Exploring preservice teachers’ digital skills, stress, and coping strategies during online lessons amid covid-19 pandemic in Ghana

Abstract There are concerns about the association between the level of students’ digital skills and the amount of stress they experience. However, evidence from both cross-sectional and longitudinal studies is limited. The main aim of this study was to address this gap in the literature by examining the relationship between preservice teachers’ digital skills and stress. The study employed an online cross-sectional survey, which was completed by 661 pre-service teachers enrolled in colleges of education in Ghana. The findings showed that preservice teachers’ low digital skills were associated with higher stress levels during the period of online teaching and learning necessitated by the COVID-19 pandemic. The study further revealed that preservice teachers’ limited access to resources was a significant cause of stress. Preservice teachers’ main coping strategies for reducing stress were positive framing, such as seeking advice from friends and family, seeing something good in what is happening and learning from the experience. Recommendations include providing more training on digital technologies during the initial preparation for teachers.


PUBLIC INTEREST STATEMENT
Stress among students comes from different sources; however, stress as a result of digital skills has not been explored very much. The study's objective was to explore the relationship between preservice teachers' digital skills and stress to address the gap in the literature. The study used an online cross-sectional survey to invite 661 preservice teachers in colleges of education in Ghana. The study finds that preservice teachers with low digital skills experience higher stress levels during online teaching and learning sessions. Furthermore, limited resource-related factors were a significant cause of stress among preservice teachers. The preservice teachers used 'seeking advice from family friends, making fun of the situation, watching movies, and using social media as the coping strategies to reduce stress. The paper concludes with a recommendation for the preparation of teachers to focus more on integrating ICT in training of teachers in these unknown times in a technological world.

Introduction
The novel coronavirus COVID-19 pandemic outbreak at the beginning of the first quarter of 2020 put the global community and the education system into crisis. The international academic calendar was thrown into disarray by the coronavirus outbreak. Most schools, colleges and universities were shut down due to government mandates implemented to control the spread of the virus. According to UNESCO, the world has never witnessed such large-scale disruptions in educational systems as it did during the COVID-19 pandemic (UNESCO, 2020). Consequently, during the COVID-19 pandemic, countries made efforts to utilise digital technology to support remote learning, also called distance education or online learning. During this period, most institutions promoted the use of information and communication technology (ICT) in teaching and learning to avoid disruptions in the academic year.
Due to the COVID-19 pandemic, ICT integration in education has increasingly been prioritised in both developed and developing countries. The COVID-19 pandemic has led many educational institutions to use blended modes (i.e., a mixture of online and face-to-face classes) of teaching and learning to deliver their curriculum. However, the effective use of online teaching and learning platforms requires users to have basic digital skills. Even though many higher education institutions in developed contexts have been using online platforms to deliver their curricula to students, universities and colleges in sub-Saharan Africa (SSA) have been slow to implement online teaching and learning (Ogbonnaya et al., 2020). The closure of schools during the COVID-19 pandemic accelerated the adoption of ICT by many institutions in Ghana. However, research evidence shows that if students and instructors have not been given the necessary training and if resources are inadequate, online teaching and learning become a source of stress (Agyei & Voogt, 2011;Çebi & Reisoglu, 2020). Previous studies have reported low ICT skills among teachers in Ghanaian basic schools (Adarkwah, 2021;Apori Ansah, 2019), limiting the use of digital technologies in their teaching and learning.
There are gaps in the literature related to preservice teachers' digital skills and whether these skills influence their stress levels. Current scholarship in developing contexts shows inconsistent findings on digital skills and stress amidst the push for ICT integration in teaching and learning. Additionally, the literature related to stress also shows inconsistent results on the relationship between the level of digital skills and stress, specifically related to smart devices (Barrot et al., 2021;Van der Schuur et al., 2019). Moreover, digital skills in using devices and their effects on users' stress levels have not been explored much, even in developed contexts. Huang et al. (2015) conducted a survey of 972 students to investigate the influence of emotional costs on lowerincome students' technology efficacy, academic efficacy, and computer application proficiency; the results revealed an association between digital skill level and stress. The study found that students with lower technological skills in using digital media experienced frustrations which could be a source of chronic stress. In addition, the pressure on students and teachers to use digital media to keep in touch with friends and family and participate in academic activities could also become a source of fatigue and chronic stress (Hampton et al., 2016). Stress comes in different forms, affecting a person's health irrespective of race, age socioeconomic characteristics (Williams, 2018). If not controlled, excessive stress can result in poor academic performance, addictions, dropping out of school, and crime among students. Therefore, there is a need for an empirical study to examine digital skills among pre-service teachers and investigate the relationship between their digital skill levels and stress levels in a developing country context. This study has become necessary due to the increased use of digital media in the training of teachers.
There is no prior empirical study in Ghana investigating whether pre-service teachers' digital skills have any relationship with their stress levels. A study of this kind will contribute to the design of online learning platforms as part of Ghana's ongoing teacher education reforms. Therefore, we examined pre-service teachers' digital skills and their stress levels during a period of online teaching/learning necessitated by the COVID-19 pandemic in a resource-limited context. The study was guided by the following research questions: (1) What digital devices do pre-service teachers use to access online teaching and learning platforms?
(2) What are the digital skills of pre-service teachers during online teaching and learning throughout the COVID-19 pandemic in Ghana?
(3) What is the relationship between pre-service teachers' digital skills and their stress levels during online teaching and learning?
(4) What coping mechanisms do pre-service teachers use to minimize stress?

