Evaluation of the PISA studies: influence of social background

from Wikipedia, the free encyclopedia
This article or paragraph presents the situation in Germany . Help us to describe the situation in other countries.

The OECD's PISA school studies include a two-hour performance test and a questionnaire session lasting just under an hour. In particular, the questionnaires collect data on social background. As part of the evaluation of the PISA studies , the influence of social background on the test results was examined in detail.

overview

In the international evaluation of the individual PISA cycles, one chapter is devoted to the relationship between test performance and various background variables. Essentially, only the performance in the respective focus area tested is analyzed (reading in PISA 2000, mathematics in 2003, natural sciences in 2006). The data are presented and interpreted in a similar way each time. A large number of key figures are produced and communicated in various tables and graphics. Some of these results can be read as country rankings.

In order to quantify social background one-dimensionally, an “International Socio-Economic Index of Occupational Status” (ISEI) is used, more precisely the value of the higher rated parent (highest ISEI = HISEI). The relationship between test performance and this index is above average in Germany. In 2000, the strongest HISEI gradient was determined for Germany among all 32 participating countries in the priority area reading (followed by the Czech Republic, Switzerland and Luxembourg). This unfavorable top position was clearly highlighted in the national evaluation by the German project management and reached a broad public. Because of this lasting effect, which is also reflected in extensive literature, the following presentation concentrates on the results for Germany.

The following results in particular attracted attention in Germany:

  • Strong relationship between test performance and social class;
  • In an international comparison, extremely strong influence of the migration background , which, however, mostly runs parallel to unfavorable social conditions;
  • considerable social condition of choice of school even with the same test achievement;
  • East-west and north-south differences in a comparison of the federal states, which in view of the sheer number of partial results do not give a clear picture.

methodology

Performance data

Each test participant works on around 45 to 60 test items, most of which are only rated “right” or “wrong”. According to the number of correct problem solutions, each subject is assigned a point value per task area, which is interpreted as “competence”. Critics point out that PISA results reflect not only subject-related skills, but also, for example, motivation and test ability (familiarity with the task format , qualified guessing, timing, stress resistance).

PISA tests not only school knowledge, but also the ability to reflect on this knowledge and apply it to everyday questions. This educational goal is called literacy in English ; in the German reports the term is taken over without translation. To test literacy , all math and science tasks begin with, in some cases, extensive introductory texts. Since the entire test takes place under considerable time pressure, all performance results, not just in the reading area, depend to a large extent on the ability to read quickly and comprehensively. In an international comparison, this leads to distortions because the task texts in different languages ​​are of different lengths and different degrees of difficulty to read. For example, the math problems in German contain 16% more letters than in English.

Social data

The questionnaire collects background data of several hundred bits per test participant. Some evaluations (for example on the migration background) refer to very specific bits. Other evaluations follow the consortium's approach of summarizing information on social background into a global key figure.

In PISA 2000, social background was assessed solely on the basis of a socio-economic index value of parental occupation ( HISEI ).

In PISA 2003 a newly defined "index of economic, social and cultural status" ( ESCS ) was used, which is determined as the main component ( eigenvector for the greatest eigenvalue of the matrix of correlation coefficients ) of the following three sub-indices:

  • The already mentioned “international socio-economic index of occupational status” ( ISEI ) according to Ganzeboom (1992), whereby only the value of the parent rated higher in this regard is taken into account (“highest ISEI” = HISEI);
  • the training duration of the longer trained parent, derived from the student information on the parents' training qualifications;
  • a rapidly scaled indicator that summarizes the household's equipment with individual cultural goods (dishwasher, calculator, internet access, books of poetry, works of art, ...).

The authors of the German reports express theoretical reservations about the ISEI, which they consider to be poorly founded and unpredictable, and above all to the ESCS, which they accuse of obscuring the actual social inequality and that it is neither from country to country, still comparable from test cycle to test cycle.

For its own evaluations, the German consortium prefers a more coarse classification of parental professions into seven social classes. With reference to Erikson / Goldthorpe / Portocarero (1979), these layers are also referred to as EGP classes. In PISA 2000, the following class distribution of student parents was found for Germany:

EGP Social class father mother
I. Upper class of service 20.7% 7.4%
II Lower class of service 16.5% 22.8%
III Routine services in trade and administration 4.9% 39.4%
IV Self-employed 12.5% 5.9%
V, VI Skilled workers and workers with management functions 26.0% 7.3%
VII Semi-skilled and unskilled workers, farm workers 19.5% 17.1%

In PISA 2003 no EGP class could be determined due to missing information; in PISA 2006 the proportion of self-employed and routine service providers (IV, III) decreased slightly, that of the upper class increased.

