c e p s Journal | V ol.11 | N o 4 | Y ear 2021 213 Self-Concept in Immigrant School Children and the Impact of Length of Residence: Evidence from PISA 2015 for Current Educational Practice Sandra Figueiredo* 1 , João Marôco 2 , Margarida Alves Martins 3 and Odete Nunes 4 • Comparative analyses of the Programme for International Student Assess - ment between immigrant and native students place immigrant students in an unfavourable position in schools, with disadvantageous indicators re - garding socioeconomic and professional paths. However, the Programme for International Student Assessment assesses a series of dimensions that involve constructs that have been little studied in the school immigrant population and that relate to self-concept and school adjustment. Based on the Programme for International Student Assessment’s most recent edi - tion, Portugal’s database of 7,325 15-year-old students was analysed. We se - lected 438 immigrant cases with two objectives: (1) to evaluate the impact of the length of exposure in the host country on three dependent variables of school adjustment: sense of belonging, perceived loneliness and atti - tudes towards school (expectations of educational and professional oppor - tunities); (2) to evaluate the differences in results for the same dependent variables, but considering the first and second generation of immigrants in Portugal. For the data analysis, sampling weights and plausible values were analysed with the International Database Analyzer. The results show that students who have been in the country for a year or less have greater dif - ficulties and increased significant differences compared to other migrant groups in the referred indices of self-concept and inclusion. However, oth - er groups, especially those with periods of long-term residence between four and five years, also face substantial levels of school maladjustment. Keywords: PISA assessment, immigrants, generations, self-concept, length of residence, school adjustment, sense of belonging 1 *Corresponding Author. Department of Psychology, Universidade Autónoma de Lisboa Luís de Camões, Portugal; sfigueiredo@autonoma.pt. 2 Instituto Universitário de Ciências Psicológicas, Sociais e da Vida, Portugal. 3 Instituto Universitário de Ciências Psicológicas, Sociais e da Vida and Centre for Educational Research, Portugal. 4 Department of Psychology, Universidade Autónoma de Lisboa Luís de Camões, Lisbon, Portugal. doi: 10.26529/cepsj.859 214 self-concept in immigrant school children and the impact of length of residence Samopodoba otrok priseljencev in vpliv dolžine prebivanja: podatki PISA 2015 o trenutni izobraževalni praksi Sandra Figueiredo, João Marôco, Margarida Alves Martins in Odete Nunes • Primerjalne analize Programa mednarodne primerjave dosežkov učen - cev (angl. Programme for International Student Assessment ) med prise - ljenskimi in domačimi učenci uvrščajo otroke priseljence v neprijeten položaj v šolah, in sicer z neugodnimi indikatorji glede družbenoeko - nomskih in poklicnih poti. Program za mednarodne primerjave dosežkov učencev pa ocenjuje število razsežnosti, ki vključujejo konstrukte, ki so bili pomanjkljivo preučeni v šolski priseljenski populaciji in ki zadevajo samopodobo ter prilagoditev na šolo. Na podlagi najnovejših podatkov Programa za mednarodne primerjave dosežkov učencev je bila analizirana portugalska podatkovna zbirka 7.325 15-letnih učencev. Izbrali smo 438 primerov priseljencev z dvema ciljema: 1) da ocenimo vpliv dolžine iz - postavljenosti v državi gostiteljici na tri odvisne spremenljivke prilagaja - nja na šolo, tj. na občutek pripadnosti, zaznavo osamljenosti in na odnos do šole (pričakovanja glede izobraževalnih in poklicnih priložnosti); 2) da ocenimo razlike v rezultatih za iste odvisne spremenljivke ter pri tem upoštevajoč prvo in drugo generacijo priseljencev na Portugalskem. Za analizo podatkov smo z Mednarodnim analizatorjem podatkovnih zbirk (angl. International Database Analyzer ) utežili podatke glede na velikost vzorca in analizirali verjetne vrednosti. Rezultati kažejo, da imajo učen - ci, ki so bili v državi leto ali manj, večje težave, obenem pri njih prihaja do izrazitejših pomembnih razlik v primerjavi z drugimi priseljenskimi skupinami glede na omenjene kazalnike samopodobe in inkluzije, ven - dar pa se te druge skupine, zlasti tiste z obdobji dolgotrajnega prebivanja med štirimi in petimi leti, vseeno spoprijemajo z občutno stopnjo nepri - lagojenosti na šolo. Ključne besede: PISA ocenjevanje, priseljenci, generacije, samopodoba, dolžina prebivanja, prilagojenost na šolo, občutek pripadnosti c e p s Journal | V ol.11 | N o 4 | Y ear 2021 215 Introduction The performance of immigrant students has been little studied in the context of the Programme for International Student Assessment (PISA) of the Organisation for Economic Cooperation and Development (OECD). One of the goals of PISA is the comparative analysis of the performance of 15-year-olds from different countries (OECD, 2016) in three domains: reading, mathematics and science. At triennial intervals, this international research allows a presenta - tion of indicators of the evolution of the literacy of students from more than 70 countries (OECD, 2016). The analysis predicts the academic and professional success of student generations as they approach the end of compulsory school - ing. According to the most recent database, in 2015, PISA was deployed in 72 countries and economies. In addition to assessing academic content, PISA tests are comprehensive in that they allow us to examine other variables that have been neglected in com - parative scientific analyses of students’ PISA performance, especially immigrant populations. These variables are mainly related to the conditions of the host coun - tries; specifically, the schools and teaching methods perceived by the migrant stu - dents. Self-concept, school adjustment and students’ expectations regarding their performance and professional future in the country in which they are living are crucial. There is also a need to consider covariates such as the students’ country of origin, languages spoken at home, mother tongue (L1), and the generation of immigration to which they belong (first or second). All of these variables appear in the OECD database and the 2016 report (OECD, 2016). Based on previous PISA assessments, studies such as that by Cummins (2008) have reflected on the impact of the L1 variable on immigrant outcomes. These studies have concluded that proficiency in L1 is an important factor to consider in schools (valuing students’ learning at home and on a continuous basis) and is associated with better performance, rather than total immersion in the second language (L2) (Agirdag & Vanlaar, 2018). One of the main concerns, especially for cases like the Portuguese, has been the performance in reading. More recent PISA-based studies have focused on the immigrant popula - tion and its importance for the variability of general results by country, that is, how to differentiate native students and immigrant students in the same tests. The latest OECD (2016) and Factbook (2016) reports show more optimistic averages (above the OECD performance average) than previous PISA studies for immigrants in specific areas of problem solving, mathematics and reading. However, these results have mostly placed migrant minority students in lower performance positions (Hippe & Jakubowski, 2018). 