Using the draw and tell method with adolescents as part of an interpretive descriptive study

Introduction

Methodology is a complex concept, encompassing two components of the research approach – design and methods – that guides the process of inquiry (Durdella, 2019; Schwandt, 2015). It is crucial that researchers give due consideration to methodological choices, as even the most seemingly insignificant variation in design may lead to vastly different outcomes (Wynn and Borrie, 2020). Although research papers contribute potentially invaluable information to a field, some research may lack a robust methodology (Chaudhary, 2020), calling into question conclusions offered (Belaroussi et al., 2020; Chaudhary, 2020). Owing to these pitfalls, it has been suggested that “robust research methodology must be the guiding tenet around which important decisions on publications are made” (Chaudhary, 2020, p. 347).

However, despite the importance of methodology, it is important that researchers do not engage in what Lowenberg (1940) referred to as “methodolatry”, a combination of method and idolatry (Janesick, 1994), where methodology is viewed in an almost sacred manner and placed higher than the phenomenon under investigation (Chamberlain, 2000; Charmaz, 2017; O'Connor, 2012). Several authors have recommended avoiding the practice of methodolatry, as it forces the researcher into an unquestioning attachment to the method, thus distancing them from the substance of their findings (Janesick, 1994; O'Connor, 2012; Regelski, 2002; St. Pierre, 2018). Rather than aligning oneself to a particular methodology with blind devotion, one which may not be entirely appropriate in answering a specific research question, researchers should instead adopt more critical and creative perspectives, identifying alternative and more apposite methodologies.

The methodological approach adopted may alter study outcomes (Wynn and Borrie, 2020), and so the design adopted warrants deliberation. We aimed to conduct a study exploring adolescents' perceptions of mental health services, the results of which have been published elsewhere (Goodwin et al., 2022). Adolescents with experiences of mental health services were excluded from this study; as such, their “general population” position precluded us from using traditional methodologies, such as phenomenology, given that experience is generally central to such approaches (Thorne, 2016). While discourse analysis would have illuminated “how” discourses around mental health services are organised, highlighting the semantic intention of adolescents (Alba-Juez, 2009), we were also interested in exploring “what” this population group spoke about. This prompted us to adopt a content analysis approach. However, we were conscious of the position of content analysis as “method” rather than “design”; as such, the methods used to collect and interpret data warranted underpinning by an appropriate research design. Furthermore, owing to the fact that adolescents can be a challenging population to engage in the research interview process (Bouchard, 2016; Boyle, 2007; Hultman et al., 2017), we proposed the utilisation of a supplemental strategy in addition to the traditional interview. Ultimately, we encouraged study participants to draw their perceptions of mental health services and positioned the research within an interpretive descriptive design.

The aim of this paper is to provide an account of a novel approach to engaging adolescent research participants in the interview process. We report on how a creative way of collecting data was well situated in an interpretive descriptive framework. First, an overview of interpretive description is presented, with a rationale provided for its utility in a study on adolescents' perceptions of mental health services. Secondly, the use of drawings with adolescents – specifically, Williams et al.’s (1989a, b) draw and tell method is addressed.

Interpretive description: an applied methodological approach

Owing to the way the world is textured, there are some complex social phenomena that traditional academic methods of inquiry do not adequately capture (Law, 2004). With this is mind, several researchers have called for new, creative methodological approaches that would facilitate exploration of research questions not easily achieved using traditional methodologies (Hunt, 2009; Sandelowski, 2000; Thorne, 2016; Toms and Wilson, 2016; Wiersma, 2011). Although there are several methodological approaches available to researchers (such as phenomenology and ethnography), in the 1980s, it became apparent that, in order to adequately answer questions which would further advance healthcare research, rigorously adhering to existing methodologies may have been limiting (Thorne, 2016, 2020). At the same time, other disciplines associated with qualitative research (such as education and social science) showed less concern for strict adherence to methodological frameworks and were able to adapt/modify conventional approaches so that their unique needs could be more satisfactorily met. Health-related research was one of the few qualitative-related disciplines which continued to demonstrate strict methodological rigidity. There was a great deal of pressure put on health researchers to locate themselves within a particular tradition, to make explicit references to “branded” (e.g. grounded theory) methodologies and to cite the recognised authorities to justify their work (González Rey, 2020; Thorne, 2011, 2016).

