LINGUISTIC INTEGRATION OF MIDDLE SCHOOL IMMIGRANT CHILDREN IN CZECHIA

Acquiring the language of the host country is an important factor that influences the process of integration of immigrants into society. For the children of immigrants, proficiency in the official language of their new country is crucial for their success and their smooth integration into the education system and the wider community. This article analyses Czech language proficiency of immigrant children enrolled in middle school 1 in Czechia and seeks to examine how their language skills develop over time and what factors impact the process of Czech language acquisition. The authors created a special diagnostic instrument to test students’ communicative competence. Subsequently, they used it in a pilot study they carried out on a sample of immigrant children enrolled in selected middle schools in Prague. The findings of this pilot study confirmed that the linguistic distance between Czech and the students’ mother tongue had a significant influence on how and how quickly they mastered the Czech language.


Introduction
Integration of immigrants into host societies has always been an issue for both policy makers and scientists in countries with a long history of immigration like the USA (Bernard 1967).However, it also became an important issue in the Western Europe of several decades past, because if we use Heckmann and Schnapper (2003) words, the "illusion of temporary migration has disappeared" (p.9).Czechia has been for a long time rather a country of emigration, while in direct relation to changes after 1989, the country experienced transition from an emigration to immigration country (Drbohlav 2011).Thus, the country had to adjust its institutions and policies to this changed situation in a relative short time span (Drbohlav et al. 2010) in order to deal with the problem of integration.As Berry (2002a) has pointed out, it has been believed for a long time that immigration in general can lead to both psychological and social problems of immigrants in the host country.Nowadays, he claims, it is widely believed that the outcome of immigration is very much individual: while some of the immigrants adapt to the new situation very well, others may experience deep problems.In reality, factors on both the group and individual levels (Berry 2002b) come into play.They create a unique mix of conditions for each individual immigrant that is either conducive or unconducive to success.Such a situation directly calls for research that can shed light on these various factors and their interplay of which can under different conditions either aid in the successful integration of immigrants or make it more difficult.
Various factors and their influence on the process of integration of immigrants into host societies have been already studied in a great detail.Researched factors are various: the long list includes structural changes in economy (Bevelander 2000), institutional setting and migration policies (Heckmann, Schnapper 2003), social capital (Tillie 2004), legal system (Carens 2005), interplay between origin, destination and community effects (Van Tubergen, Maas, Flap 2004), religion (Foner, Alba 2008), cognitive skills (Suarez-Orozco 2007), adaptation strategy chosen by immigrants (Berry 2002b) or language acquisition (Chiswick, Miller 2001).Some of the factors underlying the reasons for immigration have already been studied in Czechia as well.Drbohlav (2011) summarized several key findings from the empirical research on immigration integration in Czechia that has been achieved so far.He namely emphasized the role of ethnicity and the citizenship of immigrants (Drbohlav, Dzúrová 2007), the assimilation strategy (Drbohlav, Dzúrová, Černík 2007), and the interplay of such factors influencing the integration of immigrants.Janská et al. (2007), using the model of a case study, studied in one middle school on the outskirts of Prague the role of language acquisition, familial relations and role involvement during the process of integration of Vietnamese children within their own ethnic community.
The aim of this article is to contribute to the analyses of factors that influence the integration of immigrants in Czechia by studying the process of language acquisition and its interplay with other factors like gender, ethnicity or exposure to the Czech language.We conducted a case study with immigrants who enrolled in middle schools in Prague, and primarily devoted attention to Czech language acquisition.Furthermore, we used our own diagnostic test (Kostelecká et al. 2013).For the first time in Czechia, we were able to receive not only more detailed information on language acquisition by school-aged immigrants, but also to gather more reliable information than that based off of the evaluation of language acquisition by the immigrants themselves.

