TEDS Data Dictionary

Introduction

The data collected in the 16 Year GCSE Study presented several challenges for data entry:

• The results included a large range of different qualification types, in addition to the expected GCSEs. Some of these types were unexpected in the first cohort, but were easier to predict subsequently.
• The qualifications included some very rare types, and some types that were difficult or impossible to equate to GCSEs in terms of grade and workload.
• Twins sometimes recorded qualifications in an approximate or confusing way. For example, the qualification type was sometimes recorded wrongly, or ambiguously; separate module or unit results were sometimes recorded in addition to final grades; important details such as level or award type were sometimes omitted; occasionally the same qualification was duplicated.

The cohort 1 exam results form (pdf) was very simple and relatively unstructured. This gave twins the flexibility to record all qualification types. However, data entry was difficult and slow due to the lack of structure, particularly for GCSEs where subjects had to be selected from a long list.

Experience of data entry in cohort 1, and analysis of the data, enabled improvements to be made. The cohort 2/3/4 exam results form (pdf) was more structured, with the most common GCSE subjects set out in a list, and with the more common other qualification types having their own tables. The data fields were re-structured in line with this new form, allowing faster and simpler data entry.

The information under the headings below explains aspects of the main qualification types that are being entered from cohort 2 onwards, and summarises the fields used to store the raw data.

Qualification Types

The table below summarises the more common types of qualifications that were recorded by twins at age 16. The listed types were all accredited by Ofqual at the time of data collection; see the 16 Year measures page for references. The tabulated types were all sufficiently common that data entry structures were designed for them.

Various qualification types at age 16 that were much rarer (and hence not tabulated here), but that were nevertheless accredited by Ofqual, include Foundation/Higher Extended Project, Foundation/Intermediate FSMQ, NCFE, CACHE, IFS and others. Because of the rarity of these types, they are recorded only in an approximate way at data entry (see data entry fields below).

NQF Levels

UK government guidelines set out the level at which any accredited qualification can be recognised in England, Wales and Northern Ireland. Most qualifications at age 16 are at levels 1 and 2 (GCSEs or exams at an equivalent level). A few are at entry level (below GCSE level), for low achievers at age 16. A few more are at level 3 (above GCSE level), for high achievers who have progressed beyond GCSE level by age 16. The levels relevant at age 16, and the main qualification types at each level, are tabulated below.

Note that many common qualification types (including BTEC, OCR Nationals, Functional Skills, etc) are available at more than one level. It was therefore important to recognise the level when entering the data.

Grades

GCSE passes are graded from A* to G; other qualification types (e.g. NVQ, Functional Skills) may have a simple ungraded pass at a given level; other qualification types (e.g. BTEC, DIDA) have their own distinctive pass grades. In the TEDS data, an attempt has been made to unify all the various types of pass grades using a single, ordinal, numerical code with values 1 to 17. GCSE grades G to A* (or equivalents in other qualification types) are coded with values 4 through to 11; sub-GCSE Entry Level Certificates are coded with values 1 to 3; qualification types above GCSE level are coded with values 12 to 17. The table below summarises the coding conventions used in the TEDS data, and the types of pass grade for each common qualification type.

Workload equivalence

Different qualifications at the same level, even if they have equivalent grades, may not be equivalent in workload. For example, a short-course GCSE is equivalent in workload to half a single-award GCSE. In the TEDS data, workload equivalences are useful for computing composite variables such as mean grade or total point score across different qualification types.

For some qualification types, the GCSE workload equivalence was explicitly given in official documentation, for example a level 2 NVQ was stated to be equivalent to 5 GCSEs, while a BTEC Certificate was equivalent to 2 GCSEs. For other qualification types, where no such guidelines could be found, the workload could be inferred from documented parameters such as the number of "guided learning hours". For example, a single-award GCSE typically has 120-140 guided learning hours, so a Certificate in Preparation for Working Life with 60-80 guided learning hours can be assumed to be equivalent to approximately half a GCSE. The table below summarises the workload equivalences that are assumed in the TEDS data for different qualification types.

