CHAPTER 3. USING VISUAL METHODS TO EXPLORE SELF-HARM
3.8 D ISCUSSION
3.8.3 Part three: A critique and recommendations for future research
3.8.3 Part three: A critique and recommendations for future research
3.8.3.1 Is the method restrictive - Is it accessible to everyone?
Throughout the research process I was prompted to consider whether some people might be more inclined than others to participate in a visual methods study. Though this is arguably a consideration for all types of research approaches it felt as though the visual element of the task was quite significant. Like Frith and Harcourt (2007) pointed out in their study, this type of approach would not appeal to all and so those
participants who do consent to participate may be more familiar with, or receptive to, the idea of taking photographs. Almost certainly, capturing images to represent experience of self-harm was more of a challenge for some than others – see p. 163;
obviously this could have been due to a number of factors, not least the topic area.
Other factors such as personal preference and perceived creative ability seemed relevant, and although the method was intended to give participants control and facilitate their expression, for some it might have been perceived as a measure of their ability. More specifically, resembling the participants in Mannay’s (2010), Packard’s (2008) and Frith and Harcourt’s study (2007), on several occasions participants apologised for their images and seemed to lack confidence when showing their
images, or a lack thereof. Some seemed embarrassed and perhaps felt under pressure to produce several images of great interest, which in turn might have inhibited their ability to express their experience of self-harm. Like Packard (2008) I also experienced the discomfort of looking at some of the images with participants which were of poor quality. For example, Nicola seemed particularly embarrassed when looking at some of her images which she had taken whilst intoxicated. Some of the images were blurred and of such poor quality she was unable to identify their purpose. This particular point
Nicola, it was sometimes questionable whether she had capacity to consent to
participate, and it became apparent that whilst taking some of the images (generating the data) she was intoxicated. This led to questioning whether some of the images should have been excluded or whether additional consent to use all of her images should have been sought. Though she did not express a wish to exclude any of her images I’m not confident that this was the most appropriate way to handle that
particular data. So, whilst I would agree this type of approach is beneficial and valuable in changing the research dynamic and establishing rapport, I would also add, like Packard (2008), that it might also inadvertently embarrass participants and inhibit communication.
So, despite my intention to design a research task which did not require any level of skill or expertise, this was not always how it was perceived or experienced. Taking pictures is considered a familiar method, however on reflection that wasn’t the issue.
Taking pictures to represent difficult experiences in a difficult context wasn’t familiar and that was the issue that required more thought. Using pictures to represent experiences of self-harm is assumedly not something participants would do
spontaneously, instead it required effort, abstract thinking and reflexivity (Drew et al., 2010) which perhaps some people struggled with more than others. The apparent prerequisite to be self-reflexive and able to symbolise experiences might relate to some of the previous discussions around barriers and acquisition of an emotional lexicon (p.60). However, in spite of these potential limitations participants did approach the task differently. Some people captured images which at first glance were quite banal and more concrete in form, and appeared to require little skill and expertise in terms of reflexivity and symbolic representation. For example, a cup, a chair, a bed, yet they unveiled interesting and complex narratives just as much as those images that were more metaphorical and abstract in form.
3.8.3.2 Capturing the intangible, or can you?
The assumption that affective dimensions of experience can be captured through language is questionable. Firstly, emotions and feelings are not always obvious to those experiencing them and so they will not necessarily be disclosed verbally; and secondly, affect is often described as ineffable, something which is not always amenable to verbal description (Cromby, 2011), hence the exploration of the utility of visual methods in accessing the affective experience, and others, with people who self-harm. However some of the findings from the study might suggest that visual methods might also struggle to capture the ineffable and the intangible. Two participants spoke
of the difficulties associated with capturing the intangible features of their experience, like their varied affective states or memories. For example, Tori reported, “finding images for stuff like emotions and things like being angry it’s like, I just, I don’t, I don’t understand how I can take a picture of anger, like I guess I could take a picture of something that causes the anger which I did”. Tori’s challenge of capturing her affective features of her experience was similar to what Bancroft et al (1979), Rodham et al (2004) and Michel et al (1994) reported. When participants were asked to describe in their own words the reasons for their self-harm, they were more likely to articulate concrete antecedent events, rather than affective experiences which served their self-harm.
Similarly, like many other studies, findings related to some of the more conceptual functional models of self-harm didn’t feature. For example, experiences relating to the dissociation or the boundaries models. It is possible that for some the method is
restrictive in its ability to capture the more difficult and intangible internal processes and affective experiences; the camera faces outwards and perhaps limits the scope of what some people might consider capturing when tasked with this sort of exercise. On a similar note, some of the more embarrassing and ‘positive’ functions of self-harm didn’t feature either, for example none of the participants spoke of any sexually gratifying experiences associated with their self-harm.