216 self-concept in immigrant school children and the impact of length of residence The evidence has associated the low performance of immigrant adoles - cent students with the unfavourable socioeconomic contexts of their families, the school environment (as perceived by non-native students) and low expecta - tions of school and professional success, especially job prospects after complet - ing compulsory schooling (Hippe & Jakubowski, 2018). The same authors also found that from 2006 until the last PISA tests (2015), the differences between migrant and non-migrant groups in a considerable number of European Union countries remained significant, with immigrant students being in a disadvan - taged position. The most predictive variation was according to the host country. Other dependent variables have arisen, such as the language spoken at home and expectations that incorporate self-concept (such as the sense of be - longing to the school). Another study by Marôco et al. (2016) found that socioec - onomic status did not explain the differences between immigrant students (from the second generation) and non-migrant students, with immigrant students being better positioned than their native peers. On the other hand, immigrant populations have interested researchers in the fields of psychology and education (and economics) because of the difference in their performance between gener - ations. Generations are identified as first (born in another country) and second (descendants of the first generation, born in the host country) taking into ac - count variations in school policies in the host countries (Cattaneo & W olter, 2015; Volante et al., 2019). On the one hand, studies reveal that the latest generation has shown im - provements in results since the last two PISA studies due to the new welcoming and preparation measures adopted by European Union (EU) and non-EU coun - tries to promote the inclusion of these school populations (Šori et al., 2011; Y eas - min & Uusiautti, 2018). On the other hand, in Portugal, the second generation scores were lower than those of the first generation (OECD, 2016), as opposed to Northern Europe countries and the Netherlands, for example. Regarding read - ing, in Portugal, the second generation surpassed the first, yet native students continue to outperform their migrant peers in class (OECD, 2016). The main reason for coming to Portugal was family reunification, as in most immigration cases (Di Liberto, 2015; OECD, 2016), and immigration was on a permanent basis and in high numbers (9,607,397 individuals, OECD, 2016). Whereas Portuguese immigration declined in 2013, the number of refugees received increased twofold in the 2015–2016 timeframe (OECD, 2016). Promoting the inclusion of migrant minorities and encouraging the ed - ucational equity proclaimed in the United Nations sustainability goals (OECD, 2018) is hampered by inequality of performance in immigrant groups (Welch, 2018). This difficulty is also attributed to school environment factors (Agasisti c e p s Journal | V ol.11 | N o 4 | Y ear 2021 217 & Zoido, 2018) and a lack of awareness that different groups of students involve different pedagogical approaches (Sakellariou, 2018; Woessmann, 2016). Fur - thermore, these groups are in different stages of development regarding their adaptation because they have different exposure times to the conditions of the host country and therefore different maturation. One of the factors that has been less studied is exposure time or length of residence (LOR). In studies of the second language (L2), bilingualism and migrant populations, this terminology can also be identified with the “age of onset of acquisition” (referring to the age of the acquisition of L2 in the case of migrant populations). Immigrant (and refugee) students with lower levels of exposure tend to be more exposed to higher levels of anxiety and malad - justment to the host school compared to colleagues who, irrespective of their origin, entered the country more than five years earlier (Kia-Keating & Ellis, 2007; Salmela-Aro et al., 2018). In the literature, the five-year period for immigrant status is considered “long term” (Grubanov-Boskovic et al., 2017). Students with the lowest LOR are among those facing more difficulties in PISA tests at an early stage, especially regarding problem solving (Martin et al., 2012). This was also verified in studies not related to the OECD (Jang et al., 2013). On the other hand, the last years of compulsory schooling are more associated with anxiety symptoms and lower rates of wellbeing in the general student population (Wang et al., 2015). Thus, the case of the immigrant school population of the same levels of education (especially between grades 10 and 11) is considered to be a group more subject to such symptoms and to wellbeing problems due to the need to adapt to new contexts and a new school environment. Regarding the length of exposure or arrival date in the host country, several studies have examined the relationship between recent entry into the country and negative (school) performance as perceived by immigrant stu - dents (Bozick et al., 2016; Martinez-Taboada et al., 2017). Length of residence is associated with the likelihood of greater or lesser employment opportunities in the host country, with immigrants with more than five years of permanent residence in the country having an advantage. Portugal is part of the group of countries where this correlation has occurred, and, negatively, it is also one of the countries where immigrants perform professional functions requiring low literacy skills (Grubanov-Boskovic et al., 2017). Portugal was also the object of analysis by Schnell and Azzolini (2015), based on data from previous PISA tests. In Portugal and in three other cases in Europe, the position of immigrants is fragile, especially those who arrive later (with lower LOR) and after the age of six years. The study highlights the 218 self-concept in immigrant school children and the impact of length of residence disadvantage variable of immigrant families who take on less specialised oc - cupations than they could given their high level of education obtained in the country of origin (Schnell & Azzolini, 2015). If 15-year-olds have a poor self-concept and underperform in tests, the likelihood of positive expectations about the labour market in the destination country decreases, as does their effective employability in specialised jobs requir - ing high educational qualifications. LOR may also influence other dimensions that are assessed in the PISA test but have not been adequately examined in the literature, such as school adaptation. Adjustment to school is measured in PISA (2015) according to items such as self-assessment of the sense of belonging and integration in the school of the destination country. To a large extent, this adjust - ment determines the level of