Personal Narratives of 10-Year-Old Children in Iceland: Verbal Productivity, Word Diversity, and Topics Open Access

Personal narratives or accounts of personally experienced events make up a large part of daily conversations [1]. In everyday life, people share events and experiences as an important part of being understood and understanding the feelings and thoughts of others. Telling personal stories is a universal feature in human discourse, but the topic chosen could be influenced by cultural differences [2] or current events happening in the surroundings, such as the COVID-19 pandemic. Iceland is known as the land of stories, and storytelling is an important cultural tradition that can be traced back to the Icelandic sagas written in the 13th and 14th centuries [3]. The sagas continue to be a vital part of modern Icelandic society and are even used to teach critical thinking in compulsory schools [4]. However, there is limited knowledge about the storytelling abilities of modern Icelandic children.

Between 2019 and 2021, researchers around the world investigated the feasibility of using a protocol labelled the Global TALES (Talking About Lived Experiences in Stories) [5] protocol for eliciting personal stories. The initial feasibility of using the protocol was evaluated originally with 249 ten-year-old children from 10 different countries. The main finding was that children around the world share many commonalities regarding topics of conversation, but within each country, there was great individual variability in verbal productivity (number of utterances and number of words) [6].

We used the same Global TALES protocol [5] in the current study. However, the data collection was conducted online because of the COVID-19 pandemic. Our study aimed to explore personal narratives among monolingual Icelandic children using the Global TALES protocol. Because of the circumstances, we were able to evaluate the feasibility of collecting data online through tools like the Zoom platform.

The initial Global TALES study was conducted by members of the Child Language Committee of the International Association of Communication Sciences and Disorders (IALP) [6]. The researchers from 10 countries (one country, Israel, included two groups of Arabic- and Hebrew-speaking children) recruited monolingual children from mid-socioeconomic areas, who were performing well at school and were aged around 10 years. All countries used the same protocol for eliciting the stories, and all children were seen individually (face-to-face) in a quiet location. The protocol included six scripted prompts for eliciting personal narratives, each about a different type of experience, characterised as: excited, worried, annoyed, proud, problem situation, or important. The child’s verbal productivity was measured with two metrics: total number of utterances (TNU) and total number of words (TNW). Across international data, the mean TNU ranged from 44 utterances (Israel Arabic) to 81 utterances (Cyprus). The range within each country between the child with the lowest and highest number of utterances was quite large. The mean for TNW ranged from 160 words (USA) to 622 words (Australia). The range within each country between the child with the fewest words and the child with the most words varied from 153 words (New Zealand) to 2,190 words (Taiwan). The responses to each of the six different protocol prompts were similar across countries (on average 5–10 utterances), but differences were observed within each country. The proportion of children who received extra follow-up prompts across all languages was low, ranging from 11% on protocol prompt 1 (excited) to 34% on protocol 6 (important). However, some differences were noticed between the countries.

In the Global TALES feasibility study, researchers also aimed to investigate whether 10-year-old children worldwide discuss similar topics in their personal stories or whether there are differences in the topics of interest across cultures and languages. Similar topics were popular across countries: family event was the most frequent topic in narratives responding to prompt 1 (excited), siblings/peer relationships was the most frequent topic following prompt 3 (annoyed), and personal achievement was in the top range for prompts 4 (proud) and 6 (important). School tasks were frequently reported to be something to worry about (prompt 2) and peer/family relationships were mentioned as problems to be solved (prompt 5).

Personal storytelling is an important aspect of everyday life and can offer valuable insights into cultural practices and beliefs. Using a standardized protocol to elicit personal narratives, researchers can compare storytelling abilities and topics across different cultures. Since Icelandic is a highly inflected language, it is important to test the feasibility of using the Global TALES protocol to elicit personal narratives from Icelandic 10-year-olds. Evidence of cultural differences or similarities in topic selection would be of interest as well.

