Annex A. Technical notes on sampling procedures, response rates and adjudication for TALIS 2018
The objective of the Teaching and Learning International Survey (TALIS) in 2018 was to obtain, in each participating country and economy, a representative sample of teachers for each ISCED level in which the country and economy participated. Moreover, a representative sample of teachers teaching students of the appropriate age in schools selected for the Programme for International Student Assessment (PISA) in 2018 was required for each country and economy that opted to participate in the TALIS-PISA link. TALIS 2018 identified policy issues that encompass the classroom, teachers, schools and school management, so the coverage of TALIS 2018 extends to all teachers of each concerned ISCED level and to the principals of the schools where they teach. The international sampling plan prepared for TALIS 2018 used a stratified two-stage probability sampling design. This means that teachers (second stage units, or secondary sampling units) were to be randomly selected from the list of in-scope teachers in each of the randomly selected schools (first stage units, or primary sampling units). A more detailed description of the survey design and its implementation can be found in the TALIS 2018 Technical Report (OECD, 2019[1]).
A teacher of ISCED level 1, 2 or 3 is one who, as part of his or her regular duties in their school, provides instruction in programmes at that ISCED level. Teachers who teach a mixture of programmes at different ISCED levels in the target school are included in the TALIS universe. There is no minimum cut-off for how much teaching these teachers need to be engaged in at any of the three ISCED levels.
The international target population of TALIS 2018 restricts the survey to those teachers who teach regular classes in ordinary schools and to the principals of those schools. Teachers teaching to adults and teachers working in schools exclusively devoted to children with special needs are not part of the international target population and are deemed out of scope. Unlike in TALIS 2008, however, teachers working with special needs students in a regular school setting were considered in scope in TALIS 2013 and 2018. When a school is made up exclusively of these teachers, the school itself is said to be out of scope. Teacher aides, pedagogical support staff (e.g. guidance counsellors and librarians) and health and social support staff (e.g. doctors, nurses, psychiatrists, psychologists, occupational therapists and social workers) were not considered to be teachers and, thus, not part of the TALIS international target population.
For national reasons, participating countries could choose to restrict the coverage of their national implementation of TALIS 2018 to parts of the country. For example, a province or state experiencing civil unrest or in an area struck by a natural disaster could be removed from the international target population to create a national target population that does not include these provinces or states. Participating countries were invited to keep these exclusions to a minimum by keeping the national survey population to at least 95% of the teachers.
TALIS 2018 recognised that attempting to survey teachers in very small schools can be inefficient and difficult. For each ISCED level, surveying teachers in schools with no more than three teachers at a specific ISCED level and those teaching in schools located in geographically remote areas could be a costly, time-consuming and statistically inefficient exercise. Therefore, participating countries were allowed to exclude those teachers for TALIS 2018 data collection, thus creating a national survey population different from the national target population. The national project manager (NPM) for each country was required to document the reasons for exclusion, the size, the location, the clientele, etc., of each excluded school. This documentation was required for each ISCED level in which a country participated. The school exclusions for the TALIS-PISA link were the same as those used in PISA 2018.
Within a selected in-scope school, the following categories of teachers were excluded from the sample:
teachers teaching in schools exclusively serving special needs students
teachers who also act as school principals: no teacher data collected, but school principal data collected
substitute, emergency or occasional teachers
teachers on long-term leave
teachers teaching exclusively to adults
teachers who had taken part in the TALIS 2018 field trial.
For each ISCED level, the same requirements for sample size and precision of estimates were established. To allow for reliable estimation and modelling, while allowing for some amount of non-response, the minimum sample size was set at 20 teachers within each participating school. A minimum sample of 200 schools was to be drawn from the population of in-scope schools. Thus, the nominal international sample size was a minimum of 4 000 teachers for each ISCED level in which a country or economy participated. Participating countries and economies could choose to augment their national sample by selecting more schools, by selecting more teachers within each selected school or by increasing both. Some countries and economies were asked to increase the within-school sample to counterbalance the effect of selecting too many schools with fewer than 20 teachers. The sample size requirement was reduced for some participating countries and economies because of the smaller number of schools available for sampling. In a few cases, because the average number of teachers in the schools was fewer than expected in the international plan, the number of schools sampled was increased to maintain a minimum total number of participating teachers.
In many countries, the separation of grades in ISCED levels does not correspond to a physical separation of school buildings or administrations: schools that offer grades 8 to 12 straddle ISCED levels 2 and 3, but all of ISCED level 2 would not be covered by those schools. In countries and economies that participated in more than one ISCED level, arrangements were made with the NPM and their team to optimise the selection of the school sample by either minimising the overlap of the respective samples (one school is selected for participation in only one ISCED level) or maximising the sample overlap (a selected school contributes to all concerned ISCED levels). However, in the case of maximised overlap, teachers who taught at more than one level would be asked to participate in only one.
