Tele-Assessment of Oral Personal Narratives in Arabic- and Hebrew-Speaking Children Using the Global TALES Protocol

In recent years, especially due to the COVID-19 pandemic, many professions across healthcare, rehabilitation, and education sciences, including speech and language pathologists (SLPs), have been obligated to make a quick transition to tele-assessment (TA) and tele-treatment to meet service provisions so as to prevent the spread of the COVID-19 virus (e.g., [1, 2]). However, evidence regarding the validity of using TA and tele-practice for language assessment is limited [3]. This study aimed to investigate the feasibility of TA for gathering personal oral narratives, a complex language skill that encompasses a wide range of language, cognitive, and social abilities [4]. The study also explored potential differences in personal narrative skills among Arab and Jewish children, two distinct language/culture groups residing within the same country. Additionally, the study design enables to investigate whether there were interaction effects between languages/cultures (Arabic/Jewish) and the assessment mode (TA or face-to-face [FTF]). The research team consisted of two SLPs, one Jewish and one Arabic, both residing in Israel and experienced in both cultures. Furthermore, the Arab researcher is proficient in both Arabic and Hebrew languages.

Tele-practice or online treatment in speech-language pathology has been established for at least 25 years [5, 6]. The COVID-19 pandemic, however, has further necessitated its implementation in educational institutions and has required allied health professionals to rely on tele-services for many activities that were previously FTF [7, 8]. Tele-services, including TA, offer several advantages, such as increased accessibility, reduced waiting lists, reduced time and travel costs, and assistance with infection control by minimizing FTF contact during a pandemic [9]. However, challenges in communication, rapport and engagement, emotion identification, visual and physical cues, maintaining children’s attention, and technological issues have been identified (e.g., [10, 11]). Additionally, families from low-resource communities may have limited access to telecommunication technologies, such as videoconferencing and/or computerized software programs [12], and in some rural or remote communities, unreliable electricity supply and low bandwidth may impede the successful implementation of tele-practice [13]. Despite these challenges, the feasibility and validity of using tele-practice to conduct language assessments are still not fully established.

Most standardized language assessment tools have been developed and normed for FTF administration, raising the question of whether there is agreement between FTF and TA scores in children. However, studies comparing FTF and TA scores on children’s language performance have generally found no significant differences between the two modes, whether in receptive or expressive language abilities on standardized norm-referenced language assessments [3, 14‒18]. Furthermore, Ciccia et al. [14] examined the feasibility of speech, language, and hearing screening for TA in children up to 6 years of age via video chat and found similar scores to FTF assessments.

The feasibility of the TA of children’s spontaneous language skills has also been investigated. Recently, Manning et al. [2] examined the feasibility, validity, and reliability of using video chats with the families’ own devices to conduct language sampling for children. The study reported no significant differences across language variables measured in person in the laboratory or via video chat at home. Similar results were found by Brennan et al. [19], who compared language assessment between FTF and videoconferencing TA conditions using a story retelling procedure in 40 adults with brain injury. In a different study, Burchell et al. [1] examined the validity and reliability of TA of conversational, fictional narrative, and expository discourse in school-aged children. Although there were no modality effects in the narrative context, the children produced longer conversation samples during the FTF assessment. The authors attributed this to disruptions in conversational flow caused by varying internet connection quality, which inevitably affected the duration of tele-practice sessions.

Several studies have shown that TA can be both feasible and valid for atypical populations. For example, Nelson and Plante [3] compared the administration of the Test of Integrated Language and Literacy Skills (TILLS) via TA and FTF in children with language/literacy disorders, and Brennan et al. [19] investigated story retelling in adults with brain injury. Although Nelson and Plante reported that the nonword repetition subtest scores could be slightly depressed in the TA condition, both studies reported equivalent diagnostic results for both modes. However, Conner et al. [20] observed disengagement in elementary children with autism spectrum disorder during the TA of language, reading, and writing abilities, which may indirectly affect children’s performance and hence, assessment outcomes. To the best of our knowledge, there has been no previous investigation into the impact of assessment modality on children’s personal narrative skills. This aspect is particularly important as it involves maintaining the listener’s attention and making inferences based on the narrator’s own knowledge of the world [21].

