International psychological research addressing the early phase of the COVID‐19 pandemic: A rapid scoping review and implications for global psychology – Wiley

Working off-campus? Learn about our remote access options
Corresponding Author
Martin Obschonka
School of Management, Queensland University of Technology, Brisbane, QLD, Australia
Qiyue Cai
Department of Psychology, Arizona State University, Tempe, AZ, USA
Athena C. Y. Chan
Department of Family Social Science, University of Minnesota Twin Cities, Minneapolis, MN, USA
Scott Marsalis
University Libraries, University of Minnesota Twin Cities, Minneapolis, MN, USA
Sydni A. J. Basha
Department of Psychology, Arizona State University, Tempe, AZ, USA
Sun-Kyung Lee
Department of Psychology, Arizona State University, Tempe, AZ, USA
Abigail H. Gewirtz
Department of Psychology, Arizona State University, Tempe, AZ, USA
Department of Family Social Science, University of Minnesota Twin Cities, Minneapolis, MN, USA
Corresponding Author
Martin Obschonka
School of Management, Queensland University of Technology, Brisbane, QLD, Australia
Qiyue Cai
Department of Psychology, Arizona State University, Tempe, AZ, USA
Athena C. Y. Chan
Department of Family Social Science, University of Minnesota Twin Cities, Minneapolis, MN, USA
Scott Marsalis
University Libraries, University of Minnesota Twin Cities, Minneapolis, MN, USA
Sydni A. J. Basha
Department of Psychology, Arizona State University, Tempe, AZ, USA
Sun-Kyung Lee
Department of Psychology, Arizona State University, Tempe, AZ, USA
Abigail H. Gewirtz
Department of Psychology, Arizona State University, Tempe, AZ, USA
Department of Family Social Science, University of Minnesota Twin Cities, Minneapolis, MN, USA
We thank Michael Beckstrand, PhD, LATIS, University of Minnesota, for creating the network visualisations, and Kate Carlson, U-Spatial, University of Minnesota for creating the choropleth maps. Funding for the use of COVIDENCE was provided by non-sponsored funding at the University of Minnesota to Dr. Gewirtz.
Give access
Share full-text access
Use the link below to share a full-text version of this article with your friends and colleagues. Learn more.
Share a link
In March 2020, the World Health Organization (WHO) declared COVID-19 a global pandemic. Given that keeping abreast of international perspectives and research results is of particular importance for such massive global emergencies, we employed a scoping review methodology to rapidly map the field of international psychological research addressing this important early phase of the pandemic. We included a total of 79 studies, with data mostly collected between March and June 2020. This review aimed to systematically identify and map the nature and scope of international studies examining psychological aspects of the unfolding COVID-19 pandemic. We mapped key research themes, subfields of psychology, the nature and extent of international research collaboration, data methods employed, and challenges and enablers faced by psychological researchers in the early stages of the pandemic. Among the wide range of themes covered, mental health and social behaviours were the key themes. Most studies were in clinical/health psychology and social psychology. Network analyses revealed how authors collaborated and to what extent the studies were international. Europe and the United States were often at the centre of international collaboration. The predominant study design was cross-sectional and online with quantitative analyses. We also summarised author reported critical challenges and enablers for international psychological research during the COVID pandemic, and conclude with implications for the field of psychology.
Over the more than 120-year history of psychological science as an empirical field, international collaboration and exchange have played a vital role in the progress of the field (Sabourin & Cooper, 2014; Silbereisen et al., 2014; Stevens & Gielen, 2007). The process of internationalisation has made significant contributions to psychology and beyond (e.g., education and practice), yet various barriers hampering such international efforts still exist for the psychological research community (Berry, 2013; van de Vijver, 2013). While many of these barriers are not limited to psychological science but transcend the behavioural sciences (Henrich et al., 2010), psychology as the scholarly investigation of mind and behaviour, has a special mandate to be representative of humanity and the respective relevant contexts, both in terms of the diversity and universality of psychological phenomena. At the same time, scholars exploring the internationality of the field continue to stress that “psychology still has a long way to go to become a science truly representative of human beings” (Thalmayer et al., 2021, p. 116).
Among the plethora of topics and themes that have been addressed in international psychological research are those studies that represent a relatively small yet particularly important niche in the literature—those devoted to the topic of major global crises (e.g., Doherty & Clayton, 2011). As stressed by Stevens & Zeinoun (2013, p. 758), “although generally overlooked by the discipline, international psychology maintains interest in such global concerns as inter-group conflict and conflict resolution, societal transformation and national development, threats to the natural environment, physical and mental illness.” In this field, of particular urgency is research addressing ongoing global health crises affecting (and affected by) psychological processes and outcomes across populations and contexts (e.g., Liamputtong, 2013; Murphy-Berman & Berman, 1993). International perspectives and research designs are, for example, of particular relevance for understanding and fighting immediate health urgencies that transcend national and cultural borders (Rosenfeld et al., 2021). Psychology as an international, global science has been highlighted as “essential science” in the prevention and treatment of urgent global health crises (Kazak, 2020). This has become particularly evident in the context of the outbreak and early phase of the COVID-19 pandemic (e.g., Gelfand et al., 2021), as an unprecedented, massive and uncontrolled health threat to humanity (World Health Organization, 2020).