Of course it was only one study with a small sample and perhaps more participants might have yielded such discussions. Even so, it might have proved more
embarrassing to capture a picture of something sexually related and positive compared with ticking a box, especially given the fact that the picture was knowingly going to be viewed by others and discussed. So perhaps, using images in this ‘public’ way only allows access to some of the hidden and the private, and those experiences which might be perceived as more embarrassing or unusual remain private and inaccessible.
Exploring the content of images where the owner remains anonymous on the other hand might offer us access to something different. This will be explored in the next chapter.
3.8.3.3 Reflections on the method
Sample size: The sample size used in this study might be considered a limitation in that I cannot be confident of saturation, or transferability to the population as a whole.
Nonetheless, a smaller sample might prove useful still in future research and enable a more focussed in depth analysis of the data, after all delving into the particular is said
heterogeneous sample of eight participants yielded a vast amount of rich and distinct visual and textual accounts of self-harm, which were analysed individually before looking at the data as a whole. Through this process the vastness and uniqueness of the accounts became apparent and given the exploratory nature of the study all of the data (over ten hours of dialogue and 143 pictures) were attended to in detail. Thus, although the sample may be considered small, it was diverse, in depth and rich which seemed an acceptable compromise.
Presenting the analysis: The analysis involved an exploration of each individuals account followed by a journey into the data across participants. Presenting these results in totality within the thesis was a challenge; the individual accounts offered richness through context and depth, some of which was lost in the presentation of the group analysis (although this richness was inherent within the analysis). However, presenting all the individual accounts was unwieldy. The compromise in the presentation was to offer only two full individual narratives to give a flavour of the analysis. It may be that such a presentation loses the richness of the individual but hopefully the group analysis is useful in capturing that which is common. It may be that for future research a more definite case study approach could be adopted to capture the complexity.
Multiple interviews: Data from some participants was collected on more than one occasion, though this was usually at the wish of the participant and prior to any formal data analysis. Alternatively, designing research which aims to collect data from all participants on more than one occasion might offer a further opportunity to those who didn’t present with images or those who presented with fewer images to engage with the method. Moreover, the formal process of analysis usually begins after data collection has ended which subsequently limits what we might explore in subsequent interviews. Our initial thoughts about what might be considered most poignant within the data can often change as the analysis matures, as can our thoughts about what questions are being raised in the data. Further interviews following preliminary analysis might allow for further exploration of particular phenomena. For example, further exploration of some of the cross cutting themes such as, the conflicts between the public and the private, would have proved useful.
Number of images: Not having posed a restriction on the number of images
generated by participants allowed for participants to explore all different aspects of their experience, including those distal and proximal. However, in hindsight this may have been overwhelming. An abundance of images proved difficult to work with during the interview for both the participant and the researcher, and proved difficult during the
reduced any uncertainties and anxieties about the task. Yet researchers need to be mindful of whether this might inadvertently restrict access to content.
Exploring the Process: In the closing section of the interview participants’ were asked questions relating to the method, however there was little focus on their responses which in retrospect could have been explored further. More emphasis on the actual process of taking images would have proved valuable when attempting to
comprehensively critique the method. For example, Frith et al (2005) wrote about the importance of how people manage the process of engaging with the camera, such as, context and circumstances of the fabrication of photos, what their choices were
regarding what to make visible and why, especially when asked to capture experiences they wouldn’t usually photograph. Given some of the discussions about accessibility as a method in terms of people and topic, perhaps this is something future research would benefit from.
Analysis of visual data: An adapted polytextual thematic analysis was employed due to its ability to incorporate the analysis of visual images to understand people’s
experiences of self-harm (Gleeson, 2011). Whilst it was developed to analyse visual images, Gleeson (2011) acknowledged the value of drawing on supporting materials where available to help contextualise those images. Similarly, others have also
suggested that images in isolation can be problematic for research, rather, they should be surveyed with the series of images to which they belong and any other contextual annotation (Frith et al., 2005, Collier, 2002). So, pictures can usually only be
understood when accompanied by a commentary. Whilst I would agree with this to an extent, I also think there might be some value in looking at images separately from text.
Whilst carrying out my analysis I noted how the analysis of images felt superficial because of the difficulty I had in removing myself from the text that accompanied the images, whereas surveying the images separate from the narrative seemed to help develop the analysis and still provide an analytic account which was grounded within the data. So whilst the analysis of images without other contextual information might be disadvantageous, it might also be a further way to learn more about the personal experience of self-harm. The findings from this study suggested that self-harm, like images, is sometimes used as a visual language when other means of expression have proved too difficult. Thus, exploring images as a means of communication in and of themselves might prove useful.
Ethical challenges
This next section will discuss how some of the ethical procedures outlined in section 3.3.2, which are considered ‘standard’ in word or number based research became more problematic, though not irresolvable, throughout the course of the research.
