Trends in the study of Modern languages in Swedish lower secondary school (2000 – 2018) and the impact of grade point average enhancement credits

ABSTRACT

This paper investigates trends in the study of Modern languages or second foreign languages (SFLs) in Swedish lower secondary school between 2000 and 2018. Over the last decades, SFLs, i.e. French, German and Spanish, have been the target of several policy measures as a response to a declining interest. However, few reports on their impact have been published. We report the proportion of pupils studying an SFL at the national level and across demographic municipality groups and analyse a possible impact of one policy measure, grade point average enhancement credits (GPAEC, meritpoäng) for SFLs. We found an increase in the proportion of pupils studying an SFL, but mainly in urban areas. Moreover, we found important differences between SFLs. Spanish is the most widely studied SFL in all municipality groups. French is most popular in urban areas while German is more chosen in rural areas. If the increase of the number of pupils studying and SFL can be ascribed to GPAEC remains still to be seen, but if so, the GPAEC mainly had an impact on highly urban areas where the proportion of pupils studying an SFL was the highest already before introducing this policy measure.

KEYWORDS:

• Sweden
• Modern languages
• French
• Spanish
• German
• Sweden
• policy measures
• lower secondary school
• grade point average enhancement credits

1. Introduction

Over the last decades, the school subject labelled Modern languages, i.e. second foreign languages (SFL) like French, German and Spanish introduced after English in lower secondary education, has been the target of several policy measures legislated by the Swedish government.¹ The policy measures aimed to increase the number of pupils choosing to study an SFL, to decrease the number of pupils abandoning the subject (i.e. drop-outs) and to increase the completion rate at the end of lower secondary school in year 9. Furthermore, the government aimed at increasing the number of pupils who continue to study an SFL in upper secondary school.

One of the measures initiated was a reward system for studying SFLs in school. In 2007 (revised on several occasions during 2010–2015) grade point average enhancement credits (GPAEC) (meritpoäng) for SFLs were introduced, meaning that pupils who continue studying their SFL in upper secondary school and complete higher levels can receive extra credits that will raise their GPA and their chances of being accepted to university. In 2014, a similar measure came into force, which targeted the transition from lower secondary school to upper secondary school. Pupils who complete an SFL in year 9, the last year of lower secondary school, can include GPAEC when applying to upper secondary school (gymnasium).

It is a relevant question to ask if these policy measures have had an impact on the proportion of pupils studying an SFL, and if that is the case, on which pupils. Some researchers have previously discussed trends in SFLs and this research indicates a rather stable number of pupils in secondary school studying an SFL during the period from 1994 to 2010 (Börjesson & Bertilsson, 2010; Tholin, 2017, see below). However, to our knowledge, there has not been any detailed investigation of study trends of different SFLs for the period after 2010, in particular with respect to the period before and after the introduction of the SFLs reward measures. Therefore, the first objective of the present paper is to provide a closer analysis of trends in proportions of pupils studying different SFLs in year 9 (end of lower secondary school) during the period 2000–2018.

Furthermore, from the available statistics in previous research and reports, we know that pupils’ preferred choice of SFL has changed substantially in Sweden since 2000 (Bardel, Erickson, & Österberg, 2019, and references within). In the beginning of this century, German was the most widely chosen SFL, followed by French. Spanish was only introduced in lower secondary school in 1994, but has since then, and in particular, during the first decade of the 21^st century, outgrown both French and German. Today there are more pupils in secondary school studying Spanish than French and German together.

Apart from a nation-wide dominance of Spanish, not much is known, however, about regional or even local preferences for particular SFLs. Reporting on a survey that included questions on SFL options in secondary schools, Granfeldt, Sayehli and Ågren (2019) found that while German and Spanish seem to be offered to a nearly equal extent across the country, French was significantly less offered in the northern parts of Sweden. Several aspects might have been contributing to this difference (e.g. gender, ethnicity or socio-economic status). One aspect, however, that characterises the north of the country and for which register data is readily available, is its low population density. Therefore, a second objective of the present study is to investigate more closely the proportion of pupils studying SFLs across different demographic municipality groups. Existing categorisations based on population size and commuting patterns will be used to analyse the proportion of pupils studying an SFL in year 9 in three major municipality groups (see below for details). We will focus on two sets of research questions reflecting the objectives of this study:

Research question 1. What are the general trends for the study of SFLs in year 9 between 2000–2018 in Sweden?

Research question 2. Are the trends in the study of SFLs in year 9 different when the period before the introduction of GPAEC (2000–2007) and the period thereafter (2008–2018) are compared?

1. Are there differences between SFLs (French, German and Spanish)?

2. Are there differences between different demographic municipality groups?

1. Are there differences between SFLs (French, German and Spanish)?

2. Are there differences between different demographic municipality groups?

Section 2 below gives a brief overview of the organisation of SFLs in the Swedish school system and accounts for previous studies focusing on different aspects of SFLs in school. Thereafter, in section 3, we present the methodology of the study. Section 4 presents the results concerning general trends in the study of SFLs at the end of Swedish secondary school (Year 9) in relation to the different languages involved and to different types of municipalities. Finally, in section 5, we pinpoint the recent trends emerging from the data analysis and discuss possible interpretations and implications.