The outbreak and spread of COVID-19 has become an immediate research priority for the psychological research community around the globe. It is, however, unclear to what extent this manifested for international research. While important work has been published in this field, anecdotal evidence also suggests that (potential) immediate international psychological research (e.g., international research collaborations and study designs) might have faced particular and new challenges during the outbreak and early spread of the pandemic, factors that may not only accentuate and go beyond the structural barriers typically discussed in international psychology (Berry, 2013; Thalmayer et al., 2021; van de Vijver, 2013), but particularly affect the research community in low-to-middle-income countries (Cai et al., 2021; Fry et al., 2020; Obschonka et al., 2021).
We followed the scoping review framework established by Arksey and O’Malley (2005) and extended by Peters et al. (2020). Employing the methods specified by the Preferred Reporting Items for Systematic Reviews and Meta-Analyses Extension for Scoping Reviews (PRISMA-ScR; Tricco et al., 2018), we established a review protocol before conducting the literature search and pre-registered the project in the Open Science Framework (OSF) Registries (
The searches were conducted in three bibliographic databases: APA PsycInfo (Ovid), Scopus, and the Social Science Citation Index and Emerging Sources Citation Index segments of Web of Science. A combination of keywords and subject heading queries was customised in each database to identify studies of interest. Boolean and proximity operators, as well as truncation and lemmatization as available, were utilised to develop a highly sensitive search. Generally, the strategy used terms in two categories: (a) COVID-19 OR pandemic; AND (b) cross country differences OR regional differences OR international. In Scopus and Web of Science, the search was limited to the Psychology subject segment. In APA PsycInfo, we also included a strategy to identify records where more than one geographic code was applied. Appendix 1 provides a narrative further explicating our approach as well as the complete strategies for each database, including the list of search terms used. Literature search results were imported into Covidence (, a software programme that supports deduplication, collaboration among reviewers during the screening process, recording decisions and creation of a PRISMA flow diagram upon screening completion.
Included studies met the following inclusion criteria: (a) published from 2020 until date of search in peer-reviewed journals; (b) had at least an English-written abstract, and the full text written in languages that our team could read: English, Chinese and Korean; (c) empirical studies using quantitative, qualitative or mixed-methods research designs; (d) involved human subjects with no limitations on age; (e) considered more than one country in their data analysis; and (f) included explicit investigation of the psychological implications/effects of the COVID-19 pandemic. Studies were excluded if they (a) did not provide an English abstract, and/or the full text not written in languages our team could read; (b) were purely theoretical/conceptual in nature (no empirical study); (c) involved non-human subjects only; (d) were based on data from a single country; (e) did not compare participants from different countries (e.g., treated the sample as a whole and analysed the data in aggregate); or (f) investigated only non-psychological aspects of COVID-19.
The literature search was conducted between 28 March and 2 April 2021, and we identified 956 records after removing duplicates. These records were imported into Covidence for two-stage screening. At the title/abstract screening stage, each record was randomly assigned to and screened by two screeners independently after pilot screening and training. Disagreements were discussed and resolved by all five screeners. The title/abstract screening yielded 152 records with inter-coder reliability of 89.4%. At the full-text screening stage, four reviewers were first trained by screening a subset of articles. This process was used to refine the criteria and establish consensus among reviewers. After achieving adequate reliability, reviewers screened the articles independently with each article being reviewed by at least two reviewers with inter-coder reliability of 76.8%. Disagreements were resolved by consensus with the rest of the team. All exclusions during title/abstract and full text screening were recorded and summarised in Figure 1. A total of 79 articles constituted the final sample in this review. Forty-two were published in 2020 and 37 were published in 2021. Most studies reported they collected data from March to June 2020, and only three studies collected data after October 2020.
