Knowledge and Skills for Life

OR

GA

NI

SA

TI

ON

F

OR

E

CO

NO

MI

C

CO

-O

PE

RA

TI

ON

A

ND

D

EV

EL

OP

ME

NT

EXECUTIVE SUMMARY

Programme for International Student Assessment

EDUCATION AND SKILLS

Knowledgeand Skills for LifeFIRST RESULTSFROM PISA 2000

What is PISA? (Programme for International Student Assessment)

A unique collaboration betweencountries to monitor educationaloutcomes.

PISA was co-ordinated bygovernments of participatingcountries, through theOrganisation for Economic Co-operation and Development(OECD).

Leading international expertsworked to develop an assessmentwhose results are comparableacross different national andcultural contexts.

PISA improves internationalinformation on student outcomes,giving countries benchmarks andregular updates on how studentsperform against them.

A new three-yearly survey of theknowledge and skills of 15-year-olds in the principal industrialisedcountries.

The survey, conducted first in 2000,will be repeated every three years.

265,000 students from 32 countriestook part.

Students sat pencil and paperassessments in their schools.

Students and their principals alsoanswered questionnaires aboutthemselves and their schools. Thisallows PISA to identify what factorsare associated with better and worseperformance.

The countries taking partIn 2000, 28 OECD Member countries and four other countries carried out the first PISA survey. A further 13 countries willconduct the same survey in 2002, and the two OECD countries that did not take part in 2000 will participate in the secondsurvey in 2003.

2

Countries reported on here (assessed in 2000):

Countries conducting the same survey in 2002:Albania, Argentina, Bulgaria, Chile, China, Hong Kong -Special Administrative Region of China, Indonesia, Israel,Lithuania, Macedonia, Peru, Romania, Thailand.

A new way of looking atstudent performance.

PISA assessed young people’scapacity to use their knowledgeand skills in order to meet real-life challenges, rather thanmerely looking at how well theyhad mastered a specific schoolcurriculum.

PISA assessed literacy in reading,mathematics and science.

Students had to understand keyconcepts, to master certainprocesses and to apply knowledgeand skills in different situations.

Information was also collectedon student attitudes andapproaches to learning.

OECD members: Australia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Japan, Korea, Luxembourg, Mexico, Netherlands, New Zealand, Norway, Poland, Portugal, Spain, Sweden, Switzerland, United Kingdom, United States.

Non-OECD members: Brazil, Latvia, Liechtenstein, Russian Federation.

OECD countries joining the survey in 2003:

Slovak Republic, Turkey.

Colour key:

Results and findings - What PISA 2000 tells us…

This brochure summarises the results ofPISA 2000 in terms of studentperformance, and wider findings aboutwhat lies behind these results. The fullreport, Knowledge and skills for life –First results from PISA 2000, can beobtained from the OECD (available inEnglish, French and German) andfurther information can be found atwww.pisa.oecd.org. See also back cover.

…about student literacies…

PISA 2000 assessed students’ capacitiesto apply knowledge and skills inreading, mathematics and science.These capacities are referred to asreading, mathematical and scientificliteracy and widely seen as essentialprerequisites for students to be wellprepared for adult life.

PISA does not measure literacy as an“all or nothing” set of knowledge andskills. Rather, each student receives ascore on a continuous scale. The scoreobtained by a student indicates themost difficult type of task that thestudent is likely to perform correctly.Each PISA task is associated with ascale score.

The PISA scale for each literacy areawas devised so that across OECDcountries, the average score is 500points, and around two-thirds ofstudents achieve between 400 and 600points.

For reading, five levels of literacy aredescribed, Level 5 being the highest.

The results for reading literacy aresummarised on pages 4-9.

The results for mathematical literacyare summarised on pages 10-11.

The results for scientific literacy aresummarised on pages 12-13.

…about what students are like aslearners…

Today’s young people will need to goon learning in different waysthroughout their lives. PISA looked atsome aspects of what they are like aslearners at age 15. It considered theirmotivation and engagement in learning,and aspects of their learning strategies.

The findings on what students are likeas learners are summarised on pages14-15.

…about gender differences inperformance and engagement…

PISA’s results show differences inperformance between males andfemales. They also show differences intheir engagement in school and theirlearning strategies.

The findings on gender differences aresummarised on pages 16-17.

…about the relationship betweenstudent performance and familybackground…

Students with more advantagedfamily backgrounds tend to performbetter educationally. PISA allows thestrength of this link to be looked atmore closely, showing how it differsbetween school programmes andbetween countries.

The findings on family background are summarised on pages 18-19.

…about the relationship betweenschool differences and familybackground…

Can schools help to moderate theeffect of family background? Theanswer depends partly on the degree towhich students with differentcharacteristics attend different schools,and on how much the results of theseschools differ.

The findings on school differences are summarised on pages 20-21.

…and about the characteristics ofschools where students do well

Schools differ in terms of resources,policies and classroom practices. PISA2000 found that students do better, onaverage across countries, in schoolswith certain characteristics.

The findings on school effectsare summarised on pages 22-23.

3

How students perform in reading literacy

Students taking part in PISA wereasked questions based on a variety ofwritten texts, ranging from a short storyto a letter on the Internet andinformation presented in a diagram.They were assessed on their capacity toretrieve specified information, onwhether they could interpret what theyread, and on how well they could reflecton and evaluate it, drawing on theirexisting knowledge. For each of thesethree aspects of reading literacy,students were given a score based on thedifficulty of the tasks that they couldperform. A combined score shows theiroverall reading performance.

On the basis of these scores, eachstudent was assigned to one of fivereading levels (see legend on facingpage). Examples of tasks at the differentlevels are given in the following pages.

The figure (below) shows thepercentage of students who areproficient at each level in thecombined OECD area and in each

Percentage of students by highestlevel of reading proficiency

100

80

60

40

20

0

Countries are ranked by percentage of students at least at Level 3 * Non - OECD country

2. In

divd

ual c

ount

ries

1. C

ombi

ned

OEC

D a

rea

OECD av

erage

country. Here, countries are ranked bythe percentage of students who areproficient at Level 3 or above. About60% of 15-year-olds in the combinedOECD area can complete such tasks(see Page 7 for examples). But thisfraction varies from above three-quarters in Finland and Korea, tobelow half in Brazil, Latvia,Luxembourg, Mexico, Portugal and theRussian Federation.