Immigrants at Czech middle schools and their integration
After 1989 as increasing numbers of immigrants moved to the country, Czech schools experienced a notable increase in the number of pupils whose first language was a language other than Czech.After the country joined the European Union, there was a particularly sharp rise in the number of immigrants.While in 1985 there were just 37,000 foreign nationals dwelling on the territory of Czechia, a figure equal to approximately 0.36% of the total population, in 2008 the Ministry of the Interior registered a record number of 437,000 foreign nationals, practically a twelvefold increase over that period (ČSÚ, Cizinci v ČR -monthly basic data) and equating to more than 4% of the total population.Although the economic recession which began in 2008 reduced the attractiveness of the Czechia for immigrants, the number of registered foreigners remained basically unchanged (ČSÚ, 2014).Drbohlav, Dzúrová and Černík (2007) analysed in detail the available statistical information about foreigners enrolled in elementary and middle school in the early 2000s, showed that the number of immigrant students grew similarly quickly.During the school year 2002-2003, 3,592 foreigners were enrolled in preschools, 12,770 in elementary and middle schools, and 3,592 in high schools (p.164).Data from school statistics shows that despite the economic crisis the number of immigrant students grew at every level of the educational system between 2003 and 2011.During the 2010-2011 school year, there were 4,223 children of immigrants in preschool facilities in Czechia or 1.4% of the total number of children in institutions of this type; there were 14,109 children of immigrants enrolled in elementary and middle schools, making up 1.7% of all elementary and middle school students; there were 9,020 children of immigrants at high schools, or 1.5% of all students at this level (Statistická ročenka školství 2010/11).There were 37,688 foreign students in Czech post-secondary institutions (10.2% of all students at this level); however post-secondary students are a specific group due to a very selective enrolment process whereas only those applicants that fulfil the requirements of the given school, including language proficiency, are admitted.Many programmes at the post-secondary level are moreover taught not in Czech but in another major language.Ukrainians, Vietnamese, Slovaks, Russians and Chinese represent the largest immigrant minorities in the Czech educational system.
How education systems and teaching curriculum respond to the integrative needs of the children of immigrants and to the ethnic and cultural diversity of society has been the subject of studies in intercultural and multicultural education (e.g.Berry 1992bBerry , 1997)).Although immigration is a relatively new phenomenon in Czechia, a number of theoretical studies have already been conducted on the integration of children of immigrants into the educational system and the factors that can influence this process (e.g.Průcha 2007Průcha , 2010Průcha , 2011;;Šindelářová 2005, 2008;Hájková, Strnadová 2010).
For our study, the article of Drbohlav, Dzúrová and Černík (2007) who analysed in detail how foreigners in elementary and middle schools integrate into schools and Czech society is of a special relevance.The abovementioned authors combined three main theoretical concepts of integration: Heckmann's (1999) concept of integration, which distinguishes and defines four types of integration: structural, cultural, social, and subjective forms of integration; Portes' segmented assimilation theory (2001); and Berry's concept of the adaptation strategy chosen by immigrants (1992b).Drbohlav, Dzúrová and Černík (2007) conducted an empirical study using a mixed quantitative-qualitative design to test a series of hypotheses regarding how integration proceeds and which factors influence the integration.They conducted a survey with 80 foreign students and 47 Czech students who served as a reference group.They found that the process of integration was very complex: factors that influence the success of integration were various and included psychological, economic and factors related to identification and self-identification as well.They discovered that the level of integration as well as the level of life satisfaction of immigrant students depends on cultural background.Moreover, it was shown that the integration of Asians was influenced by a different set of factors than the integration of students of Slavic origin.The authors explicitly mentioned cultural and language similarity/distance as an important factor.
Similarly, a series of empirical studies published to date and based on research conducted in schools (e.g.Kocourek 2001Kocourek , 2002;;Kostelecká et al. 2010Kostelecká et al. , 2011Kostelecká et al. , 2012Kostelecká et al. , 2013) ) found that a key factor for the successful integration of students is proficiency in the Czech language.Although many immigrant students master basic communicative skills in the Czech language relatively quickly after arriving to Czechia (they have what Cummins (1979) calls 'basic interpersonal communicative skills'), it is not at a proficiency level that enables them to cope with the educational demands placed on them at school since that would require what Cummins calls a higher 'cognitive/academic language proficiency' (ibid.).