A standardised Qualtrics form with coding manual was developed to extract key study characteristics. The form was pilot tested with 17 (21%) randomly selected articles, and each was independently coded by four coders. Disagreements were discussed at the group meeting with all authors and the form was refined. After finalising the data extraction form, four coders were each randomly assigned 12–19 articles and independently coded them. After the data extraction process, the second author (QC) combined data and reviewed all key variables and discussed any changes with the group. The final extracted data can be found in Table S1 in the appendix and under
Extracted data included article information (i.e., APA-style citation), international collaboration (i.e., countries of the authors’ affiliated institutions, countries/regions of study sample), core research interests of the study (i.e., research question/purpose of study, psychological subfield and COVID-19-related measurement), methodology (i.e., data collection period, study population, basic demographics, sample size of overall studies and maximum/minimum number by country/region, data collection methods, data analytic strategies), cultural aspect (i.e., cultural-related measurement) and challenges of the research processes and future direction for conducting COVID-19 pandemic studies. For our analysis, we used the United Nations’ (UN) country list ( Since some papers reported the number of total countries involved in the data analysis but did not report all countries’ names, we could only code the reported countries (following the UN country list).
The synthesis was both quantitative (i.e., descriptive statistics including frequency and Means) and qualitative (i.e., content analysis; Elo & Kyngäs, 2008), mapped in tables and other visualisations (i.e., networks, choropleth, word cloud). We addressed the research questions (RQ) as follows:
(1) Key aims and scope of international psychology research. First, a word cloud using words from the article authors’ keywords was generated to visualise the key themes/topics. Second, we categorised the main research questions in the articles into different (combinations) of psychological subfields, such as clinical/health, social, personality, educational, measurement, organisation/industry.
(2a) Internationality of the author team. First, we described the national makeup of the research teams (authors), for example, whether the research team was international and the number of countries with respect to the main affiliation of each coauthor in a study. Network analysis was conducted in Gephi (Bastian et al., 2009; to display the international collaborative networks. Network analysis can map the connections (edges) among entities (nodes, in our case are countries), which allows us to identify the members within the network, the connectedness of collaboration and the most frequently appearing countries within the network. We signified these as undirected networks since the connections among countries did not have directions, and we reported the number of nodes, level of connectivity (edge weight) and centrality of the network. The number of nodes was used to identify how many countries were included in the network. Connectivity was measured using average and median degree, path length and density. Degree of a country is defined as the total number of links connected to the particular country. Path length is defined as the distance (i.e., shortest path) between two countries, and it is measured by the number of connections between them. Network density is defined as the ratio between the existing numbers of connections compared to all connections possible. Three types of centrality were reported, including degree centrality, betweenness centrality and closeness centrality. Degree centrality measures how many connections a country has, betweenness centrality measures how many times a country lies on the shortest path between other pairs of countries, and closeness centrality measures the mean distance from a specific country to other countries.
(2b) Internationality of the data analysed. We mapped the countries considered in the data in the articles (e.g., via a global heatmap). As in (2a) we employed the same network analysis process to analyse and map international collaborative networks with respect to the data used.
(3) Cross-cultural measures. Although all included studies presented data from multiple countries, we suspected not all took an explicitly cross-cultural perspective, for example by measuring culture. We first coded whether the articles took an explicitly cross-cultural approach, if so, we also coded how they measured cultures.
(4) Study methodology. We reported the descriptive statistics of the key variables, including sample demographics, study design, methods of data collection and data analysis. Coders derived summarising text from each article regarding the method of gathering international data. We then used qualitative content analysis on this text to identify key themes and concepts to categorise the strategies that psychologists used to gather international data.
(5) Challenges in the research process itself during the pandemic and future directions. Coders extracted text from each article describing specific study challenges associated with the unfolding pandemic. Content analysis was used to analyse these responses to identify and categorise common challenges and future directions.
We extracted 376 authors’ keywords from the original articles. To better understand the study aims and psychology subfields involved, we removed words referring to the COVID-19 pandemic (e.g., “COVID-19 (pandemic),” “pandemic,” “SARS-COV-2,” “corona” and “coronavirus”), and country/region-related words (e.g., “European countries,” “Arab countries and MENA region” and other special country names). We used free online software ( to generate a word cloud (Figure 2) of all keywords used. The most frequently used words (included in more than five articles) were “anxiety,” “mental health,” “social,” “depression,” “fear,” “health,” “cross-cultural,” “behaviour” and “stress”. Hence, most studies were concerned with mental health and social behaviours, but there was a wide range of other topics covered as well.
Furthermore, we coded the main research questions mentioned in the studies and then categorised studies into different fields of psychology, including clinical/health, social, personality, educational, measurement, cross-cultural, political and industrial/organisational. Consistent with the word cloud, most studies can be assigned to the field of clinical or health psychology (n = 63, 80% of all studies; for details see Table S1 in the appendix). Several example topics included the impact of the COVID-19 pandemic on mental health (e.g., Al Omari et al., 2020) and mental health services (e.g., Thome et al., 2020), COVID-19-related fear (e.g., Caycho-Rodriguez et al., 2021) and exercise behaviours during the pandemic (e.g., Brand et al., 2020). The second most popular field was social psychology (n = 22), and examples included social influences on preventive behaviours (e.g., Tunçgenç et al., 2021), people’s perceptions, attitudes and behaviours towards the pandemic (e.g., Galasso et al., 2020), and prosocial behaviours during the pandemic (Jin, Balliet et al., 2021). Other subfields of psychology included personality psychology (n = 8), educational psychology (n = 4), methods (n = 2) and others (n = 2).
