|Twin birth dates:||Jan-94 to Aug-94||Sep-94 to Aug-95||Sep-95 to Aug-96||Sep-96 to Dec-96|
|Web study mailings:||September 2010||June 2011||(not applicable)|
|Families in web study:||2382||3899||(not applicable)|
|GCSE results mailings:||August 2010||August 2011||August 2012||August 2013|
|Families in GCSE study:||2318||3801||3319||1174|
|Behaviour/LEAP study mailings:||Wave 1:
|Families in behaviour/LEAP study:||2340||3864||3462||1208|
|LEAP-2 study mailing:||June 2012|
|Families in LEAP-2 study:||426||543||587||217|
|Approximate mean twin age (years):||16.5 (web)
Timing of data collection
Collection of GCSE and other school qualification data was timed to coincide with the release of school examination results in August each year, at the end of the twins' final compulsory school year (when they had reached age 16).
Initial plans for the web study, in cohort 1, were to contact families soon after twins had completed their GCSEs and other school examinations, at the start of the summer holiday. Unfortunately, the web tests were not ready in time for this, and the start of the web study had to be put back to September. In cohort 2, however, families were contacted from mid-June onwards, to coincide with the end of GCSE exams.
The behaviour/LEAP study was administered separately from the web study, and was timed to start soon after the web study finished. In cohort 1, the web study officially ended in January 2011, so the behaviour questionnaires were mailed in mid-February in time for the school half-term holidays. In cohort 2, the web study officially ended in November 2011, so the LEAP questionnaires were mailed in early December 2011 in time for the school Christmas holidays. The cohort 3 and 4 LEAP questionnaires were sent simultaneously with cohort 2.
The LEAP-2 study sample selection was based on analysis of the data collected in the behaviour/LEAP study, so data collection did not start until roughly six months after the (wave 2) LEAP study had ended. The start of the LEAP-2 study was timed to coincide with the beginning of the school summer holiday in 2012, for all cohorts.
No consent form was used for the GCSE results. The examination results forms were mailed to parents, and were completed either by parents or by twins themselves, with the implicit consent of the parents.
In the web study, on line consent forms were used. Families were mailed login details and an information sheet, but no paper consent form. Each family was sent three logins: one for the parent(s), and one each for the twins. Parents were required to log in first in order to complete the on line consent form (with the SES questionnaire); twins were prevented from proceeding with web activities until the parental consent was submitted. Once parents had consented, twins were able to log in and start their activities. The parent consent form was linked to a web version of the information sheet (pdf); the parent information sheet also contained links to read-only versions of the twin environment and wellbeing questionnaires. Twins were not required to complete a consent form themselves, but they were provided with a twin web version of the information sheet (pdf), linked to their activity page.
In the behaviour/LEAP and LEAP-2 studies, a paper consent form was sent to families along with the parent and twin booklets and an information sheet. This consent form was designed to collect consent from the twins as well as from a parent, by means of dated signatures. No other information was collected on the consent form.
In August of the appropriate year, just before the official publication of GCSE results, families were mailed a pack containing a letter, two GCSE results forms (one for each twin), and a return envelope. The results forms were designed to capture the results of all common types of school qualification at age 16, including not only GCSEs but also vocational qualifications (BTEC, OCRN, Key Skills, etc) and any AS-levels that had been completed after an early start to post-16 studies.
The design of the exam results form was changed from a simple but flexible cohort 1 version (pdf) to a more detailed cohort 2 version (pdf), which was re-used for cohorts 3 and 4. Further details of qualification types and how they were recorded can be found on the 16 year exam results page.
Families that had not returned their results forms within a month or so were telephoned in order to collect as many results as possible (see telephone calling below). Because all contactable families were phoned in cohorts 1, 3 and 4, no written reminders were sent. However in cohort 2, not all families could be phoned so the remainder were sent a written reminder with a web login allowing them to enter results on line if they had lost their paper forms.
The web study
At the start of the web study, each family was mailed a pack containing a letter with the family login details, an information sheet, and a flyer to publicize the prize draws (see below for details of rewards). Families were asked to log in immediately to give consent (see above) and start the activities.
As described below, families that had not completed the activities within a month or so were telephoned by web callers to encourage them to do so. In addition, a written reminder was also sent shortly before the date of the final prize draw.