It has previously been shown that those foreigners enrolled in middle schools in Czechia with a low level of Czech language proficiency tend to have difficulties at school and not just with their Czech lessons, but also with many other subjects (Kostelecká et al. 2013).Not only does this significantly reduce their chances of obtaining a good education, but they may also suffer feelings of social isolation and be ostracised by others.This can have a negative effect on their self-identity and trust in others.These students are in a difficult situation, because the Czech education system has had little practical experience over the years with pupils and students speaking Czech as a second language (Janská et al. 2011).
In countries with a long history of immigration, many studies have been conducted on the linguistic integration of adults from minorities into society.The linguistic aspects of integration into American society have been studied, for instance, by Espenshade and Fu (1997); Beenstock (1996) examined the linguistic integration of immigrants in Israel; and Van Tubergen and Kalmijn (2008) focused their research on immigrants in the Netherlands.Empirical studies conducted outside of Czechia have demonstrated a clear link between the proficiency of adult immigrants in the official language of a society and their overall success in the labour market (Böhlmark 2009) which is connected to their level of earnings (McManus 1985).
However, there are comparatively far fewer studies on the educational integration of children.One of the problems that scholars interested in researching this issue encounter is the lack of data on the level of the language skills of children in primary and middle schools in relation to other personal characteristics.While in some countries (e.g., United States or France) standardised tests of children's language skills are used to measure the proficiency of school-age children in the dominant language of the country, the results of such tests are intended mainly to benefit children and their teachers and are not usually used for any other purpose.Hence, no other information on the children or their family background is gathered during testing.
Even when empirically exploring proficiency of adult immigrants in the dominant language, authors of studies rarely use standardised diagnostic tests for measuring skill levels but rather rely either on observations and evaluations of the researcher (Li 2007) or, more commonly, on the self-ratings of the research subjects (Bachman, Palmer 1989, 2009;Redstone Akresh 2007;Ender and Straßl 2009;Cort 2010).It is somewhat rare to find the results of standardised language tests being used for this purpose (Bang et al. 2009).
Czech language acquisition among immigrant children in schools in Czechia has not yet been the subject of systematic research.The main reason for this is that the country has no official tests to measure Czech language proficiency among students whose mother tongue is not Czech.Because there are no such standardised diagnostic tests and nor even any self-reported information on Czech language skills (rare exceptions are the work of Drbohlav et al. 2005 andJanská et al. 2011), we cannot objectively quantify the amount of progress such students make over the course of a year or evaluate whether linguistic integration is, for example, more difficult for one ethnic minority than another.Nor can we determine how large the language barrier these students face is, or how many of these children -and in what areas -reach the 'threshold level' of communicative competence.A threshold level is reached when a student acquires the minimum level of linguistic skills necessary to function as an independent user (Šára et al. 2001;Cvejnová 2006Cvejnová , 2007;;Šindelářová 2010).Under the Common European Framework of Reference for Languages the B1 level of language proficiency is considered to be the threshold.
This article aims to present selected results of a research project that sought to overcome the limitations of exploring this subject due to a lack of any standardised tests.In the scope of our project, a special diagnostic instrument was developed for the purpose of measuring the Czech language skills of immigrant children attending Czech middle schools, and that instrument was then tested in a pilot study at selected schools in Prague (the details of this study are presented in Kostelecká et al. 2013;Vodičková 2014;Vodičková, Kostelecká 2014).An analysis of the data produced in this study made it possible to describe the process of Czech language acquisition by the children of immigrants, as well as identify some of the factors that interact with language acquisition and influence the acquisition process.Below, we will first briefly introduce the testing instrument that was used and then provide information about the students who were tested.The core part of this paper contains an analysis of the results of the language tests and how they relate to certain explanatory factors.

Research methodology
Since there is no standardised test in Czechia that could be appropriately used to measure the level of communicative competence of pupils who speak Czech as a second language, we created such a test ourselves.To create this test we drew inspiration both from existing Czech Language Diagnostic Tests for Adults (Cvejnová 2007), as well as the Czech Language Certificate Exam for Young Learners (for details see http://ujop.cuni.cz/cce-mladez) and from internationally recognised tests like Cambridge English tests for young learners.We also took advantage of test development methods described in the methodological literature (for example Harrison 1983;Davies 1990;Smith 1995;Hasselgreen 2005;McKay 2006;Hughes 2007;Bachman, Palmer 2009;Vlasáková 2009).The diagnostic tests were developed in conformity with the Common European Framework of Reference for Languages and the European Language Portfolio for the respective groups, and (in conformity with) the recommendations of the international organisations for language testing: Association of Language Testers in Europe (ALTE) and European Association for Language Testing and Assessment (EALTA).