As in previous cycles, TALIS 2018 followed the INES (Indicators of Educational Systems) data collection definition of a teacher for sampling and analysis:
“A classroom teacher (ISCED 0-4) is defined as a person who plans, organises and conducts a group of activities with the aim of developing students’ knowledge, skills and competencies as stipulated by educational programmes.” (OECD, 2018, p. 43[2]).
The basic principle that guides the adjudication is to determine, for each participating country and economy and for each of the TALIS options, whether the data released to the countries and economies are fit to provide policy relevant, robust international indicators and analysis on teachers and teaching in a timely and cost effective manner.
To establish fitness for use, a number of quality assurance processes were designed and activated throughout the survey process. Some processes relied on expert advice and opinion; some relied on qualitative information and learned judgement; some relied on quantitative information. For more detailed information, please refer to the TALIS 2018 Technical Report (OECD, 2019[1]).
During the adjudication session, each individual dataset – that is, the combination of participating countries/economies, survey options and questionnaire types – was submitted to the same examination. For the first time in a TALIS cycle, principal data were evaluated on their own. In other words, principal and teacher data received separate adjudication evaluation per TALIS option and per country and economy.
The issues evaluated concerned the questionnaire adaptation to national context, translation and verification, quality of the sampling frame, handling of out-of-scope and refusal units (i.e. teachers and/or schools), within-school sampling, data collection, data cleaning, the reports of quality observers, participation rates and overall compliance with the technical standards. Once each survey process had been assessed, a recommended rating was formulated, accounting for the participation rates, and for any unresolved issue.
The adjudication rules, based on participation rates for principals and teachers, are displayed in Tables AII.A.1 and AII.A.2. An explanation of the codes used is given below.
The following bulleted list is a simple guide aimed at helping data users appreciate the limitations on use or quality:
Good: the participating country’s/economy’s data can be used for all reporting and analytical purposes and can be included in international comparisons
Fair (A): national and sub-national estimates can be produced; some teacher characteristics may suffer from a larger standard error (s.e.), hence the warning “Fair” and no additional warnings to users appear necessary
Fair (B, only for teacher data adjudication): national and sub-national estimates can be produced; some sub-national estimates may be of lower precision (larger s.e.) if sample size is locally low, hence the warning “Fair” and no additional warnings to users appear necessary
Fair (C):
national and sub-national estimates can be produced
some sub-national estimates may be of lower precision (larger s.e.) if sample size is locally low, hence the warning “Fair”, but
a note on data quality could appear pointing to the outcome of the non-response bias analysis (NRBA)
since school participation is somewhat lower than under (B), comparing sub-national estimates should be done with care, as some of those results are based on few schools
comparing small sub-national estimates with similar groups from other countries is likely to uncover any statistically meaningful differences as s.e. are likely too large
Poor (D):
in addition to the warnings issued for the previous category, a note should warn users of indications of non-response biases in some estimates
comparisons of sub-national estimates should be limited to the groups with the larger sample sizes
at this point, the sample represents between 37% and 56% of the teaching workforce, from a rather small sample of schools
comparisons with similar groups in foreign countries would not be encouraged
Poor (E, only for teacher data adjudication): sub-national estimates would not be recommended; there should be a note pointing out the difficulty of obtaining a representative sample of schools
Poor (F, only for teacher data adjudication): limitations similar to those of line E, but there should be a note pointing out the difficulty of obtaining at least 50% participation of the selected sample of schools; risks of having a non-representative sample of schools
Insufficient: weights should not be calculated for any official tabulations; hence, data should not be incorporated into international tables, models, averages, etc.
The participation rates and the adjudication rating per participating country/economy and by ISCED level are presented in Tables AII.A.3 to AII.A.8.1
This section lists issues to be noted regarding the sampling or field operations that should be considered when interpreting the data reported for these countries.
Alberta (Canada):
TALIS data collection conducted during a labour dispute.
Non-response bias analysis shows no evidence of high risk of school non-response bias on the investigated variables for teachers or principals in ISCED 2 or ISCED 3 and, as such, their rating was upgraded from “poor” to “fair”.
Australia:
For both ISCED 1 and 2, the data collection window for both teachers and principals was extended from the end of the academic year in 2017 to the beginning of the following academic year in 2018.
For ISCED 1 principals and teachers and for ISCED 2 principals, data from Australia are located below the line in selected tables in this report and not included in the calculations for the international average. This is because Australia did not meet the international standards for participation rates, as shown in Tables AII.A.3 to AII.A.5.
