https://orcid.org/0000-0003-4306-420X
Figures
Abstract
Addressing issues in sport rehabilitation with technology-based applications is becoming more common due to improved technological solutions. This drives the need for evidence-based rationales and decisions; technology features should be informed by expert opinion and practice. The current lab protocol paper proposes the use of the framework analysis (FA), a qualitative data analysis method, for identifying practitioners’ perspectives on technology solutions aimed at aiding athlete rehabilitation. The FA allows data to be analysed in a structured and rigorous process, whilst also allowing for the flexibility associated with qualitative enquiry. Subsequently, this method is being increasingly used in healthcare and nursing research and could be applied in the context of using technology to enhance sport rehabilitation. In the current paper, FA is applied in the context of determining the usability of virtual reality (VR) in sport rehabilitation and obtaining perspectives on VR features necessary for integration into existing rehabilitation practice. The paper includes a worked example, taking it from raw data to a working theme. The use of the current lab protocol led to the identification of interconnected key themes regarding the following VR application features: how the VR application would be delivered, what the VR application would involve, and why, where, and when the VR application would be used. The lab protocol also allowed subthemes to be derived, indicating how these VR features would be met. The findings inform the ongoing development of a VR application designed to assess quick directional change in sports, potentially applied to the treatment of musculoskeletal injury in athletes. The use of the FA to derive this content is susceptible to limitations present in all qualitative data processing, such as reflexivity, the implications of levels of rigour, as well as being a time-consuming process.
Citation: Tang HKM, Lake MJ, Bezombes FA (2025) Lab protocol using framework analysis to capture practitioners’ perspectives on the usability of virtual reality for sport rehabilitation. PLoS One 20(12): e0337814. https://doi.org/10.1371/journal.pone.0337814
Editor: Wanli Zang, Soochow University, CANADA
Received: July 3, 2025; Accepted: November 12, 2025; Published: December 30, 2025
Copyright: © 2025 Tang et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All data files are available from the Zenodo database doi: https://doi.org/10.5281/zenodo.17341651.
Funding: This work was supported by Liverpool John Moores University in the form of Doctor of Philosophy funding under Grant [Student ID 854667]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
1 Introduction
Return to sport after injury can come with a variety of associated costs for many stakeholders. These may include psychological, social, financial and physical costs, with athletes facing the greatest impact [1–3]. The variety of technology-based applications developed to address issues in sport rehabilitation is growing due to the increased accessibility of improved technological solutions [4]. This drives the need for evidence-based rationales and decisions; technology features should be informed by expert opinion and practice [5]. Without a clear or valid basis for application or tool development, the technology produced may not meet the needs of the end user. This could render the technology not fit for purpose or could affect the delivery of the application and the integration of the final product into current sport rehabilitation practice [5].
Whilst recent research has begun exploring practitioner and athlete perspectives on virtual reality (VR) in sport rehabilitation, this is limited as it is an emerging field. The review of existing literature demonstrates a considerable body of evidence supporting the effectiveness of VR interventions in sporting rehabilitation. In particular, the studies highlight improved patient recovery, reflected in enhanced functional abilities and increased muscular strength [6]. However, the literature does not explicitly state whether these have been informed by practitioner perspectives and practice. Few studies document input from practitioners to shape application development. In one such example, Gouveia et al. [7] informed the development of their VR system for musculoskeletal sporting injuries by interviewing physiotherapists working in professional football (soccer). A qualitative analysis of these interviews produced a validated list of rehabilitation exercises, later translated into a football-themed VR environment. The system was then tested with 37 university students from diverse fields, who completed a full VR rehabilitation session across five customised games. Data collection included physiological measures (heart rate, perceived exertion) with subjective assessments obtained through a post-session survey.
The survey design (Gouveia et al. [7]) provided a multidimensional view of participants’ subjective experiences with VR rehabilitation. The System Usability Scale (SUS), a widely used 10-item questionnaire, measured usability by asking participants to rate ease of use, clarity, and overall system functionality. To capture psychological engagement, selected items from the Intrinsic Motivation Inventory (IMI) were used, focusing on dimensions such as enjoyment, perceived competence, and effort, which reflect how motivating participants found the VR session. Dedicated questions on immersion (Immersive Tendencies Questionnaire) and presence (Witmer−Singer Presence Questionnaire) assessed the extent to which users felt absorbed in the virtual environment. However, they did not use open-ended questions for participants to provide in-depth reflections on their experiences, enabling them to highlight valued aspects, difficulties encountered, or suggestions for improvement.