No single indicator can adequatelysummarise the performance of studentsin different countries. The figures onthe following pages show different partsof the distribution of studentperformance.

In general, within countries similarpercentages of students are proficientat each level in the different aspects ofreading literacy – retrieving,interpreting, and reflection andevaluation. However, in some countriessignificant differences exist betweenstudent performance in the more“routine” reading tasks of identifyinginformation and interpreting it, andtasks requiring reflection andevaluation (see Tables 2.1b, c and d inthe full report).

4

One in ten students completed PISA’s hardest reading tasks....

Students proficient at Level 5 on thecombined reading literacy scale arecapable of completing sophisticatedreading tasks, such as:

managing information that isdifficult to find in unfamiliartexts;

showing detailed understanding ofsuch texts and inferring whichinformation in the text is relevantto the task;

evaluating critically and building

Percentage of students at Level 5

19 18 18 17 16 14 12 12 11 11 10 9 9 9 9 9 8 8 7 6 6 5 5 5 5 4 4 4 3 2 1 1

* Non - OECD country

Rank

ed b

y pe

rcen

tage

of s

tude

nts

at L

evel

5

100

80

60

40

20

0

OECD av

erage

For exampleStudents were shown a tree diagram of a country’s working-age population, anddescriptions of the labour force status of individual workers. They had to decide inwhich category of the diagram each worker belonged. They had to work out whatcriteria to use to classify workers from the structure and content of the diagram,drawing on information in footnotes and therefore not prominent. This task isassociated with a score of 727 points on interpreting scale.

Students were shown a notice from a personnel department about a service thatwould help with job mobility. They had to work out the two ways in which thisservice could help people who lost their jobs – information that was stated indirectlyand had to be distinguished from competing information that could easily be mistakenfor the information required. Associated score: 655 points on the retrievinginformation scale.

After reading a three-page story about a woman’s adventure, students were asked tosay whether they thought it had an appropriate ending, explaining why. To obtainfull credit, they had to evaluate the ending in terms of its thematic completeness, byrelating the last sentence to central relationships, issues or metaphors in the story.Associated score: 652 points on reflection and evaluation scale.

For the full sample items see www.pisa.oecd.org.

hypotheses, drawing on specialisedknowledge, and accommodatingconcepts that may be contrary toexpectations.

Only 10% of 15-year-olds in thecombined OECD area are proficient at Level 5 in OECD countries. Thepercentage ranges from over 15 percent in Australia, Canada, Finland,New Zealand and the United Kingdom,to below 5% in Brazil, Latvia,Luxembourg, Mexico, Portugal, Spainand the Russian Federation (see Table 2.1a in the full report).

727

655

652

5

Reading literacy, as defined in PISA,focuses on the knowledge and skillsrequired to apply “reading for learning”rather than on the technical skillsacquired in “learning to read”. In linewith most contemporary views aboutreading, PISA focuses on measuringthe extent to which individuals areable to construct, expand and reflecton the meaning of what they have readin a wide range of texts common bothwithin and outside school. Thesimplest reading tasks that can still beassociated with this notion of readingliteracy are those at Level 1. Studentsproficient at this level are capable ofcompleting only the least complexreading tasks developed for PISA, suchas locating a single piece of

...One student in six could at most complete the simplest tasks...

Percentage of students at Level 1 or below


Rank

ed b

y pe

rcen

tage

of s

tude

nts

at L

evel

1 o

r bel

ow

100

80

60

40

20

0

OECD av

erage

below 335 points

408 to 480 points

For exampleStudents were shown a magazine article, written for young people, explaining theprocess and purpose of DNA testing. A multiple-choice task required them simply torecognise that the writer’s main purpose was to inform, rather than to warn, amuseor convince. This task is associated with a score of 406 points on interpreting scale.

Students were asked about a sentence in an article about sports shoes, discussingaspects of the shoe that can avoid damage to feet. A multiple-choice task requiredstudents to recognise the relationship between the two parts of the sentence: that thesecond part provided the solution to the problem stated in the first part (rather thanillustrating, repeating or contradicting it). Associated score: 402 points on thereflection and evaluation scale

After reading a short adventure story, students were asked in a multiple-choice taskwhat happened next after a brief quoted extract. The answer was stated explicitly inthe narrative and was easy to locate from the information given in the task.Associated score: 367 points on the retrieving information scale.


information, identifying the maintheme of a text or making a simpleconnection with everyday knowledge.

On average across OECD countries,12% of 15-year-olds are proficient atthis level but no higher. A further 6%could not perform these simple tasks.This does not mean that these studentscannot read in a technical sense.Nonetheless, such students haveserious difficulties in using readingliteracy as an effective tool to advanceand extend their knowledge and skillsin other areas (see Table 2.1a in thefull report).

In total, just over one student in six(18%) does not progress beyond Level

1 tasks. However, this percentagevaries considerably between countriesand it is noteworthy that everycountry has some students who couldperform only at this level. Parents,educators and policy-makers incountries where this proportion is highneed to recognise that significantnumbers of students are not benefitingsufficiently from educationalopportunities and may not beacquiring the knowledge and skills todo so in their further school careersand beyond. In only four countries,Canada, Finland, Japan and Korea, do10% or less of 15-year-olds read atLevel 1 or below. In three OECDcountries, Luxembourg, Mexico andPortugal, more than 25% do so.

406

402

367

6

...most students have neither very high nor very low reading skills

On average across OECD countries,nearly three-quarters of 15-year-olds areat Levels 2, 3 or 4. But in some countriesthe proportion falls to just over half andin others it rises to above eight in tenstudents (see Table 2.1a in the fullreport).

Students proficient at Level 4 arecapable of solving complex readingtasks, such as locating embeddedinformation, construing meaningfrom nuances of language andcritically evaluating a text.

Students proficient at Level 3 arecapable of solving reading tasks of

100

80

60

40

20

0

Percentage of students at Levels 2, 3 and 4


Rank

ed b

y pe

rcen

tage

of s

tude

nts

at L

evel

s 2, 3

and

4

OECD av

erage

For exampleStudents had to look at two letters posted on the Internet giving conflicting opinionsabout graffiti. They had to comment on which was written better, analysing writingstyle and structure of argument rather than just what the letters said. To do this,they had to draw on their understanding of what constitutes good writing. This taskis associated with a score of 581 points on the reflection and evaluation scale.