Icelandic is the main language in Iceland. It is spoken by approximately 370,000 inhabitants and is related to Scandinavian languages (Norwegian, Swedish, Danish). Icelandic is highly inflected; for example, nouns, adjectives, and pronouns have endings to mark gender (masculine, feminine, or neuter), number (singular or plural), and case (nominative, accusative, dative, or genitive). The verbs conjugate for tense, mood, person, number, and voice [7, 8].

Personal narratives have not been investigated in Iceland in relation to language development and acquisition. However, a few studies have used language samples to explore the language proficiency of Icelandic children.

The development of macrostructure or story grammar in fictional narratives has been investigated in four different age groups (ages 3, 5, 7, and 9) utilizing the Frog story “Frog, Where Are You?” [9]. It emerged that Icelandic-speaking children could first tell a story with a clear timeline by 9 years [10]. However, only 10 participants were included in each age group. Later, a study with a larger group of 5-year-old children (N = 165) demonstrated substantial individual variability within the group. Some 5-year-olds told stories similar to those of 3-year-olds, while others performed more like 7- or even 9-year-old children [11].

The mean length of utterance (MLU) as a metric was adapted for Icelandic, and the study found a high correlation between the mean length of utterance in morphemes (MLUm) and words (MLUw) [12]. Another study [7] compared language samples from Icelandic- and English-speaking school-age children (N = 42) with and without developmental language disorders (NL/DLD) in conversation, narration, and expository discourse. The results showed that for both languages, MLUm and MLUw varied systematically across the three sampling contexts, with the MLU (MLUm and MLUw) in narratives being the highest and in conversation being the lowest. While English-speaking children with DLD produced significantly lower MLU (MLUm and MLUw) than the NL group, there was no significant difference in the MLU (MLUm and MLUw) between the DLD and NL groups of Icelandic children. These results suggest that the manifestation of language impairment in Icelandic-speaking children may differ from that of English-speaking children, highlighting the importance of not relying on results from studies in other languages.

Norm-referenced data have been collected for conversational language samples of Icelandic children aged between 2 and 6 years old [8]. The data were collected from 221 children, with each sample consisting of 50 utterances. Means were calculated for seven age groups, with 6-month intervals, for the following metrics: MLUw, TNW, NDW, and proportion of words with grammatical errors. In line with studies focussing on other languages, such as English [13], the results showed that means for all metrics increased with age, with MLUw increasing from 2.7 words for 2-year-olds to 5.5 words for 6-year-olds, TNW increasing from 134 words to 273 words, and NDW increasing from 64 words to 130 words. The children produced proportionally few words with grammatical errors, and the number of errors decreased with age. However, there was considerable individual variability on all metrics observed within each age interval. As part of this study, the Icelandic computer programme Málgreinir [14] was developed and published. This programme is free of charge and can calculate the following metrics from transcribed word documents: TNU, TNW, NDW, MLUw, and proportion of words with grammatical errors. Additionally, the programme can establish a word frequency list (in Excel) by calculating how often each word appears in the transcribed document.

The limited studies conducted on the development of Icelandic narratives and conversational language samples have shown that verbal productivity increases with age as well as children’s proficiency to tell stories, although high individual variability was documented in the samples. Studies using the Global TALES protocol to elicit personal stories have shown that there is more individual variability on verbal productivity within each participating country than between the countries. Personal narratives have not been investigated systematically in Iceland, either related to productivity measures or the topics of children’s personal narratives.

In the current study, we aimed to gather data from Icelandic-speaking children to explore the similarities across languages and topics to previous studies using the Global TALES protocol [6]. The Icelandic data could contribute to the understanding of language development and topics of interest in personal narratives for 10-year-olds around the world.

The purpose of the current study, therefore, was to investigate personal narratives of Icelandic monolingual children by using the Global TALES (2018) protocol [5]. To fill gaps in the current knowledge base, the following research questions were addressed.

• 1.

  How do Icelandic monolingual children perform on measures of verbal productivity for the quantity of speech (TNU, TNW), diversity in word use (NDW), syntax (MLUw), and grammar (proportion of words with grammatical errors) in personal narratives gathered with the Global TALES protocol?

• 2.

  Were some protocol prompts more successful than others in eliciting more utterances and words in personal stories from Icelandic children, and did any protocol prompts require additional prompting?