In contrast, Lewellen et al. [8] explored VR adoption from the athlete’s perspective. Fourteen athletes with VR experience, most of whom competed at the collegiate level, were recruited to participate in the study. Data was collected through semi-structured interviews conducted via Zoom, which were audio-recorded and transcribed verbatim. Analysis utilised Braun and Clarke’s six-phase reflexive thematic framework, combining inductive coding with deductive elements informed by the Technology Acceptance Model and the “4Ws” framework (where, when, why, what). This process produced nine overarching themes that captured athlete perceptions, including enthusiasm for VR, recognition of barriers such as cybersickness, cost, and coach attitudes, and suggestions for strategies to facilitate broader use.
The first-hand accounts that were captured by Lewellen et al. [8] were analysed through reflexive thematic analysis. Their use of Braun and Clarke’s inductive thematic analysis and the deductive method of the Technology Acceptance Model provided a means of exploratory investigation as well as structure. A combination of standardised scales and open-ended feedback could offer a robust evaluative framework that assesses not only the technical performance of the VR system but also its psychological and experiential impact on users. Building on this work, qualitative and in-depth Framework Analysis (FA) can incorporate existing rehabilitative practices and the perspectives of practitioners. This can be used to highlight both the immersive benefits of VR and the practical obstacles to its adoption.
1.1 Theoretical and conceptual framework
The FA method, originally developed by U.K. social policy researchers as a practical tool for real-world investigations [9], has since been widely applied to specific research questions, particularly in healthcare and nursing, and holds promise for sport rehabilitation. FA is well suited to managing large datasets generated from interviews, focus groups, or case studies, offering a structured yet flexible approach to qualitative enquiry [10,11]. This makes it ideal for use with a purposive sample, where the findings are not necessarily representative of all rehabilitation contexts but provide ecologically valid insights from individuals in the most suitable positions [10].
Compared to other qualitative methods, FA offers distinct advantages that directly address common limitations seen in other qualitative data analysis methods. While thematic analysis (TA) is widely appreciated for its flexibility, its open-ended nature can compromise transparency, as coding criteria may be inconsistently documented, interpretations can vary significantly between researchers, and the analytical path from raw data to conclusions is often difficult to audit or reproduce [12]. Grounded Theory (GT), though valuable for generating new theories, is highly inductive and time-consuming, making it less suitable for studies with predefined questions or practical constraints [13]. Likewise, methods such as Interpretative Phenomenological Analysis (IPA), narrative analysis, and discourse analysis provide rich, nuanced insights into individual experiences and language use, but they tend to lack scalability and are less compatible with structured evaluations or applied research settings [14].
In contrast, FA integrates the benefits of different methods. Data can be described in the participants’ own subjective expressions prior to interpretation and before being summarised systematically and clearly in a matrix, allowing meaning to be preserved while making the dataset manageable [10,15]. The method integrates both deductive and inductive reasoning, enabling researchers to draw on existing knowledge and an overarching question, whilst also generating theories from observed patterns [10]. This dual approach makes FA particularly relevant for VR in rehabilitation, where documented rehabilitation practices can be combined with participant perspectives to inform application design. The process is iterative and recursive, involving continual reference to transcripts to ensure coding accuracy and authentic representation of participants’ views [9,15]. Through this, themes and subthemes can be identified, mapped, and interpreted for practical application [16].
In this study, FA is used to determine the usability of VR and identify the core features required for integration into sport rehabilitation, with findings informing the development of a VR application for assessing rapid directional change and supporting musculoskeletal injury treatment in athletes. By combining methodological rigour with practical relevance, FA emerges as the most appropriate and effective framework.
This paper proposes the use of the qualitative data analysis method FA to analyse interview data of rehabilitation specialists in order to answer the question, ‘What are the core features required for a virtual reality (VR) application to enhance sport rehabilitation?’. The method could therefore determine perceptions on usability of VR and obtain perspectives on VR features necessary for its application into existing sport rehabilitation. This protocol could be applied in any context to capture the expertise of specialists and end users to inform the development of a VR application.