Students had to use information in two diagrams, one about historic water levels inLake Chad, the other about when various animals appear in cave paintings nearby,in order to recognise that certain animals disappeared after a period of falling waterlevels. Students had to combine information presented in two different ways.Associated score: 508 points on the interpreting scale.

After reading a short extract from a play by Jean Anouilh, students had to work outwhat the play is about: one character is playing a trick on another. A multiple-choice task asked about the purpose of the trick. This required a low level ofinference to work out the main idea in the text. Associated score: 423 points on theinterpreting scale.


moderate complexity, such as locatingmultiple pieces of information,making links between different partsof a text, and relating it to familiareveryday knowledge.

Students proficient at Level 2 arecapable of solving basic reading tasks,such as locating straightforwardinformation, making low-levelinferences of various types, workingout what a well-defined part of a textmeans, and using some outsideknowledge to understand it.

The figure below shows countries rankedby the percentage of students who are at

these middle levels of proficiency.Leaving aside those countries with morethan a quarter of students performingbelow Level 2 (Brazil, Latvia,Luxembourg, Mexico, Portugal and theRussian Federation), this gives someindication of how equally performancein each country is distributed. In Korea,as an illustration, 89% of 15-year-oldsare at the middle levels, with few veryhigh or very low performers. By contrast,in New Zealand only two-thirds are inthe middle, with the remaining thirdsplit between students at Level 5 (threetimes as many as in Korea) and those atLevel 1 or below (over twice as many asin Korea).

581

508

423

7

Mean reading literacy scores

The figure (below) shows the meanscore on PISA’s reading literacy scalein each country. Since PISA is asample survey, the resulting estimatesare associated with some level ofuncertainty. The bar around the meanshows the range within which themean lies with 95% confidence. Fordetails on the comparison of meanscores between countries, see Figure 2.4in the full report.

550

530

510

490

470

450

430

410

390

Mean reading literacy scores: country similarities and differences

Points

550

530

510

490

470

450

430

410

390

Points

The estimate of each country mean is shown by the middle line

The range in which the mean can be said to fall, with 95% certainty, is shown by the bar

OECD average = 500


OECD av

erag

e

8

Mean Reading Literacy

Range of rank order positionsfor each country based on

sample (with 95% confidence)

Country RankHighest Lowestpossible possible

Finland 1 1Canada 2 4New Zealand 2 8Australia 2 9Ireland 3 9Korea 4 9United Kingdom 5 9Japan 3 10Sweden 9 11Austria 11 16Belgium 11 16Iceland 11 15Norway 11 16France 11 16United States 10 20Denmark 16 19Switzerland 16 21Spain 17 21Czech Rep. 17 21Italy 19 24Germany 21 25Liechtenstein 20 26Hungary 21 26Poland 21 27Greece 23 28Portugal 24 28Russian Fed. 27 29Latvia 27 29Luxembourg 30 30Mexico 31 31Brazil 32 32

It is evident that because the meanscores of many countries are quitesimilar, the precise ranking of countriescannot always be determined: there isoverlap in the bars showing the rangein which each country’s mean scorecan be said with confidence to fall.The table (right) shows the range ofrank order positions within which eachcountry’s mean lies, with 95%probability.

One way to summarise theperformance of each country is tocompute the mean of student scores.The figure on the previous page showsthe mean reading score for eachcountry.

Some significant differences appear inthe mean performance of studentsacross countries, with 125 score pointsseparating the lowest from the highestperforming OECD country. To theextent that such differences arepredictive of students’ career paths,this raises questions about the futurecompetitiveness of countries with largenumbers of low performers (Table 2.3in the full report).

The difference in the meanperformance between some countries islarge, but the variation in studentperformance within countries isgenerally much larger. This can be seen

by considering the range of scoresachieved by the middle half of thepopulation: the gap between the 25thand the 75th student in a group of 100,ranked by performance. The greater thegap, the more unequal are the resultswithin a country.

In all countries, the range ofperformances among the middle 50% ofstudents exceeds the magnitude of oneproficiency level (73 score points) andin Belgium, Germany and New Zealandtwice this difference (OECD average1.8 proficiency levels) (see Table 2.3ain the full report).

PISA shows that high meanperformance and low disparities can gotogether. The three countries with thesmallest range in the middle half of thepopulation, Finland, Japan and Korea,are also among countries with thehighest mean performance in readingliteracy. By contrast, one of the threecountries with the highest internalvariation in performance, Germany,scores below the OECD average.

This shows that wide disparities areneither inevitable nor a preconditionfor countries to attain a high overallperformance. On the contrary, theresults suggest that public policy canpromote equal opportunities andequitable learning outcomes for allstudents.

9

Mean mathematical literacy scores: country similarities and differences

550

530

510

490

470

450

430

410

390

370

350

330

550

530

510

490

470

450

430

410

390

370

350

330



Points Points

OECD average = 500


OECD av

erag

e

How students perform in mathematical literacy

10

The figure (below) shows the meanscore for each country on PISA’smathematical literacy scale. SincePISA is a sample survey, the resultingestimates are associated with somelevel of uncertainty. The bar aroundthe mean shows the range withinwhich the mean lies with 95%confidence. For details on thecomparison of mean scores betweencountries see Figure 3.2 in the fullreport.

Mean Mathematical Literacy




Japan 1 3Korea 2 3New Zealand 4 8Finland 4 7Australia 4 9Canada 5 8Switzerland 4 10United Kingdom 6 10Belgium 9 15France 10 15Austria 10 16Denmark 10 16Iceland 11 16Liechtenstein 9 18Sweden 13 17Ireland 16 19Norway 17 20Czech Rep. 17 20United States 16 23Germany 20 22Hungary 20 23Russian Fed. 21 25Spain 23 25Poland 23 26Latvia 25 28Italy 26 28Portugal 26 29Greece 27 30Luxembourg 29 30Mexico 31 31Brazil 32 32

It is evident that because the meanscores of many countries are quitesimilar, the precise ranking of countriescannot always be determined: there isoverlap in the bars showing the rangein which each country’s mean scorecan be said with confidence to fall.The table (right) shows the range ofrank order positions within which each country’s mean lies, with 95% probability.