2. Second foreign languages in the Swedish school system

Since 1962, English has been an obligatory school subject in compulsory school and the language enjoys very strong support among the Swedish public (Eurobarometer, 2012). Before Sweden entered the European Union in 1995, a new national curriculum for compulsory school was introduced (Lpo94) with the intention to reinforce the position of SFLs like French, German and Spanish. As a member of the EU, Sweden adheres to the “mother tongue plus two” language policy promoted by the European Commission (1995). Following this policy, a “language choice” (språkval) became obligatory in lower secondary school in 1994 where pupils had the possibility to expand their linguistic repertoire. Pupils could at the time and still today choose on the one hand an SFL, such as French, German and Spanish. On the other hand they could also choose mother tongue instruction (if other than Swedish), remedial Swedish or English (or a combination of both today labelled SV/EN) or Swedish sign language. Consequently, the choice of an SFL remains optional. Previous studies suggest that one consequence of this “language choice” system, has been that the number of pupils studying an SFL has not increased despite governmental measures and intentions (Tholin, 2017).

In the current school law that came into force in 2010, organisers, public or independent, are required to offer at least two of the three languages French, German and Spanish if at least five pupils have chosen the language. In practice, this means that SFLs are mandatory for schools, but not for pupils. The organiser decides which SFLs to offer. In a recent study on SFLs in Swedish lower secondary schools based on a survey to school leaders (N = 147), Granfeldt et al. (2019) found that 76% of the schools offered all three SFLs French, German and Spanish. The remaining 24% offered only two of the three main SFLs with Spanish and German as the most frequent combination.

Stakeholders have sometimes described the situation for SFLs in Sweden as a “language crisis” and a number of contextual factors have been suggested that might affect the current situation. First, the lack of motivation to learn an SFL is argued to be a crucial factor. According to the EU-barometer on languages (2012), 52% of the Swedes (age 15–64) responded that lack of motivation was the main reason for not learning a new foreign language. Research on Swedish adolescents has also shown that the perceived high level of English, their first foreign language, negatively affects their motivation to learn an SFL (Henry, 2012). The adolescents in Henry’s (2012) study are of a similar age as the pupils we are reporting on here whereas the respondents to the Eurobarometer are older.

Possibly linked to low levels of motivation for SFLs and to the non-mandatory status of these languages, abandon rates of SFLs are high. In a survey study of 124 schools, Tholin and Lindqvist (2009) found that around one-third of all Swedish pupils starting to study French, German or Spanish abandon the subject within three years. The authors suggest that this trend is not necessarily linked to pupils’ difficulties in SFLs, but seems instead to be their way to avoid a heavy workload and to “buy time” for studying other subjects (Tholin & Lindqvist, 2009, p. 9). The high abandon rate was confirmed in a report from the Swedish Schools Inspectorate (2010) who evaluated the quality of language teaching of SFLs in 40 Swedish secondary schools with specific focus on schools with low numbers of pupils studying SFLs. It was again found that many pupils abandon their SFL within a couple of years’ time and that schools seldom analyse the underlying causes of these abandonments.

Turning now to previous research focusing on trends in studying SFLs, Börjesson and Bertilsson (2010) found that the proportion of pupils in SFLs who receive a pass grade in year 9 of lower secondary school has been stable, around 60%, during the period 1998–2010. In a recent study from 2017, Tholin discusses the trends in studying SFLs at the end of secondary school from 1994 to 2010 and the possible effects of the first set of governmental measures introduced to reinforce the role of SFLs in Swedish schools. Like Börjesson and Bertilsson (2010), Tholin finds that the measures did not affect the number of pupils studying an SFL in compulsory school nor the proportion of pupils who receive a pass grade in year 9. Finally, Krigh (2019) in her recent doctoral thesis finds that there has been an overall increase in the proportion of pupils studying SFLs in Swedish lower secondary school between 2008 and 2016. Krigh argues that studying an SFL in lower secondary school is a “socially differentiated practice” (Krigh, 2019, p. 222). Native-born pupils choose to study SFLs to a higher extent than foreign-born pupils and pupils from working-class and from lower middle-class families tend to be the ones who do not study an SFL. Krigh also found gender differences with more girls than boys studying SFLs.