Narrative is a discourse genre that involves organizing a series of events with a temporal sequence or causal connection [22], often motivated by agents [23]. The production of a coherent narrative demonstrates the cognitive ability to sustain talk about the world beyond here and now [24]. Producing narratives requires reconstructing a sequence of events with one or more characters and utilizing knowledge of content and linguistic structure. Content pertains to knowledge of the events, while structure refers to organizing and formulating knowledge [4, 25, 26]. Narratives can summarize past personal experiences through a sequence of clauses and other linguistic units that match the sequence of events, which often include a complication requiring resolution [27]. Narratives can be presented in various forms, such as oral, written, filmed, or via illustrated media, and can be produced in the form of story retelling or story construction.

Narratives can be fictional [4, 28] or personal. Fictional narratives are stories of fictional events [29, 30]. The current study focused on personal narratives, which are accounts of personally experienced events [30]. Personal narratives assist individuals in processing and understanding their experiences, incorporating subjective thoughts and emotions about those actions and outcomes, to create meaning from these experiences [29]. Personal narrative skill develops with age [31, 32], with the accepted notion that the parental reminiscing style plays an important role in fostering children’s narrative development and autobiographical memory [29]. Personal narratives comprise more than half of children’s conversations, emphasizing their importance in daily communication. Humans use narratives to describe events, establish and maintain relationships, and express their feelings and thoughts on different topics [33]. The ability to share personal narratives helps build and maintain friendships, promotes physical and emotional well-being, supports classroom participation, and underpins academic success and vocational outcomes [34].

When analyzing children’s personal narrative skills, the analysis typically focuses on micro- and/or macro-levels of measurement [35]. Macrostructural complexity may be demonstrated by story grammar components and episodic buildup [35]. Specifically, macro-level analysis using story grammar analysis assesses the narrator’s ability to organize narratives into story grammar elements [36]. In contrast, microlevel measures focus on verbal productivity, assessed by the total number of words (TNW) and total number of utterances (TNU). Lexical skills are often measured by the number of different words (NDW) in a text [37, 38]. Syntactic complexity is often assessed by measures such as the mean length of utterance in words or morphemes (MLU), percentage of complex sentences, and/or morphological and syntactic error analysis [37, 38].

Analysis of personal narratives may also focus on the topic and/or theme, which reflects the meaning-making aspect of the narrative [32, 39]. The topic of a story refers to an explicit message that answers questions of who, when, and why [40]. Although there is a scarcity of research investigating topics in children’s personal narratives [32, 40, 41], it was reported that preschool-age children can maintain a topic in a story [32].

Both fictional and personal narratives may vary across languages and across cultural communities in three main areas: (1) narrative content, which includes ideas and topics [32, 40, 41]; (2) structural organization, flow, and macro-level patterns of stories; and (3) functions that focus on intentions, goals, and purposes [4, 42]. Differences may exist in story length, sequence of actions, amount of narrated descriptive information, and interpersonal relationships of characters [4, 43].

Children’s cultural backgrounds may also influence their personal narratives. This is reflected in the fact that children tend to produce personal narratives that reflect their cultural and sociocultural styles and norms [44]. This may further affect the topic of their personal narratives, event sequencing, inclusion of extensive background information (such as family connections), and narrative microstructures [44, 45]. Parents’ narrative styles, values, and beliefs may also affect their children’s personal narratives. Some parents may have shorter conversations about each event and provide less narrative structure, while others invite their children’s input more, and some appear to expect short factual reports as opposed to elaborate narratives [44]. There may be cultural variations in how mothers support the narration of their children [46, 47]. For example, it was shown that Korean mothers were less likely to encourage their children to introduce their own topics or contribute information, whereas Canadian mothers were more likely to encourage narrative co-creation [46]. Another study showed that Japanese mothers provided fewer evaluative comments in response to their children’s narratives and requested less detail than European-American mothers [47].