Literacy in mathematics and science isimportant for understanding medical,economic, environmental and otherissues that shape modern societies,which rely heavily on technologicaland scientific advances. PISA 2000offers a snapshot of studentperformance in mathematical andscientific literacy but assessed them inless detail than reading literacy. Eachwas assessed on a single scale, withoutdistinct literacy levels, with an averagescore of 500 points as with reading.

PISA looked at mathematics inrelation to its wider uses in people’slives. Mathematical literacy in PISA ismeasured in terms of students’ capacityto:

recognise and interpretmathematical problems encountered ineveryday life;

translate these problems into amathematical context;

use mathematical knowledge andprocedures to solve problems;

interpret the results in terms of theoriginal problem;

reflect on the methods applied; and

formulate and communicate theoutcomes.

PISA mathematical tasks varied indifficulty according to several criteria,including:

the number and complexity ofcomputational steps involved;

the need to connect and integratematerial; and

the need to represent and interpretmaterial and reflect on situationsand methods.

15-year-olds in Japan display thehighest mean scores in mathematicalliteracy, but Japan’s mean performancecannot be distinguished with statistical

significance from that in Korea andNew Zealand. Other countries that alsoscore above the OECD average areAustralia, Austria, Belgium, Canada,Denmark, Finland, France, Iceland,Liechtenstein, Sweden, Switzerlandand the United Kingdom (see Figure 3.2 in the full report).

Examples:

A difficult PISA mathematicalliteracy task

Students were presented with a diagramshowing the pattern in which different treeswould have to be planted in an orchard inorder that conifers provide sufficientprotection to apple trees. They had to workout which type of tree would increase fasterin number as the orchard was enlarged –and explain why. This required them tonotice that the number of apple treesincreased in proportion to the square of thenumber of conifers. The task requiredstudents to think mathematically andrecognise a general principle. This task isassociated with a score of 723 points.

A medium PISA mathematicalliteracy task

From a mathematical representation of thedimensions and shape of a pyramidal barnroof, students had to calculate the area ofits base. This required students to identifyand perform a straightforward calculation,understanding the overall concepts of spaceand shape. Associated score: 492 points.

An easy PISA mathematical literacy task

From a graph showing the speed of aracing car as it travelled round a track,students had to answer a multiple-choicetask about where on the track the car wentthe slowest. This required only a simpleobservation and sufficient understandingof the concept of change to realise that theslowest speed would be shown at thelowest point of the plot on the speed graph.Associated score: 403 points.

11

For the full sample items seewww.pisa.oecd.org.

How students perform in scientific literacy

550

530

510

490

470

450

430

410

390

370

350

330

550

530

510

490

470

450

430

410

390

370

350

330



Mean scientific literacy scores: country similarities and differences

Points Points

OECD average = 500


OECD av

erag

e

The figure (below) shows the meanscore for each country on PISA’sscientific literacy scale. Since PISA is asample survey, the resulting estimatesare associated with some level ofuncertainty. The bar around the meanshows the range within which themean lies with 95% confidence. Fordetails on the comparison of meanscores between countries, see Figure 3.5in the full report.

12

Mean Scientific Literacy




Korea 1 2Japan 1 2Finland 3 4United Kingdom 3 7Canada 4 8New Zealand 4 8Australia 4 8Austria 8 10Ireland 9 12Sweden 9 13Czech Republic 10 13France 13 18Norway 13 18United States 11 21Hungary 13 21Iceland 14 20Belgium 13 21Switzerland 13 21Spain 16 22Germany 19 23Poland 19 25Denmark 21 25Italy 22 25Liechtenstein 20 26Greece 25 29Russian Fed. 26 29Latvia 25 29Portugal 26 29Luxembourg 30 30Mexico 31 31Brazil 32 32

It is evident that because the meanscores of many countries are quitesimilar, the precise ranking of countriescannot always be determined: there isoverlap in the bars showing the rangein which each country’s mean scorecan be said with confidence to fall.The table shows the range of rankorder positions within which each country’s mean lies, with 95% probability.

Examples

A difficult scientific literacy task

Students were shown extracts from a 19thcentury scientist’s diary, a table with hisobservations and a commentary,discussing the post-natal death from aparticular fever of a large proportion ofmothers in two wards of a hospitalmaternity clinic. Students had to indicatewhy the evidence did not support acontemporary belief that earthquakescaused the fever. This required them toexplain the significance of different deathrates in the two wards. This task isassociated with a score of 666 points.

A medium scientific literacy task

After reading a text on the risks to theozone layer and their implications,students were asked whether each of twoquestions could be answered by scientificresearch. To answer correctly, theyneeded, in particular, to recognise thedifference between a question that requiresa political choice and one that can beanswered scientifically. Associated score:529 points.

An easier scientific literacy task

Students were asked why washing hospitalsheets in high temperatures helps reducethe risk that patients will contract a fever.In their answer they needed to apply theirscientific knowledge to this real-worldproblem by referring, for example, to thekilling of bacteria. Associated score: 467points.

For the full sample items seewww.pisa.oecd.org.

Scientific literacy was scored on a scalemeasuring students’ capacity to:

use scientific knowledge;

recognise scientific questions;

identify what is involved inscientific investigations;

relate scientific data to claims andconclusions; and to

communicate these aspects ofscience.

PISA’s scientific tasks varied indifficulty according to several criteriaincluding:

the complexity of the concepts used;

the amount of data provided;

the chain of reasoning required; and

the precision required incommunication.

Japan and Korea show the highest per-formance on the scientific literacyscale. Other countries that score statis-tically significantly above the OECDaverage are Australia, Austria, Canada,the Czech Republic, Finland, Ireland,New Zealand, Sweden and the UnitedKingdom (see Figure 3.5 in the fullreport).

13

Students need to leave school not justwith sound subject-matter knowledgeand skills, but also ready to continuelearning throughout life. In order to doso, they must be able to manage theirown learning, rather than solely relyingon teachers. This means that studentsmust, on the one hand, be interestedand engaged in learning and, on theother hand, have strategies for learningeffectively.