As far as we know, the large-scale impact of the GPAEC introduced in 2007 and 2014, respectively, has not been studied elsewhere in any detail, but it has been briefly commented on in various reports. In a follow-up study to the European Study of Language Competences (ESLC, 2011), the European Commission published a country-comparative report focusing, among many other things, on trends in studying SFLs across Europe. For Sweden, the report states that the introduction of GPAEC has increased the completion rate and the number of pupils studying SFLs, but without providing any data to support the claim (European Commission, 2014). The National Agency for Education (Skolverket, 2018) also finds it plausible that the introduction of GPAEC has had a positive effect on the proportion of pupils studying an SFL, not only in upper secondary school, but also in lower secondary school. In a small-scale interview study on six pupils in upper secondary school, Henry (2017) found that for some of the pupils the GPAEC was the main reason for continuing studying the SFL (French) in a situation where they otherwise could have abandoned their studies had the GPAEC not been there.

Granfeldt et al. (2019) revealed differences between SFLs across the country since French was offered significantly less often in the northern and less densely populated parts of the country. It was also in the north that schools more often included mother tongue instruction as part of the language choice instead of the three traditional SFLs (for example Samí, Finnish or Meänkiele). Regional differences in foreign language learning have been reported in other countries as well. In the UK, for example, the proportion of pupils studying a foreign language at the end of secondary school (age 14–16 years) has dropped dramatically (from 76% in 2004 to 40% in 2011) since the subject was removed from the compulsory curriculum (British Council, 2018). This report reveals variation in foreign language learning across regions and socio-economic strata. It is in the northern parts of the country and in socio-economically disadvantaged areas that the languages on offer and the proportion of pupils studying foreign languages are the lowest. Differences in the foreign languages on offer at schools in different types of municipalities are also observed in Sweden’s neighbouring countries Finland (Korhonen, 2006) and Norway (Doetjes & Thue Vold, 2012), even if the regional differences are not the main focus of these reports. In her article, Korhonen states that in smaller municipalities in Finland, for example, education provided in other languages than English is actually very rare. These results suggested the presence of regional and possibly demographic factors influencing the organisation of SFLs that we wanted to investigate further in the present study.

The purpose of the present study is to examine the impact of GPAEC in more detail by focusing on data from the time before and after its introduction in the Swedish educational system. We will also study if the possible impact of GPAEC is the same across different demographic municipality groups.

3. Materials and method

Data for the present study were gathered from the database Jämförelsetal made available by the National Agency for Education and collected by Statistics Sweden (SCB) on a yearly basis. The database makes available data about Modern Languages for all municipalities in the country.² It also contains historical data for each year since 1994. We are focusing in a first analysis on data from lower secondary school and on the following three variables:

1. The proportion of pupils studying French in year 9 out of all pupils in year 9

2. The proportion of pupils studying German in year 9 out of all pupils in year 9

3. The proportion of pupils studying Spanish in year 9 out of all pupils in year 9

We computed a fourth variable by adding the values of the first three:

1. The proportion of pupils studying one of the languages French or German or Spanish in year 9 out of all pupils in year 9

The reason to focus on the proportion of pupils studying each of the three main SFLs was that it could reflect a possible change in the number of pupils studying an SFL. An alternative variable would have been the proportion of pupils who receive at least a pass grade after year 9 (cf. Tholin, 2017). We considered this measure as less reliable, since there might be differences in grading between teachers. The proportion of pupils (still) studying an SFL in year 9 was therefore considered a better variable for analysing possible impact of policy measures on the proportion of pupils studying an SFL in year 9.

Each of the four variables above will be considered dependent variables in our analyses. The data were downloaded from the database Jämförelsetal in several Excel-files. The data were collected for each year during the period 2000–2018 (i.e.19 years) and for each municipality separately. We also obtained a national average for each year. We decided to take year 2000 as our starting year even though data are available from 1994 in the database, because in year 2000 new course plans for SFLs were introduced.

In total there are 290 municipalities in Sweden and we thus obtained a value for each municipality (290) plus a national average over a period of 19 years and for each of the three different languages (French, Spanish, German), plus a grand total (French + Spanish + German) which we computed ourselves. In total, the database contained 22 116 posts (291 x 19 × 4). The data were analysed SPSS®.

In order to study possible variation between (groups of) municipalities, each of the 290 municipalities was coded according to the classification by the Swedish Association of Local Authorities and Regions (SALAR) (Sveriges kommuner och regioner). The classification consists of three main groups (A, B, C) and is based on structural parameters such as population and commuting patterns (see Appendix for full definitions).

1. Large cities and municipalities near large cities (n = 46)

2. Medium-sized towns and municipalities near medium-sized towns (n = 108)

3. Smaller towns/urban areas and rural municipalities (n = 136)

3.1. Adjustment for over- and under-representation in the data

For each of the three main groups (A, B and C) a mean proportion for each of the dependent variables (see above) was computed. When computing group means for a particular year, cases (e.g. proportion of pupils studying French/German/Spanish in a given municipality a given year) were weighted against the total number of pupils in public schools for the same year using the “weight cases” option in SPSS®. Weighting group means on the basis of the size of pupil cohorts was done in order to reflect the underlying composition of the population and to control for over- or under-reporting from the different groups which vary considerably in the number of pupils.