Israel has two distinct ethnic-cultural societies: the Arabs as a minority and Jews as the majority. Arabs are a permanent, non-assimilating minority, clearly distinguished from Jews in place of residence and cultural terms, speaking their own language and adhering to their own traditions [48]. Jewish culture in Israel is usually considered an individualistic culture often related to Western Europe and North America. Arab culture, however, is usually considered a traditional collectivist culture, emphasizing ideals such as solidarity, cooperation, commitment, mutual trust, support, and a sense of belonging within the family and community [49]. Disparities exist in terms of education and income levels between the Arab and the Jewish societies in Israel, with the Jewish society generally having an advantage [50, 51]. As of 2021, the statistics revealed that 36.3% of the Jewish population and 16.2% of the Arab population held an academic degree. Additionally, 8.2% of the Jewish population and 3.7% of the Arab population possessed a high school diploma, while 41.4% of the Jewish population and 38.9% of the Arab population had completed high school. On the other hand, 14.1% of the Jewish population and 41.2% of the Arab population did not graduate high school. The data from that year also indicate that the risk of poverty among the Jewish population in Israel stood at 18.2%, whereas the risk of poverty among the Arab population in Israel was notably higher at 53.4% [52]. Moreover, in Israel, there are two separate educational systems that coexist alongside each other, catering to the Jewish and Arab populations, respectively. However, these systems are characterized by unequal allocation of resources, associated with disparities in outcomes [53]. The Arab education system faces significant shortcomings due to factors such as institutional discrimination in budgeting, limited opportunities for academic training, and minimal integration of Arab teachers into the Hebrew-speaking education system. Furthermore, reports after the COVID-19 pandemic indicated limitations in home and school communication technology infrastructures, as well as lower digital skills among teachers, parents, and students in the Arab sector [51, 54].

Side by side with Hebrew as the majority language, the Arabic language in Israel constitutes a mother tongue for the large Arabic minority [55], spoken by about one-fifth of the citizens (Central Bureau of Statistics; https://www.cbs.gov.il). Arabic and Hebrew are both Semitic languages, characterized by rich morphology, sharing structural similarities in phonology, morphology, and syntax [56]. However, the two languages differ in many aspects, such as their orthographic systems. Moreover, Arabic is known as a diglossic language that distinguishes between modern standard Arabic (MSA) and colloquial Arabic [57]. The spoken or colloquial variety is acquired naturally by native speakers as a “mother tongue” and is used as the primary language for everyday speech by everyone, everywhere for all informal functions among the Arab community [56]; MSA refers to the variety associated with literacy and education [58] and is acquired mainly through formal education and is used for conventional writing and formal speech, as in religious sermons and media [57]. There are differences and similarities between MSA and spoken varieties of Arabic in all language domains. Research has shown that the distance between the two varieties present a challenge for Arabic-speaking children in acquiring literacy and basic reading processes in Arabic [58]. Studies of narrative production in Arabic speakers [59‒63] have mainly focused on diglossia. In general, these studies suggest that Arabic narratives are influenced by diglossia, since children often choose to tell narratives that they read or were read to them at school from a written, literary Arabic medium. However, they chose to narrate their stories in spoken Arabic.

In a study conducted by Westerveld et al. [34], an initial comparison was made between Arabic- and Hebrew-speaking children regarding their personal narrative production. The study involved 249 ten-year-old children from ten different countries who spoke eight different languages. The findings showed that Arabic-speaking participants produced the lowest mean number of utterances in their narratives compared to other languages, including Hebrew. Westerveld et al. [34] also investigated the topics produced by the participants. It was found that Arabic- and Hebrew-speaking children produced almost the same topics for most prompts, with only minor differences between the two languages/cultures. A study by Kawar et al. [61] that investigated the personal narratives of Arabic-speaking adolescents also showed that these Arabic-speaking participants produced short personal narratives which mostly ended on high ending (i.e., the main point of the story), while excluding many details, personal beliefs, thoughts, and feelings [61]. Taken together, those findings suggest that there may be some differences in personal narrative production between Arabic- and Hebrew-speaking children, particularly in terms of the length of their narratives.

The growing use of TA in speech pathology practice highlights the need for empirical data to determine whether language assessments traditionally administered FTF are comparable in results to those administered through tele-practice. Personal narratives represent one of the most spontaneous and early-developing forms of discourse [30]; however, there is insufficient evidence supporting the validity of protocols assessing personal narratives through TA. Moreover, there is limited evidence regarding the effect of language/culture on narrative ability [4, 42, 43], particularly involving Arabic- and Hebrew-speaking children living in Israel.

The aims of the current study were twofold: (1) to investigate the feasibility, of TA (via Zoom) of oral personal narratives produced by typically developing school-age children living in Israel, and (2) to compare the personal narrative skills of Arabic- and Hebrew-speaking school-age children on measures of verbal productivity, semantic diversity, and mean length of utterances, and on the topics of children’s personal narratives. This study addressed the following research questions:

• 1.

  Does the assessment mode (TA vs. FTF) impact the microstructure level of personal narratives of school-age children living in Israel, and, if so, can performance variations be attributed to a language/culture effect?

• 2.