The individual tasks in the texts were designed bearing in mind that this was to be a test for young learners whose age, mental characteristics, and cognitive development must be taken into account when considering the length of the test, the selection of topics, the test techniques, the text types, the communicative situations, and the inclusion of visual aids for a test which started from the level of zero knowledge of the language up to the B1 level.This diagnostic instrument was designed to test the four language skills: reading, listening, writing, and speaking.We were unable to assess the language skill levels of children in their mother tongue, while some experts (Abdelilah-Bauer 2008;Kielhöffer, Jonekeit 2006) note that the ability to read and write in the mother tongue may positively influence the acquisition of these skills in another language; this task would have gone well beyond the scope of our research project.
The diagnostic test was developed in three consecutive steps.In the first step, we developed the test specifications (which describe the constructs of the skills tested, the intended population, the format of the test, etc.), assessment criteria, and created a first draft of the test which consisted of four subtests: Reading, Listening, Writing, and Speaking.Additionally, guidelines for the examiners were developed.The first version of the test was then submitted for review to professional reviewers.In the second step, the test was revised based on the recommendations of the reviewers, and then was submitted to a pre-testing process.Based on the results of the pre-tests, the test was revised again in order to arrive at a final version intended for testing in schools.Throughout this process great care was devoted to amending the test specifications to accurately reflect proficiency; these specifications were not only an important guideline for designing, revising, and assessing the test, but also served as an important source of information for the professional public as to who should be tested (characteristics of the target group for whom the test was intended), what should be tested, and how the testing should be carried out.
As a measurement of internal consistency of the test, the alpha coefficient (α = 0.96, see Cronbach 1951) and split-half reliability (0.91) coefficient were computed (Kuder, Richardson 1937).The alpha coefficient is higher than 0.9 equating to proof of excellent (High-Stakes) testing.The alpha coefficient is also higher than the splithalf reliabilities, which again demonstrates that this test is reliable.The quality of the test is also confirmed by the very high score in the Spearman-Brown Prophecy test (0.95).We used descriptive analysis (Horn 1993) for the description of the language test results and correlation analysis for the measurement of the relationships between language test results and selected indicators.
The diagnostic test was employed during a pilot study in 2010 on a sample of 153 children at selected middle schools in Prague.We deliberately selected schools with a larger percentage of children of immigrants making up the student body.We submitted a request for cooperation with schools in different areas within the city of Prague that have the largest shares of immigrant minorities (Ukrainians, Russians, Vietnamese, and Chinese) and where we predicted to find linguistic integration to be an issue.Given the strong similarities between the Czech and Slovak languages allowing people from a Slovak background to integrate without any difficulty into Czech society, we did not include Slovak students in our tests.The chosen schools were quite diverse in other respects, differing in size, their location within Prague, and their peripheral or central location within the city (see Map 1).Eight of the selected schools agreed to allow testing of the target group on the condition that each individual tested obtain prior written consent from his or her parents.On the agreed test date, three to four experienced testing specialists conducted an on-site visit to the individual schools in order to perform the testing exercise.All children of immigrants that were present at school on the day of the testing and whose parents agreed on their children's participation in the diagnostic testing were tested.
First, all pupils were subjected to three subtests at the basic A1 level: the subtests were in reading, listening, writing, and speaking.The group of tested students was given approximately 30 minutes to complete the A1 level test.After a break of 10 to 15 minutes, the students proceeded to complete subtests in reading, listening, and writing at the A2 level, which again took a total of approximately 30 minutes.After a second break, students were given approximately 50 minutes to complete reading, listening, and writing subtests at the B1 level.Finally, students' oral skills were tested, wherein each student spoke with an examiner face to face for about ten minutes.Each student's performance was then evaluated.

The number of tested students and the age and ethnic structure of the sample
A total of 153 children enrolled at the selected middle schools in Prague were given the test; the number of students tested at each school ranged from 8 to 38.The average age of the tested students was 13 years.The youngest child tested was 10 years old, while the oldest was 17 years old.The largest age group was of 14-year-olds.
