectiveness of food safety programmes in primary school

Abstract

Objective:

The high incidence of foodborne disease among children suggests the value of health promotion. Children are a high-risk group so far as foodborne disease is concerned, although they may be hard to reach with training programmes. This research investigated the use of drawings, compared with questionnaires, to evaluate the impact of a health promotion programme to improve knowledge and habits in the context of food handling and personal hygiene.

Setting:

Children (184) attending primary school and living in the north of Italy were enrolled in the programme.

Methods:

Qualitative and quantitative tools: pre- and post-intervention questionnaires were administered, and children were asked to produce pre- and post-intervention drawings about microorganisms and their effects on humans. An observation grid was built to code key features in the drawings.

Results:

Results clearly showed that some drawing features correlated with and predicted high scores in the questionnaire on knowledge of microorganisms. These were the use of captions, the representation of a natural context and the presence of a causal link between depicted elements.

Conclusion:

Study findings highlight the potential of the use of drawing as an effective evaluation tool. The use of drawing can support the design of strategies for the validation of health campaigns aimed at the amelioration of children’s food contamination–related (and perhaps other) risks.

Keywords

Child health food safety health promotion Italy school health

Introduction

Health promotion agencies generally prefer to target their campaigns towards population sub-groups and age levels that are believed to be capable of using acquired information to improve their own health protection behaviours: adolescents, young adults or adults. Unlike in the USA and Canada (Byrd-Bredbenner et al., 2010; Woodruff and Kirby, 2013), in continental Europe, children, although recognised as an at-risk group for food safety issues (Marcus, 2008), are usually exempt from health protection education, as this is instead provided to parents, educators or teachers (Shearer et al., 2013).

This happens for understandable reasons. Children may have difficulties comprehending scientific jargon and understanding issues that may seem distant from their everyday lives. Moreover, proper evaluation of the effectiveness of health programmes directed towards children is often problematic due to evaluator uncertainty about respondents’ potential misunderstanding of questionnaire items.

However, teaching food hygiene in schools could be an effective way to enhance development of proper food handling practices and the sharing of acquired knowledge on correct practices for use in the home setting (Haapala and Probert, 2004; Losasso et al., 2012). Although the questionnaire is the most common tool to evaluate the effectiveness of children’s health prevention programmes, efficacious questionnaire use requires careful consideration (Alderson, 1995). Easy and familiar language is needed, during both the intervention and evaluation phases, to promote concrete linkage to children’s everyday lives and to facilitate the comprehension of scientific concepts. Extreme care is needed during evaluation, due to possible discrepancies between intended response and actual response due to the wording of questionnaire items (Backett and Alexander, 1991).

To assure maximum effectiveness of children’s health promotion interventions, professionals should provide age-appropriate, fully comprehensive experiences, rather than simple cognitive knowledge. The implementation of health promotion interventions and evaluation processes also need to be specifically adapted to children.

Drawing to depict scientific knowledge

Drawing is a powerful means of accessing children’s understanding and imagery. It can improve children’s individual skills, knowledge and understandings of reality (Hayes et al., 1994; Taylor and Andrews, 1993). Many children feel more confident expressing themselves through drawing rather than using words. The psychological processes beyond the activity of drawing itself appear more similar to problem solving than to the simple reproduction of personal representations (Gan et al., 2007). When asked to depict abstract concepts or objects that are not directly visible, children organise their illustrations to produce innovative knowledge, rather than relying exclusively on what they see (Bombi and Pinto, 2000). This is because representing abstract concepts or objects is a meaning-making activity. The child artist is actively engaged in linking previous information to new requests, ideas and knowledge (Abrahams and Millar, 2008) to produce a unique and global representation that testifies to ‘deep understanding’ and personal elaboration of the experiences (Grotzer, 1999).

Drawings have been employed in health education research with children since the 1990s (Williams et al., 1989), and their use provides an empirical demonstration of the high quality of data which can be collected from young children (Pridmore and Bendelow, 1995). Despite this, the use of drawing in health promotion programmes involving young children appears to have decreased in recent years. The aims of this study therefore were to investigate the effectiveness of drawings as a tool for simplifying and consolidating children’s deep understanding of a scientific topic and also for evaluating the successfulness of a health promotion programme targeting young children.