Data Entry Fields

The table below summarises the fields used in the data entry of the 16 Year examination results, and shows the names of the item variables that relate directly to these fields. The fields are designed broadly to follow the layout of the cohort 2 exam results form (pdf). The choice of data entry fields implies that certain qualification types are recorded in an approximate way - any approximations used are described under another heading below.

All data were entered in numerically coded form: for further details, see the GCSE raw data coding (pdf). Grades and pass levels were entered using the codes described above (see grades). Awards (where variable) are entered as the number of GCSE workload equivalences (see workload equivalence above). Subjects and qualification types (where variable) are entered using predetermined numeric codes, which are recorded in variable value labels; see also subject coding below.

For the purpose of analysis, a systematic set of composite variables has been derived from the large set of raw variables tabulated below. For details of these composites, see the derived variables page and the GCSE dataset variables (pdf).

In the latter three fields, the numbers of miscellaneous 'other qualifications' are entered such that each additional other qualification is worth half to one GCSE. These fields are used for recording rare cases of accredited qualifications that cannot be accommodated in the other data fields. Examples of rare types that have been recorded by twins and that may be entered here include:

• FSMQ Foundation (level 1) or Intermediate (level 2)
• Foundation Project (level 1) or Higher Project (level 2)
• CACHE (Council for Awards in Care, Health and Education) childcare awards
• NOCN (National Open College Network) awards
• IFS (Institute of Fiscal Studies) awards
• IAM (Institute of Administrative Management) IT awards

Qualifications Not Entered

There are several categories of awards and qualifications that have not been entered in the TEDS data. These are generally rare types that were recorded by a small minority of twins.

Scottish Standard Grades were not entered because there was no obvious way of equating these with GCSEs. Families not living in England, Wales and Northern Ireland were generally not asked for their twins' examination results at age 16.

Some twins apparently completed their GCSEs early, and at age 16 had started courses that were generally designed for post-16 students. These included AS-levels and FSMQ Advanced, which were sufficiently common that they were entered in the data (see above). However, other post-16 qualifications were more exceptional and were not entered. Such qualifications were subsequently recorded at age 18, once completed (see the 18 Year Study). Examples of post-16 qualifications occasionally recorded at age 16:

• IB (International Baccalaureate) units
• Welsh Baccalaureate units
• Apprenticeships
• AS-levels, A2-levels or complete A-levels
• Level 3 vocational courses such as BTEC or NVQ
• Any qualification started but not yet completed (no grade given)
• Any component module or unit results contributing to an overall examination

Qualifications not accredited by Ofqual were not entered, because they were difficult or impossible to equate to GCSEs in terms of level, grade and workload. Such qualifications include the following types that were occasionally recorded by twins:

• BSC (British Safety Council) awards, e.g. Health and Safety at Work
• ASSET language courses
• DELF French courses
• JFO (Junior Football Organisers) sports courses
• ELC (English Language Centre) English courses
• INGOT (International Grades Open Technology) IT courses

Extra-curricular awards generally were not entered, for several reasons. Firstly, twins were not explicitly invited to record such awards, and only a minority did so. Secondly, some of these extra-curricular awards are not accredited by Ofqual, and are therefore difficult or impossible to equate to GCSEs. Thirdly, even in cases where they are accredited, these awards generally have a small workload relative to GCSEs, so they make only a minor contribution to an individual's examination achievement. Typical examples recorded occasionally by twins include:

• Graded music exams (awarded by RSM, Rock School and others)
• Graded dance/ballet/tap exams (awarded by RAD, ISTD and others)
• Graded drama/elocution exams (awarded by VAPA and others)
• Duke of Edinburgh awards
• Sports awards if not taken as accredited school examinations
• Other extra-curricular awards
• School-specific awards and certificates

Subject coding

For some qualification types, the subject is not variable and is therefore not coded or recorded explicitly. An example of such a type is a DIDA, where the subject is digital applications.

