Customizing Methodological Approaches in Qualitative Research on Vulnerable Children with Autism Spectrum Disorders

Round 1

Reviewer 1 Report

This was an interesting preliminary study and there was important information and points in the paper which reveals the potential it has to improve our understanding of ASD and application of the qualitative method with this group. Hence, there were some very important issues which need to be addressed beforehand.

The study benefited from IPA approach or interpretative phenomenological analysis with an idiographic focus, which means that this method aims to offer insights into how a given person (in this paper a child with ASD), in a given context (at the interview session with a facilitator), makes sense of a given phenomenon (questions about school wellbeing). The extracted themes are missed here.

In IPA right after transcribing the interviews, the researcher works closely and intensively with the text, annotating it closely ('coding which is generally reported in tables and sometimes with percentage and frequencies reporting') this helps with receiving and insights into the people with ASD's experience and perspective on their world. The emerging codes and reporting their frequencies might subsequently help the researcher to discover patterns in them. These patterns are called 'themes'. Themes are not mentioned here even in the findings and discussion part. I am sure reporting these in the form of a table recurring patterns of meaning (ideas, thoughts, feelings) throughout the text.

This is difficult to distinguish between the findings that are applicable to the general population (typically developing children) and children with Autism. Subheading might be helpful to improve this vagueness.

The paper has a very clear message which is stressed across the text and that is the safeguards and approaches which should be taken into account while interviewing with children with ASD. But this very “simple truth” needs more endeavour to be transferred to the reader. People with ASD are a very heterogeneous group and when it said that the participants were diagnosed with mild and moderate ASD it needs more information to understand the level of different abilities of them. I think although this study is trying to shed light on some personal aspects of the life of children with ASD in the context of school well beings, the limitations of this study need to be mentioned and also considered for the interpretation of the presented findings and for future similar research in this area. Limitations such as:

Few in the number of interviews (although this is a qualitative study but to be able to extract the themes and to justify them more interviews might be very helpful.

Justification of regarding the clinical diagnosis of ASD (in its mild and moderate forms) to show levels of similarity and differences regarding the signs and symptoms of ASD and the scales which has been used for diagnosis (ADOS, ADI-R or…)

Reporting on any comorbid conditions might also help. Finally, I recommend more information on inclusion and exclusion criteria.

Author Response

Dear reviewer,

Thank you for reviewing our manuscript and for providing very fruitful and relevant comments. We have incorporated the suggestions to the best of our abilities. In the following, we will present how we have made use of each of your reviewer comments:

Reviewer: The study benefited from IPA approach, but the extracted themes are missed in the presentation..(..)..reviewer suggests reporting the themes in the form of a table, possibly with percentage and frequencies.

Author response: We acknowledge that the presentation of the analysis would benefit from reporting more explicit the analytical themes emerging from the IPA. Therefore, we included a figure illustrating the three overall themes from the analysis: Building trust (including framing and introduction, establish rapport), Customizing communication (including accept deviant communication forms and subjects) and Considering questions (open versus structured). Far the most of the studies based on IPA that we have found, present IPA themes in figures, which we found suitable in our manuscript and thus we chose this form of illustration. We also elaborated on the paragraphs describing our analytical steps to meet reviewer’s request of more detailed description of the analysis process. And we revised the Findings subtitles in order to more clearly follow the themes from the analysis. We find that the analysis is much more clarified with these elaborations. (See 2.4. Data Analysis, p.9-10 and Figure 1, p.11. See also 3. Findings p.12ff).

We looked for inspiration from other studies using IPA concerning the possibility of reporting percentages and frequencies in themes in the analysis, but we found that ‘counts’ are rarely reported in IPA analysis. Jonathan A. Smith (one of ‘the founding fathers’ of IPA) gives in his description of IPA ‘Three good examples of IPA’ (Smith and Osborn, 2007, p. 78) and neither of these IPA examples are reporting percentages. See also Zortea et al., in Social Science & Medicine vol. 235, 2019, and Wadman et al., in Social Science & Medicine vol. 212, 2018, both being fresh examples of IPA not using counts (but figures, illustrating themes). For this reason, we believe that the elaborated description of our IPA steps and the thematic overview added in form of a figure is covering the need for description of the analysis carried out, and we find that these elaborations have improved the methodological descriptions.