The current paper forms part of a larger study with the following objectives (see Fig 1 for a visual flowchart of the main study’s current stages): 1) observe the use of instrumented equipment employed in a sport rehabilitation context, 2) conduct interviews with one Sport Physiotherapist and three Strength and Conditioners, 3) apply the FA method for processing qualitative data, 4) conduct a focus group to widen the input from practitioners with varied backgrounds including VR-rehabilitation research, 5) use the guidance on the VR application’s implementation and required benefits for RTS to inform its development. The main objective of the current lab protocol was to utilise a worked example, illustrating how the FA process was conducted and the themes reached, specifically pertaining to objectives 3, 4 and 5 of the larger study. The findings have informed the ongoing development of a VR application designed to assess quick directional change in sports. For a review of existing VR applications and how these inform the novel features proposed, as well as a development and pilot of the VR application, see Tang et al. [17].
2 Materials and methods
2.1 Ethics
Prior to commencement, the study received ethical clearance from Liverpool John Moores University Research Ethics Committee; reference number: 20/EEE/001. On commencing the interviews, individuals were provided with participant information forms and they provided written consent; this also ensured a uniform introduction to data collection. The forms provided researchers with information that could potentially identify participants, and the forms were kept in a secure location, with all further data coded to protect participant identification. The codes were stored separately from the forms and data to preserve anonymity.
2.2 Associated content
The protocol described in this peer-reviewed article is published on protocols.io, https://dx.doi.org/dx.doi.org/10.17504/protocols.io.e6nvw8r82vmk/v1 [Accessed: 20/11/2025] and is included for printing as S1 File with this article.
2.3 Duration
This study was conducted from the 18th Feb. 2020–31st Jun. 2022. Communication to request permission to conduct the observations and interviews, and for a discussion on recruitment, occurred via email on 18th Feb. 2020. The interviews were held over one day at St George’s Park [18] on 6th Mar. 2020. Final correspondence with the interview participants occurred via email on 25th Mar. 2021. Auditing of categories with osteopath and biomechanist occurred at Liverpool John Moores University Byrom Campus on 26th Nov. 2021 and 22nd May 2022. The focus group recruitment occurred on 26th May 2022. The focus group occurred on 31st May 2022. Auditing of quotations occurred with English language specialist on 6th Jun. 2022.
3 Sample data, results and discussion
The interview study identified four main themes of what, why, where and when. A summary of the key themes and their relevance to VR usability is depicted in Fig 2.
The sample data below is a worked example of how this protocol was applied in order to arrive at the themes. For further advice on how to interpret and analyse the data, see the step-by-step protocol (S1 File) where it is explained by stage and in full. The example included in this paper follows the progression of one theme: when to implement the VR application. The example demonstrates the process from raw data to theme formation and visualisation.
The data processing was originally conducted in a combination of various formats: hard copy (initial thoughts), NVivo (coding), Excel (framework), Inspiration 10 (mind mapping software), PowerPoint (presentation/diagrams of final themes and sub-themes), and Word (write-up). As explained in the step-by-step protocol (S1 File), the software format selected was based on what was optimal for that specific stage of processing. For example, a hardcopy format increased ease of reading; a primary feature of NVivo was the ability to organise and code data by participant with ease; Excel offered a matrix for tabulating data; Inspiration offered a solution for visualising categories and allowing connections, similarities and differences to be visually presented to aid the formation of themes; PowerPoint allowed for easy manipulation of objects; Word was most suitable for writing-up findings. Subjectively, transferring output from one system to another was deemed to be simple; however, if analysts determine that a particular software programme is not personally advantageous, then they should select those which they consider to be most convenient.
For the example data below, this has been collated in Word for clarity, i.e., one can see the data processing methods taken at each stage and compare the stages in order to understand how the process and output developed. Additionally, the following data is not definitive nor conclusive (dedicated publications will present final output). See Table 1 for definitions as applied in the context of this worked example [10].
Table 2 identifies the stages of data processing. The stages required to produce the data samples can be seen in the step-by-step protocol (S1 File), alongside a detailed explanation of each data processing step. This is also conveyed in the visual flowchart summarising the protocol (Fig 3).
3.3.1 Transcribe, familiarise, import to NVivo and assign codes.
After the data was transcribed and read, important ideas were written on the interview notes and data was organised by interviewee on NVivo.
3.3.2 Assign initial codes.
The following are examples of raw data extracts relevant to the final theme ‘When’. This is not an exhaustive list and is merely for illustrative purposes to demonstrate how the step-by-step process is applied: i.e., many more codes contributed to the eventual formation of this theme.