Conceptual framework
Drawing on the concept of the "second digital divide," which refers to the gap between people who do and do not have access to and the skills to use digital devices and the internet (James, 2021), this study recognizes "digital skills" as a significant component of the "second digital divide". The digital divide concept was prominent in the 1990s, mainly referring to the gap between people with and without access to digital devices. However, in recent times, digital skills and internet access have been reconceptualized as the "second digital divide" (Van Laar et al., 2017;World Bank, 2016). Digital skills are a range of abilities required to use digital devices, such as computers, smartphones, tablets, etc., as well as communication applications and networks to access and manage information (UNESCO, 2019). Digital skills play a leading role in new remote education practices (James, 2021). The United Nations' sustainable development goal (SDG) 4, specifically indicators 4.4.1 and 4.4.2, measure the "proportion of youth and adults with information and communication technology (ICT) skills by type of skills" and the "percentage of youth and adults who have achieved at least a minimum level of proficiency in digital skills", respectively (UNESCO, 2019, p. 13). The Young Lives longitudinal study conducted in Peru, Vietnam, India and Ethiopia (Cueto et al., 2018) conceptualizes "digital skills" as access to and use of digital devices and computers and internet-related skills. To extend Cueto et al.'s (2018) categories of digital skills, we categorised computer skills as "file navigation skills" and "basic Microsoft Office skills". Other studies have used these two skills (Çebi & Reisoglu, 2020;Van Ingen & Matzat, 2018) to explore ICT competencies among teachers.
To make the connection between digital skills and stress, Huang et al. (2015)) argue that stress is experienced by disadvantaged students when they struggle to gain access to digital devices and perform digital tasks to keep up with their more advantaged peers. Stress refers to the psychological and physiological disturbances that an individual experiences due to interactions with the environment, depending on personal characteristics such as sex, age and health status. We hypothesised that there is a relationship between the level of students' digital skills and the stress they experience. The Dental Environmental Stress (DES) questionnaire (Cohen et al., 1983) conceptualises stress as "academic-related" and "psychological-related". We extend this conceptualisation by adding "digital skills-related", based on the review of digital education literature (Barrot et al., 2021;Van der Schuur et al., 2019).

