A systematic review was conducted following the PRISMA 2020 criteria ( Page et al., 2021). Initially, two authors (INITIALS) performed a global systematic search across ten electronic databases: Web of Science (WOS), Scopus, LILACS, Dialnet, PubMed/MEDLINE, Redalyc, Scielo, Directory of Open Access Journals (DOAJ), Latindex, and CLASE. The search was conducted in English and Spanish between October 29, 2024, and November 30, 2024.

The search strings combined key terms related to university dropout, Indigenous populations, and the educational context. To maximize results, both Spanish and English terms were used. The search string applied was: ALL=((dropout) OR (deserción) AND ("university students") OR ("estudiantes universitarios") AND (indigenous*) OR (indígenas*)). Boolean operators were adjusted according to each database’s search criteria (using “ AND”, “ OR” “ *”, “or” ) to maximize sensitivity and specificity, ensuring the inclusion of relevant studies. Additionally, exclusion filters by study type (full articles or letters to the editor) were applied in databases that allowed it (Scopus, WOS, LILACS, Dialnet, PubMed, and Scielo).

A total of 23,403 records were identified: 21,619 in WOS, 356 in Scopus, 115 in LILACS, 743 in Dialnet, 385 in PubMed/MEDLINE, 16 in Redalyc, 53 in Scielo, and 116 in DOAJ. The searches yielded no results in the Latindex and CLASE platforms.

To ensure the relevance of the studies included, predefined inclusion and exclusion criteria were established. Studies addressing university dropout or attrition in university students (undergraduate or postgraduate) belonging to Indigenous communities were included. No restrictions were placed on age range, sex, or geographic location, and studies in Spanish or English, published as full and accessible articles, analyzing risk or protective factors related to dropout, were considered. Additionally, studies providing information on dimensions or determinants of school attrition in this population group were included.

In contrast, theoretical studies, such as systematic reviews, meta-analyses, and literature reviews, were excluded, as well as research focused on non-university school populations or vulnerable populations other than Indigenous communities.

The study selection process was carried out in three stages following the PRISMA guidelines ( Raúl, 2025). PRISMA flow diagram has been presented as extended data as well ( Raúl, 2025; See extended data as figure 1). First, an initial review of the titles, abstracts, and keywords of the 23,391 unique records identified after removing duplicates within each database was conducted. During this stage, 23,271 studies were excluded for not meeting the eligibility criteria, including 408 articles that did not address a university population, 21,873 that did not include an Indigenous population, 736 that did not focus on the phenomenon of dropout or attrition, and 306 studies of a theoretical or bibliometric nature. In the second stage, 120 articles were selected for full-text review. During this phase, 11 duplicate records across all databases were identified and removed, as they had not been detected in the first stage. Additionally, 86 studies were excluded for not meeting the inclusion criteria after a detailed full-text review ( Raúl, 2025; See the extended data as Table 1). Finally, 23 studies met all the established criteria and were included for analysis and narrative synthesis.

A standardized format was used for extracting key data, which included information on the general characteristics of the studies (author, year, country, methodological design), participants (educational level, rural context, Indigenous population), risk and protective factors associated with dropout, and methodological approaches used. Data extraction was conducted independently by two reviewers (INITIALS), with discrepancies resolved by consensus with the principal investigator (RQA).

For the description of the results, the general characteristics of the included studies were analyzed and described, reviewing aspects such as the country, sample size, proportion of Indigenous students, geographic context (rural, urban, or mixed), and type of institution (public, private, or mixed). To ensure the comparability of the data, absolute frequencies and percentages were calculated. Regarding the age of the participants, some studies only reported age ranges; therefore, it was necessary to calculate the estimated mean age using the lower and upper limits as references. Due to the heterogeneity of the studies, quantitative results were synthesized through descriptive analyses. Non-homogeneous or missing data were presented through a narrative synthesis, complemented by tables that facilitate the comparison between studies.