Raw data extracts (and further surrounding text for context) were highlighted in NVivo by line. Samples of raw data excerpts are seen in Table 3, with corresponding examples of initial assigned codes. ‘Over-coding’ was favoured to allow for a more thorough interpretation and some of these initial codes will have been merged on review. Conversely, more codes may have been added on re-reading the data in context at a later stage. Additionally, these codes will have also contributed to the formation of different categories and other final themes.
Codes were assigned deductively and systematically. Remaining data was coded inductively, open to different perspectives.
3.3.3 Example from secondary coder.
See Table 4 for an example of two (out of twelve) randomly selected data excerpts coded (blind) by the secondary coder with experience in qualitative research. These were discussed with the secondary coder alongside initial category ideas. To provide an indicator of inter-coder reliability, the Miles and Huberman (1994) [19] percentage agreement formula was applied: reliability = number of agreements/ (number of agreements + disagreements). Agreements and disagreements between coders were tallied for each extract. Codes that were similar in concept but phrased differently were considered an agreement between coders; additionally, multiple codes expressing the same concept were classed as one similarity [20]. Applying the formula gave the fractional percent of codes that agree, making it easy to compare to a desired agreement. For the coded extracts, the reliability percentage was 73%. Placing this into context, this is slightly less than the 80% minimum suggested by McAlister et al. [20].
3.3.4 Develop working analytical framework.
After all data was coded by line, the analytical framework was developed. All data that did not fit was coded as ‘other’. Table 5 presents part of the initial analytical framework: the interviewee is in the left-hand column, and the potential category is the heading ‘Return to outdoors’.
This was eventually placed in the existing coding index or treated as emergent codes, expanding the coding index. As a result, the remaining analysis was more data-driven and inductive. For example, ‘When’/‘Timing of application’ was not on the list of interview topics asked of the rehabilitation specialists. However, grouped codes of similar and interrelated ideas were used as initial categories. In this instance, ‘Return to outdoors’ became an initial category. In this way, this potential category emerged from the interview responses.
Once all of the codes were collated into the thematic framework, there was a consistent revision of the original text, with the addition, adaptation, merging and deletion of codes.
In this example, there are four data sets, a small number. However, when there are more data sets, potentially hundreds, this process is essential to reduce information, aid clarity and allow for a coherent summary to be extracted.
3.3.5 Apply the analytical framework.
The codes were indexed by category, i.e., placed in the thematic framework (for demonstrative purposes; see Table 5 as an example that follows from the data set above).
After all codes were indexed, they were discussed and reviewed by two primary researchers until a coherent summary arose. See Table 6 for a summary of interviewees’ data for the category of ‘Time of VR application’. N.B. This was formed with reference to the full data set in an iterative process to ensure that accurate meaning was conveyed.
3.3.6 Chart data into framework matrix.
Data was summarised by category, first broadly (Table 7), and then in a more refined form with quotations (Table 8). These summaries were placed in a further spreadsheet matrix (Microsoft Excel).
3.3.7 Interpret the data.
As characteristics and differences of data were identified to form themes, categories were developed and mapped on Inspiration 10, and a section of this map relating to the theme of ‘When’ can be seen in Fig 4. This process not only highlighted distinct categories, but also the links within and between these developed categories. The purpose of this diagram is to present the mind map in its working state. At this time of processing, the connections between themes and subthemes were still being developed.
3.3.8 Participant checking and auditing.
Dialogue was maintained with interviewees and representative quotations were approved. Ideas were discussed with an auditing biomechanist and an osteopath to review potential themes. This involved drawing on their existing knowledge regarding human movement and rehabilitation to remove experimenter bias and address topics in which the primary researcher would not otherwise have knowledge or experience.
3.3.9 Refine and process to convey results.
When a category was fully established, it was identified as a theme. The final theme of ‘When’ was identified and the following themes and subthemes were drawn on PowerPoint (Fig 5). The solidification of themes is best illustrated by the transition from the diagram in Fig 4, where categories were formed on Inspiration 10, to Fig 5, where the theme is shown in its final state drawn in PowerPoint.
Each diagram that presents a theme identifies a requirement of the VR application and how this will be specifically addressed to benefit the athlete. The sample data provided in this paper illustrates the evolution of the analytical process and demonstrates how the codes were refined into specific ways of meeting requirements. As a result, codes were categorised into themes that subsequently emerged as the overarching requirements.
The themes and subthemes are explained in full in a final dedicated publication. This worked example is a simplified version of one of the simpler themes. The theme of ‘When’ was less hierarchical in nature, with only one primary branch. For themes with more diverse subthemes, significant processing was required when developing the categories in order to better define the individual components.
4 Limitations, future work and conclusions
Gale et al. [10] identified the primary issues associated with the FA. For example, there are reflexivity, rigour and quality issues present in the use of the FA as with other qualitative methods. Additionally, although the tool itself is neither inductive nor deductive, the research may be impacted depending on how inductive or deductive the research question and researcher’s intentions are. The process is also time consuming. Additionally, the analysis should be overseen by an experienced researcher; however, Gale et al. [10] highlighted that this does not preclude those new to qualitative research from participating in the analytical process.
The study involved a limited number of participants, recruited through convenience sampling, resulting in a relatively niche cohort. Participants were selected based on their professional roles and place of employment, specifically as practising rehabilitation specialists working with athletes at a national level. Given their shared context, participants may have held similar perspectives, potentially limiting the diversity of viewpoints represented. This likely resulted in potential biases in practitioner perspectives, in particular, the focus on football in elite athletes, with experience in the treatment of injuries common to football players of this level being overrepresented in the data. Moreover, the advanced equipment available in national-level settings may not be accessible in community or local sporting environments, further constraining the transferability of findings. To enhance generalisability, future research could incorporate the perspectives of rehabilitation specialists operating in other contexts, as well as those of athletes themselves.
The reliability calculated as a percentage was slightly lower than that advised [20]. This may indicate that between the coders, the coding was not ideally consistent. Slight variations may have resulted from differences between coders as opposed to interviewees’ responses; however, the data may have been open to multiple interpretations. As the percentage was still relatively high at 73%, the threat to validity may not undermine the findings. However, the results should be read taking this into consideration.
The protocol could be validated with larger, more diverse samples, for example, those with contrasting perspectives from different practitioner roles, individuals that do not work for the same organisation, and athletes as patients and potential headset users. In combination, physiotherapists can contribute expert input during the design phase, with their interviews systematically transcribed and content-analysed to guide the development of clinically grounded VR systems. The dual practitioner-patient perspective could underscore the promise of VR for rehabilitation while also drawing attention to the methodological diversity and challenges that must be addressed for effective implementation in sport contexts.
Future work could integrate experiential insights from FA with established usability models and standards to enable a more comprehensive evaluation of VR applications. Once end users such as clinicians or athletes engage with the system, Nielsen’s heuristics [21] may be applied as a diagnostic tool to assess adherence to established design principles (e.g., visibility of system status, consistency, error prevention), or a modified ISO 9241−11 [22] could provide a complementary measurement model that defines usability in terms of effectiveness, efficiency, and satisfaction within a specified context of use. In a proof-of-concept study, post-task reflections could be collected using a ‘System Usability Scale’ [23] that adapts these heuristics for VR and captures dimensions such as psychological and emotional fidelity, spatial awareness, interaction quality, presence, and decision-making, such as the example seen in Table 9. Alongside these reflective measures, real-time user experiences could be captured through a ‘think-aloud’ protocol [24], where user-related events can be reported in real time, enabling participants to verbalise thoughts and difficulties during interaction, thereby yielding immediate insights into decision-making processes and emerging usability issues. Integrating quantitative usability metrics into the findings could offer a fuller perspective into application strengths, limitations and development.
Supporting information
S1 File. Step-by-step protocol, also available on protocols.io.
https://doi.org/10.1371/journal.pone.0337814.s001
(PDF)
Acknowledgments
The authors would like to thank the staff at Game Changer, St George’s Park.
The authors would also like to thank the following individuals based at LJMU, Liverpool, UK: Héloïse Debelle (osteopath and biomechanist) and Karl Gibbon (biomechanist), who audited the categories and themes; Bex Walker (biomechanist), who coded randomly selected data excerpts; Paul Wilson (English language expert) who audited the quotations.