Digital skills, stress and coping strategies
During the COVID-19 pandemic, there has been a heightened use of digital technology across the globe. Consequently, individuals lacking digital skills must either catch up quickly or risk being left behind. Young adults and educational institutions use several platforms that require knowledge of and skill with digital technologies. Online teaching and learning require basic digital skills and access to digital devices to meet educational demands (Cueto et al., 2018). The Young Lives study on digital access, use and skills among students revealed higher usage of smartphones and the internet than computers in Vietnam, India and Per, however this was not true in Ethiopia (Cueto et al., 2018). Furthermore, students in Vietnam and Peru had relatively more digital skills than students in India and Ethiopia (Cueto et al., 2018). In exploring digital skills, Çebi and Reisoglu (2020) observed differences among pre-service teachers. Their study found that Turkish pre-service teachers' digital skills in information and data literacy, communication and collaboration were higher than their skills in safety, content creation and problem-solving. This study also found stronger skills among male than female pre-service teachers. According to UNESCO (2020), women in most countries are 25% less likely than men to leverage ICT for fundamental purposes, such as calculating simple arithmetic formulas in spreadsheets. This gender gap in digital skills is more pronounced among older, less-educated women and women residing in rural areas in developing countries (UNESCO, 2019). A qualitative study exploring tertiary students' perceptions of e-learning during the COVID-19 pandemic in Ghana identified significant challenges associated with online learning, including lack of access to ICT tools, little exposure to online modes of learning, and unreliable internet access (Adarkwah, 2021). Ogbonnaya et al. (2020) investigated the experiences of pre-service teachers enrolled in a highly rated technological university in Ghana during the COVID-19 lockdown and found that about 10% of the students had limited digital literacy. These findings are consistent with the results of a study by Apori Ansah (2019) that explored basic school teachers' ICT skills in Ghana. The study found that most basic school teachers had adequate computer file navigation skills, such as opening files/folders, moving files/folders, deleting files/folders, and using MS Word. However, a reasonable number of the teachers had limited skills in performing tasks with MS Excel and using the internet.
A lack of digital skills among youths, especially pre-service teachers, results in the exclusion of these teachers and their learners in future digital societies, which are known to offer better opportunities and welfare (James, 2021). The literature shows a lack of empirical evidence investigating the association between digital skills and stress among students, especially during online teaching and learning sessions. Stress could result from using a computer for long hours, exposure to computer and smartphone screens, and long hours of using social media, etc. (Mheidly et al., 2020). However, stress resulting from the struggle to perform digital tasks has received very little attention in the digital education literature. In this study, we focus on stress resulting from using digital devices. Teachers with higher technological competencies can be assumed to have lower stress and anxiety levels than teachers with low technological competencies, who experience more frustration and, thus, stress and, as a result, tend to hesitate to use computers (Sang et al., 2010). A study in Ghana that explored the will, skill and tool model among prospective teachers and practicing teachers revealed that in-service teachers experienced some stress. The study found that teachers with high levels of technological skill experienced less computer-related stress than teachers with lower levels of technological skill (Agyei & Voogt, 2011). Another form of stress from digital devices stems from online classes and social media. A study of undergraduate students in Indonesia (Kumalasari & Akmal, 2021) during the COVID-19 pandemic found a significant to moderate correlation between academic-related stress and satisfaction with online learning.
A similar study in Japan (Jung et al., 2012) found that students experienced stress when they had to perform many digital tasks and other internet-related activities to stay on par with other students during an online programme. On the other hand, a study by Hampton et al. (2016) with a sample of adults in the USA found that the use of digital media was not directly related to high levels of psychological stress. Most participants in Hampton's study were educated individuals who, it can be assumed, had strong digital skills. However, as indicated by Huang et al. (2015)), people with lower technological competence could experience chronic stress brought by daily use of digital devices. Several studies on students' stress have cited academic workload, examinations, family relationships and financial problems as sources of stress among students. However, we know little about the relationship between digital skill level and stress in the context of online teaching and learning, especially during the COVID-19 pandemic. Uncontrolled stress can be harmful to an individual's physical and mental health (Jung et al., 2012;Van der Schuur et al., 2019). Prolonged stress can result in physical problems, including high cholesterol, hypertension, arthritis and heart disease (Lim, 2020). Therefore, individuals must learn strategies to cope with stress, especially higher education students, who are consistently confronted with psychosocial and academic challenges and anxiety about the future. Stress coping strategies are conscious mechanisms used by individuals to reduce stress and its effects (Baloran, 2020). A study on college students' stress in two institutions in Illinois, USA, identified social support, religious support, and positive reappraisal as sources of support to reduce stress among students. Awoke et al. (2021) investigated perceived stress among undergraduate health science students and found that seeking emotional support from family and friends was the main coping strategy used by the students to reduce their stress levels while participating in remote classes during the pandemic. These findings are similar to those of studies (e.g., Bamuhair et al., 2015;Pierceall & Keim, 2007) conducted in other parts of the world, which showed that when students experienced stress from using digital technology, they sought support from friends who had a higher level of digital skill. However, it is worth noting that not all coping strategies are positive. Students also use negative coping strategies to reduce stress, such as using alcohol and drugs, venting their anger towards others and avoiding social interactions (Bamuhair et al., 2015). Our study also explored the strategies that pre-service teachers used to cope with stress during the period when remote/ digital education was necessitated by the COVID-19 pandemic.