This systematic review included 21 records, detailed in Table 1. Most of the studies were conducted in South America (9/21), followed by Central America (4/21), North America (2/21), Oceania (5/21), and Europe (1/21). Among the 226,049 subjects included in the 21 records, only 2,340 corresponded to Indigenous university students, representing 1.03% of the total sample. Of the total records, 2 studies did not provide disaggregated information on Indigenous university populations ( De Gracia Vega & Vega de Martínez, 2024; Kirby et al., 2020; Saldarriaga Isaza et al., 2019). The average age of the participants was 24.19 years, considering only the 9 records that reported this information. In some cases, the age range was adjusted to calculate these values ( Bergamaschi et al., 2016; Calegare & Sales, 2023; Gutiérrez et al., 2015; Li & Jackson, 2024; Meneses Pardo, 2011; Oliver et al., 2015; Rafael Riveros et al., 2024; Rodríguez Gallegos & Alvites-Huamaní, 2023; Salazar Cóndor, 2022).

In these studies, 63.06% of the participants were men, and 36.95% were women. However, 11 out of 21 studies did not report the sex of the participants ( Benitez & Ramírez, 2023; De Gracia Vega & Vega de Martínez, 2024; Gutiérrez et al., 2015; Kirby et al., 2020; Li & Jackson, 2024; Loureiro et al., 2024; Lydster & Murray, 2019; Rodríguez Gallegos & Alvites-Huamaní, 2023; Salazar Cóndor, 2022; Saldarriaga Isaza et al., 2019; Walton et al., 2020), and only 2 records focused exclusively on Indigenous women ( Rafael Riveros et al., 2024; Rodríguez & Ossola, 2021).

Regarding the type of higher education institution, the majority of studies (17/21) were conducted on students from public universities ( Benitez & Ramírez, 2023; Bergamaschi et al., 2016; Calegare & Sales, 2023; De Gracia Vega & Vega de Martínez, 2024; Gutiérrez et al., 2015; Hearn et al., 2019; Kozlova et al., 2022; Loureiro et al., 2024; Meneses Pardo, 2011; Oliver et al., 2015; Rafael Riveros et al., 2024; Rodríguez Gallegos & Alvites-Huamaní, 2023; Rodríguez & Ossola, 2021; Salazar Cóndor, 2022; Saldarriaga Isaza et al., 2019; Schmidt Araneda et al., 2023; Walton et al., 2020), while only one study focused on students from private universities ( Lydster & Murray, 2019). Additionally, three records analyzed populations from both public and private universities ( Kirby et al., 2020; Li & Carroll, 2020; Li & Jackson, 2024).

In Table 2, the cultural characteristics of the participants are presented. Of the total studies, 4 were conducted in Indigenous communities in Central America (Mexico: Tseltal, Wixarika, Nahua, Mazahua, Purépecha, Otomí; Panama: Ngäbe and Buglé); 2 out of 21 records in North America (Canada: Métis, Inuit, and First Nations); 6 in South America (Colombia: Indigenous peoples of the Amazon; Argentina: Wilchi and Kolla; Chile: Mapuche; Peru: Quechua and Aimara; Brazil: Tuyuca, Tukano, Bará, Munduruku, Yahua, Sateré-Mawé, Tariana, Kaingang, Guaraní, Fulniô, and Juruna; Ecuador: Achuar, Andwa, Kichwa, Sapara, Shiwiar, Shuar, and Waorani); 4 in Oceania (Australian Aboriginals and Torres Strait Islanders); and 1 in Europe (Russia: Komi-Permyak). Although 4 studies did not report this information ( Kirby et al., 2020; Li & Jackson, 2024; Loureiro et al., 2024; Meneses Pardo, 2011). Regarding place of residence, in 15 out of 21 studies, participants reported living in rural areas ( Benitez & Ramírez, 2023; Bergamaschi et al., 2016; Calegare & Sales, 2023; De Gracia Vega & Vega de Martínez, 2024; Gutiérrez et al., 2015; Kirby et al., 2020; Li & Carroll, 2020; Li & Jackson, 2024; Lydster & Murray, 2019; Meneses Pardo, 2011; Oliver et al., 2015; Rafael Riveros et al., 2024; Rodríguez & Ossola, 2021; Saldarriaga Isaza et al., 2019; Schmidt Araneda et al., 2023), while 5 out of 21 lived in both rural and urban areas ( Hearn et al., 2019; Kozlova et al., 2022; Loureiro et al., 2024; Salazar Cóndor, 2022; Walton et al., 2020). One study did not specify the area of residence ( Rodríguez Gallegos & Alvites-Huamaní, 2023).