Materials and methods

Study design

The study design is described in detail elsewhere (Faccio et al., 2013; Losasso et al., 2014). In brief, between November 2011 and March 2012, a health programme called Mission to the Invisible World was developed to reach students, aged between 9 and 11 years, who were attending compulsory state primary schools. Programme aims were to improve participants’ knowledge of foodborne disease and to teach students the importance of correct food preparation and storage to reduce the risk of foodborne illnesses.

Twelve north Italian schools were enrolled in the study, six urban and six rural. The participating classes, one from each school, were randomly divided into two groups, a theoretical and a practical group based on two different teaching approaches previously described (Faccio et al., 2013; Losasso et al., 2014).

A between-groups experimental design was used (Cipolletta and Faccio, 2013; Faccio, 2013; Faccio and Costa, 2013; Faccio et al., 2012; Iudici et al., 2015; McBurney, 1983) with random assignment of classes to theoretical and practical groups. Pre-intervention data were compared to post-intervention data in order to analyse differences in and between groups before and after the health promotion programme.

Experts in food safety (veterinarians and microbiologists) participated in the teaching together with experts in science communication and psychology. The programme was delivered via two lessons of 2 hours each.

Written consent for each child was obtained from parents or tutor, and the Academic Ethics Committee of the University of Padua approved the study.

Questionnaires and outcomes

To evaluate the effectiveness of the health promotion intervention, a pre- and post-intervention questionnaire (Losasso et al., 2014) was administered to both practical and theoretical groups. The questionnaire was initially tested on a sample of students aged 9–11 years.

The questionnaire was divided into three sections. In the first of these, information was collected on respondents’ characteristics including gender and ethnicity. The second section focused on knowledge in each of the following nine areas, each measured with seven true-false items: (1) bacteria in the environment, (2) differences between viruses and bacteria, (3) relationships between bacteria and the human body, (4) hand hygiene, (5) correct raw meat handling to reduce foodborne disease risks, (6) correct fruit and vegetable handling to reduce foodborne disease risks, (7) bacteria and food technology, (8) food handling hygiene and (9) insight into influenza and antimicrobial resistance. A third section focusing on behaviours, was composed of eight questions on the following topics: (1) hand hygiene before eating, (2) hand hygiene when returning home, (3) hand hygiene after stroking animals, (4) hand hygiene after handling eggs, (5) hand hygiene after handling meat, (6) hand hygiene after eating, (7) hand hygiene after using the toilet and (8) the use of hands to cover the mouth when coughing.

Items in the knowledge section were coded as dummy variables by assigning a value of 1 to the correct answer and 0 otherwise. Each item in the behaviours section consisted of four possible answers (‘never’, ‘sometimes’, ‘often’ and ‘always’) coded as 0 (worst habit, corresponding to the answers ‘never’ and ‘sometimes’) or 1 (best habit, corresponding to the answer ‘often’ and ‘always’). Thus, the total possible scores were 9 and 8, for the knowledge and behaviours sections, respectively. Differences between pre- and post-intervention overall scores were examined using paired t-test.

Drawings and outcomes

All children were asked to draw their personal representation of the world of microorganisms and its relationship with humans (Faccio et al., 2013). Drawings were requested immediately before the questionnaires were administered both before and after the intervention.

The task was aimed at allowing children to represent their ideas about microorganisms and their effects on humans in a creative and familiar way, giving them space to express ‘uneducated’ theories or false beliefs (Inagaki and Hatano, 2002). Teachers communicated the drawing title to students (‘Microorganisms and me’) and a generic definition of microorganisms. (Microorganisms are living beings that cannot be seen, but they live among us, they move and they can have different shapes. Moreover, they really do all kinds of things!) Children were given no further instructions.