PISA looked at both – motivation andengagement as well as learningstrategies, as reported by students. Theresults can be related to students’performance in reading, mathematicaland scientific literacy in order to providesome idea of what kinds of learners showhigher performance at age 15.

Key Findings

Given substantial investment ineducation and its importance tosocieties’ and students’ future well-being, it is disappointing that asignificant minority of 15-year-oldsdisplay negative attitudes towardslearning and a lack of engagementwith school. In most countries, morethan a quarter of all students say thatschool is a place they do not want togo – ranging from less than 20% inDenmark, Mexico, Portugal andSweden to more than 35% of studentsin Belgium, Canada, France, Hungary,Italy and the United States (see www.pisa.oecd.org).

Student interest in reading andmathematics also varies widely and isclosely associated with performance.About half of 15-year-olds aregenerally positive about reading.However, this varies across countries,with fewer than one-third of studentsin Belgium and Korea but around two-thirds in Denmark, Mexico andPortugal agreeing that reading is funand that they would not want to give itup. Interest in mathematics varies evenmore between countries, with in mostcountries only a small proportion of15-year-olds seeing mathematics as

What are 15-year-olds like as learners?

worth pursuing because they consider itrelevant to their future (see Table 4.2in the full report). Not only does lackof interest tend to be associated withpoorer student performance (see Tables 4.1 and 4.2 in the full report),but other research shows that studentswho are disaffected with learning atschool will also be less likely to engagein learning activities, either inside oroutside school, in later life.

Students who are engaged inreading beyond what is required forschool tend to be better readers, butthe relationship is not straightforward.In some countries the association isstronger than in others, and somecountries with fewer keen readers stillperform well, on average (see Table 4.3in the full report).

Looking at various aspects of howstudents learn, PISA finds that:

The aspect most closely associatedwith performance is “controlling thelearning process”. Students were askedto what extent they:

figure out exactly what they need tolearn;

work out as they go what conceptsthey still have not really understood;

look for additional informationwhen they do not understand;

force themselves to check whetherthey remember what they havelearned; and

make sure they have rememberedthe most important things.

In every country, the quarter ofstudents who say they adopt suchstrategies the most have readingscores significantly above thequarter that adopt them the least.In some countries, the gap is as muchas a full proficiency level – forexample, the difference between theaverage student with reading literacy atLevel 2 and the average student atLevel 3 (see Table 4.5 in the fullreport).

In assimilating new knowledge,students use both memorisation andelaboration strategies. (The latterrefers to strategies to process newinformation, integrating it into alearner’s prior knowledge base.) The results of PISA 2000 show thatstudents who report placing emphasison memorising information do notalways achieve better results, whilethose who elaborate what they learntend to do well (see Tables 4.6 and 4.7in the full report).

Co-operative and competitivelearning strategies both have theirplace in effective learning. Studentswho report adopting competitivelearning strategies tend to do betterthan those who do not. Likewise,students who report that they learn co-operatively tend do better than thosewho do not. This finding suggests thatco-operative and competitive learningcan each help in certain situations:they are complementary rather thanalternative strategies (see Tables 4.8and 4.9 in the full report).

Without further analysis, it cannot beconcluded that certain learningstrategies cause better performance: itmight simply be that students who dowell for other reasons are more likely toadopt the most effective strategies.Nonetheless, the findings do showwhat kind of learning is associated withsuccess. Schools and parents need toconsider how to help students to adopttechniques that allow them to regulatetheir own learning effectively.

14

Reading for enjoyment and reading literacy

Students were asked how much theyread for enjoyment. A substantialpercentage in every country say not atall. The proportion varies from amajority of students in Japan, to below20% in Brazil, Latvia, Mexico, Portugaland the Russian Federation, as shownin the first column of the table. Ofthose students who do read forenjoyment, most do so for under anhour a day. The second column showshow many – 11% on average acrossOECD countries - are keener readers,spending 1-2 hours a day reading forenjoyment.

Are keener readers also better readers?On average in each country, theanswer is yes. For example, the finalcolumn of the table shows thatstudents in Australia who read 1-2hours a day for enjoyment score 92points higher on average on thereading scale than those who do notread for enjoyment. This is equivalentto more than one proficiency level –for example, the difference betweenbeing able to perform Level 3 readingtasks and Level 4 reading tasks – seeexamples on page 7 above.

However, it is important to note that:

Whereas, within countries, there is aclear positive association betweenreading for enjoyment andperformance, countries with fewerkeen readers do not necessarilyproduce fewer good ones. As anillustration, Japanese students areleast likely to read for enjoyment, butshow high average readingperformance. Conversely, in Latviaand Greece, with relatively manykeen readers, mean performanceremains below-average.

It is not clear to what extent readingfor enjoyment leads to higher readingliteracy, or the other way around, orto what extent some other aspect ofstudents’ background contributes toboth. Nevertheless, the associationbetween engaging in reading andbeing good at it is an important one,indicating that it may be productiveto encourage both.

Country

AustraliaGermanySwitzerlandFinlandCanadaNew ZealandIcelandSweden

United KingdomLatviaFranceNorwayIrelandCzech RepublicAustriaPortugalUnited StatesBelgiumDenmarkSpainHungaryRussian Fed.PolandKoreaItalyNetherlands

JapanLuxembourgBrazilGreeceMexico

3342352233303036

291830353326411841422732261924313143

5538192214

1298

18101079

920118

12139

128999

13171612136

812172012

9284837977767372

706767656563636259595754534949414140

272525196

Percent reading for enjoymenta) Not at all b) 1-2 hours

a day

More than one proficiency level% %

Between half and one proficiency level

Less than half of one proficiency level

Difference between average reading scores of

students in a) and b)

15

Different results by gender

Policy-makers have, historically, givenconsiderable priority to issues of genderequity in education, with particularattention to disadvantages faced bygirls and women. PISA’s results pointto the success of many countries’efforts, but also to a growing problemfor males, particularly in readingliteracy. In mathematical literacy, thereremains a measurable disadvantage forfemales in about half of the countrieswhile in scientific literacy genderdifferences tend to average out.