• 3.

  What are the topics that children talked about in response to the six prompts in the protocol?

We hypothesized based on previous work in Iceland as well on studies using the Global TALES protocol that the microstructure metrics on verbal productivity (e.g., number of utterances, number of words), diversity in word use, syntax, and grammar would show a large individual diversity within the Icelandic group. However, we assumed that the calculated mean of these measures would be in line with the outcomes of other countries using the same protocol. We also assumed that the Icelandic group would need similar amounts of scripted follow-up prompting for eliciting the stories as the children in other countries. Since Iceland is a modern society, similar to many of the participating countries utilizing the Global TALES protocol, we assumed that the topic of children’s responses would be in line with the results of that study [6].

Following the ethics permission obtained through the Human Research Ethics Committee at Griffith University (HREC; No 2018/273) in Australia, this study was reviewed and accepted by the Scientific Ethics Committee at the University of Iceland (June 9, 2020). All participants and their parents provided informed consent before participating in the study.

The recruitment of the participants followed the procedure of the initial Global TALES study [6]. The inclusion criteria were monolingual Icelandic-speaking children approximately 10 years of age with no history of language and learning difficulties and currently not receiving speech and language therapy or special education. The recruitment used social media because of the COVID-19 pandemic restrictions. The participants were 27 children, all born in 2011, 13 boys and 14 girls, living in different parts of Iceland. The mean age of the participants was 10 years and 2 months (SD = 0.4, range 9;7–10;9). The parents reported average or above knowledge in Icelandic and Mathematics for their children. Table 1 shows the demographic data for the Icelandic sample.

Prior to the data collection, the first author translated the protocol and the documents and ensured that the translation was according to the original protocol. The investigators viewed a demonstration video and practiced administering the elicitation protocol. The first author pretested the translation of the prompts to elicit the six personal stories (excited, worried, annoyed, proud, problem-solving, and important) with five children and added a brief addition to the word worried which some of the Icelandic children had difficulties understanding. The addition included that the children were asked to tell a story about a time when they were worried and/or did not know what to do. The second author pretested the protocol with three children.

The sample was a convenience sample obtained using a snowball sampling strategy and social media. That is, the investigators asked parents of 10-year-old children whom they knew before the study to share the Facebook announcement with other parents who might be interested. Interested parents contacted the investigators and received all the necessary documents related to the study (information letter, consent form, and background information) by email. They replied by returning the consent form and background information to the second author (speech-language pathology master’s student). The second author interviewed all participants by using the Zoom platform.

All children were interviewed in their homes, with parents present to assist them in starting the Zoom platform. Whereas two children chose to have their parents with them during the interview, 25 children were interviewed via Zoom without a parent present. The data collection took place between July 2021 and January 2022. Consistent with the Global TALES protocol [5], the children were asked to tell a personal story administered in the same order and with the same protocol prompts and follow-up prompts as in the original protocol (see https://osf.io/ztqg6/). The duration of the interviews varied from 7 to 20 min.

The interviews were audio recorded and transcribed based on standard Systematic Analysis of Language Transcripts (SALT) conventions. The segmentation of utterances was based on communication units (C-units), with all reformulations, repetitions, false starts, and filler words put in brackets and not included in the analysis. Only complete and intelligible utterances were counted as C-units. Grammatical errors, such as the incorrect use of singular/plural, gender, or declension of nouns, adjectives, pronouns, and numbers, were specially marked with quotation marks. However, these errors were only counted at a word level, not at the utterance level. For example, wrong word order in a sentence was not counted as an error. The Icelandic programme Málgreinir (http://malgreinir.herokuapp.com/) was used to calculate the TNU, TNW, number of different words (NDW), mean length of utterances in words (MLUw), and proportion of words with grammatical errors.

The second author transcribed all the samples. The first author relistened to all the samples and ensured that the segmentation was according to C-units, following the same procedure as described in the Global TALES study [6]. An independent SLP re-transcribed samples (all six prompts) from six participants (randomly selected) or 22%. Reliability of the transcription was sufficient across the metrics, between 80% and 100% [15].