Finally, the studies reported a great linguistic diversity. Identified languages included Andean languages (Quechua, Aimara), Amazonian languages (Achuar, Andwa, Sapara, Shiwiar, Shuar, Waorani, Yahua, Tuyuca, Tukano, Bará, Munduruku, Sateré-Mawé, Tariana, Fulniô, Juruna), Southern Cone languages (Mapuche, Wilchi, Kolla, Guaraní, Kaingang), Mesoamerican languages (Tseltal, Ch’ol, Nahuas, Mazahuas, Purépechas, Otomíes, Wixarika, Huichol), North American languages (Métis, Inuit, First Nations), European languages (Komi-Permyak), and dominant languages (Portuguese) ( Benitez & Ramírez, 2023; Bergamaschi et al., 2016; Calegare & Sales, 2023; Gutiérrez et al., 2015; Kozlova et al., 2022; Rafael Riveros et al., 2024; Rodríguez Gallegos & Alvites-Huamaní, 2023; Rodríguez & Ossola, 2021; Salazar Cóndor, 2022; Saldarriaga Isaza et al., 2019; Schmidt Araneda et al., 2023). In 5 studies, the participants were bilingual ( Benitez & Ramírez, 2023; Rodríguez & Ossola, 2021; Salazar Cóndor, 2022; Saldarriaga Isaza et al., 2019; Schmidt Araneda et al., 2023); while in 10 records, this information was not reported ( De Gracia Vega & Vega de Martínez, 2024; Hearn et al., 2019; Kirby et al., 2020; Li & Carroll, 2020; Li & Jackson, 2024; Loureiro et al., 2024; Lydster & Murray, 2019; Meneses Pardo, 2011; Oliver et al., 2015; Walton et al., 2020).

Then, we analyzed the methodological aspects of the studies. According to Table 3, 20 studies examined university dropout at the undergraduate level; of these, 17 out of 20 analyzed a wide variety of university programs ( Benitez & Ramírez, 2023; Bergamaschi et al., 2016; De Gracia Vega & Vega de Martínez, 2024; Gutiérrez et al., 2015; Hearn et al., 2019; Kozlova et al., 2022; Li & Carroll, 2020; Li & Jackson, 2024; Loureiro et al., 2024; Meneses Pardo, 2011; Oliver et al., 2015; Rafael Riveros et al., 2024; Rodríguez Gallegos & Alvites-Huamaní, 2023; Rodríguez & Ossola, 2021; Salazar Cóndor, 2022; Saldarriaga Isaza et al., 2019; Schmidt Araneda et al., 2023).

The identified areas include Education Sciences, Exact Sciences and Engineering, Health, Social Sciences, Humanities, Agricultural Sciences, Administrative Sciences, Economics, Tourism, Arts, Environmental Studies, and Cultural and Educational Development. However, fields such as Architecture, Technology, and Information Sciences had lower representation. The only study that analyzed dropout at the graduate level ( Calegare & Sales, 2023) focused exclusively on Education and Anthropology programs. However, three records did not report the participants’ professional fields ( Kirby et al., 2020; Lydster & Murray, 2019; Walton et al., 2020).

Regarding the methodologies used, 7 out of 21 studies employed qualitative approaches using semi-structured interviews and observations to explore individual and community experiences associated with dropout ( Benitez & Ramírez, 2023; Calegare and Sales, 2023; Gutiérrez et al., 2015; Lydster & Murray, 2019; Oliver et al., 2015; Rafael Riveros et al., 2024; Rodríguez & Ossola, 2021). Meanwhile, 6 out of 21 studies applied quantitative methodologies incorporating statistical analyses of institutional databases and surveys ( Li & Carroll, 2020; Li & Jackson, 2024; Rafael Riveros et al., 2024; Rodríguez Gallegos & Alvites-Huamaní, 2023; Saldarriaga Isaza et al., 2019; Schmidt Araneda et al., 2023). Finally, 8 studies used mixed methodologies, combining interviews with structured data analysis ( Bergamaschi et al., 2016; De Gracia Vega & Vega de Martínez, 2024; Hearn et al., 2019; Kirby et al., 2020; Kozlova et al., 2022; Loureiro et al., 2024; Meneses Pardo, 2011; Walton et al., 2020).

