Article
School absenteeism among students in Germany, Japan, Sweden, and the United Kingdom: a comparative study using PISA data
Ulf Fredriksson
Stockholm University
Email: ulf.fredriksson@edu.su.se
Maria Rasmusson
Uppsala University
Email: maria.rasmusson@edu.uu.se
Åsa Backlund
Stockholm University
Email: asa.backlund@socarb.su.se
Joakim Isaksson
Stockholm University
Email: joakim.isaksson@socarb.su.se
Susanne Kreitz-Sandberg
Stockholm University
Email: susanne.kreitz-sandberg@edu.su.se
Abstract
School absenteeism is a challenge in many countries. Still, there are few comparisons between countries, which is partly due to a lack of shared definitions of concepts. This article makes use of PISA data to compare self-reported student absenteeism in Germany, Japan, Sweden, and the United Kingdom (UK). Three data sets are used, from 2012, 2015, and 2018. The self-reported absenteeism, which is referred to as truancy in the PISA studies, was measured as having skipped a whole school day at least once in the two full weeks before students completed the PISA student questionnaire.
The results show great variation between the studied countries, from 24.4% in the UK in 2015 to 1.5% in Japan in 2012. The percentage of students who reported having skipped school is much higher in the UK than in the other three countries. The differences between the countries concerning the percentage of students reporting having been absent from school are significant for all years, except between Sweden and Germany in 2015. Germany and the UK have a similar pattern in development, with the highest percentages in 2015, while Sweden and Japan have small (albeit not significant) increases from 2012 to 2018. The UK is the only country where the changes between 2012 and 2015 as well as between 2015 and 2018 were significant.
It is not possible to see any obvious patterns between the countries that might be linked to differences related to their welfare regimes, education systems, or values. To find such patterns, it may be necessary either to include more countries in a study or to conduct more in-depth studies on each country.
Keywords: school absenteeism, truancy, PISA, Germany, Japan, Sweden, the United Kingdom
Introduction
In recent years, the issue of students’ absence from school has been increasingly discussed in many countries. Most societies offer education free of charge, based on the assumption that education is something that helps children and young persons to develop both their cognitive and social skills. Education is characterized as a universal human right (UN, 1948; 1989). Some countries, like Sweden, have even declared education to be not only a right but also an obligation (Ministry of Education and Research, Education Act, 2010, chapter 1, 1 §; chapter 7, 2 §). Many countries report high rates of school absenteeism (e.g., Akbaşlı et al., 2017; Gren-Landell, 2021). If schools are to provide every child with an education, school non-attendance needs to be counteracted and prevented (Thornton et al., 2013).
Internationally, school non-attendance has long been in focus (Reid, 2008, 2012, 2016; Ricking, 2003; Kearney et al., 2022). Different concepts are used in the international literature, such as ‘truancy’, ‘school absenteeism’, ‘persistent school absenteeism’, and ‘school refusal behaviour’. In a review of international literature, Kearney (2008, p. 452) notes that ‘A key problem in the literature is that truancy, school refusal, and school phobia are used interchangeably or defined inconsistently’. Reid (2016) notes that there is a lack of agreed consensus about the use of critical definitions. School absenteeism seems to more generally refer to any kind of absence from school, while truancy more specifically refers to non-authorised absence. Examples of definitions in the literature of these concepts could be Ricking´s and Hagen´s (2016) description of school absenteeism as a concept including various behavioural patterns of the illegitimate absence of multi-causal and long-term origin and Ricking´s (2014) definition of truancy as a pattern of behaviour in which pupils disapprove of school and regularly make this clear for example by absenteeism. In this article, absenteeism is defined in line with the measure available from PISA; that is, the self-reported frequency of skipping school the two weeks before the PISA test. The concepts of truancy and skipping school will be used interchangeably in reference to this measure.
Considering that non-attendance seems to be a growing problem in many countries, an international comparative approach to this issue may offer a way to gain new insights into the size of this problem. Interesting questions to explore from an international comparative perspective are whether there are big differences between countries concerning the percentage of students skipping school and whether similar trends can be observed between countries. However, there are few international comparative studies on school non-attendance (e.g., Grewe, 2005). A challenge when comparing school absenteeism between countries is that comparable data are largely missing. International educational statics provided by UNESCO (UNESCO Institute for Statistics, 2023) and OECD (2022) do not contain such statistical information. This is due to the lack of agreed international definitions and the use of different national definitions. Data is not registered in the same way. In some countries, all absenteeism is registered while others make distinctions between different types of absenteeism. The information collected is made available in different ways. Available national statistics are not wholly possible to compare (Kreitz-Sandberg et al., 2023). Also, official statistics are based on how schools and/or teachers register absences rather than on students´ reports. In this article, to find comparable data, we follow Keppens and Spruyt (2018) and explore whether data collected as part of a large international study of students’ school achievements can be used. Keppens and Spruyt used PISA data from 24 EU countries, looking specifically at the relationship between truancy and different types of education systems.
This article will focus on and compare Germany, Japan, Sweden, and the UK. These countries have been selected as they all seem to have recognised school absenteeism as a growing problem and they represent different types of welfare states (e.g., Esping-Andersen, 1990, 1999; Goodman, 1998), different types of education systems (e.g., Dupriez et al., 2008; Janmaat & Mons, 2011; Mons, 2007) and societies with different values (e.g., Hofstede Insights, 2021). It can also be noted that the selected countries do not seem to be often compared, as many comparisons are limited to English-speaking countries. When countries are compared one should be aware of regional differences within countries, but national statistics and international comparative statistics do not always provide such detailed information from all countries.
In Germany, we find much research on school non-attendance (Hillenbrand & Ricking, 2011; Ricking, 2017; Sälzer & Heine, 2016). In Japan, school refusal has been problematized since the 1980s and an analysis of various discourses on this problem was conducted already 20 years ago (Yoneyama, 2000). In Sweden, school non-attendance has become a frequent topic in the media since the publication of a government inquiry (Ministry of Education and Research, 2016). Although media and public agencies frequently write about what is described as a serious problem, there is still very little research available in Sweden (Ekstrand, 2015). Swedish data on school non-attendance are extremely scarce compared to some other countries. In the UK, there have been many studies on school attendance problems from the perspectives of different actors and on the different parts constituting the UK (Reid, 2008, 2010, 2012).
