Systematic review of character development in low- and middle-income countries

ABSTRACT

Character strengths and development is one of the most important and influential areas of positive psychology. Yet most of this research involves – as per the field generally – research in high income countries, with relatively little awareness of its dynamics in low- and middle-income countries (LMICs). However, there is a burgeoning research base in such countries, but this has hitherto been neither consolidated nor widely disseminated. To remedy this lacuna, this paper reports on a systematic review of the peer-reviewed journal articles on character development in LMICs, analysing 164 articles selected from 1,507 initial abstracts. Our analysis shows that research into character in LMICs is developing rapidly, noting various significant trends with respect to countries, populations, outcomes, and interventions. We also point where further efforts are needed, showing how research on this important topic can continue to prosper in regions that have been relatively neglected in the literature.

KEYWORDS:

• Character strengths
• character development
• systematic review
• LMICs
• developing countries
• positive psychology

Introduction

Character broadly refers to an individual’s attributes, traits, and dispositions (American Psychological Association, n.d.). Character development, inclusive of character education, concerns the shaping of these features (Nucci et al., 2014). Studies on this topic typically explore the effects of character on outcomes such as happiness or productivity, the factors that influence that formation of character, interventions that seek to improve character, and cross-cultural differences in the meaning and salience of character traits.

For most of its history, character development has been primarily studied within philosophy. Within the so-called ‘Western’ canon, it is common to trace the systematic investigation of character to Aristotle, who regarded character and virtue as integral to ‘eudaimonia’; an optimal state for humans that is sometimes translated as well-being, happiness, and flourishing (Durant, 1961). Only in recent decades has character development been investigated using modern scientific methods by researchers in psychology. This interest in character was spurred by the positive psychology movement, beginning in the late 1990s (Seligman & Csikszentmihalyi, 2000), which emerged as a critique of the emphasis on illnesses in psychology as a discipline. As a result, there has been a growing interest in researching character strengths and behaviour that can lead to a good life.

A notable theoretical framework within positive psychology, the Values in Action (VIA) classification, was developed by Peterson and Seligman (2004) to categorise character strengths. According to them, creating a ‘manual of the sanities’ (du Plessis, Saccaggi, & de Bruin, 2015: 28) was necessary for accelerating discussions and empirical investigation in psychological health, just as the Diagnostic and Statistical Manual of Mental Disorders (DSM) had achieved in advancing psychological illness research. Aspiring to capture virtues and character strengths that are valued universally for human thriving, Peterson and Seligman (2004) reviewed the literature with an interdisciplinary and cross-cultural focus. Taking inspiration from Smart (1999)’s conceptualisation that China, South Asia (mostly India), and the West (the ‘Great Three’) represent the most influential traditions throughout human history, the two psychologists and their colleagues then directed their attention to analysing historically dominant spiritual and philosophical schools of thought from these traditions. They identified six moral characteristics that they believed to be recognised as valuable across cultures: wisdom, courage, justice, humanity, temperance, and transcendence. Each virtue encompasses a set of character strengths which manifest as traits. For example, virtue of ‘wisdom’ is associated with creativity, curiosity, judgement, love of learning and perspective (Peterson & Seligman, 2004).

Powered by theoretical development in positive psychology like the VIA classification and their accompanying measurement tool, VIA-Inventory of strengths (VIA-IS), a profusion of empirical studies on character development have been published in recent decades. This research identified the salience of character development across diverse contexts including in education (e.g. Gustems & Calderon, 2014; Proctor et al., 2011; Wagner & Ruch, 2015), health (e.g. Hausler et al., 2017; Leontopoulou & Triliva, 2012; Proyer et al., 2013) and the workplace (e.g. Gander et al., 2020; Martínez-Martí et al., 2020; Moradi et al., 2014).

Most studies on character development, however, have been carried out predominately in high-income countries (HICs). This significantly weakens claims of cross-cultural relevance in character development research. As noted by Peterson and Seligman (2004, p. 51) themselves, cultures differ in the values considered essential for living a good life. Since VIA aspired to reflect ubiquitous values, Peterson and Seligman (2004, p. 51) excluded virtues they believed to be culturally particular. One example of this is ‘duty’, which Peterson and Seligman (2004) identified as endemic to only some traditions and which was therefore excluded from VIA. Yet Peterson and Seligman’s (2004) undertook their attempt to define apparently universal values in an academic context in which character strengths in LMICs had been understudied. While their method included a review of historical texts across multiple traditions, it is plausible that the depth of Peterson and Seligman’s (2004) understanding of values outside of HICs was comparatively limited. There is therefore a strong case that VIA’s claims to relevant universal values is tenuous without further research internationally.

Subsequently, efforts to expand research on character development into low- and middle-income countries (LMICs) have received growing attention. For example, the positive youth development (PYD) framework, originally developed in the US, is a popular theoretical framework that has been recently applied and tested in LMICs (e.g. Abdul Kadir & Mohd, 2021; Cao et al., 2022; Desie, 2020; Hull et al., 2021). There are also efforts to adapt the PYD framework to local contexts, such as the development and validation of the Chinese PYD scale (Chai et al., 2022). There is a growing body of studies on other aspects of character in certain LMICs such as China (e.g. Cheng et al., 2020; Chi & Cui, 2020; Duan & Guo, 2015) and Indonesia (e.g. Edwita et al., 2020; K. Hidayati et al., 2018; Komalasari & Masyitoh, 2022; Manaf et al., 2020) with a focus on education and health.

Nonetheless, efforts to consolidate character development research in LMICs have been limited. There are only a handful of existing reviews of the current state of the literature (Catalano et al., 2019; Ghosh & Deb, 2016; Khan & Ungar, 2021). Moreover, these reviews are narrow in scope regarding the areas of character development and the geographical region covered. For example, while Ghosh and Deb (2016) reviewed positive psychology research broadly, their focus was limited to studies published in India. Conversely, while Khan and Ungar (2021) and Catalano et al. (2019) widen their search to LMICs, these reviews are restricted to the protective factors against self-harm in marginalised groups and the impact of PYD interventions respectively.