Handling of missing data

As in almost all survey-based social studies, data analysis in PISA is made more difficult by a not inconsiderable proportion of missing answers. Some students abandon the test before or during the questionnaire session, others do not answer all questions. If all participants who did not answer any question were excluded from the analysis, the data set would be so greatly reduced that one would have to reckon with unforeseeable distortions. A better approach would be to exclude only those students from each partial evaluation who are missing the specific information required. However, each partial result would then be based on a different partial sample, and considerable biases would still have to be expected from case to case.

Instead, missing answers are imputed in PISA : It is assumed that the answers given are sufficient to characterize the social background of a student so precisely that the missing information can be replaced by random numbers, the frequency distribution of which is based on the answers of the other students comparable background oriented. Specifically, the ESCS index is imputed in the international data set if only two of the three required sub-indices can actually be calculated based on the student information. The German PISA consortium goes beyond this and imputes the ESCS even if there are fewer than two sub-indices. The reason is probably that a particularly thorough random sample was drawn in Germany, including special schools, and the primary data set is therefore incomplete above average. The remaining uncertainty of the data after imputation of the missing information does not seem to have been estimated quantitatively so far.

Reliability of student information

French government agencies have warned the OECD against using the ISEI because student information about their parents' work and education is too unreliable. The OECD only accepted this objection for the country from which it was raised: "In the case of France, questions remain about the reliability of students' responses regarding parental occupation and education".

In Germany, the student information was validated by a survey of parents. When it came to the information on the parents' school-leaving qualifications, the agreement between pupils and parents was around 70% for the most frequent qualifications, and significantly lower for “exotic” qualifications. If the parents had the technical college entrance qualification, only 27% of the children stated this correctly (in this case mothers are more likely to have the secondary school leaving certificate, fathers the Abitur). For the vocational qualification and the four-digit coded occupation, the agreement between pupil and parent information was around 40%; the resulting two-digit ISEI is around 45%. Nevertheless, Maaz summarize among other things that pupils can be seen as “reliable informants for the collection of educational and occupational characteristics of parents”.

Measures of connection: quantile differences, gradient, correlation coefficient

In order to quantify the dependency of the test performance on the social background and to make it comparable in the form of ranking lists, various measures of association are used in the PISA reports. The German consortium prefers to use the non-technical term coupling , as a kind of umbrella term for various mathematical measures of connection.

The specification of quantile differences is particularly clear and easy to understand . For example, the test subjects can be divided into four groups of equal size according to their HISEI index value and the mean performance can be calculated for each of these four quartiles. Then you can form the difference between the mean performance values ​​of the first and last quartile and thus obtain a one-dimensional measure that can be communicated in the form of ranking lists for the strength of the relationship between parental occupation and student performance. To make this measure even clearer, the bottom quartile is referred to as the “working class” and the top quartile as the “upper class”. The data of the two middle quartiles are not taken into account in this analysis.

Two other measures of association, gradient and correlation coefficient , are based on linear regression : in a coordinate system , the dependent variable , i.e. the test performance, is plotted against the independent variable, the respective social index under consideration . A data point is now entered for each student. A point cloud is obtained whose more or less diagonal frequency distribution reveals a more or less strong relationship.

By this point cloud you draw a regression line , and such that the mean square vertical distance between the points is minimized by the straight line. The slope of this straight line is called the gradient; in the special case as a social gradient . If a straight line rises not only because of statistical randomness, but also because of a direct or indirect causal relationship, then the gradient expresses the strength of this relationship.

The correlation coefficient R 2 is obtained from the gradient by dividing it by the spread of the independent and dependent variables. This gives a dimensionless number that takes values ​​between 0 and 1. In the PISA reports, the correlation coefficient is regularly referred to as explanation of variance : for example, if a point cloud yields the value R 2 = 0.16, then it is said that social background “explains” 16% of the variance in test performance. This "explanation" is to be understood as statistical technical language; it presupposes that there is a direct causal relationship between the independent and the dependent variable.