Annells (2007) comments that, owing to such pressures, some authors claimed to use a certain research approach, but what they ultimately produced bore little resemblance to what they purported they were initially using. Although such practices continue into the present day (Charmaz and Bryant, 2021), there has been more confidence placed in applied methodologies, where traditional branded methodologies have been re-framed so that their design logic is more appropriate to the complex and distinct requirements of the various disciplines (Shea et al., 2019; Thorne, 2016). These applied methodologies provide researchers with a more diverse palette of methodological colours from which to choose; essentially, rather than the methodology dictating how a question should be answered, the research question dictates the way in which a methodology should be applied.

One such applied methodology, or methodological approach, is interpretive description, a noncategorical alternative for developing knowledge in qualitative research (Thorne et al., 1997). Traditional qualitative description (e.g. Brown, 2021; Sandelowski, 2010) often involves entering the field with a carefully constructed set of questions, with, for example, responses to interview questions comprising the data. It could be argued that such an approach may be limited and may not sufficiently answer the relevant research question in depth. In this sense, traditional qualitative description may be viewed as limiting. Interpretive description, on the other hand, encourages a more inductive approach to research, producing description, but also pushing for the “so what?” question and uncovering the deeper meaning within the data (Lundin Gurné et al., 2021; Thorne, 2011, 2016). The suggestion to move away from traditional description is not a criticism of this approach; certainly, there is a place for descriptive research within the qualitative domain, if the phenomenon under investigation demands that level of analysis (Chamberlain, 2000); however, description is important in bringing awareness to phenomena and in generating new questions. In interpretive descriptive studies, empirical knowledge is generated through contextual understanding of human phenomena (Thorne, 2016). Such knowledge is important in advancing healthcare professions and is inherently complex and diverse.

Researchers engaging in interpretive description are provided with the flexibility to critically move beyond the constraints of faithfulness to conventional structured methodologies, whilst also acknowledging the various techniques that have emerged from various qualitative approaches over the years (Thorne, 2016). Law (2004) commented on the futility associated with employing fixed or stable approaches when endeavouring to capture “the ephemeral, the indefinite, and the irregular” (p. 4). We anticipated a great deal of uncertainty and irregularity during interviews as a consequence of adolescents' lack of experience with mental health services. This is what made this interpretive description so attractive to a study on adolescents' perceptions of mental health services: the facility to adapt existing strategies and develop the appropriate tools to engage an adolescent population in the interview process. As will be discussed later, participant-produced drawings were completed during research interviews. Such an approach is well suited to interpretive description, where data collection takes many forms – indeed, Thorne (2016) acknowledges that “the possibilities are infinite for creative researchers” (p. 135) who are encouraged to “enter into a creative discovery process” (Thorne, 2016, p. 153).

Acknowledging philosophical underpinnings

However, despite certain methodological liberties afforded when engaging in interpretative description, the researcher is encouraged to think critically about the study design and its philosophical underpinnings. Interpretive description differs from other eclectic approaches that slur methods and result in the formation of a hybrid methodology that has little consideration for contextual factors (Thorne et al., 1997, Thorne, 2016). While method-slurring is not advised in interpretive description, when the researcher is cognisant of the subtleties of the various tools and techniques used to address the research question, it may be appropriate to combine elements of different approaches, provided they are philosophically compatible, both ontologically (what exists in the world) and epistemologically (how we know about what exists) (Bender and Holmes, 2019; Thorne, 2016). The interpretive descriptive researcher should build a research framework by developing an understanding of the paradigm in which they intend to operate, taking into account ontological and epistemological factors, and being aware of the values associated with the various methods before committing to a particular design (Thorne et al., 2015; Thorne, 2016).

Although researchers may not explicitly state their philosophical underpinnings, these can be deduced by the reader because the researcher always brings a certain set of beliefs or assumptions to their study (Creswell and Poth, 2017). It is important to understand the philosophical basis to a study, so that one might avoid methodolatry and instead allow for a more appropriate focus on the relevant phenomenon (Janesick, 1994; O'Connor, 2012); the application of paradigms becomes important here.