inclusion and sense of belonging in the host society after completion of compulsory schooling (Schachner et al., 2017). Studies such as Schachner et al. (2017) indicate that host countries with more supportive and mediating inclusion measures for immigrant students (with Portugal appearing in the group of countries with moderate support measures) have better-adjusted and self-determined immigrant students, who also have better performance. When related to specific ethnicities of migrant minorities, poor school adjustment is significantly associated with dropout rates and low levels of sense of inclusion in the school environment (Ali & Larsson, 2018). Adjustment to school occurs in several ways; for example, it is related to motivation vis-à- vis the school environment. However, the specific studies based on the perfor - mance and self-concept scores of immigrant students in the PISA tests are still not consensual about the evidence on school adjustment in immigrant minori - ties (Burgess & Heller-Sahlgren, 2018), and school adjustment in terms of sense of belonging remains poorly studied (Ham et al., 2017). Loneliness in school perceived by the immigrant students also has a cor - relation with low school performances (Mai & Asma’A, 2016). The loneliness perceived by the subjects is also evaluated by PISA (OECD, 2016) and it is im - portant to understand that these emotional aspects (such as perceived malad - justment) integrate the construct of sense of belonging (Chiu, Pong, Mori, & Chow, 2012). However, we have no knowledge of any study on the relationship between perceived loneliness and LOR for immigrant student groups. Concerning professional expectations after schooling, that is, with re - gard to students’ attitudes towards school (Chiu et al., 2012), little is yet known about the expectations of immigrant students and the observable expectations among different ethnic groups. In comparison with native peers and first-gen - eration immigrants, however, immigrant subjects have higher rates of pos - itive educational expectation, with the main variable that differentiates this c e p s Journal | V ol.11 | N o 4 | Y ear 2021 219 expectation being the host country and the respective conditions (Burgess & Heller-Sahlgren, 2018; Chykina, 2019). Furthermore, the educational expecta - tion of second-generation immigrant students is strongly related to their will - ingness to enter higher education and to have better professional opportuni - ties than their parents (Childs et al., 2017; Figueiredo et al., 2016). Educational preparation, especially the specialisation of skills in schools, leads to better pro - fessional opportunities for immigrants (Rangel & Shi, 2019). Regarding the educational and professional expectations of young im - migrants, Nygard (2017) confirmed that the optimism of immigrant status ex - plains the high aspiration of immigrants to good levels of education. It also accounts for the diversified mobility of this population in the context of the plu - rality of educational opportunities. From an early stage, other authors (Bertschy et al., 2009) have reported that PISA indicators help us to understand students who are more or less suited to particular job opportunities. In contrast, a study by Nygard (2017) found that the early tracking of educational competencies and trajectories has underlined the perception of disadvantage, especially socioec - onomic, of immigrant groups because of their difference in performance from that of their native peers (Feliciano & Lanuza, 2016; Fernández-Reino, 2016). The present study has two objectives: (1) to evaluate the impact of the length of exposure in the host country (the age of immigration with the pur - pose of obtaining permanent residence in Portugal) of immigrant students aged 15 on three dependent variables related to school adjustment: sense of be - longing to the school, perceived loneliness, and expectations of educational and professional opportunities after the completion of compulsory schooling; (2) to evaluate the differences in results for the same dependent variables, but consid - ering the first and second generation of immigrants in Portugal. Method Participants Sample selection followed the PISA sampling methodology based on the protocol determined by the Educational Testing Service (ETS, Prince - ton), which has defined the different versions of the PISA test (according to the countries and their official languages), the codification of items and the creation of databases. The main criteria for selecting the sample were that the students were 15 years of age and studying above the 7 th year of schooling. In 2015, 7 ,325 students from 246 schools participated in Portugal. The mean age of Portuguese students was 15.8 years (see Marôco et al., 2016). The type, size 220 self-concept in immigrant school children and the impact of length of residence and administrative nature of the school were criteria considered with the aim of covering the national population in an equitable way (mainland and islands). Students who had been exposed to the Portuguese language for less than one year were excluded (in the case of immigrants). Students with cognitive, functional or intellectual disorders were also excluded. Aspects related to the lim - its for the careful exclusion of PISA were previously determined so as not to ex - ceed school exclusion rates by country, which would bias the data (OECD, 2014). The choice of the age of 15 years is related to the expected age for attend - ing the cycle (high school) prior to the end of compulsory schooling. It is also the age group most likely to obtain information about professional prospects and actual performance in school, with applicability in two contexts of problem solving: in school and out of school. On the other hand, 15-year-old students are distributed by distinct school years, including grade-retained students. For the present study, only the immigrant population in Portugal was selected: 438 cases, of which 205 declared they were female and 217 male. Of the female group, 108 (41.1%) belonged to the first generation and 97 (31.4%) to the second generation of immigrants. Of the male group, 128 (42.1%) were part of the first generation and 89 (30.5%) belonged to the second generation of im - migrants (Table 1). Regarding the time of exposure in the host country (LOR), the duration varied from 1 to 14 years (M = 8.6; SD = 4.5). Table 1 Demographic data from the Portuguese database of PISA 2015: natives and immigrants Gender Immigration N of cases SD Female Native Second-Generation First-Generation 3372 97 108 .560 .314 .411 Male Native Second-Generation First-Generation 3385 89 128 .483 .305 .421 Instrument Four blocks of items were used in the four literacy domains: reading, mathematics, science and problem solving, according to the PISA assessment protocol, with an estimated duration of 30 minutes to fill each block. Following Item Response Theory (IRT), the items were analysed regarding the difficulty index and the discrimination index. c e p s Journal | V ol.11 | N o 4 | Y ear 2021 221 The coding of the items had a 92% agreement for all countries in which