Results of the international comparison

Reading focus 2000

Germany had in the reading test 2000

  • the greatest variation (111 points, followed by New Zealand 108, Belgium 107; OECD-wide standardized to 100; lowest values ​​Japan 86, Spain 85, Korea 70);
  • the greatest difference between the first and fourth HISEI quartiles (also 111 points, followed by Belgium and Switzerland with 106 each; lowest values ​​Iceland 50, Korea 33, Japan 27);
  • the strongest HISEI gradient (45.1, followed by the Czech Republic 42.9 and Hungary 40.0; lowest values ​​again Iceland 18.7, Korea 15.3, Japan 9.2);
  • the second largest correlation coefficient (0.41, behind Hungary 0.43);
  • the second largest underperformance of students from immigrant families who do not speak the test language at home (over 110 points, just behind Belgium). So the test performance in Germany was more strongly correlated than anywhere else with the socio-economic level of the parents' job. This bad news was received by the public as a main result of PISA and has stuck in the collective memory, although the subsequent rounds produced significantly more favorable results.

However, an above-average strong correlation should not lead to an above-average causal relationship. Rather, the above-average correlation between test performance and migration status suggests that social, cultural and linguistic factors make a decisive contribution to the strong variance in German test performance. The PISA evaluators therefore construct elaborate path models . However, these models contain so many parameters that the results can no longer be meaningfully communicated in the form of country rankings.

Mathematics focus 2003

In 2003, Germany had the second largest difference between the first and fourth HISEI quartiles (102, after Belgium 108 and before Hungary 98); the other HISEI measures of association were not reported, as the analysis was mainly based on the ESCS index. In the ranking of the ESCS gradients, Germany (47) was at the upper edge of a broad midfield; the deviation from the OECD average (42) was assessed as not statistically significant (Hungary and Belgium were each 55 and Slovakia 53; at the end Mexico and Portugal were 29 each and Iceland 28). On the other hand, the correlation coefficient (standardized gradient, referred to as "explanation of variance"; 0.23; third largest value after Hungary 0.27 and Belgium 0.24) was classified as significantly above average.

In addition to the parental professional level, the highest parental educational qualification and the provision of cultural goods in the household also contribute to the ESCS. In this context, the German consortium regularly quotes Pierre Bourdieu and James Samuel Coleman , who coined the term “cultural capital”.

A broad distinction is made between three levels of educational qualifications: (1) no qualification up to secondary school leaving certificate, (2) completed apprenticeship, high school diploma or similar, (3) technical college degree, master craftsman examination, university degree. The OECD average difference in skills between levels (1) and (3) is 88 points. The difference is very small in Finland (42 points) and Portugal (44 points). It is very large in the Slovak Republic (144 points difference). In Germany, too, it is relatively high at 106 points.

Equipping the parents' household with cultural goods (works of art, classical literature, poems) is also positively correlated with the test performance. The OECD average difference in math proficiency between students in the quarter of families with the most cultural goods and students in the quarter of families with the fewest cultural goods is 66 points. In Germany, too, the difference is 66 points. The smallest differences are found in Iceland (34), Switzerland (35), Canada (42) and Finland (44). The largest in Hungary (86), Belgium (81), Denmark (81) and Sweden (81).

Science focus 2006

In 2006 the German consortium made its dissent with the international project management public and based its report exclusively on the HISEI, while the OECD continues to rely on the ESCS. Both the HISEI gradient of the German test performance and the associated correlation coefficient are at the upper end of a broad midfield and are not significantly different from the OECD mean; The Czech Republic, Luxembourg and France have unfavorable values. For the first time, a trend statement was also made: in reading , which was additionally tested in 2003 and 2006, the HISEI gradient has decreased from 45 to 35 and is now behind the Czech Republic, Luxembourg, Portugal, France, Belgium, Hungary and on a par with the Netherlands, Austria and Slovakia. The authors attach no importance to most of the fluctuations and explain them with changes in the sample utilization rate and missing values.

Results of Germany-specific evaluations

As soon as the evaluation goes beyond simple correlation measures, an international comparison, especially in the form of one-dimensional rankings, is no longer possible. Questions and models must be based on the specific circumstances in the individual countries. The majority of the data evaluation by the German consortium therefore concentrates on the German part of the international data set. In some cases, data from the German supplementary examinations ( PISA-E ) are also included.

Social origin and participation in education

The social disparity in educational participation is particularly evident in Germany when attending different types of school. In PISA 2006, as in PISA 2000, school attendance was broken down into EGP classes:

EGP Social class secondary schools secondary school high school Other
I. Upper class of service 9% 26% 52% 13%
II Lower class of service 13% 25% 41% 20%
III Routine services in trade and administration 20% 24% 30% 26%
IV Self-employed 23% 31% 23% 23%
V, VI Skilled workers and workers with management functions 24% 25% 21% 29%
VII Semi-skilled and unskilled workers, farm workers 28% 22% 14% 36%
total 19% 25% 31% 25%

The difference in school attendance essentially reflects a decision made at the end of primary school (usually after four school years); Changing schools during lower secondary level has negligible effects on the overall picture.