Two key reasons drive the need to consolidate character development research in LMICs and expand the scope of the available reviews.

First, a more comprehensive review will contribute to broadening the research agenda beyond the HIC and WEIRD context. The acronym WEIRD – Western, Education, Industrialised, Rich and Democratic – was coined by Henrich et al. (2010a, 2010b). to describe the characteristics of the population who have been typically represented in psychological research. Despite making up around 80% of the world’s population (World Bank Open Data, n.d.), LMICs have generally received very little attention in positive psychology research, although these trends may now gradually be changing. For example, Hendriks et al. (2019) conducted a bibliometric analysis of randomised controlled trials of positive psychology interventions, and found that although 78.2% of the 188 studies identified were conducted in Western countries, there was ‘a strong and steady increase in publications from non-Western countries since 2012’, indicating an encouraging ‘trend towards globalization’ of wellbeing research (p.489). Through the lens of decolonisation, the effort to consolidate research in LMICs is critical for challenging the Euro-American hegemony in knowledge production and understanding realities from the perspectives of the Majority World (G. Adams et al., 2015).

Second, context matters in character development. Using India as a case study, Ghosh and Deb (2016) warn of the limitations of generalising Western notions of well-being and flourishing to other cultures, and consequently call for culturally appropriate investigations that concentrate on endemic behaviours and norms. Similarly, J. C. Christopher (2014) argued that there is an inherent bias in positive psychology research due to its roots in individualist cultures. In the case of character development research, this bias manifests as framing character strengths as one’s internal possession rather than considering them in relation to the external context. Given that how character manifests is context-dependent, consolidating research across diverse contexts will enrich our understanding of character development.

In this study, we attempt to facilitate further research in LMICs by consolidating the literature on character development in these contexts. The aims of this study are to map the empirical evidence produced to date in character development with a focus LMICs; identify key funding bodies and researchers in character development to aid collaboration and knowledge production in LMICs; and identify gaps and opportunities for further research in character development in LMICs. The study is therefore guided by four overarching questions:

1. What (and who) is being studied?

2. How are these studies undertaken?

3. Who is conducting and publishing this research?

4. What are the key findings?

Methods

Search terms

In developing search terms, we were guided by a broad definition of character used by the Templeton World Charity Foundation (TWCF) as well as our experience with the literature:

Character strengths are learned and learnable attributes, virtues, skills, habits, or capabilities that enable individuals to lead better lives. Examples include compassion, forgiveness, gratitude, honesty, humility, and kindness … [as well as] local character strengths, such as the East African concept Utu (shared humanity, humanness).

The search string included common terms associated with character development, such as character strengths of skills, ‘strengths-based’, virtues, moral or ethical development, and positive youth development. These terms were refined after several pilot searches and analyses. The search string also included the names of countries in the specified categories and common terms for these regions (e.g. ‘developing countries’). During the piloting of the search terms, we decided to exclude demonyms as these returned a large proportion of studies on immigrant populations in HICs which was beyond the scope of our research question. However, after the conclusion of the review, we became aware that a significant number of articles in LMICs use only the demonym and not the name of the country. These studies have therefore not been included in the review. The implications of this are discussed in the limitations section of this paper. The complete list of search terms is available in the annex.

Databases

We searched for peer-reviewed journal articles in four major databases: Scopus, PsycInfo, Medline, and PubMed. The search criteria did not include any limitation on publication date. We included all published articles up until the date of the search (26 August 2022).

Inclusion criteria

Studies were included based on the criteria in Table 1.

Table 1. Inclusion criteria.

Category	Criteria
Study design	All empirical studies, i.e. research that used empirical data from individuals was included. Studies that was solely philosophical or theological (without empirical data) were excluded. Studies that focused on the development of measurement instruments, and studies that were literature reviews, were included and analysed separately.
Type of publication	The search was limited to peer-reviewed journal articles. Unpublished studies, conference proceedings, book reviews, media articles, full-length books, and book chapters were excluded.
Language	The search only included articles written and published in the English language.
Population	Articles that dealt substantively with character development in low-income, lower-middle income, and upper-middle income countries (as per list provided in OECD 2022) and focus on individuals were the main subject of the research. Research conducted in high income countries was excluded. Research focused on entities other than individuals (e.g. organizations) was excluded.
Topic	Studies that substantively referenced the literature on character development were included; see further discussion below.
Outcomes	The review was not restricted to specific outcomes.

Three researchers (PN, EL, and AP) applied the inclusion criteria to separate batches of abstracts identified by the study. Researchers were trained to apply the criteria consistently through successive pilots of sample abstracts. Abstracts for which the decision was unclear were discussed with DJS and TL. In addition, DJS reviewed a random sample of 10% of each researcher’s abstracts to ensure inter-rater reliability. Three researchers (PN, EL, and JL) then each screened the full text of potentially eligible articles to assess whether they met our eligibility criteria. As before, researchers were first trained on successive pilots of sample articles. Unresolved eligibility decisions were clarified through a review DJS and TL, and a sample of each researcher’s articles were reviewed by DJS to ensure inter-rater reliability.

Of these criteria, the most difficult to apply consistently was whether a study was ‘substantively’ about character development. Character development could be defined very broadly; at an extreme, it could encompass all psychology and anthropology. We limited our review to focus on studies that situated themselves in the literature on character development. As an approximate guideline, studies that only referred to the character development literature in two or fewer sentences in the full article were excluded.