The behaviour/LEAP and LEAP-2 booklet studies
The behaviour/LEAP booklet study was administered in two waves: wave 1 (cohort 1 families) in spring 2011, and wave 2 (cohort 2, 3 and 4 families) in winter 2011/12. In all communications with families in both waves, the booklet study has been referred to as the Behaviour Study. However, within TEDS, wave 2 of the booklet study is officially called the LEAP Study (Longitudinal Experiences and Perceptions). The LEAP Study is sometimes also referred to as LEAP-1, to differentiate it from the later LEAP-2 Study.
The LEAP-2 booklet study was administered in a single wave, in June 2012, for all cohorts. In communications with the families, this has been referred to as the Behaviour Study Part 2, but within TEDS its name is the LEAP-2 Study.
The measures included in Behaviour, LEAP and LEAP-2 booklets overlap to a large degree. The measures in the parent booklets changed very little: the LEAP booklet was identical to the Behaviour booklet except for the addition of an asthma measure, while the LEAP-2 parent booklet is simply a shortened version with several measures removed; see parent booklet wave 1, wave 2 and LEAP-2 versions (pdfs). The measures in the twin booklet changed more substantially, with some measures completely removed and other new measures added; however many measures remained unchanged - see child booklet wave 1, wave 2 and LEAP-2 versions (pdfs). The measures page has full details of the measures included in the various booklets.
At the start of each wave of the Behaviour/LEAP booklet study, each family was mailed a pack containing a letter, an information sheet, the family's booklets, a consent form (pdfs), and a return envelope. Also included were two sealable envelopes in which the twins could place their completed booklets, to help them keep their responses private from the rest of the family. Families were asked to return their booklets and consent forms as soon as possible.
The materials mailed to families for the LEAP-2 booklet study were similar, except that the letter and consent form (pdf) were combined into a single document, and there was no information sheet.
Twins were rewarded for completing and returning their booklets - see below for details. In each wave of the study (Behaviour, LEAP and LEAP-2), a single written reminder was sent to families that did not return their booklets promptly.
No rewards were offered for completion of the GCSE results forms. However, both vouchers and entries into prize draws were offered as rewards in the web and booklet studies.
In the web study, each twin was rewarded with a £10 voucher on completion of the battery of web tests. Directly after the last test in the battery, twins were given a choice of two types of voucher (iTunes or Love2shop). Vouchers were sent to families by mail. In this study (unlike previous TEDS web studies), parents were not sent a voucher to cover internet connection costs.
There were four prize draws, on publicized dates at roughly monthly intervals, during each cohort of the web study. Each twin was given one prize draw entry for each web activity completed, plus a bonus entry for completing the entire battery. Entries were carried forward to all prize draws after the entries were awarded; hence there was an incentive for twins to finish the activities early. For each twin drawn as a winner, the cotwin was given the same prize (even if the cotwin had not started the activities).
In the behaviour/LEAP and LEAP-2 booklet studies, each twin was rewarded with a £10 voucher on return of the completed booklet. Twins were asked to indicate (on the front of the booklet) whether they preferred iTunes or Love2shop vouchers. Vouchers were sent by mail.
A single prize draw was carried out at the end of each wave of the behaviour/LEAP and LEAP-2 studies, on a publicized date. Each twin who had completed and returned the booklet before the date of the prize draw was given one entry into the draw. For each twin drawn as a winner, the cotwin was given the same prize provided that s/he had also returned the booklet.
Prize draw winners were contacted by telephone; with their permission, their names were published on the TEDS web site; they were also asked for photos to publish on the web site.
Some families were contacted by telephone at various times. Attempts were made to call selected unresponsive families in each cohort of the web and GCSE studies (selection was based on recent activity in other studies). Families generally excluded from calling were withdrawn and opted-out families, known phone problems, medical exclusions and special cases. Calling was not used at all in the behaviour study. In both the GCSE and web studies, families were given some time to return data under their own steam before callers were allocated. The same team of callers was used for both jobs, and where families were called for both purposes (in cohorts 1 and 2) they were allocated to the same callers. Callers were TEDS employees, and calling fell into two categories:
- GCSE results calling. The main role of these callers was to collect twin GCSE/examination results. They were not expected to act as web callers, although they were asked to mention the web tests by way of a reminder. In cohorts 1, 3 and 4, GCSE callers were allocated to 1111, 1679 and 638 families respectively: these were all the families that had not yet returned the GCSE results at the time of allocation. In cohort 2, GCSE callers were allocated to 1107 families: these were families that had not returned the GCSE results, but had been active in other studies at ages 16, 14 or 12. In some cases, callers were able to prompt families to return the paper results forms that they had been sent; in other cases, callers collected the results directly by telephoning, recording them onto fresh copies of the forms. There was no prioritization of the families allocated to each caller - callers were asked to do their best to contact all their families.