We want to explore whether the prevalence of students skipping school is similar in the studied countries, whether there is a similar development over time in the studied countries, and whether different developments can be linked to differences between countries. According to a study by Keppens and Spruyt (2018) on truancy in 24 European countries based on PISA data from 2015, it was found that there are great differences in truancy rates between countries. The researchers found a strong relationship between truancy and early school leaving, and suggested that truancy may provide an indicator of school disengagement. They also noted ‘higher truancy rates in comprehensive and individualised systems’ (p. 422). Other possible differences between countries that may have an impact on student absenteeism can be related to the different types of welfare systems as described by Esping-Andersen (1990). Peter et al. (2010) argue that the results they found support Esping-Andersen’s ‘three worlds’ typology in that the level of between-school educational inequality is the highest in conservative welfare states and is the lowest in social democratic countries. Claes et al. (2009) found some evidence that truancy occurred more often in countries with strong income inequality. It should although be kept in mind that welfare systems have changed in many countries since the 90s. For example, in Sweden, dramatic increases in school segregation have been reported in the last decade (Skolvärlden, 2022; Lärarnas Riksförbund & Lärarförbundet, Sveriges Skolledarförbund, 2022). Historic and cultural differences may be another reason behind differences. This could be considered related to the values embraced in societies as described by Hofstede Insights (2021). We will return to this in the discussion of the results.
As it is difficult to compare national statistics concerning student absenteeism this article will use one source that contains such data. The PISA study will be used for the comparison. This has been done earlier by for example Keppens and Spruyt (2018) and by Sälzer and Heine, (2016). This article is limited to a smaller number of countries than in Keppen´s and Spruyt’s study but has a broader perspective. Sälzer and Heine looked at Germany and this article will include Germany in comparison with other countries. PISA[1] is often described as the largest educational study in the world. In PISA 2018, about 600,000 15-year-old students from 79 countries participated. As PISA is based on a representative sample of students, these 600,000 students are intended to represent 32 million 15-year-old students in the 79 participating countries (SNAE, Swedish National Agency for Education[2], 2019). PISA has been criticised from different perspectives (e.g., Hopfenbeck et al., 2018; Fredriksson et al., 2018) and we recognise that the use of PISA data in this context has some limitations (see section 6 on limitations), but as we have an interest in making international comparisons of school absenteeism we also acknowledge that PISA data may be a unique source when you wish to do such comparisons.
Aim and structure of the article
This article aims to compare student-reported absenteeism in Germany, Japan, Sweden, and the UK to explore whether and how PISA data can be used for further comparative analyses to provide input into the international discussion on student absenteeism. Germany, Japan, Sweden, and the UK were selected because of the reasons given above and because this article is part of a larger project focusing on these four countries where the intention is to both compare available data and to do a more in-depth analysis of actual practice in schools.
The intention of the study is expressed in the following research questions:
- Are there differences between the countries concerning the rate of students reporting that they skip school? 'What are the rates and differences of students reporting having skipped school according to PISA data in Germany, Japan, Sweden, and the UK?
- Can similar or different trends be observed between the countries? What trends can be seen in how the rate of students reporting they have skipped school has changed over the years?
- What could explain possible differences or similarities between the countries? Is it possible to link any differences between the countries concerning the rate of students reporting having skipped school and trends related to skipping school to broader differences between the countries related to welfare regimes, education systems, and values?
Available PISA data on truancy
PISA data is used in this study because it is difficult to find any other comparable international data that address school absenteeism. UNESCO´s official statistics on education (UNESCO Institute for Statistics, 2023) and OECD´s statistics on education - Education at a Glance (OECD, 2022) do not contain such information. Other large-scale student assessments such as PIRLS (Mullis et al., 2017) and TIMSS (Mullis et al., 2020) do not contain any information on school absenteeism. PISA seems to be one of few, and perhaps even the only set of data, that provides any statistical information on students´ school absenteeism in different countries.
PISA is a large-scale survey of student assessment that has been administered every third year since 2000. Its purpose is not related to measuring school absenteeism, but to measuring 15-year-old students’ knowledge and skills in reading comprehension, mathematics, and science. PISA has taken place seven times to date: 2000, 2003, 2006, 2009, 2012, 2015, and 2018.
To provide opportunities to analyse student results in-depth, information about the students is collected not only from the test of reading comprehension, mathematics, and science but also through a student questionnaire. This questionnaire contains questions related to student absenteeism.
In the student questionnaire for PISA 2018, the following question was asked:
In the last two full weeks of school, how often did the following things occur?
- I skipped a whole school day.
- I skipped some classes.
- I arrived late for school.
For each statement, the student could choose between the following responses:
Never; One or two times; Three or four times; Five or more times
The same question was also asked of students in PISA 2015 and 2012 (see Appendix A)[3]. In PISA 2000 a similar question was asked, but phrased as follows:
How many times in the previous two school weeks did you:
a) miss school?
b) skip classes?
c) arrive late for school?
For each question, the students could choose between the following alternatives:
None; 1 or 2; 3 or 4; 5 or more
It is possible to look at students’ school absences based on the same question for the years 2018, 2015, and 2012. It is also possible to get information about absenteeism for 2000, but the question was worded in another way that year which makes it difficult to compare with the other years. In this article, absenteeism is defined in line with the measure available from PISA; that is, the self-reported frequency of skipping school the two weeks before the PISA test. The concepts of truancy and skipping school will be used interchangeably in reference to this measure.
PISA contains data from the four countries of interest in this article (Germany, Japan, Sweden, and the UK). These four countries have participated in all the PISA studies since 2000. This means that it is possible to look at truancy in all these countries for at least the period 2012–2018.
Truancy in PISA reports
In PISA reports the concept of truancy is used to describe the students' answers to the question in the student questionnaire whether they have skipped at least a whole day of school at least once in the two weeks before the PISA test (OECD, 2019b, Volume III, figure III 4.1). We are aware that a concept such as truancy could be defined in other ways, but as we use the PISA data we will use this definition in the following parts of the article.
This information on student truancy has been collected in the different PISA studies, and the results have also been published and analysed in the reports based on them.