Finally, the risk and protective factors associated with Indigenous university dropout were analyzed (see Table 4). First, the prevalence of Indigenous university dropout was estimated at 33.59%, based on data from seven studies that provided this information ( Bergamaschi et al., 2016; Loureiro et al., 2024; Meneses Pardo, 2011; Rafael Riveros et al., 2024; Salazar Cóndor, 2022; Saldarriaga Isaza et al., 2019; Schmidt Araneda et al., 2023). From these studies, an average Indigenous University Dropout (IUD) rate of 33.59% was estimated. However, 66.67% of the studies did not report this data, limiting the ability to conduct a more comprehensive analysis of dropout rates. The university dropout rate ranged between 20% and 66%, except for a single record that reported a lower rate ( Loureiro et al., 2024). Similarly, the demographic variable analysis reported that IUD was higher among women (28.57%), considering data from six studies ( Bergamaschi et al., 2016; Hearn et al., 2019; Meneses Pardo, 2011; Rafael Riveros et al., 2024; Rodríguez Gallegos & Alvites-Huamaní, 2023; Walton et al., 2020); However, in three studies ( Loureiro et al., 2024; Salazar Cóndor, 2022; Saldarriaga Isaza et al., 2019) higher dropout rates were reported among men (14.29%). In this regard, it is important to highlight that 52.38% of the studies did not report sex as a variable associated with dropout ( Benitez & Ramírez, 2023; Calegare & Sales, 2023; De Gracia Vega & Vega de Martínez, 2024; Gutiérrez et al., 2015; Kirby et al., 2020; Kozlova et al., 2022; Li & Carroll, 2020; Li & Jackson, 2024; Lydster & Murray, 2019; Oliver et al., 2015; Rodríguez & Ossola, 2021; Schmidt Araneda et al., 2023).

Regarding the risk and protective factors identified in the included studies, economic factors were observed as the main obstacle to university retention, reported in 85.7% (18/21) of the records. Other significant factors include linguistic barriers, present in 47.6% of the studies (10/21), and geographical location, mentioned 33.3% of the time (7/21). On the other hand, another identified risk factor was the lack of social support, reported in only 23.8% of the records (5/21).

On the other hand, among the protective factors, financial support was the most frequent, identified in 76.2% of the studies (16/21), followed by mentoring or academic guidance, reported in 47.6% (10/21), as well as social support, mentioned in 42.8% (9/21). The least frequent factors were cultural reaffirmation and an inclusive environment, both reported in 28.6% of the studies (6/21).

This study features crucial restrictions which researchers need to view when undertaking assessment of both results and their application range. Methodological differences between the included research studies create major difficulties for analysis. Research methods together with data collection procedures and the list of analytical methods introduce variability which hinders observation of shared results and understanding of established patterns. The wide cultural and economic diversity across geographic areas makes this observation especially crucial for research on Indigenous university dropout. One of the continuous challenges within the studies under analysis involves the insufficient breakdown of data. Many study records failed to include critical variables that identified gender, age and subject majors of participants. The absence of detailed participant data hinders advanced research examinations which restricts scientists from comprehending how various subgroups from the Indigenous population experience dropout variations. Some studies fail to provide information about the student proportion among Indigenous students which indicates limited attention to proper documentation of this population. The territorial distribution of studies in this review appears irregular since different geographic areas show various populations. Most of the research belongs to Latin America and Oceania whereas regions in Asia and Africa lack representation despite their substantial population of Indigenous communities cultural and demographically speaking. The geographical imbalance diminishes the worldwide usefulness of research results and hinders researchers from running multinational studies that could broaden our understanding about this subject. The dearth of investigations regarding graduate-level educational research displays a scholarly trend toward studying primary stages of higher education. Social and economic obstacles that face Indigenous students seem more prevalent in their education level progression. Active research procurement concerning graduate education leads to substantial knowledge gaps identifying the obstacles Indigenous students encounter from beginning to end in their academic journey. Insufficient data about risk plus protective components restricts researchers from achieving a full understanding of what factors lead to university termination or continued academic participation. The collection of non-standardized data hinders accurate identification of optimal approaches to solve these problems despite documenting generic characteristics. Standards for methodological research design and more detailed data acquisition need establishment because these issues affect the current ability to study university dropout or retention in higher education.