Protocol registration

The methodology of this review has been described in detail in a previously published review protocol registered with International Platform of Registered Systematic Review and Meta-analysis Protocols (INPLASY) (registration number INPLASY2022100117) on 29 October 2022. A link to the registered protocol is available here: https://inplasy.com/inplasy-2022-10-0117/. The protocol was registered prior to data extraction.

Data extraction and analysis

The research questions were divided into sub-questions to structure data extraction and analysis. The sub-questions are reported in Table 2.

Table 2. Research questions.

Key questions and sub-questions
What (and who) is being studied?
(1) What are the contexts of these studies? a. Which countries are being studied? b. What is the site of the study? (e.g. schools, communities, religious institutions) (2) Which populations are being studied? a. Which demographics? (Gender, age, income, minorities, disabilities) (3) What aspects of character development are being studied? a. Which character strengths? b. Including interventions – intervention design, duration, dosage.
How are these studies undertaken?
(4) What are the methodological designs of these studies? a. What measurement tools are used in these studies? b. Are these developed for use (or otherwise adapted for use) in these regions? c. What is the quality of these studies?
Who is conducting and publishing this research?
(5) Who is funding and conducting this research? (6) When and where are the studies being published?
What are the key findings?
(7) What are the key findings of these studies?

To facilitate extraction, a customised data collection excel spreadsheet was developed based on the research questions. The character strengths that constituted the focus of each article was coded using the VIA framework, with additional character strengths added as these emerged through the review if these were not reflected in the original framework. After an initial pilot, small modifications to the layout of the data extraction spreadsheet were made and prompts added before extraction continued. Data were extracted by three researchers (EL, PN and JL). Ambiguities and discrepancies were discussed and resolved by the research term.

Data extraction included an assessment of study quality. Quality was assessed using frameworks developed by the research team based on well-established frameworks for different study types: qualitative (Blignault & Ritchie, 2009), quantitative observational (Von Elm et al., 2007), quasi-experimental and randomised controlled trials (Schulz et al., 2010), systematic and non-systematic reviews (Page et al., 2021), and development of measurement tools (Terwee et al., 2007). Rather than offering a detailed assessment of each study, the frameworks developed by the research team assessed whether a study provided the basic information necessary for a full assessment of quality. As such, the quality assessment reported in this study is only an approximation of quality. However, this approach was appropriate for the research aims, given that the focus was to provide an overview of the state of the literature rather than a meta-analysis of the data. Consequently, articles were not excluded based on quality.

Results

Figure 1 shows the flow of articles during the review process. The search identified 2,258 and 1,507 remained after removing duplicates. After screening titles and abstracts, 929 records were excluded. This was followed by the full text screening of the remaining 578 articles after which 414 were excluded to give us a final list of 164 eligible studies for analysis. This included 145 empirical articles, 17 tools (one of which was also analysed as an empirical article), and three reviews of empirical articles.

Figure 1. PRISMA flowchart.

Findings

What (and who) is being studied?

This first section reports what and who are being studied; that is, the contexts of the studies (RQ1) and the populations (RQ2) and aspects of character (RQ3) being studied.

RQ1.

What are the contexts of these studies?

Figure 2 represents the frequency of articles by each country and the distribution of articles by geographical region. East Asia and South East Asia account for more than half of the articles published; the three most studied countries are China (41 articles), Indonesia (24 articles), and Malaysia (8 articles).¹

Figure 2. Location of studies by number of articles.

Only 73 articles reported whether their participants were based in urban or rural areas. Table 3 reports the frequency of each regional classification. Most articles included urban or peri-urban participants (65 articles), while only a minority included rural participants (21).

Table 3. Frequency of articles in urban and rural areas.

Classification	Number of articles	Percentage of articles reporting data on location
Urban	50	68.49%
Rural and urban	12	16.44%
Rural	8	10.96%
Urban, rural, peri-urban	2	2.74%
Peri-urban	1	1.37%
N	73

Most studies reported on the sector being researched (135 studies). As indicated in Table 4, education is by far the most common (93 studies), followed distantly by health (19 studies).

Table 4. Number of articles by sector.

Sector of intervention/research	Number of articles	Percentage of articles reporting data on sector
Education	98	72.59%
Business	12	8.89%
Health	19	14.07%
Other	9	6.67%
N	135

RQ2.

Which populations are being studied?

One hundred and two articles reported the gender of their participants. Of these articles, male and female participants accounted for approximately equal numbers (Table 5). No studies reported non-binary participants (and this may be because this option was not offered to participants), although a small minority of participants declined to report their gender.

Table 5. Reported gender of research participants.

Gender	Number of participants	Percentage of participants
Male	52,457	48.84%
Female	54,004	50.28%
Gender not reported	953	0.89%
N	107,414

With two exceptions, all studies reported the ages of their participants. Adolescents (52 articles), emerging adults (51 articles), and adults (56 articles) are approximately equally represented; children, although represented in fewer articles than other developmental stages, are nonetheless reported as participants in 36 articles (Table 6).

Table 6. Number of articles by developmental stage.

Developmental stage	Number of articles		Percentage of articles
Children (0 to 12 years old)		36	24.83%
Adolescent (13 to 17 years old)		52	35.86%
Emerging adult (18 to 25 years old)		51	35.17%
Adult (25+ years old)		56	38.62%
N		145

The sum of the number of articles exceeds N as one article may report multiple developmental stages.

Only 25 articles reported the income of their participants. Of these, 8 articles classified participants according to low, middle, of high income; 7 articles reported by average income of participants; 5 articles identified all their participants as low-income, and 5 articles reported income brackets with currency values. Only 16 articles reported the religion of their participants: 5 articles included participants following Islam, 5 articles included participants following Christianity, and 6 articles included participants with multiple religious affiliations.

Only one study reported participants with disabilities. No studies reported data on the sexual orientation of participants.

RQ3.

What aspects of character development are being studied?