Ontological and epistemological stances, along with methodology, are the components that comprise paradigms. Paradigms are a worldview or basic set of beliefs/practices which provide lenses, frames and processes to guide the researcher through their investigation (Turyahikayo, 2021). Farghaly (2018) states that it is generally accepted that there are two main paradigms: positivism and naturalism, where the former's deductive cause-and-effect approach appeals to quantitative researchers, and the latter's approach to constructing knowledge through inductive means appeals to qualitative researcher (it should be acknowledged that other paradigms are also employed, including, for example, post-positivism, pragmatism, post-modernism, etc. (Turyahikayo, 2021)). Naturalism is the paradigm most associated with interpretive description (Thompson-Burdine et al., 2021). Naturalists emphasise the importance of interpreting the world to promote understanding. More so than other paradigms, naturalism is concerned with people and how they make sense of the social world, by seeking in-depth meaning (Scotland, 2012). Bosacki (2016) argues that people begin to interpret and understand the world at an early age, “guided by tenets of relativism, constructivism, narrative, and self-agency” (p. 19). Adolescence, in particular, is a time when significant advances are made in terms of how social reality is perceived and processed (Steinberg, 2019); therefore, adolescents are an important population for naturalistic research to consider.

The naturalistic paradigm connects with the ontology of relativism, or the appreciation of multiple realities. Whereas the ontological concept of realism is concerned with objective truths, relativism is associated with meaning that is contextually bound (Scotland, 2012). Indeed, the contextual nature of knowledge is core to interpretive description (Thorne et al., 1997). Boghossian (2007) uses Galileo's discovery of the truths about motion to explain relativism, i.e. the motion of an object is relative to a frame of reference. In this sense, “truth” is a relative concept, contextually bound and dependent on a variable frame of reference. It has been argued that a relativistic perception of the world first emerges during adolescence (Chandler, 1987). A child's frame of reference changes as they enter adolescence, influenced by structural biological changes and ongoing language developments, meaning adolescence is an “ontologically distinct experience” (Béhague, 2018, p. 78). It should be noted that relativism can be understood in extreme (“hard relativism”) terms; such a perspective is not particularly helpful, because the unique position of each participant positions would need to be considered, meaning commonalities would be difficult to identify. “Softer” forms of relativism – where “truths” become subject to external criticisms (Alexander, 1986) offer the researcher more scope to analyse qualitative data and report commonalities/patterns.

Stemming from the ontology of relativism is the epistemology of interpretivism. The interpretivist recognises that knowledge is socially constructed (Billsberry et al., 2019; Bopp and Bercht, 2021). It is their role to explore human perspectives and interpret how others view the world; this, according to the interpretivist, is how scientific knowledge is produced (Bopp and Bercht, 2021; Creswell and Poth, 2017). Developing a heightened awareness of the subjective nature of knowledge was beneficial when exploring adolescents' perceptions of mental health services. It is during adolescence that people begin to make sense of the social world (Steinberg, 2019), and as such, adolescents can often hold contradictory perceptions about various phenomena (Simonds et al., 2014; Watson et al., 2020). Acknowledging the contradictions that arise from the exploration of multiple constructed realities is a fundamental aspect of interpretive description (Thorne, 2016). Such insights were crucial when interpreting what participants in our study spoke about, and facilitated a curious approach, fostering deeper questioning and reflexive analysis.

Innovation in qualitative research: use of drawings

Although interviews have been established as a successful way of eliciting responses from research participants (Burkette, 2021; Fadyl and Nicholls, 2013; Guest et al., 2017), innovative approaches to data collection methods are warranted to advance qualitative research (Travers, 2009). Innovation is central to interpretive descriptive research, where the researcher adapts existing techniques so that they serve to address the research question (Burgess et al., 2021; Thorne, 2016). To be innovative, one must be creative, not simply applying old methods to new questions, but adapting existing intellectual resources (Burgess et al., 2021; Travers, 2009). Interpretive description is one way in which this progression can be realised, assisting researchers to create a design consistent with the research questions/aims and work outside the confines of traditional branded methodologies (Thorne et al., 2004). Taking this into account, we incorporated a drawing element into the interview process.