Both authors independently coded the topic of children’s responses using the same coding system as described in the initial Global TALES study represented in Table 5 and Appendix S2 [5]. The authors agreed on 91% of the coding of the topics. The authors then met and discussed the coding of the topics until they reached a full agreement.

All statistical calculations and figures were produced in Excel.

The first research question asked about the verbal productivity and semantic diversity of Icelandic children. Table 2 shows results for measures of quantity of speech, including the TNU and words, as well as the diversity in word use (measured by NDW), syntax (measured by MLUw), and grammar (measured by the proportion of words with grammatical errors).

The TNU ranged from 26 utterances (P6) to 115 utterances (P1), and the mean was 62.1 utterances. The data showed that 18 participants out of 27 (67%) told the stories by using 42 to 82 TNU or within +/− 1 SD from the mean; four participants (N6, N11, N23, N24) were >1 SD below the mean, and five participants (N1, N2, N3, N7, N13) produced TNU >1 SD above the mean. The TNW ranged from 183 words to 989 words, and the NDW ranged from 90 to 336 words. Most of the participants (19 or 21 out of 27) produced TNW and NDW within +/− 1 SD from the mean. The mean of MLUw was 8.6 words, with utterances ranging from 6.0 words to 12.1 words. Four participants produced sentences >1 SD from the mean (N3, N8, N13, and N25), but none were >1 SD below the mean. The participants produced proportionally few errors in spontaneous speech, with the range of proportion of grammatical errors per TNW from 0 to 2.2%.

The second research question asked whether any of the six prompts elicited responses with greater numbers of utterances or different words. The results for this analysis are represented graphically in Figures 1 and 2, which illustrate the TNU and TNW by prompt. The number of utterances across the six stories and participants varied from 2 utterances (question 6 [Q6], participant 6) to 32 utterances (question 5 [Q5], participant 1). The mean number of utterances was similar, however, across the six prompts, ranging from 9.96 utterances (Q4) to 12.89 utterances (Q5) (see Fig. 1). A one-way ANOVA showed no significant difference between the six stories (F(5, 156) = 1.02, p = 0.41) for the TNU.

Similarly, the TNW varied across participants, and stories varied from 9 words (Q1, P6) to 301 words (Q5, P1). The mean number of total words ranged across the six questions from 75 words (Q1) to 111 words (Q5) (see Fig 2). A one-way ANOVA showed no significant difference between the six stories (F(5, 156) = 1.41, p = 0.22) for the TNW

In accordance with the initial Global TALES study [6], we also investigated if scripted follow-up prompts were needed to elicit a response. Results showed that 33% of the children did require a scripted follow-up prompt for protocol prompts number 2 (worried) and 6 (important), 22% for protocol prompts number 1 (excited) and 4 (proud), 30% for protocol number 5 (problem-solving), and 11% for protocol prompt number 3 (annoyed).

In response to the third research question, the topic codes for the protocol prompts are shown in Table 3. Personal achievement was the most frequent topic for the first protocol prompt (excited), but almost the same number of children chose to talk about a family event. For the second protocol prompt (worried), the children talked about a range of topics, such as new challenges, safety concerns, and illnesses. Sibling/peer relationships was the most frequent topic (85%) in response to the prompt, annoyed. For protocol prompt 4 (proud), 93% of the children select to talk about personal achievement and particular sporting achievement. Most of the children selected personal growth or contribution for protocol prompt 5 (solving a problem) and for protocol 6 (important) family event or support. Only two answers could not be coded. One child answered to be excited (prompt 1) by telling about watching an exciting movie, and one child could not provide an answer to prompt 5 on solving a problem. It is notable that the children in Iceland never mentioned COVID-19 in relationship to the six stories.