Table 7 reports the most studied aspects of character. Most articles study aspects of character that are reflected in the Values in Action (VIA) frameworks; aspects of character that are not reflected in VIA are indicated with an asterisk. The most studied aspect of character is kindness (45 articles), followed by self-regulation (36 articles), love (35 articles), spirituality (35 articles), leadership (34 articles), social intelligence (33 articles), and teamwork (30 articles). The most studied aspects of character that are not reflected in VIA are moral development (26 studies) and resilience (14 studies). Although it may be unclear whether ‘moral development’ is an aspect of character, as it was cited as such in several studies, we have included it as a character strength in our review.

Table 7. Aspects of character studied.

Number of articles	Aspect of character
45	Kindness
36	Self-regulation
35	Love, spirituality
34	Leadership
33	Social intelligence
30	Teamwork
26	Hope, moral development*, perseverance
25	Curiosity, honesty, love of learning
24	Bravery, fairness, judgement
23	Creativity
22	Perspective
21	Prudence
20	Gratitude
19	Forgiveness, zest
17	Humour, humility
16	Appreciation of beauty and excellence
14	Resilience*
9	Responsibility*
8	Self-efficacy*
6	Self-esteem*, respect*
5	Purpose in life*
3	Integrity*, independence*, tolerance*
2	Growth or strengths mindset*, politeness*
1	Filial piety*, mindfulness*, self-compassion*, altruism*, patience*, adaptability*, entrepreneurialism*

Asterisks indicates aspect of character not included in the VIA framework.

Forty-six studies used the Positive Youth Development (PYD) frameworks to study character (Table 8). The most popular frameworks were the ‘5Cs’ (24 studies), followed by the 15 PYD attributes (13 studies) and developmental assets (11 studies).

Table 8. Articles using positive youth development (PYD) frameworks.

PYD Qualities	Number of articles	Percentage of total articles
5Cs	24	16.55%
Developmental Assets	11	7.59%
15 PYD attributes Catalano et al. (2004)	13	8.97%
Total	46	31.72%

Fifty-two articles studied interventions relating to character development (Table 9). The most common interventions by far focused on courses and curricula (35 studies). While most articles did not report the duration or dosage of the intervention, of those that did, most interventions were shorter than one month (10 articles) and consisted of only one dose (13 studies).

Table 9. Frequency of intervention characteristics.

Intervention characteristics		Number of articles	Percentage of studies on interventions
Type of Intervention	Courses and curricula	35	67.31%
	Teacher professional development	5	9.62%
	Project-based training and mentorship interventions (i.e. not at educational institutions)	4	7.69%
	Sports-based interventions	3	5.77%
	Religious practices	2	3.85%
	National Youth Service Programme	2	3.85%
	Participating in community service programmes	1	1.92%
Duration	Shorter than 1 month	10	19.23%
	1–6 months	7	13.46%
	longer than 1 year	6	11.54%
	Unclear	29	55.77%
Dosage	Once	13	25.00%
	Twice	2	3.85%
	Unclear	37	71.15%
	N	52

How are these studies being conducted?

This section reports how these studies were conducted; that is, the prevalent research designs and measurement instruments used (RQ4).

RQ4.

What are the methodological designs of these studies?

Table 10 reports the frequency of research design type. Most articles used non-experimental correlation research design (89 articles). Experimental designs, whether randomised and controlled (14 articles) or quasi-experimental (10 articles), account for a minority of studies.

Table 10. Frequency of research design.

Research design	Number of articles	Percentage of total articles
Non-experimental correlation design	89	61.37%
Qualitative	35	24.14%
Randomised Controlled Trial	14	9.66%
Quasi-experimental	7	4.83%
N	145

In our initial analysis, only 42 articles (29%) provided information that would be sufficient to do a detailed assessment of quality. Further analysis identified that most studies failed a single criterion: 109 studies did not report when and where the data was collected. Excluding this criterion, 101 articles (70%) provided sufficient information for a detailed quality assessment. Table 11 reports the number of articles providing sufficient information by research design. Articles reporting on non-experimental correlation designs were most likely to provide sufficient information, whereas articles reporting on quasi-experimental or RCTs were the least likely.

Table 11. Studies reporting sufficient information for detailed quality assessments.

	Studies reporting sufficient information (revised)	Total studies	Percentage of studies reporting sufficient information (revised)
All studies	101	145	69.66%
Non-experimental correlation design	68	89	76.4%
Qualitative	21	35	60%
Quasi-experimental	4	7	57.14%
RCT	8	14	57.14%

We identified 19 tools for measuring character which had been developed solely for the purpose of the study in question, and 86 tools which had been used in previous studies. Of these 86 ‘established’ tools, more than half were developed in the United States of America (45 tools). This is followed distantly by China, which was the origin of 12 tools (Table 12).

Table 12. Origin of tools.

Country	Number of tools	Percentage of tools
United States of America	45	52.94%
China	12	14.12%
Turkey	5	5.88%
Brazil	4	4.71%
Canada	3	3.53%
Norway	2	2.35%
Various (web-based)^2	3	3.53%
Countries with a single tool cited^3	11	12.94%
Unclear origin	3	3.53%
N	86

The most popular tools to measure character development were the Chinese Positive Youth Development Scale (11 articles); VIA including VIA Youth (VIA-Y), VIA Inventory of Strengths (VIA-IS), and VIA-72 (9 articles); and the PYD-Short Form and -Very Short Form (8 articles). The next most popular tools were the Defining Issues Test (DIT) (6 articles); Connor-Davidson Resilience Scale (5 articles), and the Developmental Assets Profile (5 articles).

Who is conducting and publishing this research?

This section reports the most prolific researchers and key funders identified in this review (RQ5), and the most popular journals (RQ6).

RQ5.

Who is conducting and funding this research?