  Does the assessment mode (TA vs. FTF) impact the chosen topics of children’s personal narratives, and, if so, can performance variations be attributed to a language/culture effect?

Based on previous literature indicating the validity of TA of language skills [1, 3], this study predicted a similar performance between TA and FTF assessment, regardless of the children’s language/culture. In addition, based on a previous study comparing Arabic- and Hebrew-speaking children in Israel [34], it was predicted that the Hebrew-speaking children would produce longer personal narratives than the Arabic-speaking children, regardless of the assessment mode. Finally, it was predicted that both assessment mode and language/culture would not impact the chosen topic narrated by the participants [34, 40].

A total of 89 typically developing 10-year-old (9:6–10:11) children were recruited through the researchers’ networks and through local schools and community leaders. Some of the participants (n = 39) who were assessed for FTF (20 Hebrew-speaking children and 19 Arabic-speaking children) in the current study were the same children who participated in the initial feasibility study [34].

To establish the developmental history of the children, parents were asked to complete the brief demographic questionnaire that was developed for the Global TALES protocol (see https://osf.io/ztqg6/ for a complete copy of the project protocol). The inclusion criteria for participants were as follows: (a) no history of speech and language difficulties; (b) currently not receiving specialist services, such as speech-language therapy; (c) attending schools located in a middle- to high-income area as determined by the Ministry of Education – we adhered to this principle to prevent possible confounding of SES with cultural/linguistic group; (d) being residents of Israel; and (e) being native Arabic speakers belonging to Arab society or native Hebrew speakers belonging to Jewish society. Fifty-one participants were native Jewish Hebrew speakers, and 38 were Arab native speakers of Arabic. Forty-six children were female, and 43 were male. Forty participants were assigned to the TA group (narrated via Zoom), and 49 to the FTF assessment group (Table 1). Assignment to the assessment modality conditions (TA or FTF) could not be made randomly because 39 of the participants who were assessed FTF (20 Hebrew-speaking children and 19 Arabic-speaking children) in the current study were the same children who participated in the initial feasibility study [34]. There was no significant difference in age between the two assessment mode groups, as indicated by the t test for independence (t(87) = 1.419, p = 0.159). Additionally, there was no significant difference in gender between the assessment mode groups, as calculated by the χ² test for independence (χ²(1) = 0.320, p = 0.572).

All parents provided written consent for their children to participate in the study, and the children provided verbal assent at the start of the session. Forty-nine children were seen FTF in a quiet location, either at the child’s school or in the child’s home, and 40 were tele-assessed via Zoom while they were at their homes in a quiet room. The assessments at home, FTF, and via Zoom were conducted in the presence of one of the parents.

To elicit and collect the samples, the Global TALES protocol (see https://osf.io/ztqg6/ for a complete copy of the project protocol) – Arabic and Hebrew translations: version 1 was administered by the authors, who are SLPs, and by research assistants who were fourth-year students in the department of Communication Disorders at Tel-Aviv University in Israel. In this protocol, children first received an explanation of the task. The children were then asked to “tell a story” in response to each of the following six prompts: (1) Tell me a story about a time when you felt excited or really happy; (2) Tell me a story about a time when you felt worried or confused; (3) Tell me a story about a time when you were really annoyed or angry; (4) Tell me a story about a time when you felt proud of yourself; (5) Tell me a story about a time when you had a problem and you had to fix it; and (6) Tell me a story about something that has happened to you that was very important to you. For each topic, the examiner used a scripted follow-up prompt if the child did not respond. If the child provided only one or two sentences, the following generic encouragements were allowed: “Can you tell me more? Can you explain what you mean by that? Is there anything else you can tell me?” Finally, to encourage the child to continue talking, the examiner used additional neutral encouragements, such as “uhuh.” All protocol prompts were asked (read aloud by the examiner) in a set order and presented simultaneously in print on laminated cards in FTF assessments and on a computer in TA.

All sessions were audio recorded for transcription and analysis purposes. All samples were transcribed in their native language by the authors using standard Systematic Analysis of Language Transcript (SALT) conventions [64]. All reformulations, repetitions, false starts, and filler words (e.g., uhm) were placed in brackets and not included in the analysis. Utterance segmentation was based on communication units, which were defined as independent clauses with modifiers. Only complete and intelligible utterances were counted. Unfinished and interrupted utterances were not included, but elliptical responses were counted as separate utterances. The following productivity measures were calculated manually:

• TNW in response to each protocol prompt and all protocol prompts combined.