An observation grid was built to read the features and the qualities of drawings that correlated with a high grade of active comprehension and re-construction of scientific concepts, evaluated via the questionnaire. All drawings were analysed for the presence of five elements: shape (the depiction of microorganisms, for example, humanised, undefined or ‘standard’ shape), subject (the main subject of the drawing: the microorganism, the child or both), context (the environment in which the drawing was framed, for example, natural, domestic or human body), captions (the presence of captions or balloons) and any causal links between the represented objects (e.g. between the action of the microorganism and the consequences on the human body).

The post-intervention drawings were analysed for the presence of three additional elements: emphasis on theoretical content or on the microorganisms (when the drawing drew on information related to theoretical teaching or the action of the microorganism), hygiene practices (indicating the presence of behaviours aimed at assuring hygiene standards) and actions performed by microorganisms (the depicted actions of the microorganisms caused negative consequences, such as contagion, or actions that led to positive use, such as the use of yeast to leaven bread). These three elements were absent, and so were not analysed, in pre-intervention drawings; they were added to the analysis of the post-intervention drawings as they were likely to have been suggested as a result of the health promotion programme itself.

Differences between pre- and post-intervention data were tested using Pearson’s Chi-Squared test.

Data entry and analysis

Children filled in the questionnaires by themselves and responses were entered into an electronic database (Access 2009, Microsoft Corporation, Redmond, WA).

Data were analysed using Partial Least Squares Regression (PLS-R). PLS-R is a multivariate technique that generalises and combines features from Principal Component Analysis (PCA) and multiple regression. It predicts or analyses a set of response variables (Y variables, that is, the knowledge or behaviour scores) from a set of independent variables or predictors (X variables, that is, the drawing features), with the regression coefficients expressing the strength of the association between the dependent variable (the score) and the explanatory or independent variable (the item). PLS-R is particularly suited when the matrix of predictors have more variables than observations and in the presence of multicollinearity among X. PLS-R, as a form of PCA, reduces the dimensionality of the data space searching for few orthogonal components, which are linear combinations of the original variables, forming a lower dimensional subspace in which information on the predictor variables (X) useful in explaining the responses (Y) is summarised.

Due to these features, PLS-R allows the visualisation of the observations on a bidimensional subspace, generated by pairs of principal components (scoreplot), and investigates similarities and differences between observations by simply looking at their distances on the plane. A biplot combines the information contained in the scoreplot with the information on the relevance of the original variables to determine the position of the observations in the plane. The biplot shows samples, displayed as points, and the projection of the original axis (variables) on the plane, displayed as arrows from the origin. Another useful plot can be obtained by displaying projections of components of both predictor (X loadings) and response variables (Y loadings). This shows the contribution of the predictors in building selected components, as well as the component capability to explain Y variables. X variables important for the i-th component fall far from the origin along the i-th axis in the loading plot. Analogously, Y variables well modelled by the i-th component fall far from the origin along the i-th axis in the same plot. The importance of a given X variable for Y is proportional to the projection on Y of its distance from the origin in the loading space.

PLS techniques make use of very general assumptions on observation distribution and do not require independence of observations (Esposito Vinzi et al., 2010), making them suitable tools for exploratory analysis, allowing a visual inspection of the relations between all the variables. A PLS-R of matrix Y on X was conducted, where Y was a matrix of dependent variables (the knowledge and behaviours scores, indicated, respectively, by K_score and B_score) and X was a matrix of predictor variables (dummy variables for each of the drawing elements) using the R-package ‘pls’ (Mevik and Wehrens, 2007).

Comparison of drawings and questionnaires

To analyse the data derived from the drawings, drawing elements with multiple modalities were transformed into dummy variables with a value of 0 when the feature was absent or 1 when the feature was present. Dummy coding enlarged the number of drawing elements from 8 (items: shape, subject, context, caption, causal link between represented objects, emphasis, hygiene practices and action) to 19 drawing features:

Shape

Standard (b1)

Insects, animals or fantasy (b2)

Humanised (b3)

Indefinite (b4)

Subject

Child and microorganism (c1)

Microorganism (c2)

Context

Natural environment (d1)

Domestic environment (d3)

Scholastic environment (d4)

Body context (d5)

No context (d6)

Captions (e1)

Causal link between represented objects (f1)

Focus

Theoretical content (g1)

Effects of microorganisms (g2)

Hygiene practices

Focused on the person (h1)

Focused on the food (h2)

Microorganisms’ actions

Positive (i1)

Negative (i2)

The categories used to analyse the drawings were suggested by two independent psychology researchers who had been trained in drawing analysis. They first examined all the drawings and established categories and then they analysed the features in each drawing. Finally, they matched the results of the categorisation, comparing pre- and post-intervention drawings for each child (Cohen’s Kappa ratio .95).

