Computer-Mediated Communication: UK and Arab Samples

Methodology

Method: How Communication Styles Were Studied

3.1. Participants and procedure

The study included 599 participants, aged 18 to 60, with 272 from the UK and 327 from Arab Gulf Cooperation Council (GCC) countries. The UK participants represented individuals from England, Scotland, Wales, and Northern Ireland, while the Arab participants were from Saudi Arabia, the United Arab Emirates, Bahrain, and Qatar. Notably, there were no respondents from Oman or Kuwait. Participants were recruited via the TGM online platform, a multi-country online research company. TGM employs a wide range of technology-enabled quality controls, including advanced digital fingerprinting to eliminate duplicate responses, trap questions to detect inattentiveness, bot checks, and extensive geo-location verification. TGM also adheres to ESOMAR regulations, including a double opt-in process for panellists, ensuring both ethical recruitment and data quality. As part of its identity procedures, the platform verifies nationality, which was used in the panel-based preselection of eligible participants. Also, survey access was restricted to individuals physically residing in the selected countries. To add an additional layer of validation, participants' country of birth and country of residence were verified through self-reported identity within the survey itself.

To be eligible, participants had to be 18 years or older, frequent users of social networking platforms (e.g., Facebook, Instagram, Twitter) and messaging services (e.g., WhatsApp, Facebook Messenger) and self-identify as British or Arab based on norms and culture. These criteria were verified through a pre-selection survey, and only those meeting the requirements were invited to complete the full survey. To ensure response quality, attention checks were integrated into the survey, and the time each participant took to complete the survey was recorded. Responses were flagged as fast if they were completed in 50% or less of the median time taken by all participants. To calculate this median time, outliers such as participants who took twice as long as the expected duration, mostly because they completed the survey over multiple sessions, were excluded. The study received ethical approval from the Research Ethics Committee at the first author's institution, and data collection took place between late October 2023 and mid-December 2023. Before starting the survey, participants were required to read an information sheet and provide electronic informed consent. They were also informed of their right to withdraw from the study at any point. Only participants who completed the survey, passed attention checks, and were not disqualified for providing excessively fast or inconsistent responses were compensated for their participation.

3.2. Dataset and study design

The study was administered using SurveyMonkey a web-based survey tool. The research team followed an iterative process to improve the clarity and comprehensibility of the survey questions. The questionnaire was translated from English to Arabic, and the back-translation method was employed to ensure consistency and accuracy. The data collection was part of a larger study that can be found available on the Open Science Framework at the link provided in the Supplementary Material section.

The questionnaire consisted of several primary sections, including an assessment of demographics (i.e., gender, age, and educational level). The study examined the use of the different communication styles for participants as mentioned earlier. A brief definition of each is shown in Table 1.

Each of the eight communication styles examined in the study, paired with a short definition.

To ensure participants clearly understood each communication style before completing the survey, we provided them with definitions, realistic examples, and follow-up assessment items. A vignette-based approach was used, presenting participants with a series of brief, contextually grounded scenarios that reflected common social and professional interactions. The vignettes were designed to reflect semi-professional and everyday communication contexts, such as seeking help, planning social events with colleagues, or managing everyday transitions. The highly personal or intimate scenarios were intentionally avoided to minimize potential cultural sensitivity. Highly formal corporate communication because formalized professional discourse often follows globally standardized norms that can obscure cultural differences in communication style. By focusing on semi-professional and everyday contexts, the aim was to elicit communication that remains recognizably authentic across cultures while still allowing culturally influenced variation to be observed. This was a deliberate methodological choice, as prior research suggests that every day and relational communication highlights culturally specific norms more effectively than formalized settings, where standardized rules can conceal these differences.

All vignettes were presented in Standard English and Modern Standard Arabic, depending on the language of the survey version. These languages were chosen to ensure accessibility and consistency across participants. The scenarios themselves were not tied to a single digital platform but were presented in a conversational format that mirrors common online and mobile-mediated communication (e.g., messaging, email, WhatsApp groups). This approach allowed us to capture communication practices that are generalizable across commonly used online environments, rather than restricting the focus to one platform. Each vignette illustrated two contrasting communication styles (e.g., direct vs. indirect, oral vs. literate). After reviewing each scenario, participants were asked to identify which communication style was being used, as summarized in Table 2.

Four paired communication-style contrasts. For each pair, a short setup scenario is followed by an example message written in each style, so participants could identify which style was used.

To ensure clarity and cultural relevance, we performed face validation with participants from both target populations. Five participants - two from the UK and three from Arab countries - provided feedback, representing both cultural perspectives. Based on the participants' feedback, several adjustments were made. We standardised the names used in the scenarios to ensure the survey avoided any potential religious or gender connotations. Initially, we employed the names "Sam" and "Noor," but later unified them to "Sam" across all scenarios for consistency. To address concerns about readability issues when presenting multiple communication styles, we used two distinct colours - blue and orange - to differentiate between paired communication styles, with each colour consistently assigned to one style (see details in Supplementary Material).

3.3. Measurements

We collected information about participants' age, gender, and education. To evaluate the use of communication styles, participants responded to two key statements per style: "I use it when communicating with others?" (referred to as 'Self-use') and "Others use it when communicating with me?" (referred to as 'Others-use'). Comfort levels with each style were measured through the following statements: "I feel comfortable using it when communicating with others?" (referred to as 'Comfort_Self-use') and "I feel comfortable when others use it to communicate with me?" (referred to as 'Comfort_Others-use'). All responses were recorded on a 6-point Likert scale, ranging from 1 ("Strongly disagree") to 6 ("Strongly agree").

3.4. Data analysis

Prior to analysis, both datasets were screened for consistency. As there were no Arab participants over the age of 60, UK participants above 60 were also excluded to maintain comparability in age range between the two groups. The data were then examined for missing values and inconsistencies, with only complete and valid cases taken for analysis.

First descriptive statistics (means and standard deviations) were computed for each communication style dimension, separately for the UK and Arab samples. This presented an initial overview of how participants evaluated both their use of each style and their comfort with it.

Next, a correlation analyses (Pearson's r) was run between all communication style ratings and comfort ratings within each group. This step was used to assess whether participants' ratings of different styles were related to each other. Although most of the correlations were modest in size, they were consistent across variables, indicating that the constructs were interrelated. Because of this shared variance, we decided to analyse all ratings together in a single model rather than running separate tests for each dimension.

For this analysis, we employed Linear Mixed Models (LMMs). This method is particularly appropriate for datasets in which each participant provides multiple ratings, as it accounts for repeated measures within individuals. In our models, each rating was characterized by three factors: Mode (whether the rating referred to the use of a style or comfort with it). Person (whether the rating referred to the participant's self-communication or the others) and Condition (the specific type of communication style (Direct, Indirect, Ordered, Unordered, Being-centred, Doing-centred, Literate, Oral)).

Mode, Person, and Condition were treated as fixed effects, allowing us to test their influence on ratings systematically. In addition, a random intercept was included for each participant to account for individual differences in overall response levels (for example, some people may generally rate all styles higher or lower than others). Interaction terms between Mode, Person, and Condition were incorporated to examine whether preferences for particular styles depended on whether participants were rating use or comfort, and whether they were rating themselves or others. Where significant main effects or interactions emerged, post-hoc pairwise comparisons were conducted with Bonferroni correction to adjust for multiple testing.

Finally, to directly test for cultural differences, we fitted additional models that included Country as a factor. These models examined whether Arab and UK participants differed systematically in their reported use and comfort of communication styles, and whether such differences varied depending on whether they were rating themselves or others. All LMM models were estimated using Jamovi version 2.6.2.