Table 1. Percentage of students who had skipped at least one class or at least an entire day of school in the two weeks before the PISA test
|
|
had skipped at least one class |
had skipped at least an entire day |
|
2012 |
18 |
15 |
|
2015 |
26 |
20 |
|
2018 |
27 |
21 |
Sources: OECD, 2014, p. 47; 2016, p. 81; 2019b, p. 78
It can be noted that, across the three PISA studies in which the students were asked the question about student truancy in a similar way, there was a change. In 2012, 15% of the students in the OECD countries reported that they had skipped at least an entire day of school in the two weeks before the PISA test. In 2015 this number had increased to 20%. In 2018 there was a further increase, with 21% of the students reporting having skipped a whole day of school at least once. Based on this, it can be concluded that during the period 2012 to 2018, there was a growing proportion of 15-year-old students across the OECD countries who reported having skipped a whole day of school at least once during the prior two weeks.
Methods
PISA data from Germany, Japan, Sweden, and the UK have been used in this study as they are made available in the international PISA database, which contains data from all PISA studies and all participating countries (OECD, 2020a). This database is public and is provided by the OECD. Individual students cannot be identified or linked to data in other databases.
The data contain the students’ self-declared truancy as described above. The question to the students about skipping a school day was part of the PISA student questionnaires in 2012, 2015, and 2018. Data are analysed to examine differences between the countries and trends within the countries over time. The demands of PISA’s complex survey design have been considered, and appropriate weights and methods of analyses (OECD, 2009) have been used. The analyses have been performed using Stata 14.2 (StataCorp, 2015) and the Stata module REPEST, developed by Avvisati and Keslair (2014) for secondary analyses of data from the large-scale assessments produced by the OECD. The analyses consist of estimates of proportions, confidence intervals, and t-tests of differences for comparisons both between countries and within countries between years (Bryman, 2015). The t-tests are used to estimate the likelihood that the differences detected in the samples hold true for the populations. It should be noted, however, that when a subgroup of the PISA sample represents less than five percent of the population of a country, the results should be interpreted with caution (OECD, 2009).
Limitations
Data from the PISA database has been used that represent four countries (Germany, Japan, Sweden, and the UK). We are aware that these countries, like all countries, have different regions, and as in Germany and the UK, these regions also have their own education systems. In the case of Germany (e.g., Kreitz-Sandberg et al., 2023; Grewe, 2005; Ehmann & Rademacker, 2003; Böhm, 2011) and the UK (e.g., Reid, 2010) there are differences in policies related to school absenteeism between the different regions. For the UK separate PISA data would be available for England, Northern Ireland, Scotland, and Wales for two of the three years covered in this study. In the case of Germany, no PISA data is presented at the level of the federal states (“Bundesländer”). As it would only be possible to do a more detailed analysis for the UK for some of the covered years and not at all for Germany, we have decided to focus on the available data on a country level as presented in the PISA database. That does not exclude an interest to further explore differences within the countries in another context.
The students who reported having been absent from some lessons or a day during the two weeks before the PISA test may not be the same ones as those who have been absent from schools during longer periods. The students who are absent for longer periods may not even have been in school when the questionnaire was administered, or may not have been to school at all during the weeks in question. Thus, the actual percentage of absent students may be higher.
The students who reported truancy during the two weeks before the PISA test could be in a process of increasing their absence, but could also have had very specific reasons for being absent on just some occasion during these weeks, not having been absent in the past and not intending to be absent in the future. However, several studies show that absence constitutes a risk for further absence (e.g. Connolly & Olson 2012; Ehrlich et al., 2014).
The information provided in the student questionnaire is based on the students’ self-reported truancy. The students complete the questionnaire at school. It may be that there are students who exaggerate or understate their truancy, or do not remember. It can also be assumed that truancy can be seen, by some students and in some contexts, as a norm-violating behaviour that is not socially desirable (Grimm, 2010; Nederhof, 1985), However, it can be assumed that in a large-scale study such as PISA, tendencies to overstate or understate may balance each other out. The PISA report (2019) addresses the time aspect, noting that the questionnaire ‘refers to the last two “full” weeks of school, this period may have been exceptional in some countries and economies, which could potentially affect students’ answers” (OECD, 2019b, p. 76). Keppens et al. (2019) point out that a short period of reference might lead to an underestimation of prevalence, but longer time frames may also undermine the reliability of self-reported data.
The wording of the item – ‘skipping school’ – carries an antisocial connotation. Students may hesitate to report truancy as they do not want to be seen as antisocial. Also, some students may have been registered with excused absence although the true reason for their absence may not be a valid excuse.
Another limitation when PISA data are used for comparison is the translations from English and French into the various national languages (in this study, German, Japanese, and Swedish). Although rigorous procedures are prescribed for the translations within PISA it cannot be disregarded that words, even when correctly translated, have different values and are perceived in different ways in different cultural contexts. One of the critical issues raised by Hopfenbeck et al. (2018) is related to the difficulties to make accurate translations from for example European languages to languages in other parts of the world.
Finally, it should be noted that PISA is a study based on a sample of students from a stratified random selection of schools in the countries, and in a second step, a random selection of 15-year-old students from the randomly selected schools. There is always a risk that a sample does not fully represent the population it is intended to represent. In the case of PISA (OECD, 2019a; Swedish National Agency for Education, 2019), this risk is generally estimated to be about five percent. Another issue to consider is the validity of the results when a study is designed and a sample is drawn for one purpose and then partly used for another one (Bryman, 2015). PISA was mainly designed for comparing student achievement, but in this article, its data are used to investigate student truancy.
Despite these limitations, the data from PISA may still be among the few datasets from different countries with comparable data on student truancy.
How many students reported that they had skipped school?
First, the actual numbers of the samples are described. Each participating country draws a representative sample of 15-year-old students. For more detailed information on how the samples have been drawn, see OECD (2020b).
Table 2 shows the number of students included in the samples from each country. As mentioned in Section 3, there were no questions about truancy in the PISA studies for 2009, 2006, or 2003. Thus, for obvious reasons, these years will not be dealt with in the further analysis and are not presented in Table 2.
Table 2. Number of students in each national sample
|
Country/ Year |
Germany |
Japan |
Sweden |
UK |
Total |
|
2018 |
5,451 |
6,109 |
5,504 |
13,818 |
30,882 |
|
2015 |
6,504 |
6,647 |
5,458 |
14,157 |
32,766 |
|
2012 |
5,001 |
6,351 |
4,736 |
12,659 |
28,747 |
|
2000 |
5,073 |
5,256 |
4,416 |
9,340 |
24,085 |
These samples are sufficiently large to generalize to the whole population of 15-year-old students in each country.