Study context
As of January 2021, Ghana had 157,220 cases of COVID-19, 154,424 recoveries and 1,404 deaths (Ghana Health Service, (2021).). On 15 March 2020, the president of Ghana instructed all educational institutions to shut down to avoid spreading the virus. As a result, many tertiary institutions adopted online teaching and learning to complete the academic year. The participants in the current study were pre-service teachers enrolled in colleges of education under the umbrella of the University of Cape Coast (UCC). These pre-service teachers used the UCC online learning management system (LMS) and other social media platforms to access lectures and participate in related academic activities during the period of remote education.

Participants and procedures
The study adopted a descriptive online cross-sectional survey design to assess pre-service teachers' digital skills, stress, and coping strategies during the period when the COVID-19 pandemic necessitated the implementation of online teaching and learning. Data were collected between October and November 2020. Third-year pre-service teachers from 10 of the 46 public colleges of education were invited to complete an online questionnaire; the total population of pre-service teachers in all the 46 public colleges of education in the 16 regions of Ghana was 32,958 in 2020/ 2021 academic year. We stratified the colleges into five regional groupings: Northern, Ashanti/ Brong Ahafo, Western/Central, Volta, and Greater Accra/Eastern. We then randomly selected (using a table of random numbers) 2 colleges from each group making a sample of 10 colleges for the study. Since pre-service teachers were not on college campuses during the data collection, we used WhatsApp Messenger to reach potential participants from the selected colleges. A questionnaire was designed in Google Forms and made available online, and a link was generated and shared on WhatsApp with third-year pre-service teachers enrolled in the 10 selected colleges. Potential respondents were asked to indicate their willingness to participate in the study before they were provided with a link to the survey. A total of 661 students completed the survey, representing a 79% response rate.
The participants (see , Table 1) consisted of 303 females (45.8%) and 358 males (54.2%). Of these, 638 (96.5%) were single, and 23 (3.5%) were married. More than 95% of the participants were younger than 25 years old, while 1.8% were older than age 25 years. Three hundred and twenty-six students (49.3%) were enrolled in the primary education programme, while 45.4% were enrolled in the junior high school education programme, with the remaining students (5.3%) enrolled in the early childhood education programme. A significant minority of the students (26.2%) had social studies as their elective course, while the least common elective was science (5%).

Study measure and data analysis
The questionnaire was divided into four sections. The first section asked demographic questions about the respondent's age, gender, programme of study, devices used, etc. The second section included 19 items that measured digital skills; this section was reduced to 10 items after the items underwent factor analysis. These items were adopted from the Young Lives international study on digital access, use, and skills across four countries (Cueto et al., 2018). Participants responded to the items on a 4-point Likert scale, ranging from 1 = strongly disagree to 4 = strongly agree. The items included "I know how to change the margins of a document (e.g., using Word)", "I know how to create a presentation (e.g., using PowerPoint)" and "I know how to open downloaded files". This section comprised two subscales: "basic Office skills" and "internet skills". These subscales are supported by the studies of Biletska et al. (2021) and Eynon and Geniets (2016). The third section comprised 17 items on the causes of stress, of which 10 items were adapted from the Dental Environmental Stress (DES) survey (Cohen et al., 1983), while the remaining items were adapted from other studies on stress and digital media (Van Ingen & Matzat, 2018). These 17 items were later reduced to 11 items after undergoing factor analysis. This section also comprised two subscales: "resource limitation-related" and "academic-related" sources of stress. The last section comprised 12 items measuring stress coping strategies; this was reduced to six items after factor analysis. These items were adapted from Bamuhair et al. (2015) and Pierceall and Keim (2007).
Descriptive statistics were computed to analyse the responses to the subscale items based on gender. Pearson's correlation analysis was performed to establish the relationship, if any, between the level of digital skills and stress. The independent samples t-test was used to indicate where there might be statistical differences in digital skills between male and female pre-service teachers. The effect size was calculated using Cohen's d (Cohen, 1988). Cohen provided benchmarks for interpreting effect size; he considered d = 0.2 a small effect size, d = 0.5 a medium effect size and d = 0.8 a large effect size. To ensure the construct validity of the questionnaire, we conducted an exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) using the Statistical Package for the Social Sciences (SPSS) version 21 and SPSS AMOS (22), respectively. A simple random sampling procedure was used to split the data set into two subsamples: a subsample of n = 330 was used for exploratory factor analysis, and a subsample of n = 331 was used for crossvalidation using CFA to investigate whether the identified EFA structure fit the data.