the tests were conducted (OECD, 2016). In addition to the blocks of tests in the areas of science, reading, mathematics and problem solving, specific variables (which will be presented as indices and not items in this study) related to school adjustment, attitudes and self-concept were analysed by PISA. Measures Independent variables: • Length of residence (LOR): the students indicate their date of arrival in Portugal, so the length of permanent residence is measured in years in PISA. In the SPSS database, we computed this variable to select only the immigrant cases and only for Portugal. • First and second generation of immigrants: the “index immigration sta - tus” of PISA evaluates whether the participants belong to one of three categories: natives, first generation of immigrants, or second generation of immigrants. Dependent variables: • School adjustment is a psychometric scale that evaluates how well the student fits the school. It is composed of 4 Likert type items (1-strongly agree to 4-strongly disagree). One such item is “I feel awkward and out of place in my school” (PISA 2015 code: ST034Q04TA) and is intended to evaluate how students feel displaced in their school environment. Stu - dents have to respond on a scale of 1 to 4, with 1 being “strongly agree” and 4 being “strongly disagree” . • Another item is the perception of loneliness and wellbeing in school me - asured by the item (PISA code: ST034Q06TA) “I feel lonely at school” with a scale of response from 1 = “strongly agree” to 4 = “strongly disagree” . • Also related to school adjustment is expectation of career opportuniti - es after compulsory schooling: expectation related to performance and professional opportunities after school is measured on a scale of respon - se from 1 = “strongly disagree” to 4 = “strongly agree” , where 1 corre - sponds to the absence of agreement with good options or professional opportunities after high school. The item is as follows: “I want to be able to select from among the best opportunities available when I graduate” (PISA code: ST119Q02NA). 222 self-concept in immigrant school children and the impact of length of residence Procedure The tests were administered by test administrators with previous training in schools for the administration of blocks of tests and for coding tests in their respective academic areas. The test administrators received their training from a supervising teacher according to OECD parameters. Whenever there were doubts, administrators turned to their supervisor (Marôco et al., 2016, idem). It was ensured that all of the selected schools were equipped with com - puters and the session dates were confirmed. The samples were checked and correctly identified by the test administrators. Each student was given a USB flash drive with the tests. After completing each test, the students submitted the answers that were stored via the web in the PISA database. Data Analysis The study focuses on variables concerning the individual’s self-concept (see section Measures) and on school aspects through a series of linear regres - sion analyses, considering the sampling weights of each student using the IDB Analyzer (International Association for the Evaluation of Educational Achieve - ment, IEA). The sampling weights variable was identified as “Final Trimmed Nonresponse Adjusted Student Weight” (FSTUWT). The syntax produced by the IDB Analyzer was then run in SPSS Statistics (v. 24, IBM, Armonk, NY) on the previous immigrants selected database. Results • Question 1 : to evaluate the impact of the length of exposure in the host country (the age of immigration) of immigrant students aged 15 years on three dependent variables related to the school adjustment: sense of belonging to the school, perceived loneliness, expectations of educatio - nal and professional opportunities after the completion of compulsory schooling; • Question 2 : to evaluate the differences in results for the same dependent variables identified in Question 1, but considering the first and second generation of immigrants in Portugal. Question 1 Immigrant students with a shorter length of residence, and therefore with less exposure to the school context and to L2, had higher means regarding c e p s Journal | V ol.11 | N o 4 | Y ear 2021 223 school maladjustment and perceived loneliness in school, but not in a linear way. In the first case, in “school maladjustment” , the subjects with one and four years of LOR are those showing less adjustment. On the other hand, with regard to perceived loneliness, four and five years of residence in the host country emerge as more critical periods, as the means are lower. See Figure 1. Figure 1 Means and standard deviation scores considering LOR x sense of belonging (blue line), perceived loneliness (yellow line) and professional expectations after graduation (grey line) Tables 2 and 3 present a summary of the descriptive analyses (means of response by index) for the immigrant group regarding the three dependent variables and the coefficients of the linear regression analyses: in models 1 and 2 – sense of belonging x LOR ( β = .051; S.E. = .011; t = 2.268; R 2 = .13) and perceived loneliness/wellbeing x LOR ( β = .042; S.E. = .008; t = 2.401; R 2 = .10) – there are positive and statistically significant coefficients that point to the im - pact of LOR on the perception of school belonging and on perceived loneliness. It was verified (Figure 1) that recent arrival in the country was not always asso - ciated with low or negative levels of sense of school belonging and of perceived 224 self-concept in immigrant school children and the impact of length of residence loneliness in the school environment (and, consequently, in the learning en - vironment). In general, however, the higher the LOR, the higher the school adjustment (mainly the sense of belonging). In the third model, shown in Table 3, the coefficients are negative and highly significant, which reveals that the higher the LOR, the less stable the per - ception of good professional and academic opportunities that the immigrant students have, unlike the students recently arrived in Portugal ( β = -.041; S.E. = .041; t = -3,597; R 2 = .22). By checking the means for each category (accord - ing to length of residence, Table 2) the subjects seem more optimistic about the possibility of choosing the best options (labour market) after high school. However, two cases were observed: individuals recently arrived in Portugal (in months <1 year) and others (14 years of LOR) presented lower means in this school adjustment index. Table 2 Descriptive statistics for school adjustment: (1) sense of belonging, (2) perceived loneliness at school, and (3) professional expectations after graduation N Minimum Maximum Mean SD Valid N (listwise) 438 <1 year 14 years 8.568 4.4807 I feel awkward and out of place in my school. 1 4 3.01 .879 Valid N (listwise) 427 <1 year 14 years 8.352 4.4632 I feel lonely at school. 1 4 3.39 .802 Valid N (listwise) 434 <1 year 14 years I want to be able to select from among the best opportunities available when I graduate. 