3.2. Data correction

In several publications from the National Agency for Education (e.g. 2011), it is observed that the SFL data reported by the municipality of Stockholm for the academic year 2008/09 are wrong. All data points (for French, German and Spanish) were therefore deleted for Stockholm 2008/09 and new values were imputed using the “replace missing value” option in SPSS®. The method used for imputing new values was mean of four nearby points in the respective data series.

3.3. Analyses

In order to answer research question 2 above, the resulting data was divided into two time series. A first series of data spanned the first period 2000–2007 (8 years) before the introduction of GPAEC for SFLs. The second series of data spanned the second period 2008–2018 (11 years) after the introduction of GPAEC for SFLs. The year 2008 was set as a cut-off point since it was the first year after the decision of introducing GPAEC (in 2007). Note that the possible effect of choosing an SFL (in year 6 or year 7) will however not appear in data for year 9 until 2010 or 2011 (depending on the starting year). On the other hand, the possible effect of a decrease in the number of pupils abandoning the SFL could already be seen in 2008.

Multiple linear regression analyses with TIME as the independent variable were carried out for both periods (2), for each of the three languages plus the total (4), for each of municipality groups and for the Sweden national average (4). This analysis yielded a total of 32 regression lines (i.e. 2 × 4 x 4) or 16 regression lines for each period. In order to test the null hypothesis that there is no significant difference between regression coefficients (before and after), a series of t-tests were performed.

Results are presented in two steps. First, we present descriptive data in relation to research question 1, and then we present the results of the regression analyses which will answer research question 2.

4. Results

We will start by looking at the general trends of pupils studying SFLs in compulsory school during 2000–2018. Figure 1 provides a first overview on a national level.

Figure 1. Proportions of pupils studying a Second Foreign Language in year 9 (2000–2018) – Sweden national average

Figure 1 shows that there is an increase in the proportion of pupils studying one of the three languages French, German or Spanish in year 9. In Year 2000 the proportion was 62,3% and by 2018 it had increased to 72,3%, an increase of 10%. Figure 1 also shows the trade-off between Spanish and German/French in the beginning of the century. Spanish became the most studied SFL in 2006 and continued growing until 2014. After 2014, the proportion of pupils studying Spanish has remained more stable according to Figure 1. With respect to German and, to a lesser extent French, the decline was the sharpest during the period 2000–2008. After 2008, the proportion of pupils studying French or German in year 9 has remained stable. German has even seen a small increase in the last three years of the period.

Turning next to the study of SFLs in different municipality groups during the same period, Figure 2 presents the data for the total (French/German/Spanish) and Figures 3–5 present the same data for each of the respective three languages.

Figure 2. Proportions of pupils studying French/German/Spanish in Sweden and in different municipality groups (means) (2000–2018)

Figure 3. Proportions of pupils studying French in Sweden and in different municipality groups (means) (2000–2018)

Figure 4. Proportions of pupils studying German in Sweden and in different municipality groups (means) (2000–2018)

Figure 5. Proportions of pupils studying Spanish in Sweden and in different municipality groups (means) (2000–2018)

When considering the proportions of pupils studying SFLs in different municipality groups over the period 2000–2018, the following patterns emerge. First looking at the total proportion of pupils studying an SFL (Figure 2) we find that the differences between different municipality groups have increased. In Year 2000, the proportion of pupils studying one of the three languages only differed very little between the municipality groups. In fact, a more or less equal proportion, around the national average of 62–64%, studied an SFL in year 9, irrespectively of the size, type and location of the municipality. In Year 2018, there are clear differences between the municipality groups. The proportion of pupils studying one of the SFLs is the highest (76%) in Large cities and municipalities near large cities (Group A), followed by Medium-sized towns and municipalities near medium-sized towns (Group B) (65,9%), whereas in Smaller towns/urban areas and rural municipalities (Group C) the average remains at 62%. It is clear from Figure 2 that in particular the A-group municipalities but also the B-group municipalities have seen an increase in the proportion of pupils studying an SFL (see below).

Turning next to the trends for the individual languages, Figures 3–5 illustrate some differences and some similarities between the languages.

French (cf. Figure 3) is mostly studied in Large cities and municipalities near large cities (Group A). This has been the case over the whole period studied here, but the difference with respect to the other two municipality groups (B and C) has increased over the period. In municipality group A, the proportion of pupils studying French decreased in the first period (2000–2007) but has remained stable since 2009 according to Figure 3. In municipality group C, smaller towns/urban areas and rural municipalities, the mean proportion of pupils studying French dropped below 10% for the first time in 2016 (Table 1).

Table 1. Number of municipalities without registered pupils in French, German or Spanish (2018)

	Large cities and municipalities near large cities (A)	Medium-sized towns (B)	Smaller towns/urban areas (C)	Total
French	0	11	30	41
German	0	2	2	4
Spanish	0	4	4	8

Unlike French, German (cf. Figure 4) is most widely studied in municipality group C, Smaller towns/urban areas and rural municipalities. This has been the case since year 2000, but the differences compared to municipality groups A and B are smaller in 2018 than they were in the beginning of the century. After an initial steep decline, the mean proportion of pupils studying German stabilised in all municipality groups during the second period (2008–2018) and has even seen a small increase in the last few years in all groups.