Authors’ countries were coded using the country names of each author’s first affiliation from the author information in the journal articles. Authors of the 79 studies came from a total of 68 UN-recognised countries and 58 studies (73%) had international author teams (i.e., authors’ first affiliations were from different countries). Most international teams involved two (n = 24) or three countries (n = 13). The maximum number of countries represented by the authors in a single study was 20 (in two studies: Gloster et al., 2020; Thome et al., 2020).
Figure 3 displays the network of countries represented by the authors in the 58 studies with international author teams, using MultiGravity Force Atlas2 layout (Jacomy et al., 2014) in Gephi, coloured by continent (see the legend in Figure 3). All connections were undirected. The data underlying this figure referred to the collaborations within each study (how international each study was). For Figure 3 we analysed and displayed those 30 countries that were represented three times or more. The size of each node is scaled based on its degree, that is, how many papers a country was included in the author team out of the total 79 papers. The positions of the countries were set based on the centrality measurements, in other words, central author countries were those countries (a) that had many connections; (b) that had relatively strong connections (higher weight); and (c) where the nodes (countries) they were connected to also scored highly on points (a) and (b).
Note: Countries in the same continents are identically coloured, using the MultiGravity Force Atlas2 layout in Gephi. The percentage in the legend refers to the frequency of nodes (countries) in the network. The position of the country in the network shown indicates how central/peripheral the country was at the network, based on the centrality measures. The size of the country indicates the frequency of the total number of links connected to the particular country. The width of the links between the countries indicates the frequency of connections between the two specific countries.
After filtering, the average degree for each node (country) was 13 (median = 12.5, range = [1, 24]), meaning that on average, one country was connected to 13 other countries. The average path length was 2, indicating the average shortest path between two countries was two edges (i.e., via another country). The network density was 0.255, which refers to the ratio between the existing numbers of connections compared to all connections possible. To find the most “important” countries in the network, we calculated the degree centrality, betweenness centrality and closeness centrality. The top countries with the highest degree centrality were Italy, Spain, the United States, UK (all tied), followed by Finland, Hungary, Ireland, the Netherlands and Portugal (all tied); those with the highest betweenness centrality were the United States, followed by Spain, Portugal, Finland, Canada, Italy and UK, and those with the highest closeness centrality were Italy, Spain, the United States, UK (all tied), followed by Portugal, Finland, Ireland and the Netherlands (all tied). Thus, Western European countries and the United States were relatively central in international author collaborations within the 58 studies in our review with international author teams.
Data from a total of 114 countries were included across the 79 studies (see Table S1 in the appendix). When focusing on the mere global distribution of the data analysed across the studies (illustrated in the heatmap in Figure 4), one can see that most studies analysed data from North America, Europe, Australia and China, while few studies analysed data from other continents and regions (e.g., Africa). We used World Bank (2020) data to cluster these countries represented in the data into high income, upper middle income, lower middle income and low income countries. About half were high-income countries (n = 51), 35 were upper-middle-income countries and 22 were lower-middle-income countries. Five low-income countries (Uganda, Afghanistan, Burundi, Ethiopia and Syrian Arab Republic) were included once or twice in the total 79 studies. Among 41 countries that were included in more than 10 studies, 29 were high-income countries, 8 were upper-middle-income countries, 4 were lower-middle-income countries, and none were low-income countries.
We then took a closer look at geographic overlap between the origin of the data and the local severity of the pandemic in this early phase. Figure 5 shows cumulative cases of COVID-19 until 30 June 2020. The specific time was chosen because most of the included studies analysed data collected between March and June 2020. The overlap was indeed substantial, many studies were conducted in countries/regions where the COVID-19 pandemic was more severe in that phase, such as North America, Latin America, Europe and India. In contrast, in Asia and Oceania, especially China, Australia and New Zealand, relatively many studies were conducted while the pandemic was less severe.