Several researchers have commented on the challenges associated with engaging adolescents in research (Nolan et al., 2018; Thompson, 2016), with Pincham et al. (2016) stating that this population group “are notoriously difficult to engage” (p. 743). When facing an interview, adolescents may initially present as withdrawn or distrustful, anticipating the uncomfortable prospect of adult interrogation (Briks, 2007), often responding to open questions with “I don't know” (Baer and Peterson, 2002). As such, innovative strategies need to be employed to heighten engagement and elicit more considered responses.

Muhati-Nyakundi (2021) and Wilkinson et al. (2021) comment that creative participatory research methods such as use of art are frequently considered “child-friendly”. Drawings are frequently used as a way to engage children in research (Günindi, 2015; Massey, 2022; Ogina and Nieuwenhuis, 2010; O'Connell, 2013). For example, Massey (2022) explored children's attitudes to mathematics, inviting participants aged eight and nine to draw emojis to communicate how they felt about this subject. The authors concluded that this approach was an appropriate way to conduct survey research with younger children. In Xie and Liang (2022)'s qualitative study, children aged 8–12 produced drawings about what they thought of chronic skin disease and were then asked to describe their drawing and offer further context as part of a research interview, facilitating deep insights into the stigma associated with being diagnosed with skin disease. There is also decades' worth of evidence on the success of using the “Draw-A-Scientist” approach, where children are asked to produce drawings based on their perceptions of people working in science (Miller et al., 2018). Although it is more likely that drawings are used with child interview participants, there is a growing body of knowledge on the use of this approach with adults, yielding promising results (e.g. Hammer et al.’s (2022) phenomenological study on cancer diagnosis or Nguyen et al.’s (2019) study on reproductive health care experiences of people with physical disabilities in Vietnam).

However, with adolescent populations, different participatory methods, including Photovoice (Volpe, 2019) and use of mobile phones (Wilkinson, 2016) tend to be utilised. It is unclear why drawings are used (almost) exclusively with children, although it is hypothesised that, due to a limited ability to fully articulate themselves, children are an ideal population with which to use drawings. However, drawings offer a potentially invaluable tool to researchers and should not be restricted to just children (Guillemin, 2004). Drawings can reveal how people view the world – its simplicities and also its complexities (Guillemin, 2004; Weber and Mitchell, 1996).

An adaptation of Williams et al.’s (1989a, b) draw and write method – draw and tell – guided the data collection process for our study. Draw and write was developed in the UK in 1989 to give researchers an insight into children's views on what being healthy meant to them. An exploratory bottom-up approach is emphasised, which facilitates an inductive approach to the research (Mair and Kierans, 2007; McWhirter et al., 2004; Pridmore and Bendelow, 1995). Participants are encouraged to draw an image, which is then used as context for a series of follow-up written questions. At first, the questions are general/broad, but gradually, these become more personal, focussing in on the participant's own perspectives. This technique gave way to draw and tell, which employs interview questions rather than written ones.

Because of the paucity of evidence around the use of drawings with adolescent populations, there was some initial scepticism about this approach. However, as a team, we reflected on our substantial previous experiences engaging adolescents in research interviews and the difficulties we encountered engaging these participants; similar difficulties are reflected in the extant literature (Bouchard, 2016; Boyle, 2007; Hultman et al., 2017). Consequently, a consensus was reached to undertake a novel approach to adolescent data collection. We agreed that, if participants outright refused to produce drawings, we would need to respect their wishes, abandon our suggested approach and collect data in a more traditional manner; fortunately, participants were open to the prospect of producing drawings.

The draw and tell method: draw

In interpretive descriptive studies, high-quality data are often derived from interviews that achieve depth; such depth is facilitated through the initial development of rapport with participants (of note, other forms of data – such as observations and documents – may also be used in interpretive description) (Thorne, 2016). It is well recognised that humour helps people to feel more relaxed (Jenkinson et al., 2019; Lockyer and Weaver, 2021; Shang et al., 2021) and can enhance adolescent engagement (Goodwin et al., 2019). Use of humour has been recommended as a strategy to put participants at ease so that initial rapport can be developed (Rubin and Rubin, 2011). We commenced each interview by making a joke about having “rescued” participants from their usual school routine, which always resulted in the desired effect of laughter. Next, we encouraged participants to provide assent, emphasising that their decision to partake in the study took precedence over their parents/guardian's previously provided consent, as recommended by Helseth and Slettebø (2004). Before asking the research questions and entering into the drawing component of the interview, we invited participants to fill out a demographic survey. In addition to providing scholars with contextual information about the sample, the collection of demographic details can serve as another warming-up exercise for participants, further creating a sense of comfort that can encourage more open answers throughout the interview (Passmore et al., 2002).

