EDUCATION INDICATORS: An International Perspective
The information presented in this report was obtained from many sources, including federal, national, international, and state agencies, private research organizations, and professional associations. The data were collected using many research methods, including surveys of a universe (such as all colleges) or of a sample (such as 15,000 eighth graders), and compilations of administrative records. In the Sources of Data section, descriptions of information sources and data collection methods are presented, grouped by sponsoring organization. More extensive documentation of a particular survey's procedures does not imply more problems with the data, only that more information is available.
The accuracy of any statistic is determined by the joint effects of "sampling" and "nonsampling" errors. Estimates based on a sample will differ somewhat from the figures that would have been obtained if a complete census had been taken using the same survey instruments, instructions, and procedures. Estimation of sampling error is discussed in the sidebar entitled Using data from sample surveys. Unless otherwise noted, all statements, cited in the text from sample surveys were tested for statistical significance and are statistically significant at the .05 level. Several test procedures were used. Which procedure was used depended on the type of data being interpreted and the nature of the statement being tested. The most commonly used procedure was multiple ttests with a Bonferroni adjustment to the significance level. When multiple comparisons between more than two countries were made, even if only one comparison is cited in the text, a Bonferroni adjustment to the significance level was made to ensure the significance level for the tests as a group was at the .05 level. In this report the emphasis is on comparisons between the United States and the other G-7 countries which required adjusting the significance level for the number of other countries represented in the table.
In addition to sampling errors, all surveysboth universe and sampleare subject to nonsampling errors. These arise, for example, when the respondents or interviewers interpret questions differently, when respondents fail to respond (completely or partially), or when respondents who should be included in a universe are not. Since estimating the magnitude of nonsampling errors often would require special experiments or access to independent data, these nonsampling errors are seldom available. Thus, estimates of survey error in statistics usually, but not always understate total survey error and overstate the precision of survey estimates.
Readers should take particular care when comparing data from different sources. Differences in procedures, timing, phrasing of questions, and interviewer training mean that the results from the different sources may not be strictly comparable. Readers should also be aware that countries are at different stages in the sophistication of their data development and collection systems, and thus, have different types of data in an easily accessible form. A limited number of countries (e.g., Denmark, Switzerland) have national registers that follow student cohorts throughout their education career. Some countries (e.g., United States, Canada) have large survey programs that collect data regarding education based on samples of schools or students. Additionally, some countries have extensive national-level data available through administrative records.
Education at a Glance: OECD Indicators
The third edition of Education at a Glance (EAG), published in 1995, is produced through the cooperation of member countries of the Organization for Economic Co-operation and Development (OECD) and the Indicators of Education Systems (INES) project. 1991-1992 data for 49 indicators of international education are provided for 27 nationsAustralia, Austria, Belgium, Canada, Czech Republic, Denmark, Finland, France, Germany (Unified, former West and former East Germany, as available), Greece, Hungary, Iceland, Ireland, Italy, Japan, Netherlands, New Zealand, Norway, Portugal, Russia, Spain, Sweden, Switzerland, Turkey, United Kingdom, and United States. Indicators cover the contexts of education (i.e., demographic, social and economic, opinions and expectations); costs, resources, and school processes; and student, system, and labor market outcomes of education. Much of the data is broken down by education level. EAG, 1995 also provides annotated organization charts of the countries' education systems.
Most of the OECD data used in Education Indicators: An International Perspective came from Education at a Glance (EAG), 1995, which reports data for 1992. In the few cases where 1992 data are not available, earlier versions of EAG were used.
Since only developed nations, mostly European, are included in OECD studies, the range of analysis is limited. However, OECD data allow for some detailed international comparisons of financial resources or other education variables to be made for this selected group of countries.
In 1990-91, a total of 20 countries assessed the mathematics and science achievement of 13-year-old students and 14 of the 20 countries assessed 9-year-old students in these same subjects. Nine of the countries participating in the evaluation of 13-year-olds also administered a geography component. Some countries assessed virtually all age-eligible children in the appropriate age group; others confined their samples to certain geographic regions, language groups, or grade levels. The definition of populations often followed the structure of school systems, political divisions, and cultural distinctions. In some countries, significant proportions of age-eligible children were not represented because they did not attend school. Also, in some countries, low rates of school or student participation mean results may be biased.
Typically, a random sample of 3,300 students from about 110 different schools was selected from each population at each age level; half were assessed in mathematics and half in science. A total of about 175,000 9- and 13-year-olds (those born in calendar years 1981 and 1977, respectively) were tested in 13 different languages in March 1991.
The mathematics and science achievement tests lasted 1 hour. The tests given to 9-year-olds included 62 questions in mathematics and 60 questions in science. Those for 13-year-olds included 76 questions in mathematics and 72 questions in science. In addition, students of each age group spent about 10 minutes responding to questions about their backgrounds and home and school experiences. School administrators completed a school questionnaire. The geography component consisted of 24 content-area questions and 14 background questions.