1 4 3.14 .675 c e p s Journal | V ol.11 | N o 4 | Y ear 2021 225 Table 3 Regression linear coefficients for indices of school adjustment: sense of school belonging (model 1), perceived loneliness at school (model 2) and professional expectations after graduation (model 3) considering Length of Residence (LOR) Regression coefficient Coefficient β (s.e.) Standardised (t-value) p Model 1 2.646 .104 25.346 Constant sense of belonging .024 .011 2.268 .051 Model 2 3.086 .091 33.751 Constant perception loneliness .020 .008 2.401 .042 Model 3 3.638 .055 66.659 Constant expectations .022 .006 .3597 -.041 Note. β – the standardised linear regression coefficients. Question 2 As for the prediction of the generation of immigrants regarding the three dependent variables (the second research question of the study), the first- and second-generation groups did not differ significantly. See table 4. Table 4 Regression linear coefficients for indices of school adjustment: sense of belonging Immigration index status Regression coefficient Coefficient β (s.e.) Standardised (t-value) p Model 1 3.032 .013 241.017 Constant sense of belonging -.007 -.080 .016 .001 Model 2 3.392 .012 278.174 Constant perception loneliness -.055 .055 .012 .010 Model 3 3.462 .011 314.391 Constant expectations -.053 -.424 .014 -.006 Note. β – the standardised linear regression coefficients. Model 1 = perceived loneliness at school; model 2 = professional expectations after graduation; model 3 = considering Immigration Index Status. 226 self-concept in immigrant school children and the impact of length of residence Discussion The study initially aimed to examine the relationship between length of residence (or duration of stay in the host country) and variables related to the school environment, especially the perception of this environment by immi - grant students in Portugal tested in the last PISA (OECD, 2016). For this ques - tion (Question 1), the results showed a significant impact of LOR on school ad - justment, specifically on sense of belonging, perceived loneliness at school and expectations about professional opportunities after completion of compulsory schooling. The impact was perceived differently depending on the children’s periods of residence in the country. A more recent arrival date (lower LOR) did not always correspond to greater school maladjustment. On the one hand, as expected, students who had only arrived in Portugal a year earlier had a lower sense of belonging, thus cor - roborating previous studies by Kia et al. (2007) and Salmela-Aro et al. (2018). On the other hand, immigrant students with four years of residence (having entered Portugal at about 11 years of age) had a higher level of maladjustment (negative sense of belonging). These results, related to Question 1 of the study, are worrying, as they indicate different critical periods for the self-concept as perceived by the immigrants to the detriment of their development, especially regarding academic results (Parker et al, 2014). Thus, there is not necessarily a correlation between less time in the desti - nation country (less exposure to L2 and less experience in school) and low self- concept in indices such as those evaluated in the PISA test. There is a variability that must also consider the chronological (and maturation) age of the children (age of entry into the country and length of residence, which are two distinct variables), instead of being restricted to the typical factors that the literature analyses, such as socioeconomic status and parental investment (OECD, 2015). The OECD report (2015) shows the negative impact of exposure time in desti - nation countries such as Portugal, but it only refers to academic performance (especially reading). Our study presents the other side of PISA assessment, whose self-con - cept indices are important to explain the variation in immigrants’ academic and cognitive performance. This variation has recently been examined in the con - text of self-concept indices of young immigrant students, underestimating the predictive effect of other factors, such as the influence of the mother tongue or the socioeconomic level (Huget et al., 2013). Still on self-concept, Weber et al. (2018) examined the perceived vulnerability of immigrant children (up to 15 years of age), the strong ethnic identity of students, and the low levels of academic c e p s Journal | V ol.11 | N o 4 | Y ear 2021 227 performance. The authors concluded that this vulnerability was associated with disturbance of the sense of belonging. With respect to vulnerability and resilience in the first and second gen - eration of immigrants, the same study by Weber et al. (2018) found that the first generation was more resilient. Probably this resilient behaviour was a result of less awareness of stereotypes, making first-generation immigrants more likely to be the least affected generation regarding the sense of belonging and adjust - ment. In our research, the two generations did not differ significantly in their sense of belonging. In the academic domain, but not in self-concept, another study (Di Liberto, 2015) analysed the difference between generations and con - cluded that the second generation has an advantage regarding performance. Volante et al. (2019) found that in the context of academic (and cogni - tive) performance, the first generation scored lower in the same tests compared to the second generation, with the socioeconomic factor being a variable dif - ferentiating immigrant groups (immigrant groups with a more advantageous socioeconomic background tend to perform closer to immigrant peers). Ac - cording to authors such as Cordero, Cristobal and Santín (2018), the differences noted in students and between countries in large-scale studies such as PISA, the Progress in International Reading Literacy Study (PIRLS) or the Trends in International Mathematics and Science Study (TIMSS) are largely due to the countries’ assessment and inclusion policies and measures, especially in rela - tion to multicultural groups. These differences are also due to the actual format of the large-scale assessment mentioned above (Cordero et al., 2018). The present study focuses essentially on the factor of length of residence to explain the differences between migrant students. Previous studies have ex - amined the age of immigration, which confers immigration status as identified in the PISA coding categories, although with greater expression in student im - migrant groups in North America (Chen, 2019). One of the studied effects of LOR has been the effect of the presence of immigrants in the classroom on the performance of native colleagues, with negative effects being reported when students with lower LOR were included in the class (Bossavie, 2018). National studies (Castigo, 2017) also found that the immigrant or de - scendant characteristics of the immigrant family affected the performance of immigrant children, in comparison with the better scores of their native peers (especially in the Lisbon area). In the present study, no significant differences between immigrants and natives could be addressed as statistically significant. Castigo (2017) refers to several other studies using multilevel analysis models with the same evidence: in PISA studies, low performance mainly in reading and mathematics is explained by immigrant status. 