Table 3 shows the percentage of students reporting having skipped some classes during the two weeks before the administration of the PISA test.
Table 3. Percentage of students who have reported that they had skipped some classes
|
Year |
Country |
None |
One or two times |
Three or four times |
Five or more times |
|
2018 |
Germany |
79 |
15 |
4 |
7 |
|
Japan |
96 |
3 |
# |
# |
|
|
Sweden |
83 |
12 |
3 |
6 |
|
|
UK |
84 |
12 |
2 |
4 |
|
|
2015 |
Germany |
84 |
12 |
2 |
3 |
|
Japan |
97 |
2 |
# |
# |
|
|
Sweden |
84 |
12 |
2 |
2 |
|
|
UK |
66 |
27 |
4 |
6 |
|
|
2012 |
Germany |
90 |
9 |
1 |
# |
|
Japan |
97 |
2 |
# |
# |
|
|
Sweden |
80 |
16 |
3 |
4 |
|
|
UK |
88 |
9 |
1 |
2 |
|
|
2000 |
Germany |
88 |
9 |
1 |
3 |
|
Japan |
96 |
3 |
1 |
2 |
|
|
Sweden |
79 |
15 |
4 |
6 |
|
|
UK |
90 |
8 |
1 |
2 |
# Rounds to zero.
The great majority of students in all four countries have not reported that they had skipped any lessons during the two weeks in question, with the largest percentage of students who reported that they had skipped classes seen in the UK for 2015 (37%). No similarities in patterns can be identified among the four countries. The percentage of Japanese students who have reported that they have skipped classes seems to be fairly stable, while in Sweden there is a weak trend towards a smaller percentage of students skipping classes over the years.
Let us now look at the students responding that they skipped a whole school day during the two weeks before completing the PISA questionnaire.
Table 4. Percentage of students who have reported that they had skipped school a whole day
|
Year |
Country |
None |
One or two |
Three or four |
Five or more |
|
2018 |
Germany |
87 |
9 |
2 |
2 |
|
Japan |
98 |
2 |
# |
# |
|
|
Sweden |
90 |
7 |
2 |
2 |
|
|
UK |
81 |
15 |
2 |
2 |
|
|
2015 |
Germany |
91 |
7 |
1 |
1 |
|
Japan |
98 |
1 |
# |
# |
|
|
Sweden |
91 |
7 |
1 |
1 |
|
|
UK |
75 |
21 |
3 |
2 |
|
|
2012 |
Germany |
95 |
4 |
# |
# |
|
Japan |
98 |
1 |
# |
# |
|
|
Sweden |
93 |
6 |
1 |
1 |
|
|
UK |
82 |
15 |
2 |
1 |
|
|
2000* |
Germany |
73 |
21 |
3 |
2 |
|
Japan |
89 |
9 |
1 |
1 |
|
|
Sweden |
62 |
29 |
6 |
3 |
|
|
UK |
65 |
28 |
4 |
3 |
* In 2000 the question was about missing rather than skipping school
In Table 4, the majority of students in all four countries report not having skipped school on any day in the two weeks before the PISA test. We can note that no more than three percent of the students who participated in the PISA study in any country have reported that they had skipped whole school days five or more times. Also, the figures for 2000 are quite different from those for the other years. This is most likely related to the question having been asked differently: ‘Skipping school’ likely means that the student did not have an ‘authorized’ reason for the absence, while ‘missing school’ may be related to different types of absenteeism, with both authorized and unauthorized reasons.
Table 5. Percentage of students who have not reported that they had skipped some classes or skipped school days
|
Year |
Country |
Not skipped any classes |
Not skipped any school days |
|
2018 |
Germany |
79 |
87 |
|
Japan |
96 |
98 |
|
|
Sweden |
83 |
90 |
|
|
UK |
84 |
81 |
|
|
2015 |
Germany |
84 |
91 |
|
Japan |
97 |
98 |
|
|
Sweden |
84 |
91 |
|
|
UK |
66 |
75 |
|
|
2012 |
Germany |
90 |
95 |
|
Japan |
97 |
98 |
|
|
Sweden |
80 |
93 |
|
|
UK |
88 |
82 |
|
|
2000* |
Germany |
88 |
73 |
|
Japan |
96 |
89 |
|
|
Sweden |
79 |
62 |
|
|
UK |
90 |
65 |
* In 2000 the question was about missing rather than skipping school
Generally, there is a higher percentage of students who have reported that they had not skipped classes than those who have reported that they had not skipped whole school days, but the figures for 2000 stand out. During that year, in all four countries, more students reported having missed a whole school day than some lessons. This might be explained by what has been mentioned earlier, the fact that students may regard the terms ‘missing school’ and ‘skipping school’ as different types of absenteeism.
In the following text, we will focus on the students reporting having skipped a whole school day or more, and we will not analyse data on skipped classes further. We will also not include the data from 2000. Our initial intention was to focus only on the group of students reporting absences of five or more days; however, the numbers turned out to be very small in this group. We also note that in the PISA reports the OECD has often looked at all students who report having skipped school at least one day. It could also be argued that if one day during these two weeks represents a similar pattern across a whole year, this would correspond to 20–25 days of absence, which is often regarded as a persistent absence[4]. Based on this, we will focus on those students who report having skipped a whole school day at least once during the two full weeks of school before they completed the PISA student questionnaire. This group will be further analysed.
Table 6 shows the number and percentage of students reporting having skipped a whole school day at least once during the two full weeks of school before they completed the PISA student questionnaire.
Table 6. Number of students and percentage reporting having skipped a whole school day at least once during the two full weeks of school before the PISA test
|
Country/Year |
Germany |
Japan |
Sweden |
UK |
||||
|
N |
% |
N |
% |
N |
% |
N |
% |
|
|
2018 |
327 |
6.0 |
123 |
2.1 |
475 |
8.8 |
2,836 |
17.0 |
|
2015 |
474 |
7.2 |
115 |
1.8 |
465 |
8.6 |
3,762 |
24.4 |
|
2012 |
218 |
4.4 |
98 |
1.5 |
334 |
7.0 |
2,607 |
17.6 |
In Table 6, there are a few hundred students who report having skipped at least a whole school day, except for the UK, where the numbers are considerably higher. This is partly related to the size of the samples from the UK, as seen in Table 2.