LatviaFinlandNew ZealandNorwayIcelandRussian Fed.ItalyCzech RepublicGreeceSwedenPoland

GermanyAustraliaBelgiumCanadaHungaryLiechtensteinSwitzerlandJapanNetherlandsFranceIrelandUnited StatesLuxembourgAustriaUnited KingdomDenmarkPortugalSpainMexicoBrazilKorea

5351464340383837373736

353433323231303030292929272626252524201714

A. Reading literacy: femalesdo better in all countries

Femalesat least half a proficiency level ahead:

Femalesless than half a proficiency level ahead:

AustriaBrazilKoreaPortugalSpainLuxembourgDenmarkGermanySwitzerlandFranceIrelandLiechtensteinCzech RepublicNorwayCanada

272727191815151514141312121110

B. Mathematics: males dobetter in half of the countries

(in other countries, no statistically significant difference)

231412

C. Science: Females do better in three countries

C. Science: Males do betterin three countries

Ave

rage

sco

re a

dvan

tage

(poi

nts)

,fe

mal

es

Ave

rage

sco

re a

dvan

tage

(poi

nts)

,m

ales

Ave

rage

sco

re

adva

ntag

e (p

oint

s), f

emal

es

Ave

rage

sco

read

vant

age

(poi

nts)

, mal

es

Ave

rage

sco

re a

dvan

tage

(poi

nts)

,fe

mal

es

LatviaRussian Fed.New Zealand

KoreaAustriaDenmark

16

191212

females in mathematical literacy, muchof this is attributable to there beingmore males among the betterperformers and the advantage of malesdisappears when comparing thenumber of low performers (see Table 5.2b in the full report). Thesefindings suggest that theunderachievement of males in readingis a significant challenge for educationpolicy, in terms both of closing thegender gap and of reducing theproportion of students at the lowestlevels of proficiency.

Some countries show that largegender differences are not inevitable.There is significant variation betweencountries in the size of genderdifferences. Indeed, the results of PISA2000 suggest that some countriesprovide a learning environment orbroader context that benefits bothgenders equally. By contrast, theenduring differences in other countries,and the widespread disadvantage nowfaced by many young males in readingliteracy, require serious policyattention.

Females tend to express greaterinterest in reading, and males inmathematics. On average acrossOECD countries, approximately one-half of all males but only a quarter offemales say that they read only whenthey have to (see Figure 5.4 in the fullreport). By contrast, females tend tohave less interest than males inmathematics. The closeinterrelationship between subjectinterest and learning outcomes suggeststhat the different habits and interestsof females and males may have far-reaching consequences for learningthat education policy needs to address(see Figure 4.1 in the full report).

Males and females tend to adoptdifferent learning strategies. Femalesare more likely to report that they takecontrol of their own learning, which isa strategy that tends to be associatedwith learning success. However,females tend to be more likely to rely

Key findings

In all countries, females are onaverage better readers than males.The most striking gender differencerevealed by PISA 2000 is that femalesconsistently outperform males inreading literacy. As the table shows,this difference is not small. In NewZealand, for example, the mean scoreof males is 507 points, close to theaverage for all students across OECDcountries, but the mean score offemales in New Zealand is 553 points,more than the mean for all students inany country (see Table 5.1a in the fullreport). Females score, on averageacross OECD countries, 45 pointshigher than males on the reflectionand evaluation scale, compared with29 points on the interpreting textsscale and 24 points on the retrievinginformation scale. These differencesmay be influenced by differences inreading interests: males report readingmore comics, newspapers and web-pages, females more novels (see Table 5.1b in the full report).

In mathematical and scientificliteracy, gender differences aresmaller than in reading. Formathematical literacy, males performon average 11 points higher acrossOECD countries, but in only half ofthe PISA countries was the differencestatistically significant. In the case ofscientific literacy there is a significantdifference in favour of either males orfemales in only six (see Table 5.1a inthe full report).

Males are more likely to under-perform in reading. But females arenot more likely to underperform inmathematics. An important policyconcern emerges from the large genderdifferences in reading literacy amongthe lowest performing students. In allparticipating countries, males are morelikely than females to be at Level 1 orbelow in reading – in the case ofFinland over three times as likely (seeTable 5.2a in the full report). Whilemales do, on average, better than

17

on memorisation strategies, which areless consistently associated with betterperformance than strategies relatingnew knowledge to existing knowledge,which males report that they prefer (see Tables 4.5, 4.6 and 4.7 in the fullreport).

The importance of family background

It is well established that students whocome from more advantaged familybackgrounds, in terms of factors such asparental education and occupation,and resources in the home, performbetter at school. PISA was able to lookinternationally both at the strength ofthis relationship and at its importancein explaining overall differences instudent performance.

Key findings

The association between familybackground and student performancediffers greatly from one country toanother. The figure illustrates thispoint in the case of average differencesin reading scores according to parentaloccupation. PISA ranked occupationson an internationally recognised scalebased on their measured economicvalue. On the figure, the left-hand endof each bar represents the meanreading score of the 25% of students ineach country with the lowest-rankingparental occupations. The right-handend shows the mean reading score ofthe 25% of students whose parentshave the highest-ranking occupations.Thus, the length of each bar shows themean difference in reading literacybetween the least advantaged studentsand the most advantaged ones, basedon their parents’ occupations. InKorea, at one extreme, there is arelatively small difference of 33 scorepoints between the top and bottomquarters, equivalent to less than half aproficiency level. In Germany andSwitzerland, on the other hand, thedifference is around 114 score points,or more than one proficiency level (see Table 6.1a in the full report).

Not all students fromdisadvantaged family backgroundsperform poorly. In some countries eventhe bottom quarter of students showmedium rather than low performancescores. For example, ranked by parents’occupational status, students in thebottom quarter in Canada, Finland,and Korea score above the average forall students across the OECD (500)(see Table 6.1 in the full report).