There were considerable differences between schools in terms of the cultural background of the students tested and the students' mother tongues.For example, at one school more than half of the students tested were Chinese, at four other schools there were no Chinese students, and each of the remaining three had just one or two Chinese students.There were also large concentrations of Russians and Vietnamese at certain schools: one-third of the total number of children tested from these ethnic groups were students at one school.Ukrainians made up the largest group in the sample of tested children.The proportion they represented among the tested students was smaller than the actual proportion of Ukrainian students enrolled at primary and middle schools in Prague.Ukrainian children make up one-third of all non-Czech children at elementary and middle schools in Prague (Kostelecká et al. 2013).In our sample, however, they accounted for just around 20% of the students.
One of the things we ascertained during the testing was how long the students had resided in Czechia.There were major differences in the length of time students had been in the country.In some schools, more than half of the students tested had resided in the country for only a short period -0 to 2 years.By contrast, at two schools only a small minority of the students tested -less than 10% -were relatively recent immigrants.Similarly, there were differences in the proportion of students who had been living in Czechia for at least 7 years: ranging from 11% to 75% of the students tested at any given school.The average length of time the students in the test had resided in Czechia ranged from 3.4 to 8.6 years.

The test results
The maximum score that a student could obtain with a perfect performance in each of the subtests area was 30 points (10 points for each of the three levels of difficulty tested).A perfect performance in all four skill areas at all three language levels would produce a maximum test score of 120 points.The actual total scores ranged from 0 to 118 points, with an average score of 92.5 points and a median score of 101.0 points.With each higher level of language skill difficulty, the students' test score averages deteriorated.The average total score for all three tests differed considerably between schools, ranging from 77.3 to 104.9 points.A clear link was observed between the results of the diagnostic tests at different schools and the length of time students had resided in Czechia: The schools whose students scored lowest were also the schools that had the largest share of children who had resided in Czechia for just a short time.Conversely, the schools whose students scored highest had the smallest proportion of children who had resided in Czechia for a short period.The tests also revealed that the children generally had more difficulty with productive skills (speaking and writing) than receptive skills (reading and listening comprehension).On average, students scored lowest on the test of written expression (21.3 points out of 30).By contrast, they scored best on average on the reading comprehension test (24.7 points).
The differences observed between schools could theoretically have been due to variations in the quality of teaching (or to the different methods schools employ to teach the children of immigrants), matters on which we have no quantifiable information; but they could also simply reflect differences in the structure of children of immigrants at individual schools.We therefore proceeded in the next step to analyse available selected individual characteristics of the tested children and how they might relate to the test results.

Factors influencing students' test results -sex, length of residence in Czechia, cultural background
In the literature on the linguistic integration of immigrants, there are numerous hypotheses as to the effect of different factors on their success in the acquisition of the official language of the destination country.Chiswick and Miller (2001) formulated a general theory of language acquisition by immigrants in the destination country (referred to as the 'destination-language acquisition framework'), in which they identify the main factors of progress in language acquisition to be 'exposure' to the acquired language, 'efficiency' in language acquisition, and the 'impact of economic incentives.' The foundation of successful linguistic integration is spending time in an environment where the destination language is used.The more time immigrants are exposed to the destination language, the faster the acquisition process.Exposure to the destination language increases the longer an immigrant resides in the destination country (Thomas 2010), while acquisition is also contingent on a number of other factors, such as opportunities for contact with the destination language prior to immigration, and in the case of children, whether or not they obtain their preschool education in the destination country, attend school regularly, and are involved in extracurricular activities in the destination country (Iddings 2009).Another important factor in a child's exposure to the destination language is family, in particular the language of communication used in the family (Dustmann 1997;Li 2007), the number of siblings who are able to speak the destination language (Thomas 2010), as well as the informal environment the child spends time in (Chiswick, Miller 2001;Asgari, Mustapha 2011).