At the beginning of the coding process, the category d2 was used to indicate the contexts into which the actions of microorganisms were put (food or food preparation). On processing the data, however, the category context-foods (d2) was observed and used to encode only a very few drawings. Therefore, it was decided to merge category d2 with the category context – natural environment (d1) – so that this category then included drawings representing microorganisms in food or food preparation situations or in the natural environment.

Results

The relationship between drawing features and questionnaire scores was examined. To identify the items mostly influencing the overall scores, a PLS-R of matrix Y (scores) on matrix X (drawing features) was conducted. The amount of variance explained by the first four components is reported in Table 1.

Table 1.

Percentage of variance explained by the first four principal components.

Component	R ² X	R²X cum
1	26.122	26.122
2	32.763	58.885
3	5.968	64.853
4	5.814	70.667

This analysis was intended to explore relationships between drawings and questionnaire rather than to make predictions; however, a two-component model, able to capture almost 60% of the variability of X, was chosen independently from its predictive capacity.

Sample

Only children who produced drawings and completed questionnaires both in the pre- and post-intervention lessons were enrolled. Altogether, 184 children were included, equally distributed between girls (49.5%) and boys (50.5%); they were 40.8% urban and 59.2% rural residents.

Questionnaire scores and drawing features

Questionnaires, assessing both knowledge and behaviours, collected at pre-intervention were summarised by the overall score and compared to those compiled by the same child during the post-intervention session (Table 2).

Table 2.

Children’s overall scores for knowledge and behaviours: quantiles, means and differences between pre- and post-intervention (estimated by paired t-test).

	Pre-intervention			Post-intervention			p-value
	First quantile	Mean	Third quantile	First quantile	Mean	Third quantile
Knowledge	37.0	40.2	43.0	42.0	45.2	49.0	<.001
Behaviour	14.0	16.9	20.0	16.0	18.7	22.0	<.001

The analysis of pre-intervention drawings did not show any kind of clustering or noticeable trend between features in the drawings and questionnaire scores.

As shown in Figures 1 and 2, when the post-intervention observations were plotted in the new subspace spanned by the first two components, the points distribution indicated that students tended to produce drawings with similar variables, as can be observed by the points superimposition (68 observations superimposed in the black circle with the largest radius and 31 observations in the green circle with the second largest radius). This effect denotes that after the intervention, 54% of children illustrated the same features in the drawings. In both cases, the clusters contained students from different schools, although 92% of the observations contained in the black circle belonged to students from the same school (Figure 1).

Figure 1.

PLS regression of knowledge scores on drawing features: X scoreplot for the first two components (first component on the x-axis absorbing 26% of the total variance, second component on the y-axis absorbing an additional 33% of the total variance). Point radii of the circles are proportional to number of observations. Colour depending on knowledge score: black for low (k_score < 42, where 42 is the 30th percentile of the score distribution), red for medium and green for high (k_score > 48, where 48 is the 70th percentile of the score distribution).

Figure 2.

PLS regression of behaviours scores on drawing features: X scoreplot for the first two components (first component on the x-axis absorbing 26% of the total variance, second component on the y-axis absorbing an additional 33% of the total variance). Point radii of the circles are proportional to number of observations. Colour depending on behaviour score: black for low (B_score < 17, where 17 is the 30th percentile of the score distribution), red for medium and green for high (B_score > 22, where 22 is the 70th percentile of the score distribution).