600

550

500

450

400

350

Differences in reading scores by parent's occupation

Average for top quarter of studentsby parental occupation

Difference in points

Narrow

est differences by student backgroundW

idest differences by student background600

550

500

450

400

350

Average for bottom quarter of students by parental occupation

18

PISA distinguished the effects of manyspecific aspects of family background.In addition to the effect of parentaloccupational status, shown opposite, itfound that:

Higher parental education andmore social and culturalcommunication between parents andtheir children are associated withbetter student performance. Studentswhose mothers have not completedupper secondary education have aparticularly strong disadvantage with,on average across OECD countries,reading scores 44 points lower thanthose whose mothers have completedupper secondary education (see Table 6.7 in the full report). Theimpact of mothers’ completion oftertiary education is weaker and lessconsistent across countries (see Table 6.7 in the full report). Parentaleducation is closely interrelated withother family background factors.However, when other familybackground factors are equal, eachadditional year of parental educationstill adds almost 4.7 points to thereading scores of students (see Table 8.2in the full report). PISA also askedstudents how they interact with theirparents in aspects ranging fromdiscussing books to eating mealstogether. Those interacting more havehigher reading scores on average –cultural communication (e.g.,discussing books) playing a greater role(on average across OECD countries 59points between the top and bottomquarters, see Table 6.6) than socialcommunication (on average 30 scorepoints, see Table 6.5).

Students from wealthier familiestend do better, but the relationshipwith possessions relating to classicalculture is stronger. Students wereasked about a range of possessions intheir homes. Students from thewealthiest families typically do betterthan students from the least wealthyfamilies, but the differences inperformance are modest in manycountries (see Table 6.2 in the fullreport). A stronger predictor of

performance is whether students haveitems associated with “classicalculture”, such as literature and works ofart, in their homes (see Table 6.3 inthe full report). While possession ofsuch “advantages” is related to otherhome background characteristics, itseffects in isolation remain consistentlystrong (see Table 8.2 in the full report).The effects are higher in readingliteracy than in mathematical andscientific literacy, emphasising theways in which educational benefitaccrues from home-based access toliterature and other culturalpossessions.

Living with only one parent is, onaverage, associated with lowerstudent performance. On averageacross OECD countries, students wholive with one parent score 12 pointslower in reading literacy than studentswho live with two parents, all elsebeing equal (see Table 8.2 in the fullreport). In some countries, there is nota significant difference. However, twoof the countries where there is thewidest difference, the United Kingdomand the United States, are also the twowhere the proportion of students livingwith only one parent is largest (seeTables 6.9 and 8.2 in the full report).

Educational disadvantage amongchildren born outside the countryvaries greatly. In several countries,non-native born students show muchlower reading literacy scores thanstudents who were born (as were theirparents) in the country. They are atleast 71 points behind native-bornstudents in 10 out of the 15 countrieswith more than 3 per cent ofimmigrant students, but in somecountries the gap is much smaller (see Table 6.10 in the full report). Asignificant difference, 26 points onaverage across OECD countries,remains even when other factors offamily background are considered (see Table 8.2 in the full report). Thesestudents may be academicallydisadvantaged either because they areimmigrants entering a new educationsystem or because they need to learn a

19

new language in a home environmentthat may not facilitate this learning. In either case they may be in need ofspecial attention. PISA finds a moremixed picture in the case of studentsborn in the country but whose parentsimmigrated. In some countries, they donot perform significantly differentlyfrom native born students, but in fourcountries – Belgium, Germany,Luxembourg and the Netherlands –there is a gap of more than one readingproficiency level (see Table 6.10 in thefull report).

Can schools compensate for socio-economic background?

School systems aim to provide equalityof opportunity, giving the samechances to children regardless of thefamily circumstances to which theywere born. In practice, children withdifferent backgrounds do not doequally well. To what extent is thisbecause of home influences, and towhat extent because of experiences atschool? PISA provides some clues, inways that can inform strategies toimprove the performance of lessadvantaged students.

Key findings

Differences in the performancebetween schools account for much ofthe variation in student performancein some countries. As shown in thefigure, there are wide differences inreading literacy performance amongschools in countries such as Austria,Belgium, Germany, Hungary andPoland where these account for muchof the variation in overall studentperformance. Conversely, in countriessuch as New Zealand and Norway,differences are mainly within schools.The countries with the greatestdifferences between schools tend to bethose that send students to differentkinds of secondary school, often on thebasis of prior performance in school(see Table 2.4 in the full report).

Between schools

Variation in student performance in reading

Within schools

Index of variation (average total variation in OECD countries = 100)

10080604020020406080

100

80604020020406080

20

Varying amounts of within-schooland between-school differences inperformance are associated with socio-economic background. Students inbetter-performing schools often comefrom more advantaged families. PISAalso suggests that the effects of socio-economic clustering are larger in schoolsystems with differentiated school systemsthan in systems in which the curriculumdoes not vary significantly among schools(see Table 8.3 in the full report).

The socio-economic composition ofa school’s student population is aneven stronger predictor of studentperformance than individual homebackground. PISA shows, for example,that two students with the same familycharacteristics going to differentschools - one with a higher and onewith a lower socio-economic profile -could expect to be further apart inreading literacy than two students fromdifferent backgrounds going to the sameschool. Although this phenomenonhas complex causes, it underlines thepotential link between the socio-economic segregation of students indifferent schools and the polarisation ofstudents by performance (see Figure 8.4in the full report).

Schools that have a morefavourable climate and are betterresourced tend, to varying degrees, tohave more advantaged students. InBelgium, Germany and Luxembourg,for example, schools with a moreadvantaged socio-economic backgroundtend to employ a greater proportion ofspecialist teachers, often because suchteachers are employed in the moreadvanced tracks leading to entry intouniversity. Schools with a higher socio-economic intake also seem to have abetter disciplinary climate, as reportedby students, particularly in Italy, Japan,Spain, the United Kingdom and theUnited States. Finally, students inschools with low economic, social andcultural status also tend not to useschool resources as regularly as studentsin better-off schools (see Figures 8.5and 8.5a in the full report).

Policy implications

PISA’s findings have important policyimplications for education systems. Insome countries, students are highlysegregated in terms of socio-economicvariables, in part because ofresidential segregation and economicfactors, but also because of features ofthe schooling system. Educationalpolicy in such countries mightattempt to moderate the impact ofhome background on studentperformance by reducing the extentof segregation along socio-economiclines, or by allocating resources toschools differentially. In thesecountries, it is important tounderstand how the allocation ofschool resources within a country isrelated to the socio-economic intake ofits school.

In other countries, there is relativelylittle socio-economic segregation; thatis, schools tend to be similar in theirsocio-economic intake. Educationalpolicy in these countries might aim tomoderate the impact of homebackground through measures designedto improve school resources and toreduce within-school segregation inaccordance with the economic,cultural and social status of students. Inthe end, of course, what will mattermost is how effectively those resourcesare used.

