TEDS Data Dictionary

18 Year Study

Contents of this page:

Introduction

The 18 Year study data were collected in the following ways:

  • Twin contact details.
    Twin address, telephone and email details, collected as twins reached the age of 18 years.
  • Twin questionnaires.
    Post-16 qualifications, plus details of university or work plans, collected on publication of A-level results.
  • Perception web study.
    A short battery of twin activities, comprising two tests and a questionnaire.
  • Bricks web study.
    A short battery of twin activities, comprising six sub-tests and a questionnaire.
  • Kings Challenge web study.
    A battery of 10 twin activities to test spatial abilities.
  • Navigation web study.
    A set of 30 related game-like activities to test navigational spatial abilities.
  • Fashion, Food and Music Preferences (FFMP) web study.
    A short questionnaire administered in a web study.

The twin contact details were collected on a weekly basis as twins reach the age of 18 years. A contact details form was sent with a birthday card to each twin. This part of the study started in January 2012 and continued until December 2014.

The questionnaire (including A-level results) was collected year by year each summer for cohorts 1, 2, 3 and 4. This study was planned to coincide with the release of examination results each August, at the end of the school year when twins reached the age of 18.

The Perception web study was carried out for cohort 1 twins (aged rougly 19 years) between October 2013 and February 2014.

The Bricks, Kings Challenge and Navigation web studies were carried out for selected twin pairs in cohorts 1 and 2 (aged roughly 20-21 years) between February 2015 and May 2016.

The Fashion, Food and Music Preferences (FFMP) web study was carried out for cohort 3 twins (aged roughly 19 years) between March and April 2015.

The measures used in the study are described in full in a separate page.

The sample

The sample selected for sending twin contact forms included all TEDS families except for the following: (a) families withdrawn from TEDS; (b) "inactive" families, who had not returned data in any previous studies; (c) families with address problems, that could not be traced. Medical exclusions and other special cases were generally included.

The sample selected for the questionnaire study was essentially the same, with the further removal of some medical exclusion families that had not been active for a long time. The latter group of families were sent letters before the questionnaire mailing, asking them to confirm whether they wished to continue in TEDS; those who did not reply were withdrawn.

Twin contact forms were mailed to 10797 of the 16810 families from the original TEDS sample from ONS. Hence there were around 6000 families not contacted. Of these, roughly 1730 had withdrawn; 2300 were inactive; and around 1970 were address problems. A further 200 families were excluded from the questionnaire mailing because of severe medical conditions or other special circumstances and lack of recent activity.

The Perception study sample involved an initial pilot study followed by the main study. The sample for the pilot study included 200 families: 100 from cohort 1 and 100 from cohort 2; these families were randomly selected from each cohort, although problematic families were avoided (for example, address and phone problems, medical exclusions, special cases and overseas families). The sample for the main study involved the entire remainder of cohort 1, with the same exceptions as for the questionnaire study.

The Bricks study sample included only families in which both twins had completed the entire battery in the 16 year web study (which involved cohorts 1 and 2 only). A small number of families had to be excluded because they had withdrawn or become address problems since the 16 year study.

The Kings Challenge study sample broadened the Bricks sample by including all families in which both twins had provided some 16 year web data (not just, as in Bricks, those who had completed the entire battery). As in Bricks, this restricted the sample to cohorts 1 and 2, and excluded families who had withdrawn or become address problems.

The Navigation study sample was broader still, including all contactable families in cohorts 1 and 2, regardless of participation at age 16, except for those who were involved in the TEDS21 feasibility studies at around the same time. Other exclusions were the same as those for the questionnaire study.

The FFMP (Fashion) study sample involved all families in cohort 3, with the same exceptions as for the questionnaire study.

The data returns for the 18 Year study are summarised in a separate page. There are further pages comparing samples and returns across different TEDS studies.