As shown in Figure 1, the higher knowledge scores were positioned towards the bottom-left, while there was no clear pattern in the case of behaviours scores (Figure 2). The radii of the circles in the figures are proportional to the number of observations (number of subjects having the same representation on the plot) and are constant in the two plots. The colour of the dots indicates the score level recorded by that specific subject in the knowledge (Figure 1) and in the behaviours questionnaires (Figure 2).

After intervention, the majority of children (83.2%) drew microorganisms in a standardised way, preferring shapes resembling those viewed during the lessons to humanised or undefined shapes (significant reduction in preference between pre- and post-intervention drawings, p < .0001; Table 3). After the intervention, students’ attention was focused on microorganisms as they drew them significantly more frequently than before the intervention. In contrast, the external context remained not well defined after the intervention, although there was an increased preference for drawing the human body (19.5% pre-intervention compared to 23.4% post-intervention, p not significant).

Table 3.

Percentage of children (n = 184) who depicted a particular feature in pre- and post-intervention drawings.

			Pre-intervention	Post-intervention	p-value
b	Shape	Standardised	–	83.2%	–
		Humanised	24.3%	0.0%	<.0001
		Undefined	68.1%	26.1%	<.0001
c	Subject	Microorganism and person	23.2%	9.2%	.0005
		Microorganism	76.8%	90.8%	.0005
d	Context	Nature	10.8%	13.6%	NS
		Domestic	6.5%	2.7%	NS
		Lecture	–	2.7%	–
		Human body	19.5%	23.4%	NS
		None	68.1%	66.3%	NS
e	Captions		19.5%	40.8%	<.0001
f	Causal link		8.6%	27.7%	<.0001
g	Emphasis	Theory	–	89.7%	–
		Action	–	28.8%	–
h	Hygiene practices	Person	–	8.7%	–
	Hygiene practices	Food	–	1.6%	–
i	Action	Positive	–	9.2%	–
		Negative	–	26.1%	–

Statistically significant differences between pre- and post-intervention are reported. Percentage (and p-values) is missing for those features that were not measured at pre-intervention.

Moreover, marked increases in inserting captions to explain the functions of microorganisms (19.5% and 40.8% pre- and post-intervention, respectively; p < .0001) and in showing a casual linkage between actions of microorganisms and their consequences on people and environment (8.6% and 27.7% pre- and post-intervention, respectively; p < .0001) were observed (Table 3).

Finally, analysing the relations between drawing variables, a strong correlation (greater than .88) between causal link, with the emphasis on action and negativity of action, emerged after the intervention. The cause–effect relationship was then represented by harmful actions executed by microorganisms.

Predictor variables influencing knowledge of microorganisms

Figure 3 shows the relation between X variables and Y responses. The predictor variables f1 (causal linkage between objects), g2 (effects of microorganisms) and i2 (negative actions of microorganisms) were very close together, indicating a strong correlation between them. Therefore, children who represented a causal linkage between objects in their drawings likely also depicted negative actions of microorganisms on food and drew the microorganisms in the natural environment.

Figure 3.

Loading plot of X variables and Y responses. Variables f1 (causal linkage between objects), g2 (microorganisms’ effect) and i2 (negative microorganisms’ actions) are close in the plane, indicating strong correlation. The importance of a given X variable for Y is proportional to the projection on Y of its distance from the origin in the loading space. The predictor variables that mostly influenced the knowledge score were e1 (presence of caption), followed by g1 (emphasis on theoretic content) and d1 (presence of a natural context).

The ability of features in the drawings to predict the knowledge score was relatively good, but prediction of the behaviours score was poor (as stated above). The features in the drawings that mostly influenced the knowledge score were the presence of a caption, followed by emphasis on theoretical content and the presence of a natural context. In Table 4, regression coefficients are reported. From the univariate analysis, the features having significantly different frequencies pre- and post-intervention were humanised shape, undefined shape, subject, caption and casual link, but the multivariate regression revealed that only the last two features were related and predicted the value of the knowledge score.

Table 4.

Regression coefficients for two-component analysis.