Of the 570 researchers identified in this review, only 49 had contributed at least two papers (Table 13). Only 17 of researchers had contributed at least three papers. While many researchers have published on the topic, the number of regularly contributors is therefore relatively small.

Table 13. Number of researchers by number of papers contributed to the review.

Number of papers contributed to review	Number of researchers
5 or more	5
4 or more	8
3 or more	17
2 or more	49
1 or more	570

Based on institutional affiliation at the time of each articles’ publication, institutions contributing the most articles to the review are based predominately in China (five institutions), with one institution each in Malaysia, Norway, and Indonesia (Table 14)

Table 14. Publications by institution.

Rank	Institution	Country	Number of publications
1	Hong Kong Polytechnic University	China	12
2	City University of Hong Kong	China	10
3	Beijing Normal University	China	7
4–5	Universiti Putra Malaysia	Malaysia	5
4–5	University of Bergen	Norway	5
6–7	Universitas Pendidikan Indonesia	Indonesia	4
6–7	Renmin University of China	China	4

Of the individual researchers who contributed to most papers to the review, seven of the top eight are based in institutions in China (Table 15). Within the top eight researchers, there are two sets of researchers who publish regularly together: Professors Daniel Shek, Xiaoqin Zhu, and Diya Dou (Hong Kong Polytechnic University), and Prof. Danhua Lin and Dr. Zhi Ye (Beijing Normal University).

Table 15. Researchers contributing the most papers to the review.

Rank	Author	Institution	Country of institution	Number of publications in review
1	Daniel Shek	Hong Kong Polytechnic University	China	10
2	Xiaoqin Zhu	Hong Kong Polytechnic University	China	7
3	Nora Wiium	University of Bergen	Norway	6
4–5	Wenjie Duan	East China University of Science and Technology	China	5
4–5	Danhua Lin	Beijing Normal University	China	5
6–8	Zhi Ye	Beijing Normal University	China	4
6–8	Jesus Alfonso Datu	Education University of Hong Kong	China	4
6–8	Diya Dou	Hong Kong Polytechnic University	China	4

Only 76 (52.41%) of articles disclosed funding. Of these, the most common sources of funding (reported in Table 16) were governments (31 articles), universities (28 articles), and philanthropies (26 articles). The Tin Ka Ping (TKP) Foundation funded the most articles in this review (9 articles). The next most active funders, each supporting three articles were Compassion International, the National Social Science Foundation of China, University Putra Malaysia, and the Wofoo Foundation.

Table 16. Type of funders.

Type of funder	Number of articles	Percentage of articles that disclosed funding
Government	31	40.79%
University	28	36.84%
Philanthropy	26	34.21%
Type of funder unclear	13	17.11%
Business	1	1.32%
Other	1	1.32%
N	76

RQ6.

When and where are studies being published?

Of the 145 studies included in this review, 75.86% (110 studies) were published in the last five years and 95.17% (148 studies) in the last 10 years (Figure 3). Most studies on this topic are therefore relatively recent.

Figure 3. Publications over time.

The most popular journal for studies published on this topic is Frontiers in Psychology, accounting for 12 articles (Table 17). Of the six most popular journals, two are in the field of psychology (Frontiers in Psychology and Current Psychology), two are in the field of children and youth (Child and Youth Care Forum and Children and Youth Services Review), one is in education (Cakrawala Pendidikan), and one is in business ethics (Journal of Business Ethics).

Table 17. Most popular journals.

Rank	Journal or Publication	Number of articles	Percentage of empirical articles
1	Frontiers in Psychology	12	8.28%
2	Current Psychology	5	3.45%
3–4	Cakrawala Pendidikan	4	2.76%
3–4	Journal of Business Ethics	4	2.76%
5–6	Child & Youth Care Forum	3	2.07%
5–6	Children and Youth Services Review	3	2.07%
	N	145

What are the key findings of the research so far?

This systematic review has provided an unprecedented and much-needed summary of the state of research on character strengths and development in LMICs. This section summarises the key findings identified in the studies analysed by the review (RQ7), focusing on the four most studied topics. We also provide a brief overview of the findings on the most common type of intervention: courses and curricula. Finally, we suggest priorities for future research identified by articles in the review.

RQ7.

What are the key findings of these studies?

Two of the most studied topics identified in the review relate to character strengths as being what drives other outcomes; that is, how character influences (a) mental health and (b) behaviour. The other two relate to character as being what is affected; that is, the factors contributing to (c) moral development, and (d) resilience.

Character as a driver of outcomes

Eighteen studies reported findings on the effect of character strengths on positive mental health. Nine of these found significant relationships between strengths within the VIA framework and positive mental health, including hope (Betancourt et al., 2011, Sutter et al., 2016, Yépez-Tito, 2022); kindness (Datu & Bernardo, 2020; Datu & Mateo, 2020; Y. Li et al., 2021); curiosity (Y. Li et al., 2021; Mohanty et al., 2015); love of learning (Datu & Mateo, 2020; Y. Li et al., 2021); self-regulation (Betancourt et al., 2011; Y. Li et al., 2021); bravery, love, perseverance, or spirituality (Betancourt et al., 2011); fairness, forgiveness, or teamwork (Datu & Bernardo, 2020); gratitude (Datu & Mateo, 2020); and vitality or zest (Duan & Guo, 2015, Yépez-Tito, 2022, Castro Solano, 2016). Nine studies reported significant relationships between aspects of the PYD frameworks and positive mental health, including internal assets (Chang et al., 2020; Lai et al., 2022; Manrique-Millones et al., 2021); external assets (Lai et al., 2022; Manrique-Millones et al., 2021); connection (Duan & Guo, 2015; Kabir et al., 2021); cognitive-behavioural competence or positive identity (Dou & Shek, 2021; Zhu & Shek, 2020a, 2020b); caring (Kabir et al., 2021); and all 5Cs (Abdul Kadir & Mohd, 2021). Seven studies reported significant relationships between positive mental health and character strengths not associated with either VIA or PYD frameworks, including purpose in life (Abdul Kadir & Mohd, 2021; Betancourt et al., 2011; Mohanty et al., 2015); a growth mindset (Ding & Liu, 2022); resilience (Chai et al., 2019); self-compassion (Chi et al., 2022); and self-esteem or patience (Betancourt et al., 2011).