Note: Data from the COVID-19 Data Repository by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University (
For the quantitative studies and mixed-methods studies (n = 73), authors analysed data from a median of 5.5 countries within a study (M = 17, SD = 26), with a median sample size of 1784 (M = 8704, SD = 20,212). Sample sizes were not evenly distributed among countries with samples sizes ranging on average from a minimum sample size of 221 (range = 1–3504), to a maximum of 759 (range = 107–11,568). Twelve studies did not report the minimum sample size and nine did not report the maximum sample size. There were six studies that explicitly indicated their minimum sample size was one, although this number might be higher given 12 studies did not report this information. This is indicative of non-representative international samples and casts a shadow on the robustness of the results. Nevertheless, some articles were able to use statistical methods to deal with that problem. For example, in a paper with 54,845 participants from 112 countries, the authors conducted sensitivity analyses using 23,278 participants representative of 24 countries to improve the representativeness and robustness of the results (e.g., Han et al., 2021). Other methods included excluding countries with small sample size (e.g., Galasso et al., 2020) and combining similar countries into areas/regions, e.g., using continent (e.g., Gato et al., 2021), Latin America (e.g., Ruiz et al., 2021), North America (e.g., Jovančević & Miličević, 2020) and Europe (e.g., Garbe et al., 2020) instead of specific countries. Qualitative research studies included smaller sample sizes; with five out of six including fewer than five countries, and the median sample size was 24 participants (M = 81, range = 3–351). One study interviewed 23 experts from 23 countries (e.g., Thome et al., 2020).
Figure 6 displays the networks of reported countries included as samples in the 79 international psychology studies. All connections were undirected. The original data had more than 4500 edges and the Fruchterman Reingold algorithm was used to get the circular shape while emphasising the core versus periphery type relationships. Filters were again used and only node frequencies >4 and edge weights >4 were displayed. Like for Figure 3, the data underlying this Figure 6 referred to the cross-country data within each study (how international each study was).
Note: Different colours denoted different continents, using the Fruchterman Reingold layout in Gephi. The percentage in the legend refers to the frequency of nodes (countries) in the network. The position of the country in the network shown indicates how central/peripheral the country was at the network, based on the centrality measures. The size of the country indicates the frequency of the total number of links connected to the particular country. The width of the links between the countries indicate the frequency of connections between the two specific countries.
After filter, the average degree for each node (country) was 33 (median = 31, range = 3–47), meaning that on average, one country was connected to 30 other countries. The average path length was 1.44, indicating the average shortest path between two countries was 1.44 edges (i.e., via another country). The network density was 0.586, which refers to the ratio between the existing numbers of connections compared to all connections possible. To find the most “important” countries in the network, we again calculated the degree centrality, betweenness centrality and closeness centrality. The top countries with the highest degree centrality and closeness centrality were Germany, Spain, UK, followed by tied France, Sweden and Romania, and then tied United States and the Netherlands; those with the highest betweenness centrality were the Germany, Spain, France, Sweden, UK and Italy. Again, Western European countries and the United States were relatively central in international author collaborations within the 79 studies in our review with international author teams.
We should be cautious about directly comparing the two network analyses (Figures 3 and 6) because they used different data, however, from a network perspective, it can be inferred that the author teams were more dominated by developed countries, especially countries in Europe and North America, while the makeup of sample nationality displayed a more diverse level of internationalisation, with more countries in total as well as more developing countries. The author network is also more compact, with a smaller group of countries, while each country is more important to the whole network. In contrast, the sample network is more spread out since there are more countries involved and each country has less impact on the whole network.
Only a few had explicitly included and analysed cross-cultural measures. For example, a small number of studies conducted cross-cultural qualitative studies (e.g., von Humboldt et al., 2020) to address research questions linked to universality and diversity across countries. Most of the quantitative and mixed-methods studies actually compared countries instead of cultures (n = 58), for example, Kochuvilayil et al. (2021) compared Australian and Indian nursing students, suggesting that culture can influence how nursing students experienced and coped during the pandemic; however, culture was not measured. Only 12 studies took an explicitly cross-cultural perspective. The constructs under measurement included individualism versus collectivism (e.g., Tang et al., 2021), cultural distance (e.g., Kapoor & Tagat, 2021) and worldviews on nature (e.g., Haas et al., 2021). Gokmen et al. (2021) found cultural differences to predict COVID-19 severity in very highly developed European countries using measurement of six cultural dimensions. Ji et al. (2020) carefully compared the meaning of suffering in Canadian and Chinese participants and found cultural differences in dialectical thinking.
Most studies used quantitative cross-sectional designs and collected data via online surveys. Seventy out of 79 (89%) studies were quantitative, 6 were qualitative (8%) and 4 (4%) were mixed-methods. Sixty-eight (86%) were cross-sectional studies while 11(14%) were longitudinal studies. Given the pandemic, it is not surprising that almost all researchers collected data online, including online surveys, telephone interviews and virtual meetings. A few studies utilised secondary data including from data repositories (e.g., Salvador et al., 2020), hospital records (e.g., Ougrin et al., 2021) and publicly available data (e.g., Gruchola & Slawek-Czochra, 2021).
The studies collected data from a variety of populations. Over half of the studies collected data from the general population and no specific limits were set, while some targeted specific age ranges from childhood to older adulthood. Professions surveyed and/or interviewed included physical and mental healthcare workers, frontline workers, general employees, police officers, teachers and university students. Other specific populations of interest included people with pre-existing medical conditions, parents and minorities such as LGBTQ+. Detailed demographic information can be found in Table S1 in the appendix.