Anonymity and confidentiality
Ensuring participant data is kept confidential and participants remain anonymous are fundamental ethical issues of most research; however the visual element of the study had the potential to impair such efforts. Though participants did not take identifiable pictures of themselves or others in most cases, they did take pictures that captured other identifiable data. For example, printed names or signage of a school attended by their child. As with textual data, to counter this problem efforts to remove identifiable data were carried out, however, I would agree with Clark (2006) in that it is practically impossible to completely achieve anonymity of place. Subsequently, where necessary only select images were used to illustrate themes or photo editing software was used.
It is worth noting however that pixilation of images can be considered dehumanising, associated with criminal activity (Banks, 2001), and futile; sometimes in attempting to disguise data the very essence of the intended message can be lost (Clark et al., 2010). Also, if one of the objectives of using images is to express the very messages we feel unable to express through text (words) then altering images might seem contradictory.
The use of participant generated visual material in particular can attract added problems. For example, a participant’s single image was due to be published and so the participant was fully notified of the planned publication and expressed a wish to be named, however given the nature of the topic area the researcher felt it necessary to ensure they remained anonymous, as per the signed agreement (consent form). While they may have wished to be named at this point in their lives, their view may change in the future (Barrett, 2004) and removal of their name could prove difficult or impossible (Banks, 2001). Such efforts to protect participants however can be interpreted as silencing participants (Walker et al., 2008 cited in Clark et al 2010). These
interpretations are both concerning and paradoxical given one of the intentions of the project design was to create an opportunity for those considered marginalised to have a voice and a sense of control over the research process.
Consent to use images
Permission, in the form of written consent, for the researcher to keep copies of and publish images was sought from each participant, including permission to use the data in future research. Nevertheless, as a researcher using participant’s images in different contexts did rouse ethical uncertainty. Though all of the participants gave their consent
for the researcher to use their images when discussing the project to a wider audience, as Clark et al (2010) and others have pointed out, statements such as these can be ambiguous in that they encompass a host of different uses i.e. different modes of dissemination to different audiences, most of which regardless of whether they are clearly stated, will not necessarily be fully understood by participants (Wiles et al., 2008).
Moreover, recent regulations within certain institutions require all PhD candidates to submit an electronic copy of their thesis; in other words, deposit their data on the internet, implications of which may not have been anticipated by participants or the researcher at the outset of their research endeavours. The deposit of visual data in particular is likely to introduce further ethical concerns which participants should be fully informed of.
Even further complexities and strategies with regards to consent have been discussed in relation to visual methods. For example, Davies (2008) on behalf of the Economic and Social Research Council (ESRC) published a toolkit for visual researchers specifically focussing on issues of informed consent. She stressed how participant generated photographs differ in their emotional charge. For example, some images may be more poignant and sensitive than others and so participants may wish for restricted use of those images. Offering participants the choice to consent to use of individual images rather than consent to use the whole of the data is demonstrable of ethical practice, though Davies (2008) does acknowledge how this can introduce concern and confusion for participants resulting in them feeling reluctant to partake.
Although participants in this study were not given the opportunity to consent to
individual images, arguably, requesting their consent to use of their visual data outside the university, as discussed previously in relation to anonymity, could be perceived as an attempt to offer participants the opportunity to withdraw their consent or restrict their consent to use certain images only in certain contexts.
Finally, some participants may present pictures of existing images. For example, images of another person’s artwork or images from a magazine. This then raises the question of ownership (copyright) of the image and how it can be used. In such cases it was deemed most ethical and legal to treat the image in the same way as an image that captured someone other than the participant.
Content of images
The nature of the topic being explored did increase the risk of capturing images of a distressing nature. This particular risk was highlighted if images were generated using a disposable camera which required development through an outside agency. Using
digital cameras where selection and/ or printing of images were carried out by the participant and / or the researcher was seen as a way of minimising any risk.
Nonetheless, it failed to consider the subsequent use of those images, such as, dissemination and the effects on the viewer. In this study, in which there were no restrictions on image content other than to avoid taking pictures of others, contrary to what people may have imagined, none of the participants took images of a distressing nature to represent their experience of self-harm, or photos of actual self-harm. This could be due to several factors, including the choice to keep certain things hidden and private, or that their images were being used to translate similar meaning rendering images of self-harm unnecessary.
Ownership and disposal
Finally, some of the literature refers to ethical concerns of ownership and disposal of visual data (Temple and McVittie, 2005). For this particular study participants were considered the owners of the images and permission to keep copies of their images was sought through written consent. Interestingly, despite having clear rules about ownership most participants were reluctant to take away their images following the interview. It is possible the images ceased to have a life beyond the project because they were made for the project and not for them, leaving them redundant. Also, due to
Finally, some of the literature refers to ethical concerns of ownership and disposal of visual data (Temple and McVittie, 2005). For this particular study participants were considered the owners of the images and permission to keep copies of their images was sought through written consent. Interestingly, despite having clear rules about ownership most participants were reluctant to take away their images following the interview. It is possible the images ceased to have a life beyond the project because they were made for the project and not for them, leaving them redundant. Also, due to