Only three studies reported null effects relevant to character and positive mental health. In their study on the relationship between virtues and post-traumatic growth (PTG) comparing participants who had been either directly or indirectly affected by earthquakes in China, Duan and Guo (2015) found that while vitality accounted for PTG among those indirectly affected, it was not significant for those who had been directly affected. In their study of undergraduate students in Ghana, Kabir et al. (2021) found that, after adjusting for age and gender, while there were significant relationships between psychological well-being and connection and caring aspects of PYD, the relationships between psychological well-being and competence, confidence, and character were not significant. In their study on PYD and delinquency among adolescents in Mainland China, Zhu and Shek (2020b), prosocial attributes were not associated with life satisfaction. Of the 18 articles on character and positive mental health in LMICs, nine were undertaken in China, two in the Philippines, and one each in Ecuador, Colombia and Peru, Ghana, India, Malaysia, Mexico and Argentina, and Rwanda (Sarizadeh et al., 2020).

Thirteen studies report findings relating to character and mental health problems. Six of these used aspects of character within PYD frameworks, and found significant, inverse relationships between mental health problems and competence (Sarizadeh et al., 2020; Sutter et al., 2016), the 15 PYD attributes as a whole (Shek et al., 2021; Zhou, Shek, & Zhu, 2020), the 5Cs (Sarizadeh et al., 2020), cognitive-behavioural competence or positive identity (Chi et al., 2020), and external assets (Fuentes-Balderrama et al., 2020). Five studies used aspects of character reflected in the VIA framework, and found significant, inverse relationships between mental health problems and curiosity (Cheng et al., 2020; Kalyar & Kalyar, 2018; S. Zhang et al., 2021); bravery or honesty (Caram et al., 2019; Sarizadeh et al., 2020); creativity (Kalyar & Kalyar, 2018; S. Zhang et al., 2021); love (Caram et al., 2019; Cheng et al., 2020); social intelligence (Caram et al., 2019; S. Zhang et al., 2021); zest (Sarizadeh et al., 2020; S. Zhang et al., 2021); fairness, kindness, leadership or teamwork (Caram et al., 2019); humour (S. Zhang et al., 2021); judgement, love of learning, or perspective (Kalyar & Kalyar, 2018); perseverance (Sarizadeh et al., 2020); and self-regulation (Cheng et al., 2020). Four studies reported findings concerning character strengths outside of the VIA or PYD frameworks, and found significant and inverse relationships between mental health problems and resilience (Huang et al., 2020; Sutter et al., 2016), moral development (Al-Shuaibi et al., 2013), and self-compassion (Chi et al., 2022).

Only two studies reported null effects concerning the relationship between character and mental health problems. In their study on Chinese nurses, S. Zhang et al. (2021) that while inquisitiveness was positively associated with mental health, caring and self-control were not. In their study of families with preadolescents in Mexico, Fuentes-Balderrama et al. (2020) found that parental involvement and communication was not associated with mental health challenges, although these practices were inversely associated with behavioural problems. Of the 13 studies on character and mental health problems in LMICs, seven were undertaken in China, while the remainder were in Brazil, Iran, Jordan, Mexico, Mexico and Argentina, and Pakistan.

Ten studies in the review reported findings on the effect of character on prosocial or problematic behaviour. Six studies reported positive findings on attributes within the PYD framework, including cognitive-behavioural competence and positive identity (Chi & Cui, 2020; Dou & Shek, 2021; Lan & Wang, 2019; Moulin-Stozek et al., 2018; Shek et al., 2022; Zhu & Shek, 2020b). Four studies reported a positive relationship between resilience and behaviour (Huang et al., 2020; Lan & Wang, 2019; Moulin-Stozek et al., 2018; Wu et al., 2018). Two studies used the VIA framework in its entirety (Qin et al., 2022; Y. Zhang et al., 2014). Two studies used the developmental assets framework, both of which reported positive associations between external assets and behaviour (Fuentes-Balderrama et al., 2020; Lan & Wang, 2019), and one of which reported positive associations with internal assets and behaviour (Lan & Wang, 2019). One study reported a positive relationship between spirituality and behaviour (Moulin-Stozek et al., 2018), and one study reported a positive relationship between purpose in life and behaviour (Wu et al., 2018). Only one study reported a null effect, which was between problematic behaviour and whether households were intact or fragmented (Fuentes-Balderrama et al., 2020). Of these ten studies on character and behaviour, eight was undertaken in China, one in Peru and El Salvador, and one in Mexico.

How other factors affect character

In studies that explored the presumed effect of other factors on character, the most studied area was the impact of professional ethic and culture and moral development. Six studies found that people’s moral sense was affected by their professional environment; most of these studies focused on accountants or auditors (Haron et al., 2015; Malagueño et al., 2020; Mohd Ghazali, 2021; Oboh, 2019), while two other studies were concerned with nurses (Ranjbar et al., 2017) and business students (Hanson et al., 2017). Two studies on this topic – among auditors in Nigeria (Oboh, 2019) and undergraduate students in Malaysia (Nga & Lum, 2013) – reported null effects.

Studies on moral development and other character strengths found a positive relationship between this strength and fairness (April et al., 2010; Wang et al., 2015), kindness (Al-Adamat et al., 2020; Wang et al., 2015), and spirituality (April et al., 2010; Chan & Ananthram, 2019). Only one null effect was reported: this concerned the relationship between spiritual or moral development and independence and impartiality among auditors (Mostafa et al., 2020).