Data collection

Summary table

School Cohort: 1 2 3 4
Twin birth dates: Jan-94 to Aug-94 Sep-94 to Aug-95 Sep-95 to Aug-96 Sep-96 to Dec-96
Twin contact mailings: Jan-12 to Aug-12 Sep-12 to Aug-13 Sep-13 to Aug-14 Sep-14 to Dec-14
Families in twin contact study: 2365 3883 3376 1173
Questionnaire mailings: August 2012 August 2013 August 2014 August 2015
Families in questionnaire study: 2323 3827 3301 1137
Perception invitations: June and October 2013 June 2013 - -
Families in Perception study: 2282 100 - -
Bricks invitations: February 2015 - -
Families in Bricks study: 827 1357 - -
Kings Challenge invitations: May 2015 - -
Families in Kings Challenge study: 1246 1796 - -
Navigation invitations: September 2015, January to April 2016 - -
Families in Navigation study: 2009 3500 - -
Fashion invitations: - - March 2015 -
Families in Fashion study: - - 3166 -
Approximate mean twin age (years): 18 (contact forms), 18.5 (questionnaires), 19 (Fashion), 19.5 (Perception), 20 (Bricks and Kings Challenge), 21 (Navigation)

Timing of data collection

The twin contact forms were sent to twins along with 18th birthday cards. Where possible, these mailings were carried out weekly, to groups of twins whose 18th birthdays occurred in the coming week (around a hundred twin pairs per week).

The questionnaire was designed to collect A-level and other post-16 qualifications, along with other information. The questionnaire mailing was therefore timed to coincide with the release of school examination results in August each year, at the end of the school year when twins reached age 18. Hence the mailings have been carried out annually in August, for school cohorts 1, 2, 3 and 4.

The Perception pilot study was launched in June 2013, to coincide approximately with the end of A-level exams (for cohort 2 twins) and the end of the university term (for cohort 1 twins). The main study was then launched in October 2013, for the remainder of cohort 1 families. In both the pilot study and the main study, data collection continued for about three months, allowing families to participate around events such as holidays, university terms and examinations.

The Bricks study was started in February 2015. Data collection continued for 4 weeks. A large majority of the twins in this study were contactable by email, so the invitations and reminders could be received even by twins away from home (for example, at university). Furthermore, the twins had all been active in the 16 year web study, and were therefore known to be willing participants. These factors allowed the study to be completed in a relatively short period of time with good data returns.

The Kings Challenge study was started in May 2015, and continued until the end of August 2015 (nearly 4 months). The initial sample comprised a subset of the Bricks study sample, namely those with viable email addresses, and they were invited by email only. However, after disappointing initial returns the sample was broadened as described above, and the second wave of invitations were sent by post.

The Navigation study was first launched in September 2015 for a first wave of 300 families, but was then halted due to technical problems. The 2nd wave of 300 families was invited in January 2016; the more serious technical difficulties seemed to have been resolved, so this was followed by the bulk of the invitations (nearly 4500 families) in February 2016. A further wave, of several hundred families held back due to TEDS21 piloting, were invited in April 2016. The study ended in early June 2016, hence it continued for nearly 5 months for most twins. Twins were invited by email where possible, or by post where email addresses were not available.

The Fashion, Food and Music Preferences study was started in March 2015. Data collection continued for 4 weeks. Invitations were sent by post. A rapid return of data allowed this study to be finished promptly.

No consent form was used for the twin contact mailing. The forms were completed either by twins themselves or by their parents, with implicit consent.

In the questionnaire study, a paper consent form (pdf) was sent to each twin along with the questionnaire. This consent form was designed to collect updated twin contact details, along with consent to use the questionnaire data.

In the web studies (Perception, Bricks, Kings Challenge, Navigation and Fashion), on line consent forms were used. Twins were sent login details by mail and/or email (see collection methods below), with a paper information sheet if contacted by mail, but no paper consent form was sent. For each web study, the on line login and consent pages each contained a link to an on line version of the appropriate information sheet. The consent form was designed to collect twin email addresses as well as consent for participation in the study. The login and consent pages are described in detail for the Perception, Navigation and Kings Challenge studies; the login and consent pages for the Bricks and FFMP studies were very similar to the latter, and are not reproduced here. The information sheet for each web study can be viewed using the links above.