As the size of the sample has varied between years and between countries, using percentages for comparison is the most appropriate way to analyse the data. The percentages given in Table 6 are illustrated in Figure 1 with confidence intervals.
Figure 1. Percentage of students who have reported that they had skipped a whole school day at least once during the last two full weeks of school before they completed the PISA student questionnaire
There are no overlapping confidence intervals for the estimates between the countries in 2012 and 2018, respectively. Thus, the shares of students reporting having skipped school are statistically different (p<.05) between the countries for these two years. In 2015 the estimates for Sweden and Germany are overlapping and thus not statistically different.
Generally, Table 6 and Figure 1 show that there are differences between the four countries in respect of the rate of students who have reported that they had skipped at least one school day during the two weeks before the PISA test. It can also be noted that the UK has the highest proportion of students who report having skipped school and Japan has the lowest. The difference between Japan and the UK is considerable.
Can we see any trends in the countries?
The development in the four countries over time is illustrated in Table 7 below.
Table 7. Differences within countries, between years, in the percentage of students reporting having skipped a whole school day
|
Country/ Year |
Germany |
Japan |
Sweden |
UK |
|
2012 –2015 |
-2.79* |
-0.23 |
-1.59 |
-6.79* |
|
2015–2018 |
1.19 |
-0.30 |
-0.21 |
7.35* |
|
2012–2018 |
-1.60 |
-0.53 |
-1.80 |
-0.57 |
* = p<.05
As seen in Table 7, there are significant differences in the percentage of students who skipped school in Germany between 2012 and 2015 but no significant difference if we look at the development from 2012 to 2018 and from 2015 to 2018. However, the results for comparison with data from 2012 in Germany must be interpreted with caution, as the proportion included in this subgroup is below five percent of the population in the German PISA studies. In Japan and Sweden, there are no significant changes in the percentages reporting skipping school between the years. In the case of the UK, there is a significant increase in the shares of students who have reported that they skipped school from 2012 to 2015 and a statistically significant decrease between 2015 and 2018, but no significant difference between 2012 and 2018.
As no statistically significant differences were found between the years in most of the countries, we cannot identify anything that could be described as a trend. That there have been no significant changes in the rate of absent students between 2012 and 2018 is of interest if we consider the growing concern internationally regarding absenteeism as well as the general figure showing a slight increase for OECD counties in general. It can be noted that the proportions of students reporting having skipped school are not significantly different between the countries in 2012 and 2018. What is puzzling is the shape of the UK curve, with a significant difference between 2012 and 2015 as well as between 2015 and 2018, but no significant difference between 2012 and 2018.
Discussion
Generally, this article has tried to answer two questions about truancy in Germany, Japan, Sweden, and the UK with the support of PISA data:
- Are there differences between the countries concerning the rate of students who have reported that they have skipped school? What are the rates and differences of students reporting having been absent from school according to PISA data in Germany, Japan, Sweden, and the UK?
- Can we observe similar or different trends between the countries? What trends can be seen in how the rate of students reporting having skipped school has changed over the years?
In addition, a more general question was also raised:
- What could be the factors that explain possible differences or similarities between the countries? Is it possible to link any differences between the countries to the rate of students reporting having skipped school and trends in skipping school to broader differences between the countries related to welfare regimes, education systems, and national values?
We can see variations between the countries in the percentage of students who report having skipped school at least one day during the two weeks before completing the PISA student questionnaire. The percentage of students varies from 24.4% in the UK in 2015 to 1.5% in Japan in 2012. Among the four counties, the UK has both the highest numbers and percentages of students who report having skipped school.
On average, across the OECD countries in 2018, 21% of students reported having skipped at least one day of school in the two weeks before the PISA test (OECD, 2019b). Generally, the countries included in this article had in 2018 a lower proportion of students who reported having skipped school than the average among the OECD countries.
As the data used in this article only cover three points of measurement over six years, it is difficult to observe any trends. Although there are similarities between Germany and the UK and between Japan and Sweden, there is no general trend that can be seen for all four countries.
Could any differences between the four countries explain the different rates of absent students? The four countries covered in the article, have different models of welfare states (Esping-Andersen, 1990). Esping-Andersen (1999) distinguishes between universal, liberal, conservative, and other types of welfare regimes. In the literature, Sweden is often regarded as a typical example of a Scandinavian ‘universal welfare state’, organizing many subsystems like pensions or healthcare universally. The UK is described as a ‘liberal welfare state’, Germany as a typical example of a ‘conservative welfare state’, and Japan as an ‘East-Asian Confucian welfare state’ (Goodman, 1998). As we could see in Figure 1, the data from Sweden and Germany are not that different, despite their being what may be described as different welfare states. The UK, and to some extent also Japan, stand out. Based on the data from these four countries, however, it does not seem feasible to assume that type of welfare system is related to rates of truancy.
It should be kept in mind that the structure of welfare regimes has gone through changes in the last 20 – 30 years. As described earlier, for example, Sweden has had increased differences between schools, which could be seen as an indication that Sweden has moved away from the social democratic welfare model. To further analyse the relationship between the welfare regime of a country and how this could influence student absenteeism would require a more in-depth analysis of the systems and their impact on young people and education.
Another factor that may impact student absenteeism could be the education system as such. An important feature of all education systems is the point at which students are separated into different tracks. Keppens and Spruyt (2018) have studied non-attendance in 24 European countries using PISA data from 2015. In their analysis they make a distinction based on Mons (2007), Dupriez et al. (2008), and Janmaat and Mons (2011) according to the level of horizontal and vertical stratification. Four types of education systems are identified: ‘(1) The separation model, (2) the comprehensive model, (3) the uniform integration model, and (4) the individualized integration model’ (p. 417). The separation model ‘refers to educational systems where students are grouped into different tracks according to their educational achievement at a young age’. The comprehensive model ‘does not use tracking to select and group students’ and all students ‘follow a common curriculum until the age of 16’ (p. 417). The uniform integration model ‘provides a common curriculum for all pupils up to the age of 16’, but unlike the comprehensive model ‘this model uses grade retention to re-group pupils with poor achievement’ (p. 417). Finally, the individualized integration model ‘refers to educational systems that offer a common curriculum for all students, with heterogeneity addressed by individual or small-group tutoring. (p. 417). If we look at the four countries selected for this study we can note that Keppens and Spruyt (2018) define the German education system as separated and the systems in Sweden and the UK as comprehensive. Japan is not classified in Keppen´s and Spruyt’s (2018) article, but following the logic of the model it seems reasonable to classify the Japanese system as comprehensive. Keppens and Spruyt claim that their ‘results showed higher truancy rates in comprehensive and individualised educational systems’. They continue, arguing that two explanations are plausible: ‘First, these models differ from the separated model in the degree of tracking. Whereas pupils in the separated model are already grouped into different educational tracks at an early age, pupils in the individualised and comprehensive models follow education altogether until the age of 16. Pupils following education in the separated model might therefore receive an education that is better suited to their specific abilities and skills at an early age’ (p. 422).