After moral development, the second most studied area within the literature on the determinants of character was the impact of community environment on resilience. Four studies found that community environment includes the influence of peers and neighbourhood factors, such as safety. These studies found that resilience was positively affected when: HIV/AIDS patients in Kenya disclosed their status to supportive communities (C. Adams et al., 2022); survivors of intimate partner violence in India had supportive personal networks (Shanthakumari et al., 2014); and ‘left-behind’ children in rural China and urban youth in South Africa lived in neighbourhoods with strong social cohesion (Chai et al., 2019; Mosavel et al., 2015).

Two studies on the relationship resilience and other character strengths found a positive correlation between these strength and bravery, hope, perseverance, and spirituality (Betancourt et al., 2011; Shanthakumari et al., 2014), while two studies found an association with love (B. G. Adams et al., 2019; Betancourt et al., 2011). No studies reported null effects on the impact of community environment on character, or on the relationship between resilience and other aspects of character.

Character development interventions: courses and curricula

Interventions focused on character-based courses and curricula in schools and universities accounted for two-third of studies on interventions (35 articles). These included 10 randomised controlled trials and seven quasi-experimental studies. These 35 articles reported that curricula-focused interventions had a significant impact on mental health (8 studies), spirituality (8), kindness (7), responsibility (6), self-regulation (5), teamwork (5) perseverance (5), and leadership (4). Five studies focused on moral development; of these, three found a significant impact (Chong & Hamid, 2016; Hull et al., 2021; Shek, Han, et al., 2014) while two found null effects (Arfaoui et al., 2016; Lin et al., 2022). Null effects were also reported for the impact of curricula-focused interventions on teamwork and altruism in preschools in China (Xu, 2014), self-esteem and positive mental health in schools in India (Khanna et al., 2021), and mental health challenges among university students in China (Y. Yu et al., 2022).

Future research

One hundred and sixteen studies reported priorities for future research. These priorities are reported in Table 18. Of these, the most common priorities were including accounting for other variables (46 studies), using larger and more representative samples (40 studies), and collecting longitudinal data (39 studies). The first of these – the inclusion of more variables – is likely unavoidable, given that any one study cannot account for all variables. However, there is clearly a need for larger studies extending over longer periods.

Table 18. Priorities for future research.

Priority identified	Number of studies	Percentage of articles identifying research priorities
Including other variables	46	40.35%
Using larger and more representative samples	40	35.09%
Collecting longitudinal data	39	34.21%
Addressing limitations of self-report data	28	24.56%
Conducting research on the same population	20	17.54%
Using multiple data sources to triangulate findings	20	17.54%
Using experimental methods	16	14.04%
Using better designed tools	15	13.16%
Incorporating perspectives of other actors	14	12.28%
Studying other outcomes	12	10.53%
Studying different cultural groups	12	10.53%
Using more sophisticated statistical designs and methods	12	10.53%
More qualitative or ethnographic research	8	7.02%
N	116

Discussion

We believe this study to be the first systematic review to consolidate current research on character development in LMICs. The review has identified several emerging topics, such as the effects of character on mental health and behaviour, how social contexts shape moral development and resilience, and the impact of interventions to improve character through changes to courses and curricula. However, we have also identified how research thus far is skewed: just three countries (China, Indonesia, and Malaysia) account for more than half of the articles in this review, and research has largely neglected people in rural areas, with disabilities, and who are LGBTQIA+. Other similar reviews in LMICs also reported a lack of research attention on some of the social groups mentioned above (Frances et al., 2023; Saran et al., 2020). More commonly adopted by global health researchers (Larson et al., 2016), the intersectionality lens is rarely used in character development research. Moreover, nearly two-thirds of articles primarily relied on the analysis of correlations between quantitative variables under non-experimental conditions while qualitative or experimental methods are infrequent. Research has also focused on character within the education sector, and the salience of character elsewhere, such as in business and health, is relatively under-explored.

As identified at the beginning of this article, much research on character development aspires to be applicable across contexts and cultures. The findings of this systematic review reflect the current state of this aspiration. These findings affirm much of what has already been identified in high-income countries; that, for example, character traits affect many aspects important to how one experiences life and behaves, and character is influenced by the contexts in which one works and lives. Far from redundant, such corroboration provides important evidence for the cross-culturally applicability of many character traits.

This review also identifies how dominant frameworks like VIA could be further developed. These frameworks may overlook traits that are salient in much of the world, such as filial piety, responsibility, and respect. Expanding our understanding of character across culture and context also requires more research beyond China, Indonesia, and Malaysia. Moreover, while many character strengths can be mapped onto dominant frameworks, further research may clarify how the meaning of certain traits differ between contexts, such as, for example, what it means to be kind. This corroborates Smith et al. (2007) and J. Christopher and Hickinbottom (2008)’s call for a culturally-embedded approach to studying character strengths. Although qualitative research may be especially well-suited to unpicking cross-cultural nuance and explicating ‘new’ traits, our review identifies that research of this nature accounts for relatively few studies. An over-reliance in positive psychology on experiments relying on quantitative analysis is one of the major criticisms of positive psychology, according to Van Zyl et al. (2023)’s systematic review.

Our findings about the types of instruments used to measure character reflect a challenge for further theoretical development. Instruments define what is being studied; consequently, the prevalence of tools developed in the USA threatens to obscure how character traits might be understood through local lenses. On top of that, researches argue that psychometric instruments in positive psychology often have dubious levels of validity and reliability (Van Zyl et al., 2023). As instruments are both defined by and inform theoretical developments, advancing cross-cultural theories of character will simultaneously entail developing new instruments. Rather than beginning from the assumption that the construct developed in the USA is the universal reference to be modified, we might ask what a common reference might look like if research in LMICs is regarded as having an equal theoretical weight as research in the USA.