Even with the same exposure to the destination language, the pace of destination-language acquisition still varies between immigrants because the efficiency of the acquisition process also varies.According to Chiswick and Miller (2001), the efficiency of language acquisition is primarily influenced by the age at which an immigrant moves to the destination country, his/her education, and the distance between the immigrant's mother tongue and the destination language (see also Beenstock et al. 2001).Many empirical studies have shown that the younger an immigrant is at the time of migration, the faster the language acquisition process (Espenshade, Fu 1997;Stevens 1999;Böhlmark 2009).While it is not possible in the case of school-age children to speak of the effect of their education level on the speed at which they acquire the destination language, parental education has been demonstrated to have an effect (Li 2007;Cort 2010).Efficiency is also significantly influenced by the distance between the immigrant's mother tongue and the destination language: the greater the linguistic distance between them, the slower the process of language acquisition (Chiswick, Miller 2005).Hedbávná et al. ( 2009) make a similar argument in the case of Czechia.
We drew on the theory of Chiswick and Miller (2001) to consider what factors might influence the speed of Czech-language acquisition by the children of immigrants at elementary and middle schools in Czechia.To explain differences in the pace of linguistic integration of children of immigrants in Czechia, we therefore identified key factors to be the differences in their exposure to the Czech language and the differences of efficiency influencing the acquisition of Czech.Our analysis was somewhat impeded by a lack of sufficient data on the children we tested, as we only had a limited amount of information available to us.We knew only the children's age, sex, length of residence in Czechia, mother tongue, and cultural background.We did not have any information on their families, how they spent their free time, or on any of the children's other personal characteristics.
Therefore, we began by formulating three hypotheses.Hypothesis 1 assumes that the process of language acquisition would be significantly influenced by gender.Many studies have shown over the long term that girls attain slightly better results in language skills, and this could also have an effect on the diagnostic test we conducted.Hypothesis 2 assumes that a child's test results would be significantly influenced by how long he/she had been residing in Czechia.The longer a child has been in Czechia, the better his/her results.Hypothesis 3 assumes that a child's test results would also be dependent on his/ her cultural background, or more specifically, by the linguistic distance between the child's mother tongue and the Czech language.We assume that children from countries in which a Slavic language is spoken are able to learn the Czech language more quickly than children whose mother tongue is not a Slavic language.
We realise that the test results must also be influenced by a number of other factors.One obvious one is the intellectual skills of the children; other influences could relate to how they spend their free time, how many siblings they have, how many Czech-speaking friends they have, their socio-cultural position, the economic background of the child's family, and the education of his/her parents.However, it was beyond the scope of our research to undertake a measurement of the intellectual skills of the children or to gain information about their families, so we do not have information on these potential factors and therefore do not address them here.

The effect of sex on language acquisition
There were 79 boys and 74 girls who took part in the language skill tests.Sinke (1999) and Abu-Rabia (1997) have documented gender differences in regards to second-language acquisition; additionally Catalán (2003) has proven the better test results of girls.Therefore, we expected the girls to do better than the boys on the tests.To test whether the average scores of girls and boys were different, we used an analysis of variance (ANOVA procedure from the SPSS package of statistical programs).While the average total test score of the boys was 87.3 points (out of a possible total of 120 points), the average score of the girls was 98.1 points.The results of the analysis confirm a statistically significant difference (F = 8.53, sig.0.004).Thus, our hypothesis was unequivocally confirmed.
In the next step we proceeded to examine whether the differences between the girls' and boys' test results were in some way related to their length of residence in Czechia.We found that the differences between boys and girls were not constant, but changed in relation to the length of time the children had been in Czechia.The biggest differences in the test results were apparent among children who had been in Czechia for just a short time.These differences diminished as length of residence increased meaning that minor differences were identified among students who had been in Czechia for at least seven years, whereas the differences between boys and girls were almost negligible (Figure 1).Source: Authors' calculations based on their own data (Kostelecká et al. 2013).
We were also interested in whether the girls' test results were better in every skill area, or if girls or boys were stronger or weaker in any particular skill area.We found that the girls generally scored better in every skill area.The biggest differences in the performance of girls and boys were found in one of the productive skill areas: writing.Looking at how sex and length of residence in Czechia impacted the results of the diagnostic tests of language skill areas, again we found that the boys' and girls' test scores converged the longer they had resided in Czechia.Among the children who had been in Czechia for at least seven years, there were almost no differences between the scores of boys and girls in individual skill areas (Table 1).

