TEDS Data Dictionary

Summary of known pilot studies

The table below lists the pilot studies for which at least some minimal documentation remains in TEDS. It briefly describes the types of participants, whether a dataset has been retained within TEDS, and whether the piloted measures were documented in any detail.

As mentioned above, test-retest datasets have been retained where possible and where these have been documented. These datasets are described in somewhat more detail under the heading test-retest datasets, below.

Test-retest datasets

This section contains brief documentation of the pilot test-retest datasets that have been retained in TEDS. The measures or variables are described. In some cases, where the test and/or retest used the same measures as in the respective main study, they are documented using links to the study pages. In other cases, an annotated documented may be provided. In all cases, the participant IDs have been randomly anonymised, so the datasets cannot be linked to other data sources.

9 Year booklet test-retest dataset

This test-retest study was carried out in February 2004. This was roughly at the end of the main TEDS 9 Year study data collection (for cohort 2) and at the start of the main TEDS 10 Year study data collection (for cohort 1). The purpose of the test-retest study seems to have been partly to check reliability for measures in the 9 Year booklets, and partly to pilot measures for the 10 Year study (see measures listed below).

Participants in the test-retest study, for whom data have been retained, included 24 older siblings of twins, along with their parents; plus 7 children from the unrelated "Tracks" twin sample (one twin per pair), also with their parents.

The test-retest study was administered by post. The initial test booklets (parent and child) were sent with a short deadline for completion; when returned, the retest booklets were sent. The interval between test and retest varied between 2 and 9 days, with a median of 6 days. The measures in the test-retest booklet were:

• CEQ (Classroom Environment Questionnaire), 28 items, for both parents and children, as used in the 9 Year parent and twin booklets.
• Behaviour measures, 16 items, for parents only. This was a small subset of the items used in the 9 Year parent booklet (including some CAST, anxiety and SDQ items).
• Classroom CHAOS measure, 16 items, for children only. This measure was not used in the 9 Year study, but was used in the 10 Year teacher questionnaire.
• Author and Book Recognition test, 45 items, for children only. This measure was not used in the 9 Year study, but a modified version was used in the 10 Year twin web study.

It appears that the child "story" written task (as used in the 9 Year twin booklet) was also included in the test-retest study. However, neither data nor documentation have been preserved for this part of the test-retest study.

The measures are documented in a list of variables in the 9 year test-retest dataset (pdf).

10 Year PIAT pilot and test-retest dataset

Before the start of the 10 Year study, the PIAT web test, which had a novel presentation with branching and discontinue rules, was piloted. The final pilot involved a set of TEDS twin pairs, from cohort 1. This pilot was carried out between November 2003 and January 2004. It used essentially the same web version of the PIAT version as was subsequently used in the 10 Year study.

After piloting, the main 10 Year web study, including the PIAT alongside other tests, went ahead with all cohort 1 twins being invited. 148 of the twins who had completed the pilot went on to complete the main study as well. These 148 twins are the cases in the test-retest dataset; the "test" being the pilot, and the "retest" being the PIAT test in the main study.

Most cohort 1 twins took part in the 10 Year web test between June and August 2004. The interval between test and retest varied between 165 and 294 days, with a median of 237 days (roughly 8 months).

The variables in the retest are as documented in the 10 Year PIAT study page. The variables in the pilot test have the same numbering (19 to 100 for the main test items) but slightly different names; the "j" prefix is omitted, and suffixes "resp" (response) and "scor" (score) are used for the items.

12 Year paper retest dataset

Retests were carried out for 15 twin pairs who had completed the web tests promptly, after the start of cohort 1 testing, in April 2006. The 15 families were asked to visit the TEDS office, where the activities were carried out on paper. The test-retest interval varied from 26 to 98 days and its median was 65 days; it is measured with a variable in the dataset.

The measures retested are documented in a description of the 12 year paper retest activities (pdf). The activities retested were as follows:

• PIAT: 81 items, numbered 19-100, as in the original web test. Retest variable names (item scores) are pi19 through to pi100 and total score pitot.
• GOAL: 60 items. This included all items from the original test booklet, not just the 36 items that were in the original web test. Retest variable names (item scores) are go01 through to go60 and total score gotot.
• Yes-No (reading fluency): 98 items, as in the original web test. Retest variable names (item scores) are rf01 through to rf98 and total score rftot.
• Author recognition: 42 items as in the original web test. Item variables have not been retained in the retest data, just the scores artot (number of correct author names minus number of incorrect author names selected), authcor (number of correct authors selected) and authwr (number of incorrect authors selected).
• Maths test: 70 items. This included two sets of questions, labelled A and B, with 32 and 38 items respectively. These questions were taken from original test booklets, not just the subset of items that were in the original web test. Retest variable names (item scores) are mta01 through to mta32 and mtb01 through to mtb38. Retest score variables are mtatot (0-32), mtbtot (0-38) and mttot (0-70).

The dataset includes the retest variables as listed above, alongside the item and test scores from the corresponding original 12 Year web tests (with variable names as documented in the 12 Year study).

16 Year "validation study" retest dataset

Retests were carried out for 24 twin pairs who had completed the main tests promptly, after the start of cohort 1 testing, in November 2010. The 24 families were asked to visit the TEDS office, where the tests were carried out. The test-retest interval, which is a variable in the dataset, varied between 12 and 50 days, with a median of 27 days.

The testing involved repeating the activities already completed in the web study. Some repeat tests were done on paper and some using the web instrument. The retest measures were exactly the same as the original measures, except that some were reformatted for paper testing. Timing and administration of the paper tests were supervised by TEDS staff. The tests repeated on paper were:

• Ravens Matrices
• Mill Hill Vocabulary
• Expressions (Figurative Language)
• Passages
• Yes/No
• Environment and Wellbeing questionnaires A, B, C and D

The tests repeated via the web were:

• Understanding Number
• Dot Number
• Number Sense (Dot Task)
• Number Line
• PVT
• Corsi Block
• Reaction Times

The measures and variables are described in detail in the 16 Year study pages. Item variables have the same names as in the main 16 Year dataset. The retest variable names have been given the prefix "v" (validation) to distinguish them from the main test study variables, but are otherwise identical.

18 Year web studies retest datasets

There were five post-18 year web studies: Perception, Bricks, FFMP, Kings Challenge and Navigation (see 18 Year study). In each of these studies, the test activities were piloted with a test-retest study. For the Perception study, the retest was carried out after the start of the main study, using a sample of twins. For the other studies, test-retest pilots were carried out shortly before the start of the main twin study, using samples of siblings of the twins. In each case, test and retest were administered via the web on the TEDS web server; the test and retest data were downloaded from the web server and were used to build the datasets.

The pilot activities are not documented in detail, but they are assumed to have been essentially the same as the main twin study activities (see 18 year web activities page). The variable names are closely matched to, or at least analagous to, the documented variable names in the main twin datasets. In the SPSS test-retest datasets, the variables are also identified with labels. In each dataset, retest variables have the same names as test variable, with the addition of "R" (retest) as a prefix or suffix. Each dataset also has a variable named "testretestinterval" (or similar) measuring the number of days between test and retest for each participant.

The table below summarises essential details of each test-retest study dataset.