The education systems in Sweden and the UK, which are comprehensive, have higher percentages of non-attendance than those in Germany and Japan. As noted above, the difference between Germany with its selective system, and Sweden with its comprehensive system, is not large. It is not possible to see any differences between these four countries that may be related to different types of education systems.
Another aspect that may be relevant in this discussion is the values in society. A general approach to mapping values in different countries is taken by Hofstede Insights (2021). Based on large-scale surveys, information on basic values concerning power distance, individualism, masculinity, uncertainty avoidance, long-term orientation, and indulgence are collected and compiled. Among these, individualism, uncertainty avoidance, and long-term orientation may be most relevant in relation to truancy. Japan stands out in this context with lower values on individualism than the other countries, but higher values on uncertainty avoidance and long-term orientation. The Japanese values on long-term orientation are not much higher than the German ones. The Swedish values are different from those of other countries with respect to masculinity, which in this context is related to a preference in society for achievement, heroism, assertiveness, and material rewards for success (Hofstede Insights, 2021). As Japanese figures on students reporting that they have skipped school are much lower than the other countries this may be related to the lower Japanese values on individualism and the higher values on uncertainty avoidance that Japan shows in the Hofstede Insights. The notion of truancy as a norm-violating behaviour that is not socially desirable may be stronger in a society where individualism is a less frequent value and uncertainty avoidance a more frequent value than in other countries in this study.
Conclusions
The article aimed to use PISA data to draw comparisons between Germany, Japan, Sweden, and the UK with the objective of analysing questions about the proportion of student-who reported they had skipped schools and differences between the countries, as well as whether any trends could be found.
As shown in the discussion, it is possible to find answers to the questions about the proportion of students who have reported that they had skipped school and to draw some comparisons between the countries. The question of trends is more difficult to answer.
There are some differences between the countries, but it is more difficult to know what may explain these differences. Based on the comparison of the four countries covered in this article, it is not possible to say anything more specific about how welfare regimes, education systems, and values may have an impact on students’ truancy, how it is reported, and what is done to prevent truancy.
More research is needed to further investigate the reasons behind the differences between countries and the role of welfare regimes, education systems, and values. It would be necessary to determine whether patterns can be found between a larger number of countries, or whether a more in-depth analysis of each of the countries in this study would reveal details on how issues related to welfare regimes, education systems, and values can be assumed to affect student truancy. It may be assumed that truancy is not a simple reflection of only one of these reasons, but related to how the welfare regimes, education systems, and values interact. It can also be noted that we have only compared national figures from PISA on absenteeism. It would be possible to break up these figures into subgroups based on regions, gender, migration, and socio-economic status. As mentioned in the limitations, separate data for England, Northern Ireland, Scotland, and Wales are available for the PISA studies in 2015 and 2018, but not for 2012. If the PISA data from 2018 are explored, it can be noted that significant differences exist between the parts of the UK. The percentage of students who have skipped school during the two weeks before the PISA test in England is significantly lower than the corresponding figures for Northern Ireland, Scotland, and Wales [17.7 % in England compared with 25.3 in Northern Ireland, 22.1 in Scotland and 28.8 % in Wales (OECD, 2020a)]. This type of more in depth-analysis may provide further information that could give a better understanding of what factors may influence student absenteeism.
Finally, to what extent does it make sense to use PISA data for this type of analysis? As mentioned in the introduction to the article, one problem in international comparisons of truancy is the differences in available national statistics (Kreitz-Sandberg et al., 2023). PISA may be one way of getting internationally comparable data. This study shows that it is possible to use PISA data for such analysis, but that there are also limitations with these data.
Do the PISA studies reflect the actual prevalence of truancy? After all, they rely solely on student-reported data and do not use data on absences reported by parents or registered by the school. Whether students report truthfully about their own school absences may be different in different countries due to cultural differences, and long-term absent students may be disregarded in school-based surveys. However, even reports from school personnel may have reliability and validity problems. In a study on students aged 15-16 years in Flanders (Keppens et al., 2019), self-reported truancy was reported to be higher than that registered by school personnel.
To gain more knowledge about truancy and school absenteeism, there is a need for shared definitions that are guided by current knowledge about, for example, thresholds and correlated sociodemographic variables. Several authors in the field speak in favour of including excused as well as unexcused absences to capture the phenomenon of students’ school absences (Birioukov, 2016; Kearney, 2016). Meanwhile, the PISA studies only ask about unexcused absences.
A disappointing discovery in approaching the PISA data was that truancy had not been covered in all the PISA studies since 2000 and that partly different wordings had been used in the student questionnaires. Instead of offering an opportunity to compare data from seven PISA studies, covering almost 20 years, the span that was possible to cover was reduced to three of the studies over a period of six years. As PISA provides a unique opportunity to collect information from students about how they experience schooling, it would be useful if the student questionnaires in these studies were made more transparent between years. It is of course also important that questions about absenteeism are included in future PISA studies and perhaps also to make a distinction between different types of absenteeism.
Despite these shortcomings, we must conclude that PISA provides a great deal of data, of which this article has only used a part. PISA data on truancy can be used to analyse the relationship between truancy and students’ performance in reading, mathematics, and science, and might also be used to further analyse which students have skipped school.
Acknowledgement
The study was conducted with the support of the Swedish Research Council (VR 2019-04639).
References
Akbaşlı, S., Şahin, M., & Mardin Yılmaz, S. P. (2017). Struggle with School Absenteeism in Compulsory Education: Different Country Approaches and Policies. Universal Journal of Educational Research, 5(11), 2107-2115. https://doi.org/10.13189/ujer.2017.051128
Kreitz-Sandberg, S., Backlund, Å., Fredriksson, U., Isaksson, J., Rasmusson, M., & Gren Landell, M. (2023). Recording and Reporting School Attendance Problems: International comparative Views on Attendance Statistics in Sweden, Germany, England, and Japan. Orbis Scholae
Avvisati, F., & Keslair, F. (2014). REPEST: Stata module to run estimations with weighted replicate samples and plausible values. Statistical Software Components S457918, Boston College Department of Economics, revised 06 Jan 2020.