Content analysis was conducted to categorise strategies authors employed to facilitate collecting international data. Two themes were identified for studies with primary data: Adaptive measurement tools and recruitment. Categories for measurement tools included language, multiple platforms in different countries and culturally adapted surveys. Among those, language might be one of the most important strategies when conducting international psychological research, and 32 (41%) mentioned that translations to native languages were done or multiple language versions were available, while 9 (11%) explicitly mentioned that no translation was needed since all data were collected in countries speaking the same official language, such as English, Spanish and Arabic. However, some studies just gathered data in English and limited international participants to those who were English-speaking, leading to less representative samples (e.g., Lippold et al., 2020). Categories for recruitment included using social media (n = 26) to disseminate information, snowball sampling, professional networks, existing international participants cohorts, and quota. Other methods to recruit population representative samples included selecting participants who met quotas based on demographics in web-panels such as Qualtrics (e.g., Taylor et al., 2020) and inviting participants from existing representative online panels (e.g., Codagnone et al., 2020).
Among those studies with international samples, the country/culture variables were used in multiple ways. Fifty-nine (75%) studies described and compared the data by country, for example, providing M/SDs for each country, t-tests, ANOVA, etc. Thirty-one (40%) included the country variable in their statistical models (usually linear or logistic regressions) as predictors or covariates, 19 (24%) conducted separate models for each country and then compared them, 11 (14%) used multilevel modelling approaches to handle the clustered data and 7 (9%) explored the moderating effects of country. Interestingly, five studies also investigated measurement invariance across countries. Other novel methods included Bayesian linear regressions and individual participant data meta-analyses.
Overall, the authors from the included articles identified four categories of challenges regarding the process of conducting international psychological research in the COVID-19 pandemic (see the Table in appendix 2). These challenges included study design, measures, cross-country considerations and sampling strategies. For study design, early pandemic studies were primarily cross-sectional correlational studies, which limited drawing causal inference or determining the directionality between variables (Al Omari et al., 2020; Bareket-Bojmel et al., 2020). The lack of pre-pandemic data collection as a baseline limited the comparison and interpretation of the pandemic’s impact (Siritzky et al., 2021). For measures, the exclusive reliance on self-report methods inflates shared method variance and also increases the likelihood of social desirability bias (i.e., overreporting or underreporting; Tunçgenç et al., 2021). Another concern noted was that newly developed COVID-19 scales may not yet accurately capture the complexity of COVID-19-related experiences (Götz et al., 2021), and, on a related note, that findings in this early stage of the pandemic may under-represent mental health concerns that might only emerge over a longer period of the pandemic (Kapoor et al., 2021).
Cross-country considerations and sampling strategies were two other concerns. Since samples in many studies were primarily recruited from Western or English-speaking countries, this limited the ability of some authors to examine cross-cultural or cross-national differences between the East and the West (Bareket-Bojmel et al., 2021). Moreover, unequal or unbalanced numbers of respondents recruited from different countries within the same study presented another concern (Lindinger-Sternart et al., 2021). Above all, the lack of consideration of the differences in contextual factors across countries, such as regional infection rates and specific government pandemic policies, limited the extent of cross-country comparisons for psychological outcomes (Faulkner et al., 2021; Gloster et al., 2020).
For sampling strategies, one major challenge was that the majority of data collection in the pandemic was online, which prevented participation by people without internet access, such as the underprivileged and older adults (Arafa et al., 2021; Gato et al., 2021). This might create non-response biases (Kochuvilayil et al., 2021). Moreover, the snowball, convenience sampling via social media platform was open to selection bias to individuals using the specific platforms and willing to respond (Curseu et al., 2021). The rapid recruitment of participants in the early months of the pandemic might have also led to sample biases, with disproportionate samples of non-Hispanic Whites (Zheng et al., 2021), younger people (Fumagalli et al., 2021), females (Faulkner et al., 2021) and those with some college education (Nelson et al., 2020). These over-representations might limit the generalizability of the psychological findings in the pandemic. Of note, a few studies did not collect data regarding race/ethnicity and rural/urban residential area (Termorshuizen et al., 2020), which limited the examination of disparity within the countries/regions.
Authors suggested future directions emerging from these challenges; including longitudinal and/or experimental designs to examine causal pathways and mechanisms (Biddlestone et al., 2020; Du et al., 2020); measurement considering COVID-19 infection rates and pandemic policies across countries (Gloster et al., 2020); replicating studies across countries and cultures (Papandreou et al., 2020); and recruitment/data collection methods that ensure more representative sampling (e.g., phone, in-person; Wijngaards et al., 2020).