228 self-concept in immigrant school children and the impact of length of residence Still related to Question 1, regarding the feeling of loneliness, in the con - text of self-concept, it was found that students with four and five years of LOR had the lowest scores, which corresponds to the higher rate of perceived loneli - ness in the school environment of the host country. Interestingly, these periods – four and five years of length of residence – are identified as critical periods for the school immigrant population, as they appear frequently and weakly in the self-concept indices analysed in this study. From another perspective, our data contradict information obtained in previous studies indicating that LOR equal to or greater than five years is less critical for positive self-concept and performance (Bozick et al., 2016; Marti - nez-Taboada et al., 2017). This evidence is in line with a study by Schnell and Azzolini (2015), in which Portugal is found to be one of the most fragile coun - tries in terms of reception, especially for children over six years of age entering destination countries (considered a critical period by Schnell & Azzolini, 2015). Recent studies (Nkemasong, 2018) have shown that immigrant students with low self-esteem and sense of belonging (to the community and school) become lonelier in school. Similarly, Chiu et al. (2012) concluded that the lone - liness perceived by non-native students is significant, but that it can be attenu - ated by the teacher-student relationship. However, migrant students tend to isolate themselves in school, which increases the rate of perceived loneliness (Petreñas et al. , 2011). PISA studies help us to understand the causes of immi - grant underperformance if this type of self-concept index, as presented in this study, is examined, as the feeling of loneliness is a maladjustment that directly explains poor performance in the tests, either at the academic or cognitive level (Chiu et al., 2012). In prior research, loneliness was more prominently perceived in first-generation than second-generation immigrants (Strohmeier & Dogan, 2012). However, in the present study, addressing Question 2, this was not observed. Although Portugal has only moderate measures regarding support for the reception of immigrants, these are not enough to explain the immigrants’ scores in academic tests of the assessed literacy domains. Schnepf (2007) found a similar deficit in other European countries that are destination countries for immigrant students. On the other hand, in terms of the satisfaction index with professional expectations after completion of compulsory schooling, immi - grant students in Portugal demonstrate expectation differences according to LOR. There are, however, two scenarios. First, those who have arrived more recently (<1 year) in Portugal have less favourable indices, as expected. Second, by contrast, students with maximum exposure time (or LOR) in Portugal are the least optimistic in this index. c e p s Journal | V ol.11 | N o 4 | Y ear 2021 229 These data support an earlier study by Grubanov-Boskovic et al. (2017), which identified Portugal as one of the countries in which the correlation be - tween LOR and professional benefits is evident. In parallel with other European countries, Portugal demonstrates fragility regarding the professional charac - teristics of the positions occupied by working-age immigrants. Regarding the topic of career expectations after completing schooling in the host country, Scandinavian studies, such as those by Hvistendahl and Roe (2004), refer to the importance of school reception measures to ensure the academic success of immigrants, who sometimes exceed the performance of native peers. When referring to academic success and professional expectations, the access of im - migrant students to higher education prior to entering the labour market in the host country is also relevant. In the last PISA (OECD, 2016), Marôco et al. (2016) reported that immigrants outperformed native Portuguese colleagues in some assessment areas. Expectations also influence academic performance, especially in generations of immigrants, which in turn influences the professional success analysed in the short term, from the early phase of transition from school to university and/or to the labour market (Bertschy et al., 2009). Conclusion The data from our study strongly indicate the predictive influence of length of residence (with the permanent duration index and not only “length of stay”) in determining immigrant school adjustment. School adjustment in - corporates the self-concept dimension. Multiple variables were used: sense of belonging, perceived loneliness and students’ professional expectations after graduation. These were intentionally selected in the research based on the 2015 PISA study. For all three variables, LOR revealed a significant impact, although not linearly, as was the case in previous literature. This means that lower LOR (referring to months or less than a year of residence in the country of destina - tion) does not entirely correspond to greater maladjustment, greater loneliness and less professional optimism. On the other hand, more evidence was col - lected regarding more critical periods of LOR (four and five years of residence in the host country, for example), which explains the variability in the self- concept of 15-year-old immigrant students assessed in the context of PISA, re - porting specifically to Portugal. Question 1 of the study was fully addressed. For the Question 2, however, there was no significant data to explore the impact of generations of immigrants. One of the limitations of the study is the lack of identification in PISA of the countries of origin of immigrant students. Country of origin is important 230 self-concept in immigrant school children and the impact of length of residence for a comprehensive analysis of crucial variables such as the individual’s educa - tional preparation prior to entering Portugal, the previous socioeconomic con - text, and the specific mother tongue (and languages spoken at home). However, the PISA database offers a high impact contribution to the educational orienta - tion of the populations in their school and professional contexts. One of the ob - jectives is to predict the trajectories of immigrant and non-immigrant students in the short and medium term, in higher education and in labour markets. This prediction is possible based on the conclusions of the performances of the students of more than 70 countries throughout the world. These performances indicate how reinforcement or maintenance measures can be taken to ensure academic and, subsequently, professional success. These objectives and studies are also included in the sustainability ad - vocated by the UN Goals 2030, specifically with regard to equity and access to quality education for all. We believe, however, that future research in the fields of psychology and education should benefit from the organised and valid con - tribution of this large-scale study (PISA) to