![]()
Assessment of primary education
What 12-year-old students learn
Methodology
Objectives and Scope
Objectives
- To ascertain and assess what students know about Spanish language and literature, mathematics, social sciences and natural sciences by the end of the Sixth year of EGB (age 12).
- To relate performance levels to contextual factors and educational processes.
Geographical scope
- All Spanish territory except Ceuta and Melilla.
Population scope
- All students registered in the Sixth year of EGB in the academic year 1994/1995, except:
- Students registered at schools with less than 10 students in Second grade of Primary or Sixth grade of EGB.
- Students with special educational needs.
- Overall, the population excluded form the survey comes to a total of 5%.
Sample design
Sampling type
- By clusters, stratified in two stages:
- 1st stage: random selection from schools in the strata.
- 2nd stage: random selection of a Sixth-year EGB classroom in each of the schools selected in stage 1.
- Stratification of schools by:
- Autonomous Region
- type of school (public or private)
- school size as determined by number of classes
- The sample includes:
- students in the selected class with no special educational needs
- teachers of the assessed subject areas in that class
- managerial staff at the selected schools
- parents of selected students
Sample precision
The purpose of the sampling process was to arrive at estimates of means and percentages with a confidence coefficient of 95% and the following margins of error:
- nationwide estimators:
- Means: ± 0.1s (s: estimation of standard deviation)
- Percentages: ± 5%
- Autonomous Regions with full responsibility for education estimators:
- Means: ± 0.2s
- Percentages: ± 10%
- pre-established sample size:
- Autonomous Regions with full responsibility for education: 34 schools
- Autonomous Regions where the Ministry of Education is responsible: 20 schools
- Pre-established total: 438 schools
Proportional sample adjusted to the population size within each stratum.
Tests and questionnaires
Tests of knowledge and attitudes
Type of questions:
- multiple choice
Question selection process
A set of questions tested in a pilot survey was taken for each area and a selection made for inclusion in the final test, on the basis of:
- their statistical features
- how they fit the Item Response Theory (IRT) model
- discrimination index, point biserial correlation coefficient greater than 0.2
- similar percentages in the selection of distractors
- coverage of all the content blocks
- coverage of content types (concepts, procedures and attitudes)
- time required for answering
Composition of tests
- Language: 61 questions
- Mathematics: 37 questions
- Social and Natural Sciences
- 24 questions on Social Sciences
- 23 questions on Natural Sciences
All tests included questions on students' attitudes to the subject matter concerned.
Context questionnaires
The purpose of these is to gather information on the general educational context and on educational processes and the educational environment. The following questionnaires were compiled:
- General Student Questionnaire
- Teacher Questionnaire
- Management Staff Questionnaire
- Family Questionnaire
Administration
Dates of administration
- May and June 1995
Administration procedure
- Morning:
- 1st session: Spanish language and literature, 60 minutes
- Break
- 2nd session: Mathematics, 60 minutes
- 3rd session: Background Questionnaire, 30 minutes
- Afternoon or morning of the next day:
- 4th session: Social and Natural Sciences, 60 minutes
Effective sample size
- 437 schools
Quality of estimations of overall performance in the four areas
- Confidence coefficient: 95%
- Margin of error: ± 1% of estimator value
Reader's note:
The overall average of correct answers in the Language test was 64%. Given a margin of error of ± 1%, we have stated that the overall average of correct answers for the whole population was between 63 and 65 per cent.
The degree of certainty of this statement is given by a confidence coefficient of 95%; this means that if many samples were to be obtained by the same procedure and the corresponding intervals calculated, 95% would contain the true average for the population.