Exploratory factor analysis
All 19 items designed to measure respondents' digital skills were subjected to EFA after the suitability of the data in the development of the subsample (n = 330) was assessed. The Kaiser-Meyer-Olkin measure (KMO) of sampling was 0.85, which exceeds the recommended value of 0.60. Barlett's test of sphericity was statistically significant (χ 2 = 1243.98, df = 45, p < .001), which indicates highly acceptable factorability of the data and correlation matrix (Tondeur et al., 2017). In the EFA, all 19 items were subjected to principal component analysis with varimax orthogonal rotation, which revealed two factors with eigenvalues over 1. Nine items had factor loadings of less than the recommended value of 0.06 (Tabachnick & Fidell, 2007); hence, they were removed from the two-factor model; 10 items remained accounting for 72.08% of the total variance (see, Table 2). The remaining items had factor loadings ranging from 0.60 to 0.80, suggesting that they provided a good measure of their respective factors (Hair et al., 2019). Furthermore, they had Cronbach's alphas of 0.92, which is above the threshold value of .70, implying that all the items in the two-factor model were internally consistent (Tabachnick & Fidell, 2007).
Seventeen items measuring causes of stress were subjected to EFA after the suitability of the data in the subsample of 330 was assessed. The KMO of sampling was 0.89 and the Barlett's test of sphericity was statistically significant (χ 2 = 2660.38, df = 136, p < .001). A principal component analysis with varimax orthogonal rotation produced a two-factor model with eigen values greater than 1. Two items had factor loadings less than 0.40 and hence were discarded, resulting in seven items under factor 1 and eight items under factor 2. These items were subsequently subjected to CFA.
Finally, the 12 items measuring stress coping strategies were also subjected to EFA after the suitability of the data in the subsample (n = 330) was assessed. The KMO of sampling was 0.74 and the Barlett's test of sphericity was statistically significant (χ 2 = 868.880, df = 66, p < .001). Principal component analysis with varimax orthogonal rotation produced a three-factor model; however, the items under the third factor had items with factor loadings of less than 0.50, so these items were discarded, resulting in a two-factor model with seven items under factor 1 and three items under factor 2. These items were further subjected to CFA for cross-validation.

Confirmatory factor analysis (CFA)
The two-factor measurement for "digital skills" extracted from the EFA was subjected to CFA using the cross-validation subsample (n = 331) with the SPSS AMOS software. The CFA had adequate goodness-of-fit statistics (χ 2 (df) = 301.97(34), RMSEA = .045, CFI = .943, TLI = .925). The following measures were reported: The root-mean-square error of approximation (RMSEA), comparative fitness index (CFI) and Tucker-Lewis Index (TLI). These measures should be greater than 0.90 with exception of RMSEA whose value should be less than 0.05 for a good model fit (Hair et al., 2019;Wolf et al., 2018). Hence, a two-factor structure comprising 10 items was maintained, as the identified exploratory factor structure fit the data. We labelled the two factors (see, Table 2) "basic Office skills" (factor 1) and "internet skills" (factor 2). The two-factor measurement (15 items after the EFA) for causes of stress was also subjected to CFA, using the cross-validation subsample (n = 331). The results of the CFA revealed five items under factor 1 and six items under factor 2 (see , Table 3) had adequate goodness-of-fit statistics (χ 2 (df) = 276.05(89), RMSEA = .048, CFI = .912, TLI = .902). Factor 1 was labelled "limited resources-related causes of stress" and factor 2 was labelled "academic-related causes of stress". Lastly, the two-factor measurement for "coping strategies" extracted from the EFA was subjected to CFA using the cross-validation subsample (n = 331). The CFA results identified three items under each of factor 1 and factor 2 (see, Table 4) that had adequate goodness-of-fit statistics (χ 2 (df) = 61.843(26), RMSEA = .044, CFI = .921, TLI = .910). We labelled factor 1 "giving up" and factor 2 "positive framing" (see ,  Table 4).