explore the impact of immigrant students’ self-concept on performance in the PISA tests. Acknowledgments This work was funded by national funds through the FCT –  Fundação para a Ciência e a  T ecnologia  – as part of the project CIP/U AL – Refª Refª UIDB/04345/2020. Translation by Carolina Peralta . References Agasisti, T., & Zoido, P . (2018). Comparing the efficiency of schools through international benchmarking: Results from an empirical analysis of OECD PISA 2012 data. Educational Researcher , 47(6), 352–362. Agirdag, O., & Vanlaar, G. (2018). Does more exposure to the language of instruction lead to higher academic achievement? A cross-national examination. International Journal of Bilingualism , 22(1), 123–137. Ali, J., & Larsson, K. (2018). Ethnic victimization as a potential explanation for poor school adjustment among immigrant youth. Clinical Psychology Program. Orebro University. Bertschy, K., Cattaneo, M. A., & Wolter, S. C. (2009). PISA and the transition into the labour market. Labour , 23(1), 111–137. Bossavie, L. (2018). The effect of immigration on natives’ school performance: Does length of stay in the host country matter? Journal of Human Resources , 733–766. Bozick, R., Malchiodi, A., & Miller, T. (2016). Premigration school quality, time spent in the United c e p s Journal | V ol.11 | N o 4 | Y ear 2021 231 States, and the math achievement of immigrant high school students. Demography , 53(5), 1477–1498. Burgess, S., & Heller-Sahlgren, G. (2018) Gabriel, motivated to succeed? Attitudes to education among native and immigrant pupils in England. IZA Discussion Paper No. 11678. https://ssrn.com/ abstract=3217496 Castigo, M. J. (2017). Desempenho dos estudantes portugueses: modelos de regressão multinível [Achievement of Portuguese students: Regression models multi-level]. (Doctoral Thesis). Portugal: Universidade do Minho. Cattaneo, M. A., & Wolter, S. C. (2015). Better migrants, better PISA results: Findings from a natural experiment. IZA Journal of Migration , 4(1), 1–19. Childs, S., Finnie, R., & Mueller, R. E. (2017). Why do so many children of immigrants attend university? Evidence for Canada. Journal of International Migration and Integration , 18(1), 1–28. Chiu, M. M., Pong, S. L., Mori, I., & Chow, B. W . Y . (2012). Immigrant students’ emotional and cognitive engagement at school: A multilevel analysis of students in 41 countries. Journal of Youth and Adolescence , 41(11), 1409–1425. Chykina, V . (2019). Educational expectations of immigrant students: Does tracking matter? Sociological Perspectives , 62(3), 366–382. Cordero, J. M., Cristobal, V ., & Santín, D. (2018). Causal inference on education policies: A survey of empirical studies using PISA, TIMSS and PIRLS. Journal of Economic Surveys , 32(3), 878–915. Cummins, J. (2008). Where immigrant students succeed: A comparative review of performance and engagement in PISA 2003. Curriculum Inquiry , 38(4), 493–497. Di Liberto, A. (2015). Length of stay in the host country and educational achievement of immigrant students: The Italian case. International Journal of Manpower , 36(4), 585–618. ETS (Educational Testing Service ). Princeton University. Factbook, O. (2016). Factbook 2015–2016. OECD. Figueiredo, S., Martins, M. A., & Silva, C. F. D. (2016). The parental investment effect on immigrant children at schools: Employment and specialization of parents as an explaining variable for tasks achievement in second language. International Journal of Advances in Psychology , 5, 23–34. Grubanov-Boskovic, S., Natale, F., & Scipioni, M. (2017). Patterns of immigrants’ integration in European labour markets. European Commission: JRC Technical Reports. https://doi. org/10.2760/408657 Ham, S. H., Y ang, K. E., & Cha, Y . K. (2017). Immigrant integration policy for future generations? A cross-national multilevel analysis of immigrant-background adolescents’ sense of belonging at school. International Journal of Intercultural Relations , 60, 40–50. Hippe, R., & Jakubowski, M. (2018). Immigrant background and expected early school leaving in Europe: evidence from PISA . Publications Office of the European Union. Huguet, A., Navarro, J. L., Chireac, S., & Sansó, C. (2013). The acquisition of Catalan by immigrant children. The effect of length of stay and family language. In J. Arnau (Ed.), Reviving Catalan at school. Challenges and instructional approaches (pp. 29–48). Lavenham Press. Hvistendahl, R., & Roe, A. (2004). The literacy achievement of Norwegian minority students. 232 self-concept in immigrant school children and the impact of length of residence Scandinavian Journal of Educational Research , 48(3), 307–324. International Association for the Evaluation of Educational Achievement (IEA). (2017). IDB Analyzer. V . 4.0.26. IEA. Jang, E. E., Dunlop, M., Wagner, M., Kim, Y . H., & Gu, Z. (2013). Elementary school ELLs’ reading skill profiles using cognitive diagnosis modeling: Roles of length of residence and home language environment. Language Learning , 63(3), 400–436. Mai, M. Y ., & Asma’ A, A. F. A. (2016). Modeling the relation between self-esteem, loneliness and engagement as factors of children achievement in science. European Journal of Social Science Education and Research , 3(1), 107–120. Marôco, J., Gonçalves, C., Lourenço, V ., & Mendes, R. (2016). PISA 2015 – Portugal. Volume I: literacia científica, literacia de leitura & literacia matemática. IAVE, I.P . Martin, A. J., Liem, G. A. D., Mok, M. M. C., & Xu, J. (2012). Problem solving and immigrant student mathematics and science achievement: Multination findings from the Programme for International Student Assessment (PISA). Journal of Educational Psychology, 104 (4), 1054–1073. Martinez-Taboada, C., Mera, M. J., Amutio, A., Castañeda, X., Felt, E., & Nicolae, G. (2017). The impact of cultural dissonance and acculturation orientations on immigrant students’ academic performance. Universitas Psychologica , 16(5), 6–19. Nkemasong, A. E. (2018). Understanding the lived experiences of immigrant students in Finland (Master’s Dissertation). University of Eastern Finland. OECD (Organisation for Economic Co-operation and Development). (2014 ). PISA 2012 results: What students know and can do . OECD Publishing. OECD (Organisation for Economic Co-operation and Development). (2015). Helping immigrant students to succeed at school – and beyond. OECD Publishing. OECD (Organisation for Economic Co-operation and Development). (2016). Programme for International Student Assessment . PISA. OECD (Organisation for Economic Co-operation and Development). (2018). Education at a Glance 2018: OECD Indicators . OECD Publishing. https://doi.org/10.1787/eag-2018-en & http://dx.doi. org/10.1037/a0029152 Parker, P . D., Marsh, H. W ., Ciarrochi, J., Marshall, S., & Abduljabbar, A. S. (2014). Juxtaposing math self-efficacy and self-concept as predictors of long-term achievement outcomes. Educational Psychology , 34(1), 29–48. Kia-Keating, M., & Ellis, B. H. (2007). Belonging and connection to school in resettlement: Y oung refugees, school belonging, and psychosocial adjustment. Clinical Child Psychology and Psychiatry , 12(1), 29–43. Rangel, M. A., & Shi, Y . (2019). Early