Reviewer: It is difficult to distinguish between the findings that are applicable to the general population (typically developing children) and children with Autism.

Author response: This is a very relevant comment and we see now that it was not made clear enough how this study is not able to compare with children in general. This study only provides insight into the methodological learnings with this particular sample of children, not with children in general, neither with nor without autism. However, there is a vast literature concerned with methodology with children in general, that we can lean upon to some extent, and we have now elaborated the text concerning this question (in 3.4. Limitations of the study, p.24-25). As we describe in our revised manuscript, and as we conclude by the end of the article, there are certain similarities in our study findings, and the existing literature on childhood studies’ methodology. For example, building trust is important, with children in general as well as with children with autism, as is researcher’s preparedness for deviant conversation forms and topics, whereas other findings are more specific for children with autism, such as the need for thorough interview preparedness and for posing structured and concrete questions.

Reviewer: People with ASD are a very heterogeneous group and when it said that the participants were diagnosed with mild and moderate ASD it needs more information to understand the level of different abilities of them.

Author response: We totally agree on this comment and have now added some more information about the children’s autism conditions, how the children are influenced by their autism, as well as we now report their comorbid conditions (in 2.1. Participants and Recruitment, p.4). We did not include even more detailed information about the children’s scores within different tests from the autism diagnostic process (for instance ADOS, ADI-R, WISC) since we did not collect such detailed information. We consider it to be outside the focus of this article to describe diagnostic test scores and the correlation with methodological issues.

Reviewer: The limitations of this study need to be mentioned, such as: few in the number of interviews, justification of regarding the clinical diagnosis of ASD, reporting on any comorbid conditions, inclusion and exclusion criteria.

Author response: We very much like the idea of including Limitations and have now added this in the end of the Findings section. Since we already in section 2.1. Participants and Recruitment unfolded the children’s comorbid conditions, and what criteria/characteristics the children in this study meet, we have focused in Limitations on other relevant aspects related to reviewer comments: limitations concerning the generalizability of the findings, limitations concerning the methodological approaches tested and limitations concerning possibilities to compare to children in general not having autism. We did not describe limitations concerning number of interviews since we believe within the qualitative research field that 22 participating children are quite many. Smith and Osborne (2007, p. 56) writes for instance.. “IPA studies have been published with samples of one, four, nine, fifteen and more. Recently there has been a trend for some IPA studies to be conducted with a very small number of participants”. We think the added Limitations section has increased transparency and readability of the findings.

Thank you again; we believe the comments and our following revisions have strengthened the manuscript significantly.

Reviewer 2 Report

This is an important and interesting manuscript. However, in general interviews with children risk resulting in an imbalance in power between the interviewer and the child. There is probably an increased risk when you interview children with autism. This imbalance can lead to that the interviewer´s interpretation of the interview will influence the analysis, i.e., you do not get the child´s perspective. In this case, the interviewer had an extensive experience of children with autism, and probably a preunderstanding of their situation. I need information about how this preunderstanding can influence the analysis in a negative way, and how the author did to prevent this influence.

Author Response

Dear reviewer,

Thank you for reviewing our manuscript and for your relevant comments. We have incorporated your comments to the best of our abilities. In the following, we will present how we have made use of your reviewer comments:

Reviewer: Interviews with children risk resulting in an imbalance in power between the interviewer and the child that can lead to the interviewer´s interpretation of the interview will influence the analysis, i.e., you do not get the child´s perspective. In this case, the interviewer had an extensive experience of children with autism, and probably a preunderstanding of their situation. I need information about how this preunderstanding can influence the analysis in a negative way, and how the author did to prevent this influence.