		Knowledge score	Behaviour score
Shape	Standard (b1)	0.09	0.04
	Insects, animals or fantasy (b2)	0.16	0.03
	Indefinite (b4)	−0.35	−0.11
Subject	Child and microorganism (c1)	0.02	0.00
Subject	Microorganism (c2)	−0.02	0.00
Context	Natural environment (d1)	1.00	0.24
	Domestic environment (d3)	−0.33	−0.08
	Scholastic environment (d4)	−0.03	−0.01
	Body context (d5)	−0.47	−0.14
	No context (d6)	−0.05	0.02
Captions	Caption (e1)	1.42	0.35
Causal link	Causal link (f1)	0.02	−0.03
Focus	Theoretical content (g1)	1.17	0.30
Focus	Microorganisms’ effects (g2)	0.21	0.02
Hygienic procedures	Focused on the person (h1)	−1.07	−0.27
Hygienic procedures	Focused on the food (h2)	−0.14	−0.04
Microorganisms’ actions	Positive (i1)	0.22	0.05
Microorganisms’ actions	Negative (i2)	−0.21	−0.08

Discussion

As the main focus of the study lay in the investigation of links between the drawing features and the questionnaire scores, it is useful to briefly introduce the elements which emerged from pre- and post-intervention drawing analyses.

The absence of any clustering or clear trend in the pre-intervention drawings may have been caused by a range of factors. Likely, most of the children involved in the study had not heard about microorganisms before. They depicted cells, little insects or worms with eyes and teeth, fantasy monsters or cartoon characters. Therefore, it was impossible to find any evidence of homogeneity in the data relating to the characteristics of the pre-intervention drawings (Figures 1 and 2).

Post-intervention representations were more homogeneous. The microorganism shapes were mostly standardised as bacteria, drawings were often characterised by captions and by causal links between objects and children paid particular attention to adverse effects of microorganisms on humans (Table 3).

Children who understood the information received during the lessons and achieved high scores in the knowledge questionnaire were more likely to add words in balloons or captions beside the figures to explain the illustrations. Words may be helpful to explain in detail what the drawing is not able to express since they can fix the scene in a specific moment; otherwise, it is difficult to represent what has happened before contagion or what will happen after it. The statements produced were not solely referrals to information given out during the lessons by the experts involved and reported as such, but they also encompassed personal reformulations and original ideas.

The presence of the natural environment suggests that children tend to organise scientific knowledge more deeply when they frame it in a context that they recognise, and, in this case, when they comprehended the effects of microorganisms on situations in everyday life. Some children explained the concept of contagion, depicting a sneeze or unwashed hands, or they associated the actions of microorganisms on food or the temporal changes on human bodies with pathogens. This ability, post-intervention, to highlight the causal link between elements, mainly associated with negative actions, indicates a good understanding of knowledge related to prevention of diseases caused by foodborne pathogens.

All the above comments refer to the correlations between drawing features and knowledge scores (Figure 1). We cannot comment on correlation of the drawings with the behaviours score since, as shown in Figure 2, it was not possible to perceive a clear pattern in the behaviours scores as related to drawing features.

In commenting on this finding, we must recall that data derived from the behaviours part of our questionnaire were difficult to interpret in this, as well as in other analyses (Losasso et al., 2014). Personal knowledge of a phenomenon consists of information, which is always available, reliable and sure. However, implementation of specific behaviours can depend on external conditions, independent of the child’s preferences. For this reason, data derived from the behaviours scores in this study may be less reliable.

The lack of correlation between behaviours scores and drawing features could also suggest a limitation of the study. If the drawing title had been ‘How to prevent the action of microorganisms’ instead of ‘Microorganisms and me’, perhaps a better correlation between behaviour scores and drawing features would have resulted. A second limitation concerns the large amount of time required for children to make drawings and the involvement of many teachers to collect them.