patterns of skill acquisition and immigrants’ specialization in STEM careers. Proceedings of the National Academy of Sciences , 116(2), 484–489. Petreñas, C., Pozuelo, M., & Redondo, G . (2011). Migrant girls in secondary education in Spain: Fighting to transform difficulties into possibilities. In Young Migrant Women in Secondary Education (pp. 125-142). University of Nicosia Press. c e p s Journal | V ol.11 | N o 4 | Y ear 2021 233 Salmela-Aro, K., Read, S., Minkkinen, J., Kinnunen, J. M., & Rimpelä, A. (2018). Immigrant status, gender, and school burnout in Finnish lower secondary school students: A longitudinal study. International Journal of Behavioral Development , 42(2), 225–236. Sakellariou, C. (2018). Selective immigration policies as screening devices and the immigrant–native educational achievement gap. Educational Research and Evaluation , 24(6-7),1–23. Schachner, M. K., He, J., Heizmann, B., & Van de Vijver, F. J. (2017). Acculturation and school adjustment of immigrant youth in six European countries: Findings from the Programme for International Student Assessment (PISA). Frontiers in Psychology , 8, 649. Schnell, P ., & Azzolini, D. (2015). The academic achievements of immigrant youths in new destination countries: Evidence from southern Europe. Migration Studies , 3(2), 217–240. Schnepf, S. V . (2007). Immigrants’ educational disadvantage: an examination across ten countries and three surveys. Journal of Population Economics , 20(3), 527–545. Volante, L., Klinger, D. A., Siegel, M., & Y ahia, L. (2019). Raising the achievement of immigrant students: Towards a multi-layered framework for enhanced student outcomes. Policy Futures in Education , 17(8), 1037–1056. Šori, I., Šušterič, N., & Gaber, S. (2011). Immigrant students’ achievements in Croatia, Serbia and Slovenia in context. Center for Educational Policy Studies Journal , 1(3), 31–51. Strohmeier, D., & Dogan, A. (2012). Emotional problems and victimization among youth with national and international migration experience living in Austria and Turkey. Emotional and Behavioural Difficulties , 17(3–4), 287–304. Wang, M. T., Chow, A., Hofkens, T., & Salmela-Aro, K. (2015). The trajectories of student emotional engagement and school burnout with academic and psychological development: Findings from Finnish adolescents. Learning and Instruction , 36, 57–65. Welch, A. (2018). Immigrant student achievement and education policy in Australia. In Immigrant Student Achievement and Education Policy (pp. 155–173). Springer. Weber, S., Kronberger, N., & Appel, M. (2018). Immigrant students’ educational trajectories: The influence of cultural identity and stereotype threat. Self and Identity , 17(2), 211–235. Woessmann, L. (2016). The importance of school systems: Evidence from international differences in student achievement. Journal of Economic Perspectives , 30(3), 3–32. Y easmin, N., & Uusiautti, S. (2018). Finland and Singapore, two different top countries of PISA and the challenge of providing equal opportunities to immigrant students. Journal for Critical Education Policy Studies (JCEPS) , 16(1), 207–237. 234 self-concept in immigrant school children and the impact of length of residence Biographical note Sandra Figueiredo, PhD, is an Associate Professor and Integrated PhD Researcher in the field of Psychology on the Department of Psychology at Universidade Autónoma de Lisboa Luís de Camões (UAL), Lisbon, Portu - gal. Her research interests are related to the development of instruments and empirical studies about the school achievement of immigrant young children. Recently completed her 6 years post-doctoral research, in Education, after the PhD in Psychology at University of Aveiro, Portugal. Professor Figueiredo sci - entific projects in coordination are funded by Foundation for Science and T ech - nology (FCT). Also she is co-Editor of the indexed Journal Psique. The themes she is developing are related to the second language learners’ assessment, cog - nition and students’ achievement, instruments validation for schools and for skills evaluation, experimental studies in educational field, chronobiology and school performance, among others. She published already around 60 papers at international publications with peer-review and supervised a multimedia (tai - lored assisted learning) project for immigrants’ education support in Portugal. João Marôco, PhD, is an Associate Professor of Scientific-Pedagogi - cal Department: Psychological Sciences| William James Centre for Research, at ISPA – Instituto Universitário de Ciências Psicológicas, Sociais e da Vida, Lis - bon, Portugal. Professor Marôco (Ph. D., Washington State University) teaches courses in Statistics, Research Methods and Data Analysis. He is invited asso - ciate professor of the doctoral program of business and administration at the Nova School of Business. He has been invited to perform more than 40 lectures and workshops in Portugal an abroad on Statistical Analysis and Structural Equation Modeling. His research interests include psychometric scale evalua - tion and development, large-scale student assessments and applications of re - gression, structural equation modeling and multivariate statistical classification methods in the biological as well as in the social and human sciences. Margarida Alves Martins, PhD in Educational Sciences from the University René Descartes Paris V and a PhD in Psychology from the Uni - versity of Coimbra. She is a full Professor of Educational Psychology at ISPA and head of the Centre for Educational Research (CIE-ISPA) funded by FCT (national research agency). She teaches several curricular units in graduate, post-graduate and doctoral courses in Education and Educational Psychol - ogy and has supervised post-doc, PhD and Master students in Education and Educational Psychology. She has participated as Principal Researcher in several c e p s Journal | V ol.11 | N o 4 | Y ear 2021 235 funded research and intervention projects. She published several papers in national and international journals mainly on early literacy and reading and writing acquisition in primary school and authored several books and book chapters on literacy. Odete Nunes, PhD in Psychology – Université de Toulouse II (Le Mi - rail). She is the Director of the Psychology Department at UAL. She is the Scien - tific Coordinator of the Research Center in Psychology (UAL) and a Member of the Pedagogic and the Scientific Council at UAL. She is an Associate Professor of the Department of Psychology at the Universidade Autónoma de Lisboa Luís de Camões (UAL), Lisbon, Portugal. She has published more than 50 scientific works in specialized magazines and in event proceedings. She has supervised 4 doctoral theses in addition to 25 master’s dissertations in the areas of Psychol - ogy and Health Sciences. Between 2001 and 2010 she participated in 9 research projects, and coordinated 1 of these. She is currently participating in 2 research projects. She works in the areas of Social Sciences with emphasis in Psychology and Humanities. Her research focused Self-Esteem, Empowerment, Person- Centered Approach, Smokers, Values, Stress, Parenting and Psychotherapy.