The present study replicated the Global TALES feasibility study by using the same protocol and method to elicit personal narratives from school-age children in a convenience sample from the Icelandic population. The main result was that the protocol appears to be similarly effective in eliciting personal narratives for Icelandic children as for the participants across the 10 other countries in the Global TALES study [6]. The verbal productivity measures showed large individual variability between the participants measured with TNU, but around 70% fell within +/− 1 SD from the mean. For the metrics TNW, NDW, MLUw, about 75% fell within +/− 1 SD from the mean. Overall, the participants made very few grammatical errors when telling the stories. There was no significant difference between the six different protocol prompts as measured by TNU and TNW, showing that all prompts were equally successful in eliciting productive stories. The study’s overall results were promising regarding the future use of the protocol for benchmarking personal narrative performance among 10-year-old children. The topics of the stories were most frequently related to personal achievement or relationships within the family or friends.

The similarity of methods used in the current study, except for using telecommunications (supported by Zoom) to gather the samples rather than face-to-face interactions, makes it possible to compare these results to those reported in the initial feasibility study [6]. First, the recruitment of participants in Iceland followed the same inclusion criteria as for the participants in the 11 groups in the initial Global TALES study. That is, we recruited children between 9 and 11 years old with no history of language difficulties. Second, like the Global TALES study, most of the participants had parents who identified their socioeconomic status as middle class. The educational status was high, with 75% reporting having studied at the university. In the earlier study, only Australia and Israel (Arabic-speaking group) included participants where proportionally more parents had finished a university degree. Third, the investigators in the current study followed the Global TALES protocol as closely as possible in administering the prompts and avoiding additional prompting as much as possible. Given these similarities, the remainder of the discussion addresses possible differences that might be unique to the Icelandic language or culture or to the use of Zoom to gather the samples.

The large ranges in verbal productivity statistics for the Icelandic samples were in line with those in the Global TALES study and other studies using language samples for examining language production [16, 17]. The mean for the TNU was 62 utterances, which is in line with countries as varied as Australia, Greece, Russia, and Taiwan. The minimum number of utterances in this study was 26 utterances, which is comparable to the Global TALES study, which ranged from 17 (Russia) to 35 (New Zealand). The maximum number of utterances was 115, which again falls within the range for the countries in the Global TALES study, which varied from 71 (Israel Arabic speaking) to 227 (Taiwan). The number of utterances is an important metric for studying personal narratives, reflecting how willing the children are to share personal events and if they can retrieve in the interview memories of something to share. Consistent with the protocol, the children were not provided with time to plan their personal event narratives, nor were they given a practice item. Future studies may consider adding a preparation component to determine if this would affect the length of the personal narrative samples.

The mean TNW for the Icelandic sample of children was 546 words, which falls towards the upper end of the range of means observed in the Global TALES study. The range for TNW in the initial feasibility study varied from 160 words (USA) to 622 words (Australia). Interestingly, the children who produced the fewest and most words in the Global TALES study all spoke English, suggesting that the TNW metric is not dependent on the language spoken. Unfortunately, in the current study, we were not able to compare the Icelandic children’s NDW or MLUw to those of the Global TALES study, as data were not reported for these measures in that study [6]. However, another study used MLUw as a measure to analyse narratives from Icelandic children with typical language development who were asked to retell a book or a movie. The result was that the average MLUw score was 7.85, with a standard deviation of 2.8 [7]. Given that the children in the current study were about a year older than those in the prior study, the mean MLUw of 8.6 words and standard deviation of 1.5 words seem consistent with other Icelandic data for this age group.

The Icelandic children in the current study demonstrated a relatively low occurrence of grammatical errors when telling personal stories which is consistent with the findings of previous studies that investigated language samples from Icelandic children [7, 12]. Grammatical errors are relatively infrequent in the spontaneous speech of young Icelandic-speaking children, with a proportion of errors per word around 4% for 2-year-olds and 1.3% for 6-year-olds [8]. A study comparing Icelandic- and English-speaking school-age children with DLD and NL also showed that both groups of Icelandic children made relatively few grammatical errors [7].