• TNU in response to each protocol prompt and all protocol prompts combined.

• Total NDW for each protocol prompt and for all protocol prompts combined.

• MLU-W, calculated as TNW divided by TNU.

• Topic of the story. After reviewing the stories and according to previous Global TALES analyses [34], the topic of each narrative was categorized into one of 12 topics. To reduce the number of topics (for the statistical analysis), we combined several similar topics formulated in the original categorization [34] into one topic. The following are the topics we addressed, with the original topics in brackets: personal achievements (personal achievements and/or achievement involving others); personal growth or contribution; school issue (school task, problem at school, or expectations of school); personal relationships (sibling, peer, family or friends relationships, and expectations of others); finding, constructing, or fixing something; family event or support, new experience, new challenge, or new item; safety, wellness, illness, injury, or death; losing someone or something; personal frustration; parental issue; cultural (see [34] for more details).

One of the researchers, whose mother tongue is Arabic (K.K.), analyzed all the Arabic narratives, both those examined through FTF and those examined through TA. The second researcher, whose mother tongue is Hebrew (S.F.), analyzed all Hebrew narratives. The Arabic-speaking researcher was also a fluent Hebrew speaker. This allowed us to examine inter-rater agreement by comparing analysis for five of the Hebrew narratives. The intraclass correlation coefficient was calculated to measure the inter-rater reliability regarding transcription and utterance segmentation, as well as each of the microstructure measures. A significantly high degree of reliability was found between the judges for transcription (0.898, 95% CI: 0.874–0.913), utterance segmentation (0.818, 95% CI: 0.676–0.876), TNW (0.997, 95% CI: 0.995–0.999), TNU (0.995, 95% CI: 0.978–0.995), MLU-W (0.821, 95% CI: 0.657–0.911), and NDW (0.802, 95% CI: 0.624–0.900). To examine agreement between judges regarding the topic measure, a χ² test for independence between judges was performed, which revealed that the distribution of topics was similar for both judges (χ²(9) = 2.813, p = 0.971), indicating agreement.

It should be noted that data for five of the Hebrew narratives were missing. This occurred because five Hebrew-speaking children declined responding to a different prompt. Three prompts were declined responding during FTF assessment, and two prompts were declined during TA. To address this issue of missing data, in the initial analyses, we filled in the missing data with values calculated as an average of the child’s group on the specific measure.

As a preliminary consideration for examining whether all prompts should be treated as one variable, we tested whether there were differences in performance between the six prompts. For this goal, a three-way mixed ANOVA was run for each measure (TNW, TNU, MLU-W, NDW) with the six prompts (1, 2, 3, 4, 5, 6) as the within-subject variables, and with mode (TA vs. FTF) and language (Arabic vs. Hebrew), as the between-subject variables. For TNW and NDW, there was no main effect; therefore, no post hoc analysis was performed. For TNU, there was a significant overall effect, with post hoc analysis showing a significant difference only between the first (excited) and sixth (important) prompts. For MLU-W, there was an overall significant effect, but post hoc analyses showed no significant differences between any of the prompts. Based on these results, statistical analysis from here on out were performed using data from all six prompts as one variable. Table 2 delineates the descriptive statistics for each of the microstructural measures.

Our first research question asked: (a) whether the assessment mode (TA vs. FTF) impacts the personal narrative performance of school-age children living in Israel on measures of productivity (TNW, TNU), semantic diversity (NDW), and MLU-W; and (b) whether, if so, there is an effect on performance that could be attributed to language/culture. Analyses were conducted on children’s performance on these four microstructure measures across all six narratives, and all four measures were found to be normally distributed (Shapiro-Wilks’ test, p > 0.05). Given that some of the correlations exceeded 0.80, a series of two-way ANOVAs were performed with the two independent variables of mode (TA vs. FTF) and language/cultural groups (Arabic vs. Hebrew), and four dependent variables: TNU, TNW, NDW, and MLU-W. In line with our research question, we will report the main effects first, before reporting the interaction effect.