Primary and secondary disparities

For further analysis, “the separation of primary inequalities covered by performance and secondary inequalities solely caused by social class is of great interest”. Secondary inequality arises because the parents' school decision is not based solely on the child's school performance, but is also shaped by the motive of maintaining intergenerational status, by different expectations of success and by class-dependent cost-benefit relationships; The counseling behavior of the primary school teacher can also be influenced by such factors.

In order to assess how significant these secondary disparities are, school attendance is broken down by social class if the PISA test performance is recorded. Different performance characteristics are used as a basis: none at all (model I), problem-solving tasks from the German supplementary test (model II) or additionally the international reading tasks (model III).

The results are reported as odds ratios . For PISA 2006, the odds for attending grammar school versus secondary school are:

EGP Social class Model I. Model II Model III
I. Upper class of service 2.7 2.5 2.2
II Lower class of service 2.1 2.1 1.9
III Routine services in trade and administration 1.6 1.4 1.3
V, VI Skilled workers and workers with management functions 1 1 1
VII Semi-skilled and unskilled workers, farm workers 0.7 0.8 0.8

In principle, the figures for model I result directly from the percentage school attendance quotas given above. An example: A young person from EGP class I has the chance 52: 26 = 2.0 to attend a grammar school instead of a secondary school. For a young person from class V / VI, however, the odds are 21:25 = 0.84. This gives an odds ratio of 2.0 / 0.84 = 2.4. This means: for a child in a managerial position, the chance of attending a grammar school instead of a secondary school is 2.4 times greater than for a child of a skilled worker.

However, the numerical value 2.4 explained in this example does not exactly match the table value 2.7. The difference is explained by the fact that all odds ratios were calculated without special and vocational school students. This shows how sensitively odds ratios can depend on the details of the sample definition. The evaluation using odds ratios is also problematic insofar as this concept is often confused with probability ratios or relative risk in the abbreviated reproduction of study results in the press .

Compared to PISA 2000, the odds ratios have decreased significantly; the highest value at that time was still 4.2. It is unclear whether the secular trend towards equalization of educational opportunities becomes apparent in this short period of time.

Test performance according to school type

Test performance at different school types (measured in "competence points")
type of school Very "low" social background "Low" social background "High" social background Very "high" social background
secondary schools 400 429 436 450
Integr. comprehensive school 438 469 489 515
secondary school 482 504 528 526
high school 578 581 587 602

It was found that the type of school attended has a major influence on skills. The students acquired the greatest competencies at the grammar school, the lowest at the secondary school. Comprehensive and secondary schools are in the middle. Statistically, at grammar school, the acquisition of skills is least linked to social origin.

Influence of individual socio-demographic characteristics

Where do immigrant students have the greatest chance of success?

With the special study Where Immigrant Students Succeed - a comparative Review of Performance and Engagement from PISA 2003 (German title: Where do students with a migration background have the greatest chances of success? - A comparative analysis of performance and engagement in PISA 2003 ) it was determined whether immigrant children in the school system are just as successful as autochthonous students.

A first result was that there is no decisive connection between the number of immigrant students in the sample countries on the one hand and the extent of the differences in performance between immigrant children and native students on the other. This refutes the assumption that a high level of immigration has a negative impact on integration.

In the country comparison of this study, Germany brings up the rear when it comes to the integration of second-generation migrant children. Although the study confirmed that the migrant children were willing to learn and had a positive attitude, their chances of success in the German education system are lower than in any of the other 17 countries examined:

  • On average, migrant children lag behind native children by 48 points; in Germany, however, by 70 points. The differences are greatest in the natural sciences and smallest in reading skills.
  • While in almost all the other participating countries the migrant children achieve higher performance points in the second generation, in Germany they drop again extremely: migrant children of the second generation are behind their classmates about two years ago. More than 40% of them do not have the basic knowledge of level 2 in mathematics and do similarly poorly in reading skills.

More detailed studies based on the PISA 2000 study show that, as a result, it is not origin as such, but rather, in addition to the language spoken at home, the educational level of the parents, especially the mother, that determines educational success - a relationship that also applies to the local population was detected.