Colombia: Non-response bias analysis shows no evidence of high risk of school non-response bias on the investigated variables for teachers or principals and, as such, their rating was upgraded from “poor” to “fair”.
Czech Republic: Trend items should be interpreted carefully due to complications arising from the translation process.
Denmark: Non-response bias analysis shows no evidence of high risk of school non-response bias on the investigated variables for teachers or principals in ISCED1, ISCED 2 or ISCED 3 and, as such, their rating was upgraded from “poor” to “fair”.
Flemish Community of Belgium: For both ISCED 1 and 2, entries on the sampling frame are administrative units and not “schools” as they are usually defined; a “school” may be comprised of one or several administrative units and the principal would be reporting for the school and not only the selected administrative unit; therefore, users should exercise care when analysing and comparing school-level statistics.
French Community of Belgium: Items regarding the share of students with special needs should be interpreted carefully due to complications that could arise from the interpretation of the definition of special needs. Students studying for a differentiated first degree, which is organised for students who did not pass their primary certificate and who receive extra support and resources, are formally identified as having learning difficulties but most of them do not suffer from any kind of disability.
Georgia:
Items repeated from the TALIS 2013 questionnaire (hereafter “trends items”) should be interpreted carefully due to complications arising from the translation process.
Some translation issues could still exist in the Georgian and Azerbaijani version of the questionnaires.
Israel: Coverage falls below 95%, after post facto exclusion of ultraorthodox schools for low response rates, making coverage identical to that of TALIS 2013. Because translation issues could still exist in the trends items, users need to exercise caution when comparing results across TALIS cycles.
Italy: Trends items should be interpreted carefully due to complications arising from the translation process.
Latvia:
Trends items should be interpreted carefully due to complications arising from the translation process.
Some translation issues could still exist in the national instruments that could affect the data.
Korea: For ISCED 2, in four schools, teacher listings were found to be incorrect; those schools were set to “non-participant”.
Netherlands:
For ISCED 1 and 2, the Netherlands had a six-week early start and extended collection window.
For ISCED 1 and 2, the Netherlands had an unapproved collection protocol that resulted in the inclusion of some 50 “national” schools that were not included in the international dataset but left on the national dataset; participation rates were computed on the international dataset.
For ISCED 1 principals and teachers, data from the Netherlands are located below the line in in the result tables of this report. This is because the Netherlands did not meet the international standards for participation rates, as shown in Tables AII.A.3 and AII.A.4.
New Zealand: Coverage was extended to small schools (four or fewer teachers). While the impact of this action on the target population of teachers was negligible, the impact on the target population of principals is important because, compared to TALIS 2013, the target population for principals nearly doubled in size. The TALIS 2018 results reported in the result tables on changes since 2013 for New Zealand were estimated after excluding from New Zealand’s TALIS 2018 sample those schools with four or fewer eligible teachers. This was done in order to ensure comparability with their TALIS 2013 results (as such, small schools were excluded from the TALIS 2013 sampling frame). Therefore, these results can differ from those reported for the full TALIS 2018 sample, especially those derived from the school and/or principal data.
Russian Federation: Coverage falls below 95% after the exclusion of Moscow.
Saudi Arabia: Coverage falls below 95% after the sampling excluded two provinces bordering Yemen.
Singapore: Coverage included both privately and publicly managed schools. Nevertheless, private schools were excluded from the sample in TALIS 2013 due to confidentiality issues. Sensitivity analyses (i.e. restricting the analysis between 2013 and 2018 to just public schools) revealed that the impact of including private schools was negligible for most of the trends results produced for the report. Nevertheless, analyses revealed that the results displayed for Tables II.3.6, II.3.10, II.3.33 and II.5.8 and Figure II.3.3 are due largely to the compositional difference between the 2013 and 2018 samples. Thus, trends results coming from these tables should be interpreted with extreme caution.
Spain: Trends items should be interpreted carefully due to complications arising from the translation process.
United Arab Emirates:
Comparisons with TALIS 2013 must be limited to Abu Dhabi.
Because of the selection of multi-level schools, the principal data were copied from the original ISCED level 2 principal questionnaire to the ISCED level 1 and ISCED level 3 corresponding forms, except for question 17 in the principal questionnaire.
References
[1] OECD (2019), TALIS 2018 Technical Report, OECD, Paris, http://www.oecd.org/education/talis/TALIS_2018_Technical_Report.pdf.
[2] OECD (2018), OECD Handbook for Internationally Comparative Education Statistics 2018: Concepts, Standards, Definitions and Classifications, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264304444-en.
Note
← 1. Tables AII.A.3 to AII.A.8 display the participation rate estimates that were the most favourable for the adjudication rating. The most favourable estimates could have been weighted or unweighted depending on the characteristics of the country/economy, the teacher and principals population and the educational level.