Results from the two-way ANOVAs revealed no significant effect of mode on any of these measures, irrespective of language/culture groups: TNW [F(1, 85) = 0.338, p = 0.563, η² = 0.004], TNU [F(1, 85) = 0.003, p = 0.954, η² = 0.0001], MLU [F(1, 85) = 0.065, p = 0.800, η² = 0.001], NDW [F(1, 85) = 0.001, p = 0.991, η² = 0.0001]. However, the two-way ANOVAs revealed a significant effect for language/culture, with significantly higher mean performance in the Hebrew than in Arabic narratives on all microstructure measures: TNW [F(1, 85) = 27.777, p < 0.001, η² = 0.246], TNU [F(1, 85) = 9.295, p = 0.003, η² = 0.099], MLU [F(1, 85) = 43.713, p < 0.001, η² = 0.340], NDW [F(1, 85) = 11.604, p < 0.001, η² = 0.120]. However, the two-way ANOVAs also revealed a statistically significant interaction effect between mode and group for TNU [F(1, 85) = 5.646, p = 0.020, partial η² = 0.062], and for MLU-W [F(1, 85) = 5.450, p = 0.022, partial η² = 0.060], but not for TNW [F(1, 85) = 2.621, p = 0.109, partial η² = 0.030], or for NDW [F(1, 85) = 0.126, p = 0.724, partial η² = 0.001] (shown in Fig. 1).

To help understand the interaction effects, a simple main effects analysis was run for TNU and MLU-W only. For TNU, there was a statistically significant difference between TA and FTF for Arabic narratives [F(1, 38) = 7.134, p = 0.011, partial η² = 0.165], but not for Hebrew narratives [F(1, 49) = 2.109, p = 0.153, partial η² = 0.041]. For MLU-W, no statistically significant difference was found between TA and FTF for Hebrew narratives [F(1, 49) = 2.210, p = 0.144, partial η² = 0.043] or for Arabic narratives [F(1, 38) = 3.576, p = 0.067, partial η² = 0.090]. Therefore, simple comparisons were run for the differences in the mean scores between the TA and FTF for the Arabic group for TNU only. For Arabic narratives, the means for FTF were significantly higher than for TA (p = 0.012) (Table 2).

To test the effect of both mode and language/culture on the chosen topic, for each topic we summed up the number of times it was used by each child across prompts. Figure 2 delineates the percentage of narratives across prompts cataloged for each of the 13 topics in each mode separately: via FTF compared to TA irrespective of language.

All topics appeared in both FTF and TA modes. Due to too many cells containing less than five observations in χ² test for independence between mode and topic, which we tried to perform, we opted to present the results descriptively. Although not tested statistically, the FTF mode elicited a higher percentage of narratives descriptively than the TA mode on the following topics: personal achievements (FTF: 29.3%, TA: 24.7%); personal relationships (FTF: 33%, TA: 28.6%); finding, constructing, or fixing items (FTF: 2.5%, TA: 0.9%); new experience challenges or items (FTF: 6.5%, TA: 5.3%); parental issues (FTF: 2.2%, TA: 0.4%); and culture (FTF: 1.1%, TA: 0.4%). However, TA elicited a higher percentage of narratives than FTF assessments on the following topics: personal growth and personal contribution (FTF: 5.8%, TA: 6.2%); family event or family support (FTF: 5.1%, TA: 9.7%); illness, injury, safety, wellness, or death (FTF: 9.4%, TA: 10.1%); school issue (FTF: 5.1%, TA: 9.7%); losing someone or something (FTF: 1.1%, TA: 5.3%); and personal frustration (FTF: 0.4%, TA: 1.8%) (shown in Fig. 2; Table 3). In Arabic, a higher percentage of narratives were narrated via FTF compared to TA on the following topics: personal relationships (42.5% vs. 36.8%, respectively) and illness/injury/safety/wellness/death (11.5% vs. 5.3%, respectively). However, a higher percentage of narratives were narrated via TA compared to FTF on the following topics: school issue (6.1% vs. 1.8%, respectively) and losing someone/something (8.8% vs. 0.0%, respectively) (Table 3). In Hebrew, a higher percentage of narratives were narrated via FTF compared to TA on the following topics: personal achievement (25.8% vs. 15.9%, respectively) and personal relationships (26.4% vs. 20.4%, respectively). However, a higher percentage of narratives were narrated via TA compared to FTF on the following topics: personal growth/contribution (11.5% vs. 9.8%, respectively), family event/support (12.4% vs. 4.3%, respectively), school issue (7.1% vs. 4.9%, respectively), illness/injury/safety/wellness/death (15% vs. 8%, respectively), and personal frustration (3.5% versus 0.6%, respectively).