Birioukov, A. (2016). Beyond the Excused/Unexcused Absence Binary: Classifying Absenteeism through a Voluntary/Involuntary Absence Framework. Educational Review, 68(3), 340-357. https://doi.org/10.1080/00131911.2015.1090400
Bryman, A. (2015). Social Research Methods (5th ed.). Oxford University Press.
Böhm, T. (2011). Erziehungs- und Ordnungsmaßnahmen in der Schule. Schulrechtlicher Leitfaden. [Eng. Education and Disciplinarly Measures in the schools: Legal Framework]. Carl Link.
Claes, E., Hooghe, M., & Reeskens, T. (2009) Truancy as a contextual and school‐related problem: a comparative multilevel analysis of country and school characteristics on civic knowledge among 14 year olds. Educational Studies, 35(2), 123-142. https:/doi.org/10.1080/03055690802470258
Connolly, F., & Olson, L. S. (2012). Early elementary performance and attendance in Baltimore City Schools' pre-kindergarten and kindergarten. Baltimore Education Research Consortium.
Department for Education (2020). National statistics announcement. Pupil absence in schools in England: 2019 to 2020. https://www.gov.uk/government/statistics/announcements/pupil-absence-in-schools-in-england-2019-to-2020
Dupriez, V., Dumay, X., & Vause, A. (2008). How do school systems manage pupils’ heterogeneity. Comparative Education Review, 52(2), 245–273. https://doi.org/10.1086/528764
Ehmann, C., & Rademacker, H. (2003). Schulversäumnisse und sozialer Ausschluss. [Eng. School neglect and- social exclusion]. Bertelsmann.
Ehrlich, S. B., Gwynne, J. A., Stitziel Pareja, A., Allensworth, E. M., Moore, P., Jagesic, S., & Sorice, E. (2014). Preschool attendance in Chicago public schools: Relationships with learning outcomes and reasons for absences. University of Chicago Consortium on Chicago School Research.
Ekstrand, B. (2015). What it takes to keep children in school: a research review. Educational Review, 67(4), 459–482. https://doi.org/10.1080/00131911.2015.1008406
Esping-Andersen, G. (1990). The three worlds of welfare capitalism. Polity Press. https://doi.org/10.1007/978-3-531-90400-9_32
Esping-Andersen, G. (1999). Social Foundations of Postindustrial Economies. Oxford University Press. https://doi.org/10.1093/019842002.001.0001
Fredriksson, U., Karlsson, K. G., & Pettersson, A. (2018). Kritiken mot PISA [Eng. The Criticism about PISA]. In U. Fredriksson, K. G. Karlsson, & A. Pettersson (Eds.), PISA under 15 år – resultat och framtidsutsikter [Eng. PISA during 15 years: results and trends]. Natur & Kultur.
Gren Landell, M. (Ed.) (2021). School Attendance Problems. A Research Update and Where to Go. Jerringfonden. https://jerringfonden.se/wp-content/uploads/2021/04/202101-Jerringfonden-Antologi-A5-sammanslagen.pdf
Grewe, N. (2005). Absenteeism in European Schools. LIT Verlag.
Grimm, P. (2010). Social Desirability Bias. John Wiley & Sons, Ltd. https://doi.org/10.1002/9781444316568.wiem02057
Goodman, R. (1998). The ‘Japanese-style welfare state’ and the delivery of personal social services. In R. Goodman, G. White, & H.-J. Kwon (Eds.), The East Asian welfare model. Welfare orientalism and the state (pp. 139–158). Routledge. https://doi.org/10.4324/9780203019160-8
Hofstede Insights (2021). web document retrieved 21-11-02 https://www.hofstede-insights.com/product/compare-countries
Hopfenbeck, T. N., Lenkeit, J., El Masri, Y., Cantrell, K., Ryan, J., & Baird, J.-A. (2018). Lessons Learned from PISA: A Systematic Review of Peer-Reviewed Articles on the Programme for International Student Assessment. Scandinavian Journal of Educational Research, 62(3), 333-353. https://doi.org/10.1080/00313831.2016.1258726
Hillenbrand, C., & Ricking, H. (2011). Schulabbruch: Ursachen, Entwicklung, Prävention. Zeitschrift Für Pädagogik [Eng. Reasons behind School Drop-out, Development, Prevention], 57, Heft 2, 153–172.
Janmaat, J. G., & Mons, N. (2011). Promoting ethnic tolerance and patriotism: The role of education system characteristics. Comparative Education Review, 55(1), 56–81. https://doi.org/10.1086/657105
Kearney, C. A. (2008). School absenteeism and school refusal behavior in youth: A contemporary review. Clinical Psychology Review, 28(3), 451–471. https://doi.org/10.1016/j.cpr.2007.07.012
Kearney, C. A. (2016). Managing school absenteeism at multiple tiers: An evidence-based and practical guide for professionals. Oxford University Press
Kearney, C. A., Benoit, L., Gonzálvez, C., & Keppens, G. (2022). School attendance and school absenteeism: A primer for the past, present, and theory of change for the future. Frontiers in Education, 7. https://doi.org/10.3389/feduc.2022.1044608
Keppens, G., & Spruyt, B. (2018). Truancy in Europe: Does the type of educational system matter? European Journal of Education Research, Development and Policy, 53(3), 414–426. https://doi.org/10.1111/ejed.12282
Keppens, G., Spruyt, B., & Dockx, J. (2019). Measuring school absenteeism: Administrative attendance data collected by schools differ from self-reports in systematic ways. Frontiers in Psychology, 10. https://doi.org/10.3389/fpsyg.2019.02623
Lärarnas Riksförbund, Lärarförbundet & Sveriges Skolledarförbund (2022) Skolsegregationen i svensk grundskola 2011–2021. [Eng. School segregation in Swedish comprehensive school 2011 – 2021.] https://arkiv.lr.se/download/18.354fb49b18418b6aaa350fc/1667558146197/Ladda%20ned%20PM%20%E2%80%94%20Skolsegregationen%20i%20svensk%20grundskola%202011%E2%80%942021.pdf
Ministry of Education and Research (2010). Skollagen, kapitel 1, 8 §; kapitel 7, 2 § [in English: Education Act chapter 1, 1 §; chapter 7, 2 §) 2010:800 SFS nr: 2010:800]. https://www.riksdagen.se/sv/dokument-lagar/dokument/svensk-forfattningssamling/skollag-2010800_sfs-2010-800
Ministry of Education and Research (2016). Saknad! Uppmärksamma elevers frånvaro och agera. [Eng. Missed! Recognise student absence and react.] Statens offentliga utredningar SOU 2016:94 [Public inquiries 2016:94]. Utbildningsdepartementet/Wolters Kluwer.