Author response: The comment is very relevant and we have given the imbalanced power relations and researcher role in this context a lot of consideration, prior to, during and after the study conduct. First author who conducted the interviews and conversations with the children in this study does have extensive experience with children with autism (from teaching in institutions), however, very little experience with the specific children in this study, prior to the study conduction. This means that first author (interviewer) is very much aware of how autism might influence a child but also very much aware of that children with autism are just as different from each other as children in general are; autism can influence very differently and two children with autism are never ‘the same’. For these reasons, we believe that the autism experience is not really a negative influence in this context, or being something that formed specific preunderstandings of the children in the study. The children were approached openly and with no prior knowledge about the specific children in the study. This information have now been unfolded in section 2. Methods and Empirical Data (p. 3), describing that ..’experiences with autism pedagogy and autism diagnostic processes means that the children have been met with a professional knowledge of their possible challenges, but at the same time, they have been met with an openness and ignorance of their specific personal living conditions. The absence of in-depth knowledge of the specific children in the study has increased the opportunities for the children to provide their own perspectives on everyday life and schooling with autism’.

Further, we have added in section 2.2. Procedure and Researcher Role (p. 6) a paragraph describing how first author (interviewer) during conversations and interviews sought to test my understanding of the children’s perspectives, by asking them to elaborate, or by presenting my understandings of their perspectives to which they would sometimes agree, sometimes they elaborated or corrected my understanding.

Thank you again; we believe the comments and our following revisions have strengthened the methodological descriptions in the manuscript.

Reviewer 3 Report

Although this kind of study is necessary and interesting, I’m not completely convinced about the qualitative studies. In fact, I think that it would be necessary to include self-reports with good psychometric properties in order to reinforce qualitative results. Moreover, it is completely necessary to include and reinforce study limitations (e.g. absence of control group, study biases, absence of self-reports…). Finally, I would recommend to the authors that include a control group with normative children, which reinforce the value of the study.

Author Response

Dear reviewer,

Thank you for reviewing our manuscript and for providing important comments. We have incorporated your comments as much as possible. In the following, we will present how we have made use of your reviewer comments:

Reviewer: I am not completely convinced about the qualitative studies..(..).. it would be necessary to include self-reports with good psychometric properties in order to reinforce qualitative results..(..)..and to include and reinforce study limitations (e.g. absence of control group, study biases, absence of self-reports…). Recommend to include a control group with normative children.

Author response: We acknowledge that qualitative studies have certain limitations concerning generalizability and due to the absence of using control groups. We believe that this is a condition and limitation inherent and unavoidable in qualitative studies, that focus instead on giving insights into specific individuals’ perspectives upon a given life condition. However, we find that your comment is relevant to meet, in order to address ambiguities inherent in the qualitative approach. Thus, we have in our revised manuscript added a Limitation section describing: limitations concerning the generalizability of the findings, limitations concerning the methodological approaches tested, and limitations concerning possibilities to compare to children in general not having autism. Furthermore, we have in our revised manuscript added more information about the children’s autism conditions, how the children are influenced by their autism, as well as we now report their comorbid conditions (in 2.1. Participants and Recruitment, p.4). We did not include self-reports concerning psychometric tests since it is outside the focus of this article to describe diagnostic test scores and the correlation with methodological issues. We think the Limitations section and elaborated information about the children’s conditions has increased transparency and readability of the findings.

The study only provides insight into the methodological learnings with this particular sample of children, not with children in general, neither with nor without autism. However, there is a vast literature concerned with methodology with children in general, that we can lean upon to some extent, and we have now elaborated the text concerning this question (in 3.4. Limitations of the study, p.24-25). As we describe in our revised manuscript, and as we conclude by the end of the article, there are certain similarities in our study findings, and the existing literature on childhood studies’ methodology. For example, building trust is important with children in general as well as with children with autism, as is researcher’s preparedness for deviant conversation forms and topics, whereas other findings are more specific for children with autism, such as the need for thorough interview preparedness and for posing structured and specific questions.

We believe these elaborations and addition of a Limitation section have strengthened the methodological description of this manuscript. Thank you!

Round 2

Reviewer 1 Report

This is an improved version of the previous manuscript. It is clear that the authors did their best to address all the issues which were raised in the review.

Reviewer 3 Report

I respect the work and time the authors have invested in carrying out their study. However, I maintain my opinion on qualitative studies. It is very important to collect quantitative data, but they must be completed with other more objective ones.