The acceleration in the use of TA due to the outbreak of the COVID-19 has raised the need to test the validity and feasibility of traditional FTF language assessment tools in a TA context. The current study investigated the feasibility of tele-administration of the Global TALES protocol by exploring the impact of assessment mode (TA vs. FTF), language-culture (Arabic vs. Hebrew), and the interaction between them on personal narrative production microstructure measures and topic codes. The Global TALES protocol was administered to elicit six personal narratives from each of the 89 typical-language 10-year-old children (either Arabic- or Hebrew-speaking) assigned nonrandomly (based on time of testing) to either the TA or FTF groups. The results showed that there was no significant main effect of assessment mode on any of the microstructure measures (both groups combined). However, there was a significant interaction effect, with post hoc analyses indicating that the Arabic-speaking children produced more utterances (TNU) in their personal narratives during FTF assessments compared to TA. All other results were not significant. There was also an influence of assessment mode on children’s topic choices, with FTF assessment eliciting a higher percentage of personal narratives on personal issues compared to TA, although we did not confirm this statistically.

The main finding of the present study is that TA did not affect the productivity of personal narratives narrated by Arabic-speaking and Hebrew-speaking 10-year-old children using the Global TALES protocol. Children in both cultural-linguistic groups produced similar numbers of words under both TA and FTF conditions. This outcome is in line with the findings of previous studies reporting no significant differences between the two modes on standardized norm-referenced language assessments [3, 14‒18]. When TNU was used as the metric rather than TNW, however, TA affected the number of utterances produced by 10-year-old Arabic-speaking children only – they used more utterances in FTF than through TA. In considering possible explanations for this finding, we hypothesize that it might be attributed to cultural characteristics of the Arabic-speaking children – that is, the collectivist nature of their culture. In collectivist cultures, as exemplified by Arab society, family and social connections are highly valued, and cooperation and collaboration are emphasized to maintain group harmony; hence, FTF interactions may be more essential for Arabic-speaking children, and perhaps, more highly practiced by them. Furthermore, social gatherings of entire families or clans are more prevalent in Arab society than in Jewish society, emphasizing the collective rather than the individual. FTF interactions enable nonverbal communication by providing cues such as eye expressions, body movements, and gestures, that cannot be achieved through telecommunication. This may explain why individuals from Arab culture may feel less comfortable in telecommunication, which may negatively affect the TNU they produce in their personal narratives.

The finding that Arabic-speaking children produced shorter (TNU) personal narratives in the tele-health condition compared to FTF could also be ascribed to limitations in home and school telecommunication technology infrastructures and limited support for using telecommunication in the Arab society in Israel [51, 54]. It appears that Arabic-speaking children may have fewer opportunities to develop telecommunication skills, which may affect the length of narratives they produce remotely. This view may correspond to Burchell et al.’s [1] suggestion that school-aged children produce longer samples in conversation during FTF compared to TA due to disruptions in the conversational flow caused by the quality of the internet connection at times. However, it is notable that, in the current study, we did not collect any data on the accessibility of participants to telecommunication technologies or the support provided to use telecommunication.

Interestingly, although assessment mode did not significantly affect the mean length of utterances in words of the Arabic-speaking children, this reached borderline significance (p = 0.067), with a medium effect size. This indicates that the Arabic-speaking children used longer sentences in the TA mode than FTF, indicating that although the Arabic-speaking children produced shorter stories in TA, they included a similar number of words through increasing their MLU-W. It is noteworthy however to mention that MLU in words does not always provide as much information about the complexity of the child’s language as does MLU in morphemes [65, 66], especially in morpheme-rich languages like Hebrew and Arabic [56].

An interesting finding was that mode seemed to affect the topics 10-year-old children chose to narrate. Although all topics appeared in both FTF and TA conditions, simple percentage comparisons showed that some were more frequent in one mode than in the other. For example, FTF assessment elicited a higher percentage of narratives compared to TA on personal issues, such as personal relationships, personal achievements, and new experiences, challenges, or items. These narratives dealt with personal dilemmas, included emotional responses, and placed the narrator at the center. On the other hand, TA elicited a higher percentage of narratives on relatively neutral issues such as family events or support, and school issues. In these topics, the narrators were less likely to talk about themselves, rather they talked about other people, and the narratives were less emotional. This preference for personal issues in the FTF condition versus the preference for more neutral, or other-focused topics in telecommunication may be due to differences in the quality of communication. If, compared to FTF mode, in TA, there is a reduction in the quality of communication due to difficulties in making and maintaining eye contact as well as in perceiving facial expressions, visual cues of body movements, and gestures, children may be less likely to talk about personal issues. In other words, TA may meet the basic conditions for good enough communication [10, 11], yet be unable to create an intimate atmosphere that may encourage to share personal issues and emotions, comparable to FTF. Further research is clearly needed to test these assumptions.