21

What can schools do to make a difference?

Home background influenceseducational success, and schools’ socio-economic status may reinforce itseffects. Equally important, PISA alsoidentified several things that schoolscan do that are associated with studentsuccess. In identifying a constellationof factors that interact to influenceperformance, this first report does notclaim to provide causal links betweenwhat schools do and how their studentsperform. Nonetheless, the initialfindings offer some clues about theconditions in schools that are mostclosely associated with success. Thefollowing findings consider the separateeffect of each factor identified, afterassociations with the other observedschool and home background factorshave been taken into account. Thefindings presented below tend to besimilar for reading, mathematical andscientific literacy.

Students’ use of school resources ismore closely associated with studentperformance than is the physicalinfrastructure of schools. Studentswere asked about their use of theirschool’s library, computers, calculators,laboratories and Internet connection.In schools where usage is relativelyhigh, mean reading scores tend to behigher, even when other factors arediscounted (see Table 8.5 in the fullreport). Deficiencies in the quality ofthe school’s physical or materialinfrastructure, as reported by theprincipal, tend to have a much weakerimpact than students’ use of theseresources (see Tables 7.10, 7.11 and 8.5in the full report).

Qualified teachers are among aschool’s most valuable resources.PISA asked school principals toindicate the percentage of teacherswith a university-level qualification intheir respective subject area. Havingmore of these teachers is associated, onaverage across OECD countries, withbetter student results. For example, inreading, a 25 percentage point increasein the proportion of teachers with auniversity-level qualification in therelevant subject domain is associatedwith an advantage of 9 points on thereading literacy scale, on average acrossOECD countries, the other schoolfactors measured by PISA being equal(see Table 8.5 in the full report).

The ratio of students to teachingstaff matters most where it isrelatively high. Among schools wherethe number of students for eachmember of the teaching staff exceeds25, the mean performance of students ismarkedly lower, the higher the ratio. Inthe more typical range, between 10 to25 students per teacher, there is a muchweaker association with performance inreading literacy. Schools with fewerthan 10 students per member of theteaching staff actually score slightlybelow the OECD average which may bebecause many such schools servestudents with special needs (see Table 8.5 in the full report).

Some aspects of school policy andpractice tend to be associated withbetter student performance. Threesuch factors, as perceived by schoolprincipals, show, on average across

OECD countries, a positive andstatistically significant association withstudent performance (see Table 8.5 inthe full report). These include:

teacher-related factors affectingschool climate, such as teacherexpectations of students;

teacher morale and commitment;and

school autonomy.

Some aspects of classroom practiceare associated with better studentperformance. Three such factors, asperceived by students, show a positiveand statistically significant associationwith student performance:

teacher-student relations;

disciplinary climate of theclassroom; and

the extent to which teachersemphasise academic performanceand place high demands onstudents.

22

The first two of these factors arestronger than the third (see Tables 7.2,7.3 and 8.5 in the full report).

Successful students are more likelyto do homework. The other schoolfactor with a close association withstudent success is homework. Withincountries, students who do morehomework are, on average acrosscountries, likely to perform better inreading literacy. The quarter who dothe most homework score on average44 points higher than the quarterdoing the least. This association isstrongest in countries where studentsdo more homework on average (seeTable 7.6 in the full report).

Further research is needed

Overall, the combined influence of thisset of school-level factors explains 31per cent of the variation in readingliteracy performance among schoolswithin countries, and 21 per cent ofthe variation among countries.

Together with home backgroundfactors, the factors explain 72 per centof the variation among schools withincountries and 43 per cent of thevariation between countries (see Table 8.5 in the full report).

These findings give a first indication ofPISA results. Much further researchand analysis will be needed to identifyhow each school factor operates,interacts with home background andinfluences school and studentperformance. Further thematic reportsin 2002 and 2003 will seek tounderstand in more detail why somecountries and schools perform betterand achieve more equitable learningoutcomes than others.

23

Are students well prepared to meet the challenges of the future? Are they able to

analyse, reason and communicate their ideas effectively? Do they have the

capacity to continue learning throughout life? These are questions that parents,

students, the public and those who run education systems continually ask.

Knowledge and Skills for Life, the report summarised in this brochure, provides

some answers. It assesses how far students near the end of compulsory

education have acquired some of the knowledge and skills that are essential for

full participation in society. It presents evidence on student performance in

reading, mathematics and scientific literacy, reveals factors that influence the

development of these skills at home and at school, and examines what the

implications are for policy development.

The report shows considerable variation in levels of knowledge and skills between

students, schools and countries. The extent to which the socio-economic

background of students and schools affects student performance varies. Some

countries have managed to mitigate the influence of social background and some

have done that while achieving a high overall mean performance. This is a

noteworthy achievement. Will other countries take up the challenge?

The OECD Programme for International Student Assessment (PISA)is a collaborative process among the 30 Member countries of the OECD and

some non-OECD countries, bringing together scientific expertise from participating

countries and steered jointly by their governments on the basis of shared, policy-

driven interests. PISA is an unprecedented attempt to measure student

achievement across all OECD countries and some non-OECD countries, as is

evident from some of its features:

The literacy approach: PISA aims to define each domain (reading, mathematics

and science) not merely in terms of mastery of the schools curriculum, but in

terms of the knowledge and skills needed for full participation in society.

A long-term commitment: Over the decade to come, it will enable countries

regularly and predictably to monitor their progress in meeting key learning

objectives.

The age-group covered: By assessing 15-year-olds, i.e. young people near the

end of their compulsory education, PISA provides a significant indication of the

overall performance of school systems.

The relevance to lifelong learning: PISA does not limit itself to assessing

students’ knowledge and skills but also asks them to report on their own, self-

regulated learning, their motivation to learn and their preferences for different

types of learning situation.

Knowledge and Skills for LifeFIRST RESULTS FROM PISA 2000

EDUCATION AND SKILLS

Further information and on-line ordering:

To order the Report (OECD code: 96 2001 14 1 P 1)www.oecd.org

Data underlying the Reportwww.pisa.oecd.org

All OECD books and periodicalsare now available on-linewww.SourceOECD.org

Documents

Knowledge and Skills for Life