In 2018, Spanish (cf. Figure 5) is the language which is the most widely studied in all municipality groups with between 35% (group C) and 42% (group A) of the pupils. Looking at Figure 5, we see that the most dramatic increase in the proportion of pupils studying Spanish took place up until 2014. As for French, the highest proportions of pupils studying Spanish are found in municipality group A (Large cities and municipalities near large cities).

Next, we will look at the number of municipalities that do not register any pupils in the respective municipalities. Figure 6 shows that at the beginning of the century 150 municipalities did not register any pupils in Spanish, but this number dropped rapidly as the interest in studying Spanish took off. In 2018 there are only eight municipalities that do not register any pupils in Spanish. With respect to German and French there are clear differences in Figure 6 that follow from the different “profiles” that these two SFLs have in Sweden (cf. above). German continues to be well represented across the country over the whole period despite the fact that it is nearly as small as French in terms of the number of pupils. Since German is, on average, more studied in municipality group C, Smaller towns/urban areas and rural municipalities, it means that there are very few municipalities overall that do not register any pupils in German. In this respect, French is the mirror image of German with the highest proportion of pupils in municipality group A and the lowest in municipality group C. The consequence is that the number of municipalities not registering any pupils in French has increased from zero in Year 2000 to 41 in Year 2018.

Figure 6. Number of municipalities registering no pupils in the respective languages (2000–2018)

This picture is confirmed when we look at the details for the municipalities without registered pupils in the respective languages in 2018 (cf. Table 1).

Out of the 41 municipalities that did not register any pupils in French in Year 2018 30 (73%) were municipalities in group C. The small number of municipalities without registered pupils in German and Spanish was equally distributed between municipality groups B and C. For reasons of population size, the municipalities in group A are not likely to be included here.

We now turn to the second research question concerning the possible impact of the introduction of GPAEC on the proportion of pupils studying an SFL in year 9. Recall that the period 2000–2018 was divided into two sub-periods (2000–2007 and 2008–2018, respectively). Multiple regression analyses were made looking at the trends for the respective SFLs and for the total both at the national level (Sweden) and across the three municipality groups (A, B and C). Table 2 presents the relevant statistics.

Table 2. β coefficients before (2000–2007) and after (2008–2018) the introduction of GPAEC

Groups	Statistics	French	German	Spanish	Total (Fr/Ge/Sp)
Sweden	β coefficient 2000–2007	−0,73	−2,00	2,76	0,32
	β coefficient 2008–2018	−0,04	−0,02	0,76	0,58
	β difference	0,69	1,98	−2,00	0,26
	t	−40,11	−54,53	17,40	−3,12
	Significance	<.000	<.000	<.000	<.01
Group A	β coefficient 2000–2007	−0,69	−1,72	2,95	0,53
	β coefficient 2008–2018	0,12	0,27	0,51	0,84
	β difference	0,81	1,98	−2,43	0,31
	t	−16,90	−24,11	18,05	−4,72
	Significance	<.000	<.000	<.000	<.001
Group B	β coefficient 2000–2007	−0,83	−2,06	3,04	0,15
	β coefficient 2008–2018	−0,12	0,03	0,22	0,25
	β difference	0,71	2,09	−2,82	0,09
	t	−20,50	−52,39	25,42	−1,15
	Significance	<.000	<.000	<.000	p >.05, n.s.
Group C	β coefficient 2000–2007	−0,62	−1,97	2,55	−0,02
	β coefficient 2008–2018	−0,27	−0,52	0,65	−0,03
	β difference	0,35	1,44	−1,90	0,00
	t	−13,70	−46,77	17,31	0,04
	Significance	<.000	<.000	<.000	p >.05, n.s.

Table 2 shows that for Sweden as a whole and for Large cities and municipalities near large cities (Group A) there are positive β-differences in slopes for the total, for German and for French, but not for Spanish (negative β-difference). All differences are significant when the two periods are compared. The β-difference patterns are the same in municipality groups B and C. In these two groups, differences between the two periods are significant for each of the respective languages, but not for the total (Fr/Ge/Sp). In other words, the period after 2007 has seen a significant increase in the overall study of SFLs in Sweden and in municipality group A, but not in municipality groups B and C.

If we consider the respective languages more closely, we can see that each of the three languages has undergone different changes when the two periods are compared. The differences are systematic across municipality groups, but they differ in magnitude. Figure 7 shows the relative differences in slopes (β) for the respective languages for Sweden and for the respective municipality groups. The value of the slope for the first period (2000–2007) was subtracted from the value of the slope for the second period (2008–2018). If the result is a positive value, it means that the language has undergone a more positive development in the second period than in the first period. If the value is negative the opposite holds. Note that this is not necessarily related to an actual increase or decrease in the proportion of pupils studying the respective languages during these two periods. Figure 7 only reports on the difference in change (i.e. the result of comparing the two slopes).