Mons, N. (2007). Les nouvelles politiques éducatives: La France fait-elle les bons choix? [Eng. New eduational policies : is France making the right choices ?]Presses Universitaires de France. https://doi.org/10.4000/rfp.1038
Mullis, I. V. S., Martin, M., Foy, P., & Hooper, M. (2017). PIRLS 2016: International Results in Reading. IEA. https://timssandpirls.bc.edu/pirls2016/international-results/wp-content/uploads/structure/CompletePDF/P16-PIRLS-International-Results-in-Reading.pdf
Mullis, I. V. S., Martin, M. O., Foy, P., Kelly, D. L., & Fishbein, B. (2020). TIMSS 2019 International Results in Mathematics and Science. IEA. https://timss2019.org/reports/wp-content/themes/timssandpirls/download-center/TIMSS-2019-International-Results-in-Mathematics-and-Science.pdf
Nederhof, A. J. (1985). Methods of coping with social desirability bias: A review. European Journal of Social Psychology, 15(3), 263–280. https://doi.org/10.1002/ejsp.2420150303
OECD (2009). PISA Data Analysis Manual: SPSS, Second Edition. OECD Publishing. https://doi.org/10.1787/9789264056275-en
OECD (2014). PISA 2012 Results: Ready to Learn. Students’ Engagement, Drive and Self-beliefs. (Volume III). OECD. https://doi.org/10.1787/9789264201170-en
OECD (2016). PISA 2015 Results: Policies and Practices for Successful schools (Volume II). OECD. https://doi.org/10.1787/9789264267510-en
OECD (2019a). PISA 2018 Results: What Students Know and Can Do. Volume I. OECD. https://doi.org/10.1787/5f07c754-en
OECD (2019b). PISA 2018 Results: What School Life Means for Students’ Lives. Volume III. OECD. https://doi.org/10.1787/acd78851-en
OECD (2020a). PISA Database. http://www.oecd.org/pisa/data/
OECD (2020b). PISA 2018 Technical Report. https://www.oecd.org/pisa/data/pisa2018technicalreport/
OECD (2022) Education at a Glance. OECD. https://doi.org/10.1787/3197152b-en
Peter, T., Edgerton, J. D., & Roberts, L. W. (2010). Welfare regimes and educational inequality: a cross‐national exploration. International Studies in Sociology of Education, 20(3), 241-264. https://doi.org/10.1080/09620214.2010.516111
Reid, K. (2008). The causes of non-attendance: An empirical study. Educational Review, 60(4), 345–357. https://doi.org/10.1080/00131910802393381
Reid, K. (2010) Management of School Attendance in the UKA Strategic Analysis. Educational Management Administration & Leadership, 38(1), 88-106. https://doi.org/10.1177/1741143209352387
Reid, K. (2012). The strategic management of truancy and school absenteeism: Finding solutions from a national perspective. Educational Review, 64(2), 211–222. https://doi.org/10.1080/00131911.2011.598918
Reid, K. (2016). Introduction. In K. Reid (Ed.), Managing and Improving School Attendance and Behaviour New Approaches and Initiatives. Routledge. https://doi.org/10.4324/9781315652108
Ricking, H. (2003). Schulabsentismus als Forschungsgegenstand. Heilpädagogische Forschung [Eng. School absenteeism as an object for research] (Vol. XXIII). https://doi.org/04.2003.22
Ricking, H. (2014). Schulabsentismus. [Eng. School absenteeism]. Cornelsen.
Ricking, H. (2017). Schulabsentismus [Eng. School absenteeism]. Zeitschrift Für Jugendkriminalrecht Und Jugendhilfe ZJJ, 28(3), 265–270.
Ricking, H., & Hagen, T. (2016). Schulabsentismus und Schulabbruch [School absenteeism and School Drop-out]. Kohlhammer
Skolvärlden (2022). Nya siffror – segregationen ökar mellan kommunala skolor [Eng. New figures – segregation increases between municipal schools]. https://skolvarlden.se/artiklar/nya-siffror-segregationen-okar-mellan-kommunala-skolor
StataCorp (2015). Stata Statistical Software: Release 14. StataCorp LP.
Swedish National Agency for Education (2019). PISA 2018: 15-åringars kunskaper i läsförståelse, matematik och naturvetenskap [Eng. PISA 2018: 15-Year-Olds' Knowledge in Reading Comprehension, Mathematics and Science]. Skolverket.
Sälzer, C., & Heine, J.‑H. (2016). Students’ skipping behavior on truancy items and(school) subjects and its relation to test performance in PISA 2012. International Journal of Educational Development, 46, 103–113. https://doi.org/10.1016/j.ijedudev.2015.10.009
Thornton, M., Darmody, M., & McCoy, S. (2013). Persistent absenteeism among Irish primary school pupils. Educational Review, 65(4), 488–501. https://doi.org/10.1080/00131911.2013.768599
UN (1948). The Universal Declaration of Human Rights. The United Nations General Assembly in Paris on 10 December 1948, General Assembly resolution 217 A. United Nations.
UN (1989). Convention on the Rights of the Child. Adopted and opened for signature, ratification and accession by General Assembly resolution 44/25 of 20 November 1989. United Nations entry into force 2 September 1990, in accordance with article 49.
UNESCO Institute for Statistics (2023). https://uis.unesco.org/en
Yoneyama, S. (2000). Student Discourse on Tokokyohi (School Phobia/Refusal) in Japan: Burnout or empowerment? British Journal of Sociology of Education, 21(1), 77–94. https://doi.org/10.1080/01425690095171
Appendix A
Questions about truancy in four languages
English
German
Japanese
Swedish