Despite these limiting factors, our findings suggest that drawing can be a powerful tool for children’s health promotion programmes. Drawing has been previously considered a means to express creativity, which has to be interpreted, more than a device to verify comprehension levels and learning quality (Butler et al., 1995). In contrast, our findings suggest the potential of drawings in health education programmes: analysis of post-intervention drawings could allow researchers to better assess the effectiveness of their teaching methods. Moreover, awareness of what information was salient from the children’s point of view could be a useful starting point for the preparation of content for further intervention programmes. Children’s drawings can also be used as part of the teaching materials, for highlighting both correct and incorrect information. The use of drawings could also help in identifying young children’s naive thinking about the biological world, which they may have acquired before receiving correct information, and which can sometimes persist even after an informative programme, if they do not receive help to reason about it (Inagaki and Hatano, 2002). Despite the fact that drawing analysis requires substantial time for the rigorous coding of drawing features, it provides researchers with information that can be used in creative ways for health promotion.

Conclusion

This pilot-study investigated the effectiveness of drawing as an evaluation tool for health promotion programmes aimed at young children. Although several studies have demonstrated the efficacy of drawings in facilitating children’s recall of information and experiences (Barlow et al., 2011; Gross et al., 2009), the majority of tools used to evaluate the effectiveness of interventions still rely on questionnaires and surveys. This kind of quantitative methodology, while ensuring rigour and reliability, may force the language and methods of verification to be in line with the expectations of the researcher. The items in questionnaires are formulated by researchers and, even if pre-tested, may underestimate the parts of the teaching programme, which make the greatest impression on the children. Drawings and other qualitative tools (including interviews and written open questions) guarantee the expression of creativity and may be more ecologically valid tools, especially for primary aged school students.

The current results suggest the importance of combining qualitative with quantitative methods, rather than excluding one or the other, in order to optimise children’s acquisition of information and knowledge. More specifically, our study suggests that children who better understood the content of the health programme were those who could express contextual knowledge of the information acquired and highlight the negative actions of microorganisms on the human body. Since children understood the cause and effect relationship between microorganisms and disease development, they also explained this causal link in words. We conclude, therefore, that these children were able to depict in their drawings all the important elements corresponding to the explanations they had received in the health programme.

Starting from these key points, it may be useful to enlarge the investigation of health programme effectiveness criteria by using this alternative means to evaluate children’s comprehension of complex scientific knowledge. It would also be useful to investigate whether the use of drawing can strengthen and consolidate any information gain by further verifying the quality of learning after more time lapse. Finally, it would be interesting to ascertain whether drawing could be helpful to those students with a visual learning style.

Several educators and researchers have urged that drawing be used more frequently to enrich science education, suggesting that children’s drawings in science can contribute to the development of individual skills, knowledge and understanding (Hayes et al., 1994). Drawings have been used in a variety of ways to probe understanding in science, especially to explore children’s ideas about abstract concepts (Brooks, 2009). There appears to have been no systematic attempt to involve drawing in general science education at the primary school level and there has been little research into the meaning-making processes through drawing (Brooks, 2005).

Likewise, to our knowledge, there is no rigorous research on the use of drawing in the evaluation of health promotion programmes aimed at children. This study clearly shows that this novel method of analysis is possible.

Footnotes

Funding

The author(s) received no financial support for the research, authorship and/or publication of this article.

References

Abrahams

Millar

(2008) Does practical work actually work? A study of the effectiveness of practical work as a teaching method in school science. International Journal of Science Education 30(14): 1945–1969.

Alderson

(1995) Listening to Children: Children, Ethics and Social Research. Ilford: Barnardos.

Backett

Alexander

(1991) Talking to young children about health: Methods and findings. Health Education Journal 50: 134–137.

Barlow

Jolley

Hallam

(2011) Drawings as memory aids: Optimising the drawing method to facilitate young children’s recall. Applied Cognitive Psychology 25: 480–487.

Bombi

Pinto

(2000) Le relazioni interpersonali del bambino. Studiare la socialità infantile con il disegno. Rome: Carocci.

Brooks

(2005) Drawing as a unique mental development tool for young children: Interpersonal and intrapersonal dialogues. Contemporary Issues in Early Childhood 6(1): 80–91.

Brooks

(2009) Drawing, visualisation and young children’s exploration of ‘Big Ideas’. International Journal of Science Education 31(3): 319–341.