Inspection of the data in Table 3 yielded no clear evidence that language/culture influenced the topics elicited differently via FTF compared to the topics elicited through TA. Although the number of topics that differed between the modes (TA vs. FTF) was higher in Hebrew (seven) than in Arabic (four), we were unable to identify an effect of the language/culture on differences between the modes regarding the contents of the topics. This could be a focus of further investigation. It should be noted that for a number of topics very few narratives were observed, if at all. A larger number of participants could possibly have yielded different conclusions.

Caution is needed when drawing theoretical and practical implications from the present research for several reasons. These include the following: (1) broader valid data from more children is needed; (2) the method of the current study does not allow for disentanglement of language from culture, and the assignment of students to assessment mode groups could not be random; (3) in the current study, the assessments that were carried out at home (and not at school) were carried out in the presence of the parents, which could have confounded the results, reducing the possibility of generalizing the findings to situations where the children are assessed by themselves; (4) we investigated the effect of TA and language/culture on children’s ability to tell personal narratives at the microstructure level and on the chosen topics, not at the macrostructure level, and we measured MLU in words only, not in morphemes; (5) our demographic questionnaire did not include questions regarding accessibility to telecommunication technologies, and experience in telecommunication; and (6) this study specifically examined the effect of TA compared to FTF assessment on the ability to narrate personal narratives among 10-year-old children in Israel, limiting its generalizability to other age, cultural and/or linguistic groups, and to other types of narratives. To address these concerns, we recommend that further studies enroll more children, be conducted in other cultural-linguistic settings, with controls on the presence of parents and in gathering information about the quality of internet connections, along with prior exposure of the participants to Internet use. Caution should be taken in generalizing the results to other types of narratives, different age groups (such as younger or older children), or adults, as well as to children living in areas with limited accessibility to telecommunication.

The findings of the current preliminary investigation provide evidence that the mode of assessment (TA vs. FTF) does not affect the productivity of personal narratives. This suggests that the Global TALES protocol is feasible for TA of personal narratives in school-age children. However, when using the protocol, it should be considered that language/culture can affect performance differently via TA versus FTF. In the current study, TA had a negative effect on the NU of personal narratives produced by Arabic-speaking children but not by Hebrew-speaking children. This difference could be attributed to the collectivist culture in which Arab children are raised, where FTF interactions with family and clan members are highly valued and therefore, these children may feel less comfortable via telecommunication. Moreover, many Arab children live in environments with limited telecommunication technological infrastructure and many of them receive limited support for using these technologies [51, 54]. These conditions may mean that Arab children are less skilled in telecommunication and, therefore, perhaps under TA conditions narrate with fewer utterances. Additionally, when using the protocol, it should also be considered that the assessment mode may affect the topic children choose to narrate about. It seems that while FTF conditions encourage producing narratives dealing with personal dilemmas, TA conditions encourage narratives on more neutral topics.

Given that we live in a time with increased reliance on TA in speech pathology assessment and practice, the current findings have important clinical implications. On the one hand, the current results highlight the feasibility of TA of personal narratives among school-aged children. However, the results also indicate the need to collect information regarding the culture to which children are associated, their telecommunication skills and their accessibility to appropriate telecommunication technology infrastructures. It is also worth considering that TA can influence the contents of the personal narratives of school-aged children.

The authors thank the students of Tel-Aviv University and of Ono Academic College in Israel who collected the data. Special gratitude is due to the participating parents and children.

This study protocol was reviewed and approved by the Human Research Ethics Committee of the department of Communication Disorders, Sackler Faculty of Medicine, Tel-Aviv University, Israel, on June 10, 2018 (the face-to-face phase) and on April 22, 2021 (the Zoom phase). Parents provided written informed consent, and the children provided verbal assent at the start of the assessment session.

The authors have no conflicts of interest to declare.

The authors did not receive any funding support.

Sara Ferman: substantial contributions to the conception and design of the work; analysis and interpretation of data for the work; drafting the work and revising it critically for important intellectual content; final approval of the version to be published; and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Khaloob Kawar: contributions to the conception and design of the work; analysis and interpretation of data for the work; drafting the work and revising it for important intellectual content; final approval of the version to be published; and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.