The effect of length of residence in Czechia on the children's test results
We attempted to verify the hypothesis that the length of residence in Czechia has a significant influence on the children's test scores.In the first stage, we looked at how much total test scores were correlated with the length of residence in Czechia. Figure 2 clearly shows a positive correlation between the children's test scores and their length of residence in Czechia.The longer the child had resided in Czechia, the higher the number of points attained on the test.This is not, however, a linear relation (Figure 2).The figure was created using the Curve Estimation procedure from the SPSS package of statistical programmes.This procedure plots a trend in a manner that as precisely as possible represents individual measurements in a single curve.Unlike the more common linear interpolation procedure, Curve Estimation can be used to plot a trend not just as a line but also as various mathematically defined curves.The testing revealed that the changes measured in the total test scores in relation to length of residence in Czechia are best captured in a graph of the logarithmic function y = a + b × ln(x), where y is the estimated score in the test, x is the years of residence in Czechia, and variables 'a' and 'b' are the calculated parameters of the function (R Square = 0.45).
The results attained from the diagnostic test do not improve evenly with the increasing length of residence in Czechia.The children's test scores improve much more rapidly during a student's first years in Czechie as opposed to later on.This suggests that the most visible progress in Czech proficiency can generally be observed and expected among children during the first several years after they arrive to Czechia.Students who had been in the country for approximately two years obtained on average a total of 80 out of 120 points.Figure 2 also clearly shows that many of the tested children attained A1 level language skills very quickly after arriving to Czechia.Source: Authors' calculations based on their own data (Kostelecká et al. 2013).

The effect of cultural background on test results
The initial hypothesis assumed that students from a Slavic-speaking language background would make faster progress with Czech proficiency than children whose mother tongue is not a Slavic language.Slavic languages have many similarities in grammar and vocabulary.We assumed that this would make comprehension much easier for Slavic-speaking children and give them an advantage over children from other language backgrounds.This assumption was confirmed.While children from a Chinese background attained an average of 70.2 points on the test and those from a Vietnamese background 85.1, Ukrainian-speaking children attained an average score of 99.7 and Russian-speaking children 101.8.The analysis of variance confirmed the statistical significance of such differences (F = 12.86, sig.0.000).Figure 3 shows the total test scores in relation to length of residence and cultural background.Like in the preceding case, the Curve Estimation was again used.Here, however, it must be noted that the trends for individual ethnic groups were plotted from calculations based on a small number of cases.The results of these calculations are therefore subject to a greater sampling error.They are to be regarded simply as indicative results and the information on trends should be interpreted with caution.Despite the methodological limitation of the method, it reveals significant differences in the development of language skills for different ethnic groups in relation to the length of residence in Czechia.Children who came to the country from Asia, particularly in the period immediately after their arrival to Czechia, had much lower proficiency levels in this period than children from Slavic-speaking countries.While Asian children had almost no Czech language skills after arriving to Czechia, the opposite was true for children of Slavic origin.They reached a high level of proficiency within the relatively short period of one to two years, while children of Asian origin needed seven to ten years.Although the level of Czech proficiency of children from different cultural backgrounds varied widely during the period immediately after arrival to Czechia, over time these differences decreased.Source: Authors' calculations based on their own data (Kostelecká et al. 2013).

Conclusion
Successful integration into society requires proficiency in the language of that society (see, for example, the study of Remennick (2003), which shows how language acquisition was the main vehicle of the successful social integration of Russian immigrants into Israeli society in the 1990s).The first step towards successful social integration is the integration of students into a country's education system.How proficient in the Czech language are the children of immigrants at Czech middle schools?This question is difficult to answer because there is no instrument in Czechia with which to objectively measure Czech language proficiency among the children of immigrants.This fact has led our research team to devise an instrument to test these language skills.In this research project, our researchers developed this diagnostic instrument and tested it in a pilot study on the children of immigrants at selected middle schoolsin Prague.The diagnostic test examined the proficiency level of children in four language skill areas: reading, listening, writing, and speaking.