We tasked participants (n = 30) with drawing a person who works in mental health services, and then a place where mental health care is delivered. Although draw and tell/write was originally conceived using pencils and crayons (Williams et al., 1989a, b), we gave participants the option of using pencils, crayons, or markers. Considering adolescents are also comfortable with digital devices (Hilty et al., 2021), participants were also given the option to draw using an electronic tablet with a stylus. Most (n = 26) participants used only black pencil and paper. Two participants used colouring pencils and crayons. Two participants used the tablet. The lack of engagement with the digital device was surprising, and may indicate that, despite their familiarity with digital devices, adolescents prefer conventional pen-and-paper approaches for the purpose of producing art during research interviews.

It was explicitly stated that the drawing ability of participants was not under scrutiny. Similar to what Guillemin (2004) and Laevers and Declercq (2018) reported, the request to draw was usually followed by nervous laughter, and the participant stating, “I can't draw”. We again stressed that artistic skill was not essential, encouraging participants to produce whatever they felt comfortable with. We found that minimal encouragement was necessary before participants began drawing, possibly as a result of the emphasis we placed on initial development of rapport.

It should be noted that the images produced were not subject to data analysis. Several researchers have emphasised the importance of the researcher having a solid foundation in art interpretation if one intends to analyse participants' drawings (Ahmed and Siddiqi, 2006; Gauntlett and Holzwarth, 2006; Søndergaard and Reventlow, 2019; Thorne, 2016). As both Gauntlett and Holzwarth (2006) and Guillemin and Drew (2010) note, the most significant interpretation of a drawing is conducted by the participants who produced it. Where researchers are lacking in art interpretation skills, their attempts to understand images produced by participants are merely “guesses and speculations” (Gauntlett and Holzwarth, 2006, p. 86). We acknowledge that the inclusion of someone with these skills on the team may have facilitated an even deeper understanding of adolescents' perceptions of mental health services, and this is something for future research in this area to consider and report on. However, we used drawings solely to facilitate discussions with participants; this constituted the next phase of the draw and tell method.

The draw and tell method: tell

A semi-structured interview guide was used to elicit answers from participants. The questions within the semi-structured interview guide were first internally tested among the research team until consensus was reached. We then expert tested and pilot tested the guide. An external researcher with expertise in use of the draw and tell method reviewed the guide to ensure that the questions were age-appropriate. Following some adjustments to questions, we employed Chenail's (2011) interviewing the interviewer technique. Another external researcher (with experience in engaging adolescent populations) assisted in pilot testing the guide, assuming the role of the interviewer and asking a member of the research team questions. Again, modifications were made to the guide.

In interpretive descriptive studies, Thorne (2016) notes that, before interview question are addressed, participants may want to share a lot of introductory/background information. This is based on the fact that interviews with people with first-hand experience/knowledge of phenomena are the mainstay of qualitative research (Fadyl and Nicholls, 2013; Thorne, 2016). As the participants in the current study did not have this direct experience, we found that it was quite easy to address the research questions at the early stage of the interview, once participants appeared comfortable.

Draw and tell requires that the participants draw an image in response to a question. The participants then take part in an interview or a focus group and are invited to discuss their drawings with the researcher (Coyne et al., 2021; Günindi, 2015). We started with some general questions, asking the participant about the person or the place in their drawing – what their professional role was, or where the place was located. As anticipated, there were times when participants became reticent, responding to both closed and open questions with “I don't know”. It was at such junctures that the use of drawings became invaluable. For example, when a participant demonstrated reluctance to discuss what they perceived the role of a mental health staff member was, we directed them towards the drawing they had produced, called attention to the clothes the person was wearing, asking them if a certain item of clothing was relevant to their position. We then asked if they perceived other staff members in similar clothes, which prompted further discussion about staff roles. Similarly, when participants referred to certain mental health care environments, we asked if what they described related to the image they drew, facilitating further conversation about this image and generating follow-up questions about other environments and how these differed to their drawings.