References
- 1. Kujala U, Orava S, Parkkari J, Kaprio J, Sarna S. Sports career-related musculoskeletal injuries: long-term health effects on former athletes. Sports Med. 2003;33(12):869–75. pmid:12974655
- 2. Furie K, Park AL, Wong SE. Mental Health and Involuntary Retirement from Sports Post-Musculoskeletal Injury in Adult Athletes: a Systematic Review. Curr Rev Musculoskelet Med. 2023;16(5):211–9. pmid:37014610
- 3. Lutter C, Jacquet C, Verhagen E, Seil R, Tischer T. Does prevention pay off? Economic aspects of sports injury prevention: a systematic review. Br J Sports Med. 2022;56(8):470–6. pmid:34598936
- 4. Seçkin AÇ, Ateş B, Seçkin M. Review on Wearable Technology in Sports: Concepts, Challenges and Opportunities. Appl Sci. 2023;13(18):10399.
- 5. Qi Y, Sajadi SM, Baghaei S, Rezaei R, Li W. Digital technologies in sports: Opportunities, challenges, and strategies for safeguarding athlete wellbeing and competitive integrity in the digital era. Technol Soc. 2024;77:102496.
- 6. Almansour AM. The Effectiveness of Virtual Reality in Rehabilitation of Athletes: A Systematic Review and Meta-Analysis. J Pioneer Med Sci. 2024;13(4):147–54.
- 7. Gouveia ÉR, Campos P, França CS, Rodrigues LM, Martins F, França C, et al. Virtual Reality Gaming in Rehabilitation after Musculoskeletal Injury—User Experience Pilot Study. Appl Sci. 2023;13(4):2523.
- 8. Lewellen JA, Baker E, Giacobbi PR Jr. Athlete perceptions of virtual reality and barriers to its use in sport: A qualitative examination. PLoS One. 2025;20(4):e0320225. pmid:40300013
- 9.
Ritchie J, Lewis J. Qualitative research practice: a guide for social science students and researchers. London: SAGE; 2003.
- 10. Gale NK, Heath G, Cameron E, Rashid S, Redwood S. Using the framework method for the analysis of qualitative data in multi-disciplinary health research. BMC Med Res Methodol. 2013;13:117. pmid:24047204
- 11. Swallow V, Lambert H, Santacroce S, Macfadyen A. Fathers and mothers developing skills in managing children’s long-term medical conditions: how do their qualitative accounts compare?. Child Care Health Dev. 2011;37(4):512–23. pmid:21375567
- 12. Roberts K, Dowell A, Nie J-B. Attempting rigour and replicability in thematic analysis of qualitative research data; a case study of codebook development. BMC Med Res Methodol. 2019;19(1):66. pmid:30922220
- 13. Cutcliffe JR. Methodological issues in grounded theory. J Adv Nurs. 2000;31(6):1476–84. pmid:10849161
- 14. Pringle J, Drummond J, McLafferty E, Hendry C. Interpretative phenomenological analysis: a discussion and critique. Nurse Res. 2011;18(3):20–4. pmid:21560922
- 15. Smith J, Firth J. Qualitative data analysis: the framework approach. Nurse Res. 2011;18(2):52–62. pmid:21319484
- 16. Braun V, Clarke V. Using thematic analysis in psychology. Qual Res Psychol. 2006;3(2):77–101.
- 17. Tang HKM, Lake MJ, Foster RJ, Bezombes FA. Design, development and pilot of a realistic virtual reality application to analyse quick directional change in sport: Avatar cutting scenario with alterable parameters. PLoS One. 2025;20(6):e0324941. pmid:40554489
- 18. Association TF. St George’s Park. 2025 [cited 2025 18/06]. Available from: https://www.englandfootball.com/england/St-Georges-Park
- 19.
Miles MB, Huberman AM. Qualitative data analysis: an expanded sourcebook. Sage; 1994.
- 20.
McAlister AM, Lee DM, Ehlert KM, Kajfez RL, Faber CJ, Kennedy MS, editors. Qualitative coding: An approach to assess inter-rater reliability. 2017 ASEE Annual Conference Exposition; 2017.
- 21. Nielsen J. How to conduct a heuristic evaluation. Nielsen Norman Group. 1995;1(1):8.
- 22.
Bevan N, Carter J, Harker S, editors. ISO 9241-11 revised: What have we learnt about usability since 1998? International conference on human-computer interaction. Springer; 2015.
- 23. Brooke J. SUS-a quick and dirty usability scale. Usabil Eval Indust. 1996;189(194):4–7.
- 24. Bernardini S. Think-aloud protocols in translation research. Target. 2001;13(2):241–63.