The children who were tested came from both large and small schools, located both on the periphery and in the inner city (Map 1).In all of the schools from the testing sample, immigrant children made up just a small percentage of the student body (between 6% and 22%).The immigrant children tested were from various cultural backgrounds: the majority were Ukrainians, Vietnamese, Russians, and Chinese.All of the students were enrolled in middle school and were between the ages of 10 and 17.The results confirmed the expected gender difference in the pace of second-language acquisition, with girls doing better than boys in the proficiency tests in the case of students still relatively new to Czechia.Length of residence in Czechia had a significant and expected effect on the test results of the children of immigrants and proved to be the most significant factor examined here.Generally, students' proficiency in Czech improves the longer they live in Czechia.Our test results agree with those of other similar studies.For example, Redstone Akresh's analysis (2007) of the New Immigrant Survey data indicates that the longer immigrants live in the United States, the more likely they are to use English with friends, at work, at home, and with a spouse.While the probability of newly-arrived immigrants using English to communicate with friends was 0.44, after 15 years in the country that probability doubled.However, the relation between length of residence in Czechia and proficiency level in Czech is not a linear one.Czech language proficiency improves very quickly during the first years in the country, but the pace of progress slows later on.Furthermore, as the length of residence in Czechia increases, the initially pronounced differences in proficiency between different categories of children, such as the differences between girls and boys, gradually fade.
The findings in this study indicate that the process of Czech language acquisition is also significantly influenced by cultural background and specifically by the linguistic distance between Czech and the mother tongue of the children of immigrants.This is again in accord with the results of other studies (Beenstock at al. 2001, Chiswick and Miller 2001, 2005;Isphording 2013) that have documented a more rapid progress in the acquisition of a host country's language by students whose mother tongue is linguistically closer to it.Children of Asian background in Czechia had almost no knowledge of the Czech language upon arriving to the country and required more time to acquire basic Czech language skills than students from countries where a Slavic language is spoken, who early on made very rapid progress with learning Czech.Over time, of course, the differences between children from different cultural backgrounds decreases, so that after residing in the country for a long time Vietnamese children have a level of proficiency in Czech that is comparable to children from a Slavic background.The diagnostic tests also showed that students from a Chinese background relatively have the most difficulty integrating into the Czech language environment and lagged behind the other students tested particularly in the area of Czech conversation skills.However, the longer they were in Czechia, the more their skills progressed towards the same level as other children.
The results of our analyses fully support the conclusion of Drbohlav (2011) who points to a need of using more of an "individual approach to designing integration policies, " that is to "systematically apply integration policies that will respect specific features of individual group of immigrants" (p.414).The study of Czech language acquisition by secondary school students indeed showed that tailor-made policies for different groups of migrants would be useful and may help to ease their integration into schools as well as into the society.
The abovementioned conclusion would not have been possible without the development of the pilot version of the diagnostic instrument measuring the level of language acquisition by foreign students.This diagnostic test is of multiple use.It can serve research purposes, as in the case presented in this paper, but it can also serve as a diagnostic tool useful for the children themselves, their parents, and their teachers.Testing the level of language acquisition may help teachers to monitor the progress of individual pupils, to identify their strengths and weaknesses, and to prepare tailor-made educational plans for them.Periodically repeated use of diagnostic tests may motivate children and provide feedback to them and their parents.The development of the pilot version of this diagnostic test represents the first step towards a more exact way of evaluation of language acquisition progress by the children of foreigners.Thus, the educational integration of students who are non-native speakers of Czech is a subject that warrants further observation and research.It is also a subject that, given the rapid increase in immigration to Czechia in the past twenty years, is a very relevant issue today.Furthermore, since the number of immigrants in the Czech population is likely to continue to grow, its relevance will only increase further into the future., 1985-2012 (31. 12.).Available at: http://www.czso.cz/csu/cizinci.nsf/t/B9003347F1/$File/c01r01.pdf Statistická ročenka školství -Výkonové ukazatele 2010/11.

Fig. 1
Fig. 1 Total test scores in relation to the children's sex and length of residence in Czechia.

Fig. 2
Fig. 2 Children's total test scores in relation to their length of residence in Czechia.

Fig. 3
Fig. 3 Children's total test scores in relation to length of residence and cultural background.