Credit points in mathematics for 15-year-old students
Students without a migration background First generation students * Second Generation Students **
OECD average 523 475 483
Germany 525 454 432
* Born abroad, foreign parents - ** Born in the survey country, foreign parents

According to this table, the fact that young people of foreign origin who immigrated themselves (“1st generation”) generally achieve better results than young people of foreign origin (“2nd generation”) would be a statistical fallacy. This is because the families of the students of foreign origin born in Germany (“2nd generation”, 432 points in the table) come mostly from Turkey, and migrants of Turkish origin come off particularly poorly in PISA. Among the young people who immigrated themselves (“1st generation”, 454 points), young people from ethnic German families are more represented. These are usually more successful at school. So one cannot say that the situation in Germany has deteriorated over the generations. On the contrary: within the individual groups of origin, the educational situation seems to be improving from generation to generation.

For each individual country of origin, young people born in Germany of foreign origin (“2nd generation”) achieve better results than young people born abroad. This is exemplified in the case of young people from the former Yugoslavia and Turkey in the field of mathematics. It applies in a similar way to other groups of origin and the areas of science and reading skills:

Family origin Migration status 1./2. generation Credit points mathematics
Former Yugoslavia Immigrated 1st generation 420
Former Yugoslavia born in Germany 2nd generation 472
Turkey Immigrated 1st generation 382
Turkey born in Germany 2nd generation 411

Reasons for the poorer performance of students with a migration background

It is possible that the poor performance of young people with a migrant background in PISA is a result of language-heavy test tasks. While Baumert and Schümer consider this explanation to be certain, Ramm, among others, come to the opposite conclusion.

In addition to the language problems, a large number of young people whose parents were born abroad have a low socio-economic status, which also leads to poorer educational results for children of parents born in Germany.

What should be said about attending grammar school or secondary school for young people with a migration background?

Young people with a migration background attend a grammar school or a Realschule less often than young people without a migration background. Among young people with a migration background, there is an educational participation that was found in around 1970 among young people without a migration background. Language seems to be primarily responsible for this. Baumert and Schümer come to the following conclusion in an analysis commissioned by the PISA consortium:

“Neither the social situation of the immigrant families nor the distance to the majority culture as such are primarily responsible for the disparities in educational participation. Rather, it is of crucial importance to have a command of the German language at a level appropriate to the course of study. For children from immigrant families, language skills are the decisive hurdle in their educational careers. With the same reading skills, children from immigrant families tend to make more use of the transition to a medium or higher level of education than those of the same age who come from German-speaking families "

Is the lower equality of opportunity in West Germany due to the large number of immigrants who are poorly educated?

This question must be answered in the negative:

The really surprising result of the analyzes is the clearly recognizable finding that the secondary social inequalities among the 15-year-olds without a migration background are not less, but tend to be greater than for the total cohort. So there can be no question of the problems of social equity in the narrower sense being a side effect of the immigration of socially disadvantaged groups of the population. [...] Lehmann, Peek and Gänsefuß (1997) reported a similar result for the first time from the Hamburg study on the initial learning situation. This means [...] that the east-west divide [...] turns out to be even steeper if only young people without a migration background are considered .

What influence does the family structure have on the PISA results? (Area Mathematics Competence)

In all OECD countries, young people living in nuclear families achieve higher mean competency scores in mathematics than young people living with single mothers or fathers. The difference is greatest in the USA. Here, young people from core families have a lead of 51 competence points. In Austria, their lead is the smallest with only 5 points. In Germany, too, the lead is small at only 11 points. Children from core families achieve 515 competence points, children of single parents 504 competence points.

Several reasons for the correlation are conceivable. So that children can grow up healthy, it is important that they are integrated into a social network and have caregivers. In the opinion of many scientists, families can do this better than single parents. Single parents often have fewer time resources, which can have an impact on performance development. It is also possible that in some countries the average level of education and origin is lower among single mothers. B. could explain the big differences between the USA and Germany.

In Germany, 16.7% of young people live with a single parent.

How does one parent's unemployment affect PISA scores? (Area Mathematics Competence)

Unemployment is an economic and psychological burden that can negatively affect the family. This is especially the case when the father is unemployed.

In Germany, 81.8% of PISA fathers were employed full-time, 7.6% were employed part-time and 5.5% were looking for work.

In all OECD countries, children with a full-time father had the highest proficiency scores in mathematics. The youth with a job-seeking father had the lowest. In the OECD average, the former have a lead of 46 points. In Germany, too, the difference in skills is 46 points. Pupils with a full-time employed father achieve 552 competence points, pupils with a part-time employed father 478 competence points and pupils with a job-seeking father 476 competence points.