Figure 7. Differences in slopes (2000–2007) vs. (2008–2018) for the three languages and for different municipality groups

In Figure 7, we can see that French and German are represented with positive values. This means that French and German have undergone a positive change in the second period when the two periods are compared. In the second period (2008–2018), after the introduction of GPAEC, both French and German have seen the end of the steep decline that characterised these two languages during the first period (2000–2007).

In Figure 7 Spanish is represented by negative values. Note that this does not mean that the proportion of pupils studying Spanish in year 9 has dropped. As we have shown above (cf. Figures 1 and 5.) Spanish has continued to grow also after 2007. What Figure 7 shows is that the relative change, when the trends during the two periods are compared, is negative for Spanish in the second period. Spanish grew significantly more in the first period (2000–2007), before the introduction of GPAEC, than in the second (2008–2018). The difference is significant in all municipality groups, but it is clearest in municipality groups A and B.

French has seen an opposite change compared to Spanish when the trends during the two periods are compared. French has had a positive trend in the second period compared to the first period. The difference is significant at the national level (Sweden) and in all municipality groups. However, the impact seems to be the strongest in Large cities and municipalities near large cities (Group A).

German patterns with French in Figure 7. German is the language that has seen the most positive change in the second period compared to the first period. The differences in β-coefficients when the two periods are compared reach 2,00 in all contexts, except for the smallest municipalities (group C). As we have seen before (cf. Figure 1), the overall trend is still slightly negative for German when the whole period 2000–2018 is considered, but there is a sharp difference in slopes in the first (2000–2007) and the second period (2008–2018). This is illustrated in Figure 8 where the two regression lines for German at the national level (Sweden) are compared.

Figure 8. Comparison of trends in the study of German in year 9 (2000–2007) and (2008–2018)

5. Discussion

In this paper, we investigate trends in the study of a second foreign language (SFL) in year 9 between year 2000 and 2018. The answer to the first overall research question (cf. above) is that there is an overall positive trend in the study of SFLs in year 9 when we consider the proportion of pupils still studying an SFL in the last year of lower secondary school. This positive trend could mean that a higher proportion of pupils chooses an SFL today than before. It can also mean that a smaller proportion than before abandons the subject after choosing it. It would seem likely that the overall positive trend is the result of a combination of the two factors.

At the same time, our results show that the overall positive trend is not equally reflected in all SFLs or in all municipality groups. As is well known, Spanish is the language which has seen a dramatic increase in the proportion of pupils studying this language. Spanish was introduced in lower secondary school already in 1994, but started to become increasingly popular in the beginning of the period studied here. Since 2006, Spanish has been the most popular SFL in lower secondary school. Francia and Riis (2013) suggest a number of factors behind the so-called “Spanish boom” in Sweden. Among them is the increasing recognition of Spanish as a world language, a relatively high proportion of Spanish-speaking immigrants in the country, Spain as an increasingly popular destination for holidays, an increased popularity of Latin-American culture and a perception among pupils that Spanish is the easiest among the SFLs. During the same time, French and especially German have seen an overall decrease in the proportions of pupils studying these two languages. It is possible that the overall positive trend of SFLs in year 9 is driven by the so-called Spanish “boom” where groups of pupils who did not choose an SFL before now choose Spanish. The idea would be that given the status that Spanish has, it attracts other groups of pupils than the ones who traditionally choose to study an SFL. This remains, however, a speculation at this stage and would need to be further studied.

We also found that the overall increase in the study of SFLs in Year 9 is not equally distributed across all municipality groups, but differ according to their demographic characteristics. In 2000, an average of around 62% studied an SFL in Year 9 in all three municipality groups (A, B and C). In 2018, the proportion had increased to 76% in Large cities and municipalities near large cities (group A) whereas it remained at 62% in Smaller towns/urban areas and rural municipalities (Group C). Increased regional differences, especially between urban and rural areas, are reported for many aspects of life in Sweden. Enflo (2016) finds that there are differences between urban and rural areas with respect to education levels, income, health, labour market and welfare resources. In these domains, the magnitude of differences is back at the same level as it was in the 1930 s according to Enflo (2016). The present study shows that there is an increasing difference between urban and rural areas also with respect to the study of SFLs in lower secondary school. This seems to be in line with findings by Krigh (2019) and her observation of SFLs as part of a “socially differentiated practice” depending on social class. Furthermore, the trends observed in our data are not unique to Sweden. They pattern with results from studies in other European countries, for example, in the UK, where foreign languages are less often studied outside larger cities and the metropolitan London area (British Council, 2018). Similar trends have been discussed in Finland and Norway (Doetjes & Thue Vold, 2012; Korhonen, 2006) even though it seems that the question has not been studied empirically in those two countries.