As found in the Global TALES feasibility study [6], all six protocol prompts could be used to elicit stories from most of the participants. The mean number of utterances and words was similar across the stories, and there were no significant differences between the prompts either for the number of utterances or for the words. Following the Global TALES protocol, a few scripted follow-up prompts were used to encourage the children to tell the stories or to remember something to share, but examiners were directed to use them only when necessary. In the initial Global TALES study, the use of scripted follow-up prompts ranged from 11% (prompt 1, excited) to 34% (prompt 6, important). Some differences were observed in the current study by prompt, but the range was similar, starting at 11% for prompt 3 (annoyed) and going up to a similar 33% for prompts 2 (worried) and 6 (important).

The topics in this study were coded using the same coding system as in the Global TALES study. The children in Iceland seemed to respond to the protocol prompts with similar topics as the children around the world with some exceptions. Personal achievement, especially in sport or football, was a very popular topic in current study for the protocol prompt 1 (excited) and prompt 4 (proud). In the initial Global TALES study, the most frequent topic in response to prompt 1 was family event, which was in second place for the Icelandic children. The children in Iceland seemed not to worry about school tasks as much as children in other countries, where it was the most frequent topic for the prompt, worry, for all countries except Brazil, Cyprus, and the USA. It could also be that the additional explanation for the word worry had influenced what the children chose to talk about. In Iceland, like in all other countries studied, relationships with sibling or peer are most frequently reported to be annoying.

Children in Iceland talked about personal growth most frequently in response to the protocol prompt 5 (problem-solving). This was similar to the children in Russia. Like children in the USA, Icelandic children talked about family events or support as most important in response to prompt 6 (important). As noted previously, the Icelandic samples were collected during the pandemic COVID-19, but the pandemic was never mentioned in the personal stories. This contrasted with the experience in Ireland, where samples were collected at a similar time and some topics were associated with pandemic experiences [6]. A possible explanation could be that during the pandemic the children (between 6 and 12 years) in Iceland attended schools most of the time. Only if they came into contact with someone diagnosed with COVID-19 were they required to quarantine at home for 14 days [18].

The current study collected personal narratives via Zoom, and it is not clear if these results would generalise to face-to-face elicitation. Future research should investigate the effect of elicitation mode on personal narratives to determine if children may be less inclined to share (personal) information in an online environment or may choose to discuss different topics (see also [19]). Another limitation of the study is the small sample size. By including a larger sample, it would be possible to obtain a more representative and comprehensive understanding of personal narratives among children in Iceland. Additionally, a third limitation relates to the sample composition, as the majority of participants were from middle-class families or shared similar backgrounds. Diversifying the sample to include participants from a wider range of socioeconomic backgrounds could provide a more comprehensive perspective on the personal narratives.

The Global TALES protocol offers a unique opportunity to compare personal narratives across different languages and cultures worldwide. The Icelandic children in this study demonstrated similar levels of verbal productivity as reported in the Global TALES study [6] for speech quantity (TNU and TNW). Additionally, metrics such as MLUw, NDW, and proportion of grammatical errors were consistent with previously published studies [7, 8]. Personal narratives are a crucial aspect of daily life, and the ability to share stories and reflect on personal events is an essential part of universal human discourse. The individual variability observed within the Icelandic group, as well as across all 11 groups examined in the Global TALES study, highlights the need for further investigation with larger participant groups. The results of the current study are promising regarding the use of the protocol to provide age reference data for personal narratives among 10-year-old children. It will be interesting to investigate personal narratives by using the protocol among those who may experience language difficulties or differences, such as children with developmental language disorder and bilingual children.

The authors gives special thanks to the children and parents who participated in this study.

The ethics permission followed a permission obtained through the Human Research Ethics Committee at Griffith University (HREC; No. 2018/273) in Australia. The Scientific Ethics Committee at the University of Iceland reviewed and accepted the research (June 9, 2020). All participants and their parents provided informed written consent before participating in the study.

The authors have no conflicts of interest to declare.

The research was funded by the University of Iceland Research fund. The funding included payment to the second author.

Data collection: Erna Þráinsdóttir. Methodology, data analyses, and writing review and editing: Jóhanna Thelma Einarsdóttir and Erna Þráinsdóttir. Writing, project administration, and funding acquisition: Jóhanna Thelma Einarsdóttir.