Device(s) used to access online platforms
Online teaching and learning activities require both tutors and pre-service teachers to possess digital devices to access online platforms. To examine the pre-service teachers' use of device(s) in accessing online media, we asked them to indicate the frequency with which they used device(s) to access online learning platforms. The participants' responses, shown in Table 5, indicate that a reasonable number of the pre-service teachers (22.8% of males and 23.2% of females) used smartphones daily to access online media for teaching and learning purposes. The table further shows that 90.8% of males and 85.2% of females never used a desktop computer, and 68.2% of males and 80.2% of females never used a laptop computer to access online learning platforms. It is worth noting that 29% of the male pre-service teachers and 40.5% of the females borrowed a smartphone daily, weekly or monthly to access online teaching and learning platforms.

Pre-service teachers' digital skills
We explored pre-service teachers' digital skills with 10 items comprising two subscales. The descriptive statistics, p-values, and effect sizes are presented in Table 6. Generally, the preservice teachers reported low average digital skills, considering the overall means (2.15-2.78) for the two subscales (i.e. "basic Office skills" and "internet skills"). The males had higher overall mean scores on the subscale "basic Office skills" than the females (M = 2.57, SD = 0.83 and M = 2.15, SD = 0.75, respectively). The highest mean scores for both males and females were obtained for the item "I know how to use a table in a document (e.g., using word)" (M = 2.67, SD = 0.95 and M = 2.22, SD = 0.83, respectively). The lowest scores for both males and females were obtained for the item "I know how to use a spreadsheet to plot a graph (e.g., using Excel" (M = 2.23, SD = 0.94 and M = 1.93, SD = 0.78, respectively). An independent samples t-test was conducted for all subscales. For "basic Office skills", there was a statistically significant difference between males (M = 2.29, SD = 0.69) and females (M = 2.83, SD = 0.81) t(657) = 8.5, p = .001 (two-tailed).
The magnitude of the difference in the means was very moderate (d = .08). For the internet skills' subscale, the overall mean score was higher for males (M = 2.78, SD = 0.64) than females (M = 2.34, SD = 0.62). The lowest mean score in this subscale for both males (M = 2.62, SD = 0.81) and females (M = 2.25, SD = 0.82) was for the item "I find it easy to decide what the best keywords are to use for online searches". The highest mean score for males (M = 3.14, SD = 0.69) and females (M = 2.84, SD = 0.76) was for the item "I know how to open downloaded files". However, there was