When region and language are combined, some interesting patterns emerge in our data. French is the most urban of the three SFLs. During the whole period, French has been more studied in Large cities and municipalities near large cities (Group A) than anywhere else. The difference with respect to the other two municipality groups seems to have increased over time, especially after the introduction of GPAEC (see below). Traditionally, French has been associated with the middle- and upper-class and has sometimes been considered the foreign language of the socio-economic elite. In this type of characterisation, German has been considered particularly useful for small-scale entrepreneurs and people working in the tourism sector (cf. for example Tischer, 2016), which could explain why the language is most popular in Smaller towns/urban areas and rural municipalities (Group C). This divide in preferences for different SFLs between urban and rural areas has consequences for the offer and ultimately for the availability of SFLs across the country. Both French and German have seen a decline in proportions of pupils studying the languages since 2000, but French more than German is disappearing altogether from the smaller municipalities (groups B and C) where there are a small number of pupils to begin with. Since organisers are only required to offer two out of the three main SFLs (French, German and Spanish), they might attempt cutting costs by closing down one of the languages. This was the case in one municipality, Kristianstad, which received quite a lot of media attention in 2018–2019 (see, e.g. Kristianstadbladet, 2019). The result is often that the least popular language is removed and because this is French in smaller municipalities, it disappears more often than German. In fact, the number of municipalities not registering any pupils in French is increasing rapidly in municipality groups B and C. If this trend continues, it might turn into an equity problem since the possibility to study French will depend on where in the country you live. When discussing the number of municipalities without pupils in the respective language, it is important to recall that there are at least two potential reasons behind these figures. Lack of interest from pupils is the first possible reason. A second is the lack of teacher resources in the municipality. All three languages, but in particular Spanish, have been reported to lack a number of trained and qualified language teachers (Francia & Riis, 2013).

The second research question targeted the possible impact of the introduction of one of the policy measures in recent years, the GPAEC (meritpoäng). It is important to stress that it is not possible with the current data to establish any causal relationship between the policy measure and any change in the proportion of pupils studying and SFL after the introduction of GPAECs. In fact, no data would. The results here can therefore merely be suggestive in this respect. Looking at the total proportion of pupils studying one of the three SFLs at the national level, the results show that there is a significant difference in the trends before and after the introduction of GPAEC. Looking at the results divided by municipality level, we found the same significant positive difference for Large cities and municipalities near large cities (Group A), but not for the other two municipality groups (B and C). Consequently, it can be argued that if there is an impact of GPAEC on the proportion of pupils studying and SFL, it is the strongest in urban areas where pupils were already studying SFLs to a higher degree than elsewhere before the introduction of this policy measure. In this interpretation, the introduction of GPAEC merely strengthened a behaviour that was already more prominent in these areas than elsewhere. Why could this be? A possible explanation is that the very nature of the policy measure is contributing to the uneven distribution. GPAEC increases the chances to enter university (by adding credits for advanced courses in SFLs in upper secondary school). Importantly, the proportion of young people who continue studying at university is generally higher in large cities and municipalities near large cities (Group A) than elsewhere (Swedish Higher Education Authority; UKÄ, 2016). It could therefore be the case that GPAECs promote the study of Modern Languages mostly among those who consider university studies as a future option. Again, this seems to be in line with findings in Krigh (2019).³

If we look at each of the three languages separately, it is even more difficult to establish any impact of the GPAEC, but there are striking differences between the languages. French has seen a decline in both periods (2000–2007 and 2008–2018), but a possible interpretation is that the introduction of the GPAEC levelled out the negative development. Again and typical for French, the impact is the strongest in Large cities and municipalities near large cities (Group A) which is in line with the general pattern (cf. above).

Spanish, on the other hand, has seen an increase in both periods, but the increase was much stronger before the introduction of the GPAEC. This might be considered a surprising result, especially in the light of the observation that Spanish is more popular in Large cities and municipalities near large cities (Group A) than in the other two groups. A possible explanation, which can be seen in the data (cf. Figure 1), is that there is a “saturation” effect in the demand with respect to Spanish. Around 2012, Spanish was studied by ca 40% of pupils in year 9 and this proportion has been stable since then. Teacher resources might also be a factor. Since there is a particularly high shortage of trained and qualified Spanish teachers, it can also be that schools cannot simply accommodate all pupils that would like to study Spanish.

German is the language that displays the most important change when the trends in the two periods are compared. During the first period, before the introduction of GPAEC, German saw a steep decline, which has levelled out in the second period. Interestingly, the change is highest in Medium-sized towns and municipalities near medium-sized towns (group B) followed by Large cities and municipalities near large cities (Group A), i.e. the two groups where German is normally less studied. Since it seems like the impact of GPAEC on the proportion of pupils studying an SFL in year 9 is the highest in these two groups overall, it could be that German has benefited from the introduction of the GPAEC. At the same time, other explanations can also be invoked. The steep negative trend might have levelled out because traditional cultural and economic bonds prevent a complete disappearance of German from Sweden. An additional possibility is that the decline in the first period was mainly caused by a trade-off effect with Spanish, but, at the same time, the decline seems to end while the raise of Spanish is still ongoing (cf. Figure 1).