By the time of the questionnaire and web studies, all twins had reached the age of 18 years and were therefore adults. It was therefore necessary to ask twins but not parents for consent. Parents only participated by responding to reminders and phone calls (for the contact forms and questionnaires), with the implicit consent of the twins. In the web studies, email reminders were sent to parents having viable email addresses, asking them to encourage their twins to participate.

Collection methods

The twin contact study took place between January 2012 and December 2014. Mailings were carried out weekly or occasionally fornightly, timed to coincide as closely as possible with the week before twins reached their 18th birthdays. Two copies of the contact form (pdf), one for each twin, were mailed along with two birthday cards and a single return envelope. This mailing was addressed to the twin pair (not to the parent as in previous studies). A single written reminder, this time addressed to the parent, was sent to families in which twins did not reply promptly; this reminder was sent with two further copies of the contact form.

The questionnaire study was started in August of each year for the relevant cohort. The mailing was timed to coincide with the official publication of A-level results. Each family in the mailing was sent a pack containing a return envelope with two copies of a letter and consent form (pdf) and two copies of the questionnaire, one for each twin. The cohort 1 version of the questionnaire was found to have some shortcomings and was improved in the cohort 2 version. The main sections of the questionnaire then remained unchanged, but an additional section of questions was added in the cohort 3 version, and this section was slightly amended in the cohort 4 version (pdfs).

The questionnaire mailing was addressed to the twins, not to the parent. For some families, the questionnaires were posted to separate addresses for the two twins. These were cases where we learned of separate addresses as a result of the twin contact study. A single written reminder, this time addressed to the parent, was sent where twins did not reply promptly. Families that had not returned their questionnaires within a month or so were telephoned in order to collect as many returns as possible (see telephone calling below). It was not possible to telephone all non-responsive families in cohort 1, so only the most active families were targetted; however, all contactable non-responsive families in cohorts 2, 3 and 4 were telephoned.

In the Perception and Bricks studies, the invitation mailing (containing a letter with login details and an information sheet) was posted to individual twins rather than to families, even where twin pairs were known to live at the same permanent address. This was done because so that parents could forward the invitations to individual twins at university where necessary. In the Fashion study, the invitation mailing was posted to twin pairs where they still lived at the same address, in the hope that most twins would be at home (the study coincided with the start of the Easter holiday). At the same time as the mailing, in each of these studies, twins with known email addresses were sent email messages (containing a link and login details) inviting them to participate; the invitation message explained that the information sheet was being sent by post, and was also available on line. In the Perception study, twins were sent up to two email reminders and a written reminder to encourage them to take part; telephone reminders were also used in a limited number of cases (see telephone calling below). In the Bricks and Fashion studies, written reminders and calling were not used; however, a single email reminder was sent, both to individual twins and to parents, in cases where viable email addresses existed.

In the Kings Challenge study, twins in the initial sample were invited individually by email (this sample only included twins with viable email addresses). This avoided the need to send postal invitations: twins were invited by email, and reminded by email. When data returns failed to meet the expected level, the sample was broadened and a second wave of invitations were sent, this time by post, and again addressed to individual twins rather than to pairs. In this study, up to three email reminders were sent to individual twins, and a single email reminder was sent to parents (the reminders could therefore only be sent to twins and parents who had viable email addresses).

In the Navigation study, around 3200 pairs of twins were invited individually by email. For those without recorded email addresses (over 2300 pairs), paper invitations were sent by post, and these were addressed to individual twins (not pairs) as in Perception and Bricks. Although no written reminders were sent in the Navigation study, a series of email reminders were used (the number of reminders differing for different invitation waves). For twins with mobile phone numbers recorded, text invitations were also tried. As in Kings Challenge, an email reminder was also sent to parents to encourage their twins to take part.

Telephone calling

Some families were contacted by telephone during the months following the August questionnaire mailing. Families were given time to respond under their own steam, after the initial mailing and the written reminder, before calling was started. Attempts were made to call either the parents or the twins of targeted unresponsive families. In cohort 1, it was only possible to call 221 families - these were prioritised families who had returned recent data in the 18 year contact study. In cohorts 2 and 3, an attempt was made to call all unresponsive families, and a total of 3886 families were allocