The study proposal cleared the ethical evaluation conducted by the All-India Institute of Speech and Hearing research ethics committee.

The study had a cross-sectional descriptive survey research design.

The convenient purposive sample of 35 Kannada English BAWS, aged 18-40 years (M = 25.31, SD = 2.50), was recruited for the study. Among the 35 BAWS, 33 were males, and two were females. Each participant was chosen from the All-India Institute of Speech and Hearing in Mysore, where they had initially enrolled for stuttering management. The participants were speakers of Kannada (L1), with English acquired as the second language (L2). The inclusion criteria were that the participants had been diagnosed only with stuttering and did not exhibit academic, cognitive, hearing, emotional, or oro-motor deficits or neurological impairments. Further, to ensure the participant’s information, a demographic questionnaire was administered to obtain information about age, gender, language history, duration of stuttering experience, family history, speech therapy, and comorbid conditions. A family history of stuttering was positive for nine participants. Furthermore, none of the participants had a positive treatment history. Refer to Table 1 for demographic details of all participants.

Prior to data collection, informed consent was taken from the participants. Further, they were willing to participate and share their self-perceived experiences and thoughts with the study’s authors.

The responses in the Google form were retrieved in an Excel sheet for further analysis. The “Yes” responses to the question’s first part (a) were considered with a score of “1” and no as “0”. The percentage of Yes or No responses was calculated. Further, the descriptive answers for the second part (b) were analysed using inductive qualitative content analysis described by Graneheim & Lundman (2004). The descriptive responses written by BAWS had an average word count of M = 42.11 (SD = 32.21).

Researchers can get deep insights into the meanings and interpretations hidden within descriptive data by using qualitative content analysis, which offers a systematic approach to data analysis (Hsieh & Shannon, 2005). Inductive qualitative content analysis (Graneheim & Lundman, 2004) is a type of content analysis used to analyse textual data to identify meaningful units, themes, and categories without preconceived theories or predetermined codes. It allows themes and categories to emerge directly from the data itself. Inductive content analysis allows for a flexible and exploratory approach to data analysis, making it particularly useful for understanding the concepts from participant’s perceptions when existing theories are limited.

During the analysis, the first author (VM), who had extensive experience in doing the content analysis, read the descriptive answers keenly, word by word, to understand the content. Further, the contents were repeatedly read to gain a comprehensive meaning related to the study’s aim. The meaning units were generated from the participants’ statements. The generated meaning units were assessed for codes. Once the codes were created, the similarities and differences between the reflections were identified. Further, the related codes were categorised to form sub-themes. Similarly, with the related sub-themes, the major themes were created. The procedure was repeated by another researcher (AJD) with expertise in qualitative content analysis to improve the reliability of the results. AJD also had experience in doing qualitative content analysis (Dsouza, Manivannan, & Maruthy, 2024). Further, the results were compared and discussed among the current study’s authors (VM & SM) and a researcher (AJD) to agree on the theme and sub-themes.

The data created one central theme (Personal growth) and five sub-themes (Communication, coping, conceptualization, social image and reputation, and psychosocial strength). Refer to Figure 1 for the theme and sub-themes created. Figure 2 represents the percentage of respondents in each sub-theme. The use of code-switching for BAWS is meant to help their personal growth. Further, BAWS regarded code-switching as helping to improve communication, social image and reputation, psychological strength, coping skills, and explicating thoughts (conceptualization). Only two adults believed their code-switching behavior was not supportive in the workplace. The examples of codes, sub-themes, and themes created are illustrated in Table 3.

Kannada-English BAWS reported CS as a helpful communication strategy in daily life because they use CS to avoid dysfluencies across situations. The current results support previous qualitative findings (Boyle, 2013b; Boyle et al., 2019; Plexico et al., 2009a), which show that AWS report perceptions of stuttering changing as a result of using a communication strategy, as well as benefits from the experience (such as increased self-assurance, willpower, empathy, and compassion for others). However, these helpful strategies for interaction may prove to be ineffective over time (Boyle et al., 2019; Manivannan et al., 2023; Medina et al., 2019) and lead to social and self-stigma (Boyle, 2013a; Carter, Breen, Yaruss, & Beilby, 2017; Constantino, 2016; Daniels & Gabel, 2004).

In addition, the AWS exhibit varied dysfluencies depending on the person to whom they are speaking, where they are standing, and how intense the situation is (Savithri, 2020). For example, if a BAWS speaks with his or her higher authorities, the dysfluencies may increase more than when speaking to family members. Further, greater dysfluencies may be noticed more in public places than at home. Also, situations like classroom presentations may have higher dysfluencies than telephonic conversation, or vice-versa. This effect becomes less severe when CS occurs. Therefore, the BAWS felt that CS built effective communication by managing the dysfluencies across persons, places, and situations. This finding was consistent with that of Das (2012), who found CS is a strategy to maintain the appropriateness of the context or situations and depends on the communication partner. Furthermore, they indicate that CS is a tool to emphasise social relationships. However, their findings mainly dealt with adults who do not stutter. Sankar et al. (2004) showed that CS behaviors could help reduce the dysfluencies during conversation across different interlocutors in MAWS.

Kannada-English BAWS, in the present study, gave a few examples that indicated their awareness of the CS effect in transforming ideas. The AWS always find difficulty in expressing their thoughts due to the interruption of dysfluencies. Sometimes, they need help to make the partners understand what they want to convey. While introducing CS behaviors, they can ease their struggle to transform information. Das (2012) also found that CS helps reiterate the messages and make the speech clearly understood by the listeners. Pioneer research (Muthuswamy, 1982) also described that CS provides precise meaning, ease of conceptualization, facility of production, and sentence completion when the adult speaks. Furthermore, the demand for CS at the workplace is high in order to make colleagues understand what is being said. Due to avoiding negative workplace experiences whenever dysfluencies occur, these perceptions were well accounted for in a previous qualitative study (Klompas & Ross, 2004).

Many BAWS experiences indeed showed CS as a coping behavior. They felt CS is a speech enhancer, which eliminates stuttering. Furthermore, they believe CS contributes to managing words they struggle to produce. The results were consistent with studies on a few other communication disorders (Ginsberg & Wexler, 2000; Goral et al., 2019; Gutiérrez-Clellen et al., 2009; Miccio, 2009; Ponce-Lawler, 2017), which explains that the bilingual individuals with communication difficulties cope with word-finding problems and production deficits using CS. Concerning stuttering and CS, only one qualitative published research has been available to quote CS as a coping strategy in BAWS (Medina et al., 2019). Furthermore, the perceptions are consistent with few quantitative results (Klugman, 2003; Saad, 2020; Sankar et al., 2004). Therefore, despite variations in cultural, societal, and economic situations as well as in beliefs and confessions, the literature highlights the point that CS being utilized as a coping method by bilingual AWS is relatively universal. Furthermore, a recent study found that coping with dysfluencies was perceived as a positive experience in AWS (Manivannan et al., 2023).

Few studies on stuttering, however, have discussed effective coping mechanisms such as utilising fluency techniques to manage stuttering, enhancing one’s perceptions of oneself and continuing speech treatment recommendations to reduce stuttering (Boyle et al., 2019; Medina et al., 2019; Plexico et al., 2009a, 2009b). Avoidance behaviors like code-switching, on the other hand, are considered to be ineffective coping mechanisms that cause listeners to react and evaluate negatively, worsening the AWS’s self-esteem (Corcoran & Stewart, 1998; Daniels et al., 2006; Klompas & Ross, 2004; Swartz et al., 2014).

Getting a social image was another benefit of CS that BAWS experienced. AWS typically suffer from identity formation because of their dysfluencies (Daniels & Gabel, 2004; Kathard, 2006; Nang et al., 2018); therefore CS was beneficial. According to the current study, CS can help people hide their social anxiety and get acceptance from others. CS supported societal values. This conclusion was in line with earlier research on AWNS (Alenezi, 2010; Das, 2012; Hegde et al., 1997; Muthuswamy, 1982; Niharika et al., 2021); the stuttering literature, however, offers no support for this finding.

A BAWS who works as a software engineer addressed the issue that switching to L1 may impact one’s reputation at work. The participant stated that he always combines his L1 and L2 to improve his speaking ability when he finds it challenging to communicate in L2, reducing his work value. Additionally, the participant desires to move to L1 to prevent dysfluencies brought on by self-restraint due to a fear of losing their value at work. In the present study, the participants are highly proficient in L1 (Kannada) and L2 (English), qualified with at least a degree, and most of them are working, as shown in Table 1. Regardless of the work field, speaking in English (L2) generally provides a reputation at work (Clement, 2018). Even though CS benefits BAWS in many ways, there were times when they wished to switch to avoid their dysfluencies but were unable to do so, such as in the workplace. However, only a single adult talked about it. More evidence from the stuttering literature is also needed to corroborate the results.

Unlike previous stuttering literature reports, the current results identified greater confidence in BAWS due to CS. CS allows BAWS to face public situations without fear, and this result was consistent with the previous study, which explains that self-devised techniques of older AWS increase self-confidence and reduce fear (Bricker-Katz et al., 2009). Moreover, the result of overcoming the bullying and mocking situations due to CS explains that delivering a fluent speech with CS helps avoid negative experiences. The issues of negative attitudes and reactions toward stuttering are commonly perceived by AWS (Alqhazo et al., 2017; Connery et al., 2020; Klompas & Ross, 2004). Current perceptions of BAWS, however, indicated that the benefit of CS is that it helps prevent negative reactions.

Further, CS is a relaxant in BAWS when expressing their struggles with family members. This will help BAWS indeed reduce their negative thoughts and attitudes toward stuttering. The conclusions drawn from this one qualitative study are unsupported by most current stuttering research.

Researchers who studied Kannada-English bilingual stutterers claimed that the frequency and severity of dysfluencies vary across languages due to language proficiency and linguistic structure differences (Kashyap & Maruthy, 2020; Maruthy et al., 2015). However, based on the present study results, the dysfluencies may vary with CS behaviors. For instance, if dysfluencies are avoided in a particular language by switching words from another, the accuracy in calculating the stuttering frequency is affected. Therefore, clinicians need to control the involvement of language-switching behavior to identify the accurate stuttering frequency and severity in the respective languages. It would be possible to ask the BAWS to limit intentional CS further to speak one language at a time (L1 or L2) while collecting samples for stuttering assessment and following them with another language. The clinicians should also observe the CS levels in the word, phrase, or phonological level. If the Kannada-English BAWS uses CS in the initial words or more with content words, they may suspect it is due to dysfluencies. It might be because Kannada-English BAWS report a higher frequency of dysfluencies on initial syllables and content words (Kashyap & Maruthy, 2020; Maruthy et al., 2015). Moreover, not all the code-switching behaviors are avoidance behaviors because even the adults who do not stutter express CS in initial words or content words; however, the frequency would be higher in BAWS (Neeraja, 2004; Sankar et al., 2004). Clinicians should ask the BAWS during stuttering assessments whether they intentionally switch to avoid dysfluencies. Further, it can be monitored during therapy. SLPs must additionally counsel BAWS to limit intentional CS while receiving therapy for stuttering in either L1 or L2, their preferred languages. Furthermore, applying cross-linguistic treatment generalisation from L1 to L2 or L2 to L1 can aid in the BAWS’s ability to speak both languages fluently (Kashyap & Maruthy, 2019).

The current study includes only two female participants due to lesser participant availability. Additionally, the objective evaluation of language proficiency was not included in the current study, which may have a significant impact. Nonetheless the study explored positive effects alone. Additionally, the authors suspect that participants’ positive responses to CS may be prompted by the question “how code-switching is helpful in stuttering”. We thus concur that this would be one of the study’s limitations and other than the triangulation approach, the authors did not use any additional measures to prevent researcher bias. Furthermore, the present study did not interrogate the use and self-perceived experiences of CS in bilingual adults who do not stutter (BAWNS) and how it differs from BAWS.

There is an existing gap in stuttering literature regarding code-mixing and code-switching (Manivannan & Maruthy, 2024). There is a need to compare code switching with other avoidance behaviors and identify its merits and downsides. Future studies should study the difference in the use of CS behaviors between BAWS and BAWNS. Furthermore, qualitative studies need to be done on BAWS to understand their perceptions and feelings towards stuttering. This will make clinicians focus on the stutterer’s needs.