6. Conclusion

In this paper, we have investigated trends in the study of Modern Languages or second foreign languages (SFLs) in Swedish lower secondary school between 2000 and 2018. SFLs, i.e. French, German and Spanish, have been the target of policy measures introduced by the Swedish government as a response to a declining interest in the study of these languages. One such policy measure is grade point average enhancement credits (GPAEC) (meritpoäng) which means that pupils who have completed higher courses in an SFL can add extra merits when applying to university. However, few studies on the impact of this policy measure have been published so far. We found that the proportion of pupils studying an SFL in the last year of lower secondary school has increased since 2000, but the increase has mainly occurred in urban areas. Spanish is the most popular SFL in all areas whereas French is most popular in urban areas. German is more chosen in rural areas. A consequence of this pattern is that French more than German is disappearing from the offer of SFLs from lower secondary schools. If there is a positive impact of GPAEC, it can only be seen in highly urban areas where the proportion of pupils studying an SFL was the highest already before the introduction of this policy measure.

The findings of this study point to increasing differences between urban and rural areas with respect to SFLs in Sweden. These differences can be seen in the overall proportion of pupils who study an SFL, which SFL they study and in the offer of SFLs they can choose from.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Swedish Research Council [2015-01088 (to Jonas Granfeldt)].

Notes on contributors

Jonas Granfeldt

Jonas Granfeldt Professor of French Linguistics, Centre for Languages and Literature, Lund University, Sweden. His areas of expertise include first and second language development, with a particular focus on French. His research has focused on many different aspects of language acquisition, including developmental sequences and the relationship between L2 development and L2 proficiency. He is the leader of the research project Learning, Teaching and Assessment of Second Foreign Languages – an Alignment Study on Oral Language Proficiency in the Swedish School Context.

Susan Sayehli

Susan Sayehli Senior lecturer, the Centre for Research on Bilingualism, Stockholm University, Sweden. Her research has focused on the acquisition of second languages and the factors affecting it. Particularly, she is interested in how the different languages a person knows affect each other in learning, producing and comprehending a language. Currently, she is working on a research project on learning, teaching and assessment of second foreign languages in the Swedish school context.

Malin Ågren

Malin Ågren Associate Professor of French Linguistics, Centre for Languages and Literature, Lund University. Her research falls within the domain of language acquisition in different settings, with a specific focus on French. She has previously worked on the acquisition of different morphosyntactic structures in speech and writing in bilingual children and adult learners of French. Moreover, she does research on French in the Swedish school context and on the teaching of foreign languages.

Notes

1. In this paper, we choose the term second foreign languages rather than Modern languages (sw. Moderna språk) in order to clearly separate between French, German and Spanish on the one hand and English on the other.

2. In the analysis of the proportion of pupils per municipality both public and independent schools were included.

3. Another possibility arguing in the same lines was suggested by one reviewer. At least in some large cities, there is strong competition for entering the most popular Upper secondary schools. It could be that GPAECs are needed for pupils to raise their average GPA to a level where they can choose the school they wish.

Appendix

Classification of Swedish municipalities from the Swedish Association of Local Authorities and Regions (SALAR) (Sveriges kommuner och regioner) (2017).

A. Large cities and municipalities near large cities (n = 46)

1. Large cities – municipalities with a population of at least 200 000 inhabitants with at least 200 000 inhabitants in the largest urban area.

2. Commuting municipalities near large cities – municipalities where more than 40% of the working population commute to work in a large city or municipality near a large city.

B. Medium-sized towns and municipalities near medium-sized towns (n = 108)

1. Medium-sized towns – municipalities with a population of at least 50 000 inhabitants with at least 40 000 inhabitants in the largest urban area.

2. Commuting municipalities near medium-sized towns – municipalities where more than 40% of the working population commute to work in a medium-sized town.

3. Commuting municipalities with a low commuting rate near medium-sized towns – municipalities where less than 40% of the working population commute to work in a medium-sized town.

C. Smaller towns/urban areas and rural municipalities (n = 136)

1. Small towns – municipalities with a population of at least 15 000 inhabitants in the largest urban area.

2. Commuting municipalities near small towns – municipalities where more than 30% of the working population commute to work in a small town/urban area or more than 30% of the employed day population lives in another municipality.

3. Rural municipalities – municipalities with a population of less than 15 000 inhabitants in the largest urban area, very low commuting rate (less than 30%)

4. Rural municipalities with a visitor industry – municipalities in rural area that fulfil at least two criteria for visitor industry, i.e. number of overnight stays, retail-, restaurant- or hotel turnover per head of population.