- Web calling. These callers had the job of encouraging families to complete the web tests. Web callers were allocated to 1180 families in cohort 1, and 1559 families in cohort 2. In both cohorts, callers were allocated to all families that had started the web activities (parent on line consent given) but not yet completed them at the time of allocation. Additionally, families were allocated to web callers if they (1) had returned their GCSE results prior to the allocation date; (2) had completed at least some of the web activities at age 14 or 12; or (3) in cohort 1 but not cohort 2, further families that had returned any other form of data at age 14 or 12. Callers were asked to make repeated telephone calls, where appropriate, to families in order to encourage twins to complete the activities; callers generally spoke to parents, but sometimes also directly to twins.
General data entry issues (for all studies including the 16 Year) are described in a separate page.
Data from the web study were effectively entered by the twins and parents themselves. As they answered items on their computers, their responses were recorded on the TEDS web server. The web server was programmed to produce, when required, "analysis files" containing the web test data. There was a specific analysis file for each twin web activity. Each analysis file is a plain text file, with comma-delimited variables, containing one row of data for each twin who completed the test. The analysis files were copied from the web server when they were needed for construction of the dataset.
Administrative data relating to the web study, for example start and end dates, and a log of whether or not various activities were completed, were recorded automatically on the web server. These data were downloaded regularly from the web server as "family status files" (csv text files), and incorporated into the TEDS admin database. The family status file also contained data from the parent SES web questionnaire; hence a copy of the family status file was copied from the web server (alongside the analysis files) when needed for construction of the dataset.
Other administrative data, such as mailing dates, return dates for GCSE results and booklets, and reported problems with web tests, were entered manually in the TEDS admin database.
The GCSE and other school examination results were nearly all returned on paper forms (either directly from families, or recorded by callers), and these data were entered by manual keying into an Access database, by staff in the TEDS office. These data were complex because of the large range of different qualification types gained by the twins; and also because of variability in the accuracy of the information written down by the twins or their parents. Manual keying of the data was therefore essential, because the data entry often required extensive interpretation and coding by the data entry staff. Details of qualification types, and how they were entered, are described on the 16 year exam results page. In cohort 2, on line data entry was set up to allow families to respond to a written reminder near the end of data collection (few families responded); while these data were entered by families themselves, some manual recoding of subject names and qualification types was necessary.
The behaviour/LEAP and LEAP-2 study booklets, for parents and twins, were designed for data entry using optical scanning. The booklets were quite lengthy, making them less suitable for manual keying. Questions were designed with categorical responses using tick boxes, which can be recognised easily by optical scanning. The twin booklets contained a very few questions with alphanumeric rather than categorical responses; however, responses were collected in boxes with one letter/digit in each box, and these also were scanned. At age 16, the twins were expected to be able to complete the tick-box responses reliably, so few scanning errors were anticipated.
The behaviour/LEAP and LEAP-2 study booklets were returned by mail directly to the TEDS office. They were then delivered in large batches to Group Sigma for optical scanning. After scanning each batch of booklets, Group Sigma returned the data in plain text files. The paper booklets and questionnaires were subsequently returned to the TEDS office. The files of raw scanned data returned by Group Sigma have been retained for future reference. After data cleaning, the scanned parent and twin booklet data were aggregated and stored in a Access database file, alongside the GCSE results data. Any booklets that were returned too late to be scanned were entered manually into the this database, directly alongside the scanned data. This Access database (which is the direct source of data for the analysis dataset) contains all the cleaned raw data from the 16 Year study, with the exception of the twin web activity data. The analysis files of twin web test data are stored separately because they are too large to be easily stored in database tables.
A systematic attempt was made in the 16 Year data to clean invalid or compromised data in the web activities. This involved attempts to identify instances of twins responding randomly in the activities, and to identify test items or entire tests where the data had been compromised because of timeouts or crashes. These cleaning processes were carried out in the dataset syntax files so that they should be clearly documented and also subject to change or reversal if necessary. See the web data cleaning page for further details.