Comments on the influence of social background in PISA

Comment of the Scientific Advisory Board for Family Issues of the Federal Ministry for Family, Seniors, Women and Youth

The Advisory Board commented as follows:

“ The PISA study impressively confirms the numerous social science findings on social disparities in children's educational success. [...] Competence development and educational success depend heavily on the type of school attended and the associated differential learning opportunities. The effect of the children's class on their test performance is significantly reduced when one takes into account the educational path they have taken. In the course of further school attendance, however, the differences in educational performance between pupils of the same age in different educational programs increase. One must therefore note a cumulative increase in social disparities in the educational success of children. Class-specific disparities in children's educational development before and during elementary school, but also class-specific decision-making behavior on the part of parents - something that was not examined in the PISA study - lead to different educational opportunities for children in the different educational programs, and this leads to further divergence the acquisition of skills and educational performance. The fact that the choice of course has to be made at an early stage, in almost all federal states after the fourth school year, exacerbates the situation. [...] A mechanism is set in motion very early on which, regardless of the social disparities that already exist, reinforces them based on the decision made. "

Comment from Maria Böhmer, Federal Government Commissioner for Migration, Refugees and Integration

Maria Böhmer commented:

“ It is regrettable that students from immigrant families do not yet participate in the PISA success. The educational success must not depend on the social origin. I especially call on the federal states and also the migrant associations to quickly implement the voluntary commitments they made in the National Integration Plan [...] We also need more teachers from immigrant families and we have to significantly reduce the repetition and dropout rate. In the National Integration Plan, the federal states and municipalities have committed themselves to do this within the next five years. The National Integration Plan is a plan for more educational opportunities and against a lack of prospects. [...] We have to strengthen parents so that they can fully live up to their educational responsibilities. "

Compared to other countries and educational traditions

In Germany, free access to universities has long been viewed as an essential prerequisite for social mobility, but the decisive prerequisite is - as has long been known in the Anglo-Saxon countries - the quality of early (pre) school education. The relationship between the previous education of the parents and the selected training is clearly higher in Germany than in other industrialized countries, including the USA. In the PR China, on the other hand, in Shanghai in particular, the often-cited connection between poverty and lack of (university) access has now been almost completely decoupled.

References

Literature cited

  • Artelt et al.: PISA 2000: Summary of central findings. Max Planck Institute for Human Development, Berlin 2001. (online) (PDF; 862 kB)
  • Baumert et al. (Hrsg.): PISA 2000. Basic competencies of schoolchildren in an international comparison. Leske + Budrich, Opladen 2001.
included:
  • Baumert, Schümer: Family living conditions, participation in education and acquisition of skills. Chapter 8, pp. 323-407.
  • Baumert et al. (Ed.): PISA 2000 - The countries of the Federal Republic of Germany in comparison. Leske + Budrich, Opladen 2002, ISBN 3-8100-3663-3 .
  • Baumert, Stanat, Watermann (eds.): Disparities in education based on origin. In-depth analyzes as part of PISA 2000. VS Verlag für Sozialwissenschaften, Wiesbaden 2006.
  • Bonnet: Reflections in a Critical Eye: on the pitfalls of international assessment. In: Assessment in Educ. 9 (3) 2002, pp. 387-399.
  • Ericson, Goldthorpe, Portocarero: Intergenerational class mobility in three Western European societies: England, France and Sweden. In: Brit. J. Sociology. 30, 1979, pp. 341-415.
  • Esser: Integration and Ethnic Stratification . Working Papers - Mannheim Center for European Social Research 40th MZES, Mannheim 2001.
  • Ganzeboom, De Graaf, Treiman: A Standard International Socio-Economic Index of Occupational Status. In: Soc. Sci. Res. 21 (1) 1992, pp. 1-56.
  • Kristen: Hauptschule, Realschule or Gymnasium? Ethnic differences at the first educational transition. In: Cologne journal for sociology and social psychology. 54, 2002, pp. 534-552.
  • OECD: Knowledge and Skills for Life. First Results from the OECD Program for International Student Assessment (PISA) 2000. OECD, Paris 2001.
  • OECD: Learning for Tomorrow's World. First Results from PISA 2003. OECD, Paris 2004.
  • OECD: PISA 2003 Technical Report. OECD, Paris 2005.
  • OECD: PISA 2006. Science Competencies for Tomorrows World. OECD, Paris 2007.
  • OECD: Where Immigrant Students Succeed - a comparative Review of Performance and Engagement from PISA 2003 ( PDF; 4 MB )
  • Prenzel et al. (PISA-Konsortium Deutschland, Hrsg.): PISA 2003: The educational level of young people in Germany - results of the second international comparison. Waxmann, Münster 2004, ISBN 3-8309-1455-5 .
  • Prenzel et al. (PISA-Konsortium Deutschland, ed.): PISA 2006: The results of the third international comparative study. Waxmann, Münster 2007, ISBN 978-3-8309-1900-1 .
included:
  • Ehmke, Baumert: Social origin - family living conditions and acquisition of skills. Chapter 7.1, pp. 309-335
  • Puchhammer: Language-Based Item Analysis. In: Hopmann, Brinek, Retzl (eds.): PISA According to PISA - PISA According to PISA. LIT-Verlag, Vienna 2007, ISBN 978-3-8258-0946-1 , pp. 127-137.
  • Scientific Advisory Board for Family Issues: The educational significance of the family - conclusions from the PISA study. Berlin 2002, ISBN 3-17-017927-6 . (Volume 224 - Series of publications by the Federal Ministry for Family, Seniors, Women and Youth.)
  • Wuttke: The insignificance of significant differences: PISA's claim to accuracy is illusory. In: Jahnke, Meyerhöfer (Ed.): PISA & Co - Critique of a Program. 2., ext. Edition. Franzbecker, Hildesheim 2007, ISBN 978-3-88120-464-4 .