Recommendations

While the review identified the most frequent recommendations for further research identified by researchers, this section offers our own recommendations. Further research in LMICs will be best served by supporting research capacity within these contexts. It bodes well for the field that the research institutions which produce the most research on character development in LMICs are based in LMICs; such institutions are well-placed to work closely with participants and support researchers intimately familiar with these contexts. Nonetheless, nearly all of the most productive institutions are based in China. While this has propelled the development of popular tools, such as the Chinese Positive Youth Development Scale, research capacity also needs to be supported elsewhere. We therefore recommend that initiatives to build research capacity focus on institutions and researchers in Central and South America, Africa, the Middle East, and South Asia.

Moreover, researchers included in this review identify the need for longitudinal research with larger and more nationally representative samples. Studies of this nature would do much to answer key questions in the field regarding the long-term effects and determinants of character. This will move away from the over-reliance of cross-sectional studies using correlational design with self-report measures in positive psychology (Cabanas, 2018; Diener, 2012; Held, 2018; Kristjánsson, 2010). Large, longitudinal studies of character in LMICs would reflect a new generation of research in this field. The benefits of research of this kind, especially for consolidating theoretical developments and making compelling policy recommendations, has been illustrated in other fields, such as in the Young Lives study of poverty and child development.

Limitations of this review

We identify four key limitations of our review.

First, our search string was limited to country names and common terms for LMICs but did not include demonyms. While this review likely reflects a very large proportion of the literature, it is nonetheless not exhaustive. We do not anticipate that the inclusion of demonyms would significantly change the key findings of the review, as it is probable that the decision to use demonyms rather than placenames is relatively random across the literature in aggregate. However, the review may have missed some prolific researchers, if it is likely that a researcher’s non-use of country names is consistent between articles. Future systematic reviews on research on LMICs may consider the use of both placenames and demonyms.

Second, our analysis is at the country-level and does not consider income variation within countries. Consequently, the insights of this review may be skewed by the over-representation of studies that were conducted in cities that may have populations with higher incomes than the rest of the country.

Third, as with any literature review, our analysis is vulnerable to the bias towards publishing research that has statistically significant findings. Thus, while our review suggests that the effects and determinants of character are frequently significant, the reader is reminded that studies with null effects are under-reported and thus under-represented in this review.

Finally, our inclusion criteria limited our search to publications in English. While we did not find indications (such as through citations) that we had overlooked a large non-English language literature, our review has likely under-counted the number of studies in non-Anglophone regions.

Conclusion

This review provides an overview of character development research in LMICs regarding research context, population, character strengths of focus, methodological design and impacts of character development. The paper also identifies key funding bodies and researchers in character development to aid collaboration and knowledge production in LMICs and recognizes gaps and opportunities for further research. To our knowledge, this is the first effort to comprehensively consolidate research on character development from LMICs.

The review makes some notable contributions to the literature on character development.

By focusing on LMICs in a comprehensive way, the paper contributes on the understanding of character development outside of HICs, such as the effects of character on mental health and behaviour, and how social contexts shape moral development and resilience. In doing so, it addresses the existing disparity and biases in character development research within positive psychology which has been predominately shaped by research in HICs. Building a wider perspective can make research on this topic more universal. It can also accelerate the cross-pollination of ideas between HICs and LMICs and potentially benefit populations that have been much neglected in research and interventions.

Our findings advance the role of contexts as a key dimension to understanding character development – in terms of the types of character traits valued (e.g. religiosity, filial piety, resilience), the type of interventions, and the tools and frameworks used to measure these in LMICs. In addition to illustrating the dangers of considering research from HICs as universal by default, our findings call for more culturally appropriate context-specific research in the future. For example, our findings present a strong case for the implementation of character-based interventions in the LMIC contexts, particularly for improving mental health outcomes.

The paper identifies how research on LMICs thus far is skewed, such as through its focus on specific countries, social groups & regions, and sectors. We also showcase the predominance of quantitative analysis while qualitative or experimental methods are sparsely used in the evaluations.

Considering these, the paper highlights existing opportunities and broad directions for future research. The concentration of current research efforts in and around China presents rich opportunities for collaboration and knowledge exchange. Equally, it spotlights the need to support research and build research capacity in other countries from Latin America, sub-Saharan Africa, the Middle East and North Africa, and Southeast, to broaden research focus to include socially and economically diverse participants (e.g. LGBTQ, rural communities, people with disabilities) and expand research to other sectors (e.g. business). From a methods standpoint, the top priorities for future research include expanding the number of variables included in studies, using a larger and more representative sample and investing in longitudinal data collection.

If research on character development aspires to be cross-culturally relevant, research across diverse countries and contexts remains essential. We recommend further research in this vein.

Open Scholarship

This article has earned the Center for Open Science badges for Open Data, Open Materials and Preregistered. The data and materials are openly accessible at https://doi.org/10.1080/17439760.2024.2322464

Acknowledgments

We are grateful to Ellen Morgan and Aulina Pandey for their contributions as research assistants in this review.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Data availability statement

The extracted data is available on request at https://doi.org/10.1080/17439760.2024.2322464

Additional information

Funding

The work was supported by Templeton World Charity Foundation.

Notes

1. For six countries, only two articles had been published (Egypt, Ghana, Nigeria, Philippines, Tunisia, Turkey). For 17 countries, only one article had been published (Belize, Cambodia, Costa Rica, Ecuador, Eswatini, Ethiopia, Georgia, Morocco, Mozambique, Romania, Russia, Rwanda, Senegal, Slovenia, Thailand, Uganda, Zambia).

2. In these studies, the tool was tested on participants from multiple countries who responded to an online invitation.

3. These are Germany, Indonesia, Israel, Malaysia, Mexico, Nigeria, Pakistan, Uganda, United Kingdom, Rwanda.