In interpretive descriptive studies, follow-up questions should not be about narrowing the focus to eliminate that which does not seem relevant, nor is it about incessant questioning in the hope that the next answer is the key to unlocking everything (Thorne, 2016). Follow-up questions require the researcher to listen very carefully to the content being communicated by participants and then improvise new questions to explore the research topic in depth. These questions can relate to what has been specifically discussed by the participant, but it is also important to listen out for gaps/omissions and to then explore these (Rubin and Rubin, 2011). At this point, is important to be reflexive and to not allow one's personality/background to steer the interview process in predictable directions. Instead of leading participants down a particular path, follow up questions should facilitate participants in exploring the topic in their own way, thus encouraging a highly inductive process (Rubin and Rubin, 2011; Thorne, 2016).

Once it is apparent that the participant is more comfortable in answering questions and has spent some time thinking about the topic, the focus of the questions can be narrowed and the line of inquiry becomes more personal (Williams et al., 1989a, b). Once we observed that participants were more comfortable, we asked more personal questions, such as “if a friend of yours needed to access mental health services, how would you feel about this person caring for them?”. We were also able to ask questions such as, “would you feel comfortable with a friend being cared for in the place you drew?” This was an invitation for participants to look at their drawing, and then engage in self-reflection, with consideration given to the real-life applications to the questions being asked. Suddenly, it seemed that these were not just “random” questions with little relevance to them – instead, they began to realise that, potentially, people whom they knew could end up being cared for in/by the places/people they drew. Towards the end of the interview, not only had we developed a sense of rapport with participants, but we also encouraged them to reflect on a subject they may not had previously given much consideration, a development that would have not been possible without the production of drawings.

Talking about mental health is often considered “taboo” (Fung et al., 2021), and certainly, there were times when participants became uncomfortable addressing some topics, particularly when the focus of questions was narrowed and they inevitably engaged in self-reflective practices. When it was obvious that participants did not wish to pursue the line of inquiry further, we suspended the questioning process, ensured that participants were not distressed and moved onto another question (had we deemed it inappropriate to continue the interview, we would have terminated the session; however, this was not necessary). Where we deemed it appropriate to continue the interview, drawings were an indispensable tool in “rebooting” the conversation, providing participants with the context they needed to speak about topics sometimes classified as uncomfortable. In the absence of these drawings, the promotion of conversation would have been difficult, and we would have been precluded from sufficiently exploring the adolescent mindset.

Conclusion

It is important that researchers continue to advance the field of qualitative research. Such advancements cannot be achieved if novel research questions are modified to fit into the confines of branded methodological designs, or if researchers engage in methodolatry. Embracing applied methodological approaches – such as interpretive description – affords researchers the flexibility to be innovative in their approaches, to develop unique insights into unchartered territory and to generate new knowledge. We have learned that, although operating outside of the realms of traditional/branded methodologies can be challenging, it also opens up possibilities to explore research questions in a novel way.

We have also learned about the value of thinking beyond the overall design and giving due consideration to the methods researchers use to elicit responses from participants in the field. While interviewing is well established as a method to successfully explore people's perceptions, we need to acknowledge that this process may be stymied in the event of a population voicing uncertainty about the phenomena or where the subject matter induces discomfort. In our study, we learned that using drawings with adolescents facilitated deep exploration into such a subject matter and enabled us to investigate their perceptions of mental health services. Participants were provided with an opportunity to reflect on their knowledge of mental health services, illuminating gaps in knowledge, confusion and preconceptions. Referring to drawings produced progressed the conversation and facilitated a deep level of self-reflection, which enriched the data collected for this study. Consequently, this is an approach we will continue to use, explore and refine in future.

While drawings are more commonly used with younger populations, we invite other researchers to consider this creative approach to enhance research interviews with adolescent populations. Our experiences from interviewing 30 young people are that drawings broke down barriers with a population group traditionally considered “challenging” to interview. Further studies are needed where adolescents' drawings are used to complement research interviews are warranted in order to develop a more robust evidence base. However, drawings should not be used casually, and researchers are discouraged from method slurring. Instead, researchers need to think critically about the use of creative data collection strategies, ensuring that these strategies are philosophically compatible with the overall study design. We also suggest including those with art interpretation skills as part of the analysis process, as this may offer deeper understanding of phenomena.