Butler

Gross

Hayne

(1995) The effect of drawing on memory performance in young children. Developmental Psychology 31: 597–608.

Byrd-Bredbenner

Abbot

Quick

(2010) Food safety knowledge and beliefs of middle school children: Implications for food safety educators. Journal of Food Science Education 9: 19–30.

10.

Cipolletta

Faccio

(2013) Time experience during the assisted reproductive journey: A phenomenological analysis of Italian couples’ narratives. Journal of Reproductive and Infant Psychology 31(3): 285–298.

11.

Esposito Vinzi

Chin

Henseler

. (2010) Handbook of Partial Least Squares. New York: Springer.

12.

Faccio

(2013) The Corporeal Identity: When the Self-Image Hurts. New York: Springer.

13.

Faccio

Costa

(2013) The presentation of self in everyday prison life. Global Crime 14(4): 386–403.

14.

Faccio

Belloni

Castelnuovo

(2012) The power semantics in self and other repertory grid representations: A comparison between obese and normal-weight adult women. Frontiers in Psychology 3(517): 1–7.

15.

Faccio

Costa

Losasso

. (2013) What programs work to promote health for children? Exploring beliefs on microorganisms and on food safety control behavior in primary schools. Food Control 33: 320–329.

16.

Gan

Scardamalia

Hong

H-Y

Zhang

(2007) Making thinking visible: Growth in graphical literacy, Grades 3 and 4. In: Proceedings of the International Conference on Computer Supported Collaborative Learning, Rutgers, The State University of New Jersey, NJ, pp. 206–208.

17.

Gross

Hayne

Drury

(2009) Drawing facilitates children’s reports of factual and narrative information: Implications for educational contexts. Applied Cognitive Psychology 23: 953–971.

18.

Grotzer

(1999) Math/Science Matter: Resource Booklets on Research in Math and Science Learning. Cambridge, MA: Harvard Project on Schooling and Children.

19.

Haapala

Probert

(2004) Food safety knowledge, perceptions, and behaviours among middle school students. Journal of Nutrition Education and Behavior 36: 71–76.

20.

Hayes

Symington

Martin

(1994) Drawing during science activity in the primary school. International Journal of Science Education 16: 265–277.

21.

Inagaki

Hatano

(2002) Young Children’s Naive Thinking about the Biological World. Philadelphia, PA: Psychology Press.

22.

Iudici

Castelnuovo

Faccio

(2015) New drugs and polydrug use: Implications for clinical psychology. Frontiers in Psychology 6(267): 1–3.

23.

Losasso

Cappa

Cibin

. (2014) Food safety and hygiene lessons in the primary school: Implications for risk-reduction behaviors. Foodborne Pathogens and Disease 11(1): 68–74.

24.

Losasso

Cibin

Cappa

. (2012) Food safety and nutrition: Improving consumer-behaviour. Food Control 26: 252–258.

25.

McBurney

(1983) Experimental Psychology. Belmont, CA: Wadsworth Publishing Company.

26.

Marcus

(2008) New information about pediatric foodborne infections: The view from FoodNet. Current Opinion in Pediatrics 20: 79–84.

27.

Mevik

B-H

Wehrens

(2007) The pls package: Principal component and partial least squares regression in R. Journal of Statistical Software 18(2): 1–24.

28.

Pridmore

Bendelow

(1995) Images of health: Exploring beliefs of children using the ‘draw-and-write’ technique. Health Education Journal 56: 219–230.

29.

Shearer

AEH

Snider

Kniel

(2013) Development, dissemination and preimplementation evaluation of food safety education materials for secondary education. Journal of Food Science Education 12: 28–37.

30.

Taylor

Andrews

(1993) The Arts in the Primary School. London: Falmer Press.

31.

Williams

Wetton

Moon

(1989) A Way in: Five Key Areas of Health Education. London: Health Education Authority.

32.

Woodruff

Kirby

(2013) The association among family meal frequency, food preparation frequency, self-efficacy for cooking, and food preparation techniques in children and adolescents. Journal of Nutrition Education and Behavior 45(4): 296–303.