In other qualification types, for example Functional Skills and Sports/Dance Leadership, the range of subjects is very limited; in these cases, subjects are coded using predetermined numeric codes, as shown in the raw exam results coding (pdf).

For some other types, each subject has its own variable, so that explicit subject coding is not needed. This includes over 20 of the most common GCSE subjects; English/literacy and maths/numeracy skills qualifications; and English, maths and science Entry Level Certificates.

For the common vocational qualifications (BTEC, OCR National, City & Guilds, NVQ), the range of subjects was very large, and was subject to change year by year. The subjects were especially varied in the 'technology' and 'vocational' categories. The subjects were been simplified into 7 groups or subject types as tabulated below. This subject classification was done during data entry.

The same 7 subject types have been used to code GCSE 'other subject' variables, for the less common GCSE subjects. These codes are used both for full GCSE 'other subjects', of which up to 4 may be recorded, and for half-GCSE 'other subjects', of which at most 1 may be recorded for each twin. The list of subject types, shown in the table below, is designed broadly to cover all categories of subject at GCSE level.

Approximations used

During cohort 1, a complex and all-inclusive data entry process was used in an attempt to capture all qualification types in full detail. This process was reviewed on the basis of an analysis of the cohort 1 data, and an evaluation of problems occurring during data entry (which had been slow and laborious). The result was a new data entry process used in cohort 2, with the data entry fields described above. The cohort 1 data have been restructured so that they are contained in the same variables as the cohort 2 data.

The data fields and subject lists used in cohort 2 were designed to simplify and speed up the data entry process, while capturing all the most common qualification types, all the most common GCSE subjects, and all achievement in the core subjects of English, maths and science. These choices imply that approximations and loss of detailed information will occur when recording some qualification types. However, analysis of the cohort 1 data suggested that each such approximation is likely to affect a small minority of twins, generally well under 1%.

Approximations leading to some loss of data will occur in the following cases:

• Certain GCSE subject sub-types (in DT, ICT and Art & Design) are each recorded under a single subject heading
• Rarer GCSE subjects are recorded as broader subject types (see above), in the 'other' GCSE fields
• In rare cases, twins may have too many qualifications to be entered in the available data fields:
  1. More than one 'other subject' short-course GCSE
  2. More than four 'other subject' full GCSEs
  3. More than three BTEC, OCR National and/or City & Guilds qualifications
  4. More than one 'other subject' Functional Skills, Key Skills, Skills for Life, ALAN or ASDAN qualification
  5. More than two AS-levels
  6. More than one NVQ
  7. Entry Level Certificates in subjects other than English, maths and science
  8. Both Sports Leadership and Dance Leadership qualifications
  In cases of this sort (most of which were extremely rare in cohort 1), the additional qualifications have to be entered in an approximate form using other available data entry fields.
• For the rarest qualification types, entered in the last four fields as 'others' at the relevant level, the only information recorded is the level and the approximate workload equivalence. Other information, including qualification type, subject and grade, is lost.

During the data entry process, certain assumptions must sometimes be made when a twin fails to record all the necessary details of a qualification. For the sake of consistency, these assumptions have been formalised as a set of general rules for data entry. Some of the more commonly-applied assumptions are as follows:

• If more than one result is recorded for the same subject (in GCSEs, or in other types), then only the highest-graded result is recorded. This may arise when twins re-take exams, or when they mistakenly record the same result twice.
• If results are recorded in the same subject (typically RE or ICT) for both short-course and full GCSE, then only the full GCSE result is recorded. This may arise when a twin takes a short course then later adds modules to convert it to a full GCSE.
• If GCSE results are recorded both for Science (core and additional) and for separate sciences (biology, chemistry, physics), then only the separate science results are recorded. This may arise in modular courses, when a twin adds modules to convert double-award science into separate sciences.
• In GCSEs and AS-levels, a grade cannot be assumed if the twin has not recorded one.
• In vocational qualification types, if a twin does not record the level and/or the grade, a level 1 pass is assumed.
• In BTEC, OCR National and City & Guilds qualifications, if a twin does not record the award type then an award equivalent to two GCSEs is assumed.