Initial analyses involved the calculation of the percentage of correct answers and standard errors for individual questions. For each population, the weighted percentage of correct answers was calculated for each question. The results of students who omitted questions at the ends of sections because they didn't reach them were excluded from the calculations for those questions. For each percentage correct, an estimate of its standard error was calculated using the jackknife procedure. Percentage and standard errors were calculated for subgroups within each population, including gender and grade. Statistics for Canada were calculated using an appropriately weighted sample of responses drawn from the individual Canadian populations. Results of the mathematics, science, and geography assessments can be found in three separate publications produced by IAEP and Educational Testing Service (ETS), entitled Learning Mathematics, Learning Science, and Learning About the World.
Reading Literacy Study
In the period 1989 to 1992, the International Association for the Evaluation of Educational Achievement (IEA) conducted a Reading Literacy Study in 32 systems of education. The study focused on two levels in each of these systems, the grade level where most 9-year-olds were to be found and the grade level where most 14-year-olds were to be found.
To obtain comparable samples of students, multistage sampling was used in each country and schools or classes were typically drawn with a probability proportional to the size of the school or class.
Three major domains or types of reading literacy materials assessed at both age levels were as follows:
1. Narrative prose: Continuous texts in which the writer's aim is to tell a storywhether fact or fiction. They normally follow a linear time sequence and are usually intended to entertain or involve the reader emotionally. The selected extracts ranged from short fables to lengthy stories of more than 1,000 words.
2. Expository prose: Continuous texts designed to describe, explain, or otherwise convey factual information or opinion to the reader. The tests contained, for example, brief family letters and descriptions of animals as well as lengthy treatises on smoking and lasers.
3. Documents: Structured information presented in the form of charts, tables, maps, graphs, lists, or sets of instructions. These materials were organized in such a way that students had to search, locate, and process selected facts rather than read every word of continuous text. In some cases, students were required to follow detailed instructions in responding to such documents.
To obtain raw scores, all correct answers were totaled for each student in each domain. The Rausch procedure was used to produce scales for each domain. Each scale was given a mean of 500 and a standard deviation of 100.
Computers in Education Study
The Computers in Education (Comped) study, conducted by the IEA, was designed to evaluate how computers have been introduced in education and are being used in schools around the world. Data from 21 school systems in the following 20 countries were included in the study: Austria, Belgium (Flemish and French schools evaluated separately), Canada, China, France, West Germany (former), Greece, Hungary, India, Israel, Italy, Japan, Luxembourg, Netherlands, New Zealand, Poland, Portugal, Slovenia, Switzerland, and the United States. Data collection for the first part of the two-stage study was completed in 1989, and data collection for the second stage was conducted in 1992. The aim of stage 1to collect data at the national, school, and teacher level focusing on how computers are used, the extent and availability of computers in schools, the nature of instruction about computers, and estimates of the effects that computers have on studentswas achieved through the completion of school and teacher questionnaires. The analysis of stage 1 data is presented in the IEA publication entitled The IEA Study of Computers in Education: Implementation of an Innovation in 21 Education Systems. Stage 2 of the study consists of two parts; the first part is a follow-up of stage 1, to assess longitudinal change, and the second part involves evaluating effects of school variables, teacher, and teaching variables on student outcomes in the domain of computer usage in schools.
The Luxembourg Income Study (LIS) began in 1983 as a joint project sponsored by the government of Luxembourg and the Center for Population, Poverty and Policy Studies (CEPS) in Walferdange, Luxembourg. Created to compile and provide access to an international database containing social and economic data, the project receives its current funding from CEPS/International Networks for Studies in Technology, Environment, Alternatives, Development (INSTEAD) and the national science foundations of its member nations. Along with its office in Walferdange, divisions of LIS are housed at Syracuse and Harvard Universities.
As of 1993, LIS membership consisted of 23 countries in Europe, North America, and Australia, with applications pending for Korea, Finland, Mexico, Portugal, and Taiwan. Data are provided by individual nations and cover the period from 1968 to 1989. Each study conducted by LIS is produced in the form of a working paper, of which there are more than 100. LIS reports are also published in books, articles, and dissertations.
Statistical Abstract of the United States
First published in 1878, the Statistical Abstract of the United States is an annual publication containing statistics on finance, education, industry, health, and population for the United States. Although it primarily presents national data for the United States, each volume contains some data at the state, regional, and metropolitan level. Current volumes also include a small section on international comparative statistics. Most of the data used in the publication is taken from the household survey information of the U.S. Census. Other data is provided predominantly by other divisions of the U.S. Department of Commerce and by other government agencies.
Office of Productivity and Technology
The Office of Productivity and Technology's unpublished tables entitled "Comparative Real Gross Domestic Product Per Capita and Per Employed Person" present national data for 13 OECD countries and Korea. The tables provide two sets of comparisons, based on purchasing power parities (PPPs) benchmarked to 1985 and 1990 studies. The studies were conducted jointly by the OECD and EUROSTAT (the Statistical Office of the European Community) as part of the United Nations International Comparison Project (UNICP). Information for each benchmarked year includes data for GDP, GDP per capita, and GDP per employed per person, indexed to the United States and in U.S. dollars. PPPs and relative prices are also given, with PPPs for GDP and comparative price levels indexed to the United States. The tables also present GDP trends, implicit price deflators for GDP, and population and employment measures.
Results of the 1985 and 1990 studies differ dramatically; this is most
likely attributed to weighing patterns, the change in aggregation method, or
possible measurement errors.