Footnotes and individual references

  1. ↑ OECD 2001, 2004, 2007
  2. ↑ While the international reports speak cautiously of "performance", the test results in the German reports are simply referred to as "competence" or even "competence acquisition".
  3. ↑ This index comes from a meta-study by Ganzeboom et al. (1992). Wuttke (2007) points out that the ISEI was expressly designed only for men, but that PISA is also applied to the professions of mothers, that Ganzeboom falls back on outdated sources from the 1960s, that the correlation with the most recent Allensbach professional prestige index is only 0.06, and that several ratings are obviously absurd: power plant operator far below electricity reader, musical instrument maker far below dentist receptionist, conductor far below dancer, manager far below political scientist, member of parliament far below army officer.
  4. ↑ OECD 2001, Appendix B1, p. 283, Table 6.1a
  5. ↑ Artelt et al. 2001, pp. 40f.
  6. ↑ In truth, PISA uses a probabilistic model of student behavior and therefore does not assign one competence value to each subject , but rather five. These various estimates of “plausible” personal parameters are only averaged at the end of each evaluation. See methodology of the PISA studies .
  7. ↑ OECD 2001, p. 14.
  8. ↑ Prenzel et al. 2004, p. 48, p. 64.
  9. ↑ Wuttke 2007, p. 181ff.
  10. ↑ Puchhammer 2007, p. 132.
  11. ↑ OECD 2005, pp. 316f.
  12. ↑ Baumert, Schümer 2002, p. 328, and the same in the subsequent reports
  13. ↑ Baumert, Stanat, Watermann 2006, p. 9.
  14. ↑ Ehmke, Baumert 2007, p. 314.
  15. ↑ On the other hand, in the Zeitschrift für Erziehungswissenschaften (8 (4) 2007, pp. 521–540), Ehmke assessed the ESCS as a “valid and theoretically comprehensive index”; on this contradiction see also: Thorsten Stegemann: Dispute about the Pisa study. on: heise.de December 4, 2007.
  16. ↑ Baumert, Schümer 2002, p. 328 and p. 338f.
  17. ↑ Baumert, Schümer 2002, p. 338.
  18. ↑ Ehmke, Baumert 2006, p. 324.
  19. ↑ If not a single one of the thousands of students leaves a specific question unanswered, as was the case in Poland in 2003, then there is a suspicion of manipulation (Wuttke 2007, p. 125.)
  20. ↑ Hagemeister (in Jahnke, Meyerhöfer: PISA & Co - Critique of a Program. 1st Edition. 2006, p. 269) shows an example of how country rankings change when incomplete data sets are excluded from the analysis.
  21. ↑ very briefly described in OECD 2005, p. 316.
  22. ↑ Wuttke 2007, pp. 189ff.
  23. ^ Bonnet 2002
  24. ↑ OECD 2001, p. 221.
  25. ↑ Maaz, Kreuter, Watermann, in Baumert, Stanat, Watermann 2006, pp. 31–59.
  26. ↑ Maaz, Kreuter, Watermann, in Baumert, Stanat, Watermann 2006, p. 55.
  27. ↑ Baumert, Schlümer 2001, p. 381.
  28. ↑ Eckert ( Relative chances, risks and odds ratios for the description of educational participation. In: Empirische Pädagogik. 20 (1) 2006, pp. 91–97) points out that this labeling, intended as reading aid, is inappropriately indicated in the subsequent reports ESCS quartile has been applied.
  29. ↑ Baumert et al. 2001, pp. 107, 384f., 395
  30. ↑ z. B. Baumert, Stanat, Watermann, pp. 244f.
  31. ↑ Ehmke et al. 2004, pp. 236, 249.
  32. ↑ Ehmke et al. 2004, p. 233.
  33. ↑ Ehmke, Baumert 2007, pp. 318, 321, 323.
  34. ↑ Ehmke, Baumert 2007, p. 329.
  35. ↑ Baumert, Schümer 2001, p. 355.
  36. ↑ Baumert, Schümer 2002, p. 167 f.
  37. ↑ Baumert, Schümer 2001, p. 354.