Relationship between digital skill level and stress
To examine the relationship between digital skill levels and stress, we explored the causes of stress among the pre-service teachers during the period of remote online teaching and learning (see, Table 7). The causes of stress were categorised as "limited resources-related" and "academicrelated". The highest mean score under "academic-related causes of stress" was for the item "Inability to concentrate during online lectures", while the lowest mean score was for the item "Anxiety about performance on exams" for both males and females (M = 2.84, SD = 0.68 and M = 2.83, SD = 0.76, respectively) . The overall mean score for "academic-related causes of stress" was above average for both males (M = 2.95, SD = 0.69) and females (M = 3.07, SD = 0.70), which is similar to the scores for "limited resources-related causes of stress".
Under "limited resources-related causes of stress", the item with the highest mean score for both males and females was "Financial problems" (M = 3.38, SD = 0.78 and M = 3.28, SD = 0.81, respectively) followed by 'Limited internet access (males: M = 3.37, SD = 0.80; females: M = 3.31, SD = 0.81). The scores for these two items indicate that pre-service teachers experienced stress caused by limited financial resources and limited access to the internet. Other causes of stress included 'Worries about the future (males: M = 3.01, SD = 0.99; females: M = 2.99, SD = 1.01) and 'Limited access to digital devices (males: M = 3.14, SD = 0.91; females: 3.12, SD = 0.91). Under "academic-related causes of stress", there was a statistically significant difference between males (M = 2.95, SD = 0.69) and females (M = 3.07, SD = 0.70) t(657) = 7.5, p = .003 (two-tailed), although the magnitude of the difference in the means was minimal (d = 0.01). However, there was no statistically significant difference between males and females under "limited resources-related causes of stress".
The correlation among the items measuring the level of digital skills and the items measuring stress indicates a general negative relationship (see , Table 8). For example, there is a weak negative correlation (r = −.014, n = 661, p < .01) between "basic Microsoft Office skills" and "academic-related causes of stress". However, a strong positive correlation was found between the scores on the subscale measuring digital skills and the subscale measuring stress. Furthermore, there was a negative correlation between "internet skills" (r = −.078, n = 661, p < .01) and "limited resources-related" causes of stress. Overall, there was a negative correlation between digital skills (r = −.659, n = 661, p < .01) and stress, indicating that as the digital skills of the pre-service teachers increased, the stress associated with the level of digital skills during the period of online learning decreased.
We further asked the pre-service teachers to indicate the strategies they used to cope with stress during the period of online teaching and learning (see, Table 9). The mean scores and standard deviations indicate that the pre-service teachers used both positive framing and giving up strategies to cope with stress. For example, both males (M = 2.64, SD = 0.95) and females (M = 2.74, SD = 0.93) indicated that they received emotional support/advice from friends and family. Table 9 shows that the pre-service teachers used positive framing, such as 'seeing something good in what is happening and learning from the experience (males: M = 2.78, SD = 0.88; females: M = 2.65, SD = 0.88). The pre-service teachers generally expressed disagreement to strong disagreement in response to the "giving up" coping strategies. For example, the mean score for males (M = 1.38, SD = 0.66) and females (M = 1.46, SD = 0.75) for the item "Using tobacco/alcohol/ drugs/ to feel better" was low, indicating that the pre-service teachers did not use drugs or alcohol to cope with stress.

Discussion
This study found that most of the pre-service teachers used smartphones to access online learning platforms. Smartphones have features that make them easy to use to access online media; however, as argued by Nurhudatiana and Ce (2018), when using smartphones, users tend to switch to other websites, mainly social media, which could distract pre-service teachers from  paying attention to the content of online learning platforms. This finding is consistent with the findings of other studies (Barrot et al., 2021;Murugesan & Chidambaram, 2020), which found that most higher education students prefer to use their smartphones rather than other devices to access online learning platforms. Our data also indicate that a significant number of the preservice teachers borrowed either a smartphone or a laptop to access online media. Students' lack of access to digital devices limits their effective participation in online learning (Ogbonnaya et al., 2020), possibly resulting in students missing at least some of the content of the online sessions. In addition, during the COVID-19 pandemic, higher education students, especially in sub-Saharan Africa (SSA), have faced challenges in remote teaching/learning, including erratic power supply, unreliable internet connectivity, the cost of data bundles, and limited access to digital devices (Adarkwah, 2021;Agormedah et al., 2020).
Our data revealed that, generally, the pre-service teachers' mean scores were slightly below average, suggesting that their digital skills are below the expected digital literacy skills for trainee teachers in the twenty-first century (Çebi & Reisoglu, 2020). Furthermore, the female pre-service teachers showed both lower basic Microsoft Office (e.g., MS Word, Excel, etc.) digital skills and lower "internet skills" than their male counterparts. These findings are similar to the results of Agyei and Voogt (2011) and Çebi and Reisoglu (2020), in which stronger digital skills were observed among the male pre-service teachers than among their female counterparts, and Apori's (2019) results, which showed lower digital competencies among female teachers than their male counterparts in basic schools in Ghana. The current study revealed that the female teachers had limited skills in MS Office and internet skills compared with their male counterparts. These are essential skills that are required to perform many digital tasks in any online learning environment. Studies in India and Brazil revealed that a lack of digital skills is the fundamental reason that low-income groups are not using the internet. In one study, women were 1.6 times more likely than men to report a dearth of skills that impeded internet use (UNESCO, 2019). Stress is inevitable when students with low digital skills engage in digital activities.
Our results showed that the most prominent causes of stress among pre-service teachers during online learning were related to limited resources. The pre-service teachers cited "limited internet access", "difficulty using digital skills on the learning platforms" and "limited access to digital devices", suggesting that most of the pre-service teachers were experiencing online learning for the first time. In addition, had it not been for the pandemic, they would have continued attending face-to-face classes, which are the norm in Ghana's higher education institutions. The online learning sessions required students to download materials online and use MS Word, Excel and PowerPoint to process materials to submit assignments and projects. Hence, if students had limited basic MS Office digital skills, they would necessarily experience a lot of stress completing their assignments. Unfortunately, stay-at-home mandates leading to the closure of schools did not allow pre-service teachers to learn new digital skills collaboratively with their peers, which undoubtedly increased their stress levels.
Even though the pre-service teachers experienced stress caused by "dissatisfaction with online learning", nonetheless, one significant finding of this study was the correlation between the students' digital skill levels and the stress that they experienced. There was a strong negative correlation between digital skill levels, limited resources, and academic-related causes of stress. This implies that the more digital skills a student has, the less stress they experience, and vice versa. This finding is consistent with previous studies, which showed that pre-service teachers with more technological skills experienced less stress and anxiety compared to their counterparts with fewer technological skills, who experienced higher levels of stress (Agyei & Voogt, 2011;Sang et al., 2010). However, these previous studies collected data from pre-service teachers attending regular face-to-face classes. This means that students with fewer digital skills who are confronted with online learning situations may experience increased stress levels associated with academic pressure, anxiety about the future, performance on examinations, etc.
Despite the stress associated with online learning, most pre-service teachers used positive framing strategies to cope. This finding is consistent with other studies (Awoke et al., 2021;Baloran, 2020). In addition, students saw the situation as an opportunity to learn new skills and sought support from family and friends to avoid the negative effects of prolonged stress. On the other hand, some pre-service teachers used negative coping strategies, such as alcohol/drug use, giving up, etc., to reduce stress. This suggests that some of the pre-service teachers were not educated on managing stressful situations, which may have exposed them to prolonged stress.