This rapid scoping review of published international psychological studies addressed the early phase of the COVID-19 pandemic. It thus delivers important insights into the nature of the research and research collaborations that together advanced our knowledge on international aspects of the pandemic within a relatively short time frame after the outbreak of the pandemic. This review documents the international research efforts in psychology during a major crisis, which offers important lessons for the field.
First, it became clear that most international research was concerned with clinical or subclinical effects of the pandemic (e.g., on anxiety and depression), confirming a focus on mental health as an overarching, global research priority for the international community of psychological researchers in the early phase of the pandemic (Gruber et al., 2020; Torales et al., 2020).
A second central result of our review concerns the nature of international collaboration and international data. In our network analyses, European countries as well as the United States were the most connected and represented in terms of authors and data in the single studies. To a certain degree, this corresponds with the global map of the local severity of the COVID-19 pandemic (Figure 5), but various countries and regions with relatively high reported COVID-19 case rates (e.g., India, Russia, countries in South America) were underrepresented in this published international research. Interestingly, other than in Cai et al. (2021) who analysed COVID-19 research more broadly, not limited to psychology, we did not find that collaborative links between the United States and China dominated, from a technical network perspective, the networks of international research in psychology in our review.
We were impressed by the extent of communication between countries and authors reflected in the network analyses. However, while the studies reviewed here clearly attempted to meet the mission of international psychology, the study samples were far more diverse than the study authors, and high-income countries were over-represented in both. These disparities are unfortunately neither new nor unique to this pandemic. For example, Africa, the continent most under-represented in this review, also includes many of the lowest income countries in the world and hence poorer infrastructure and training for psychologists to conduct research (Nsamenang, 1993; Teng-Zeng, 2005). Comparing Figures 3 and 6 illustrates what seems to be a typical challenge in international psychology research: that researchers in fewer, high-income countries are conducting research across a much larger group of countries (i.e., without including those countries’ psychological researchers in the study; Henrich et al., 2010).
Third, surprisingly few studies explicitly focused on culture and measured cultural dimensions. We have to consider that we only included research referring to the early phase of the pandemic and it seems likely that such culture-focused research might be more feasible in later than in very early phases of such a pandemic (e.g., as it relies on international data on the pandemic itself, involving a larger number of countries and thus from later stages of the pandemic, e.g., Gelfand et al., 2021; Kumar, 2021). This signals a potential issue related to a time lag between the outbreak of major global health crises (and the need for immediate international research evidence) and first empirical cross-cultural research evidence addressing the role of culture in this early phase.
Fourth, our analysis is consistent with many other reviews showing that psychological research addressing the early phase of COVID-19 saw a dramatic rise of online methods to recruit and examine study participants (see also De Man et al., 2021), a rise that had, of course, already started before the pandemic so that the field could rely on pre-existing research insights on the usefulness of such methods (Sassenberg & Ditrich, 2019). It has been highlighted, for example, that such methods can facilitate rapid and larger data collections (Gosling et al., 2004), which is particularly relevant from a cross-cultural perspective highlighting human diversity and the diversity of contexts (Gelfand, 2019; Gurven, 2018). On the other hand, such methods come also with a number of potential limitations that need to be considered when interpreting such results (Anderson et al., 2019). Hence, a too strong reliance on one single method in a field has the potential to amplify method bias issues associated with this method and thus biased results in that field. For example, a strong reliance on self-report, cross-sectional data collected with the same (online) method could increase issues stemming from common method variance (e.g., biased relationships between variables) and sample selectivity (e.g., underrepresentation of individuals less likely to participate in online research).
Fifth, important lessons can be learnt from our summary of the reported challenges in the research process. Major challenges included study design, cross-country considerations, sampling strategies and measures. These challenges often had to do with a certain level of “unpreparedness” of psychological researchers, given that the crisis was unprecedented and relatively unpredictable. Research projects had to be planned and executed in the short-term, often under complicated research conditions (e.g., local lockdowns). For example, COVID-19-related measurement scales had to be developed and validated in a short time frame, and it is, for example, well-established that particularly cross-cultural validations of scales require time and considerable efforts (Van de Vijver & Leung, 2021).
The constraints of a pandemic (e.g., closed borders, as well as within-country restrictions) increase the challenges to developing cross-national research partnerships, particularly with those in countries with poorer technology infrastructure (e.g., unreliable internet for zoom meetings, or for downloading data and conducting analyses). Nonetheless, the current public health crisis provides an unparalleled opportunity to understand, across contexts and cultures, how psychological science can contribute to preventing and supporting individuals, families, communities and populations, though this requires an inclusive approach to partnership and sampling. We see several ways in which this might be accomplished: providing support and mentorship for emerging scholars in under-represented countries to join an established research team; partnerships between institutions in low- and middle-income countries and high-income countries, and establishing equitable partnerships in which researchers in under-represented countries are recognised as experts among their populations and provided leadership and first authorship opportunities.