  38. ↑ This is a conservative estimate insofar as the performance of students from different types of schools diverges over the years and the disparity at the time of the decision on the type of school tends to be underestimated if one takes performance measured at the age of 15 as a basis. (Baumert, Schümer 2001, p. 359.)
  39. ↑ Ehmke, Baumert 2007, p. 330.
  40. ↑ Data for class IV were classified as “not significant” and not reported.
  41. ↑ Email from Dr. T. Ehmke, IPN Kiel, to user Ms. Holle on January 21, 2008.
  42. ↑ Eckert in Empir. Päd. 20 (1) 2006, pp. 91-97.
  43. ↑ Around 1950 a similar odds ratio was 36 (Schimpl-Neimanns in Kölner Zs. Soz. Soz.psych. 52 (4) 2000, pp. 636–669)
  44. ↑ Ehmke et al. (2004), p. 244.
  45. ↑ a b c d Ramm et al: Socio-cultural origin: Migration. In: PISA-Konsortium Deutschland: PISA 2003: The educational level of young people in Germany - results of the second international comparison. Waxmann, Münster 2004, ISBN 3-8309-1455-5 .
  46. ↑ Esser 2001; Kristen 2002
  47. ↑ cf. New findings from the PISA study ( Memento of the original from April 1, 2008 in the Internet Archive ) Info: The archive link was inserted automatically and has not yet been checked. Please check the original and archive link according to the instructions and then remove this notice. , isoplan, May 30, 2003, with reference to a study by the Rheinisch-Westfälisches Institut für Wirtschaftsforschung; see also Michael Ready: Who's to Blame? The Determinants of German Students' Achievement in the PISA 2000 Study . In: RWI Discussion Papers No. 4; IZA Discussion Papers No. 739 . Rheinisch-Westfälisches Institut for Economic Research; IZA Institute of Labor Economics, 2003, ISBN 3-936454-04-3 , ISSN 1612-3565 (English, papers.ssrn.com [accessed August 28, 2019]). @1@ 2Template: Webachiv / IABot / www.isoplan.de 
  48. ↑ Pisa study: Migrants are hit hardest. to: spiegel.de December 6, 2007, accessed May 6, 2011.
  49. ↑ Ramm et al., P. 268
  50. ↑ Baumert, Schümer: Family living conditions, participation in education and acquisition of skills in a national comparison. In: German PISA Consortium (Hrsg.): PISA 2000 - The countries of the Federal Republic of Germany in comparison. P. 199.
  51. ↑ Ramm et al., Pp. 269-270.
  52. ↑ Ramm et al., P. 272.
  53. ↑ cf. Baumert, Schümer: Family living conditions, participation in education and acquisition of skills in a national comparison. In: German PISA Consortium (Hrsg.): PISA 2000 - The countries of the Federal Republic of Germany in comparison. P. 199.
  54. ↑ Baumert et al. 2002, pp. 171f.
  55. ↑ Baumert, Schümer, 2001, 2002; OECD, 2004; Schneewind and Pekrun, 1994
  56. ↑ Ehmke et al. 2004, p. 228.
  57. ^ Betram, 2004
  58. ↑ a b Ehmke et al. 2004, p. 230.
  59. ↑ Scientific Advisory Board f. Family questions 2002, pp. 29–30.
  60. ↑ Press releases  ( page no longer available , search in web archives )  Info: The link was automatically marked as defective. Please check the link according to the instructions and then remove this notice. Downloaded January 5, 2008@1@ 2Template: Toter Link / archiv.bundesregierung.de  
  61. ^ Rosenbaum, James E. The Hidden Curriculum of High School Tracking. New York: John Wiley & Sons, 1976.
  62. ↑ Chris Cook: Shanghai tops global state school rankings. In: Financial Times . December 7, 2010, accessed June 28, 2012 .

Additional information

Web links