Conclusions and recommendations
The COVID-19 pandemic put the teacher education sector squarely within the technological ecosphere, as digital skills and internet connectivity became non-negotiable for all educational institutions. Digital skills form the backbone of two of the indicators of the United Nations' sustainable development goal (SDGs) no. 4 (4.4.1 and 4.4.2). In the twenty-first century, these skills are indispensable, equating to numeracy and literacy in our fast-changing technological world. Our study found low digital skills and associated stress experienced during remote/digital learning among pre-service teachers. This finding highlights the barriers or hotspots in the use of ICT in preparing teacher educators and pre-service teachers to work in basic schools in Ghana. These hotspots or barriers could have gone under the radar of recent teacher education reforms. According to the 2016 teacher education reforms in Ghana, digital skills are core and transferable skills essential for all pre-service teachers to acquire to be effective teachers by the end of their training (Ministry of Education, 2017).
Furthermore, the promotion of digital literacy is at the heart of Ghana's Education Strategic Plan and the new standard-based curriculum for basic schools. However, we conclude that teacher preparation in Ghana has been slow to incorporate digital technologies to promote minimum digital proficiency among pre-service teachers, resulting in stress experienced by pre-service teachers in performing digital activities. Besides stressors associated with low digital skills in online learning, unreliable internet connectivity, the cost of data, the lack of appropriate devices, and an erratic power supply impede pre-service teachers' use of technology in learning. Therefore, we recommend that institutions provide workshops for students and instructors to help them upgrade their digital skills so that they can confidently use digital technologies, including YouTube, webinars, etc. Furthermore, Ghana's investment in technological infrastructures such as internet connectivity, Wi-Fi connections, and smart devices for students is necessary since learning must continue despite the uncertain future for the global community.

Limitations
This study collected data on pre-service teachers' self-reported digital skills through an online survey; it would have been helpful to focus on measuring the digital skills of the student teachers in real time. Also, the participants were pre-service teachers in colleges of education that offer education as a major; perhaps a university-wide study of students in different programmes would have provided different results. Therefore, we suggest that the interpretation of the findings is limited to the context of this study. We recommend that a longitudinal study following pre-service teachers from their first year to their final year be carried out to gain a deeper understanding of these students' digital skills and stress. Research could also be carried out to examine teacher educators' or lecturers' digital skills in the tertiary education landscape.