Study designs, sampling strategies and samples themselves also showed the limits of pandemic data collection. Cross-sectional designs are typical in new fields and we expect to see far more longitudinal designs in studies in 2021 and beyond; similarly, moving to multiple-method and multiple informant studies (which usually take longer to plan and are more costly) will hopefully be a feature of this next stage of psychological research. Recruitment is fastest when using social media but such approaches make for unbalanced samples. A more balanced approach to sampling would be helped by having an “on the ground” research partner who is familiar with ways to access more representative samples; these might vary by culture and region. For example, low-bandwidth technology such as texting and WhatsApp is used extensively in certain regions of the world, while recruitment via local connections (e.g., local health centres or community agencies) might be more viable in other regions.
A final observation concerns open science practices. While assessing the state of open science (i.e., “transparent and accessible knowledge that is shared and developed through collaborative networks”; Vicente-Sáez & Martinez-Fuentes, 2018, p. 434) was not an a priori research question, we noticed during our review that open science principles were implemented by many authors. Papers often were publicly available and/or published in open access journals (e.g., Prado-Gascó et al., 2020). Researchers frequently published their research protocols (e.g., Kowal et al., 2020), study materials (e.g., Salali & Uysal, 2020), anonymous data (e.g., Lifshin et al., 2020) and analysis code (e.g., Salvador et al., 2020) with the OSF ( or in other public repositories. Supplementary materials were also widely attached with the manuscripts to provide more detailed information about the study and analyses (e.g., Galasso et al., 2020). Some multinational projects such as PsyCorona Study ( had their own websites with detailed descriptions of the projects enabling other researchers to request raw deidentified data for their own research.
This was a rapid scoping review with several limitations. For example, it identified only published studies that covered the first 6–9 months of the COVID-19 pandemic and, as such, provides an indication of the focus of published research for this period alone—that is, when the pandemic was new, no vaccines and relatively few treatment options were available, and masking and social distancing laws and policies were relatively new. Moreover, our review intentionally targeted research that did not only address this early phase of the pandemic, but was also published quite rapidly (e.g., to inform the unfolding debate how to cope immediate with this major, uncontrolled crisis). Hence, our review does not cover research (on the early phase of the pandemic) that was published subsequent to our literature search. Our search was also limited to psychology journals, and thus potentially under-represents psychological research published in non-psychology (e.g., education, medical, information technology) journals during this period. Finally, we searched only journals published in the languages of our author team, likely missing articles published in in other languages (e.g., Latin American journals in Spanish). Hence, our analysis might underestimate, for example, the extent of actual international collaboration among psychological researchers. Future research could address at least some these limitations by updating our scoping review (e.g., including research that was published after April 2021 and research that was also published in other languages). Future research could also use our results to start new systematic literature review projects that zoom in on specific topics (opposed to a broad scoping reviews). Such systematic reviews could delve deeper into the actual content, key findings and theories adopted in the single studies.
International research evidence and collaborations are vital for a better understanding and proactive coping with new global crises. This review summarises what type of research projects the international psychological research community conducted and published in the wake of the unprecedented pandemic during the early months of the outbreak. It illustrates the diversity but also commonalities of rapid psychological research that is internationally-minded and that can “communicate and collaborate in an inclusive, ‘horizontal’ manner, and apply contextually informed approaches to global concerns” (Stevens & Zeinoun, 2013, p. 758). At the same time, our review also accentuates pre-existing challenges in the process of internationalisation in psychology, for example, the need to cover diverse populations and cultures worldwide and to be more representative of humanity in general (Berry, 2013; Thalmayer et al., 2021; van de Vijver, 2013). The pandemic and its global spread is an urgent reminder that this process needs to be accelerated because only then psychology can more truly fulfil its promise as “essential science” for humanity in the next major crises (Kazak, 2020).
All materials, including study protocol, template data extraction form, coding manual, a clean dataset used for all analyses, analytic code in RStudio and Gephi, and appendix are available on a repository of the OSF: (
Appendix 1. Search narrative and strategies.
Appendix 2. Table with summary of challenges and future directions.
Table S1. Complete data extracted from the studies considered in this review.
Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
Early View
Online Version of Record before inclusion in an issue
The full text of this article hosted at is unavailable due to technical difficulties.
Your password has been changed
Enter your email address below.
Please check your email for instructions on resetting your password. If you do not receive an email within 10 minutes, your email address may not be registered, and you may need to create a new Wiley Online Library account.
Can’t sign in? Forgot your username?
Enter your email address below and we will send you your username
If the address matches an existing account you will receive an email with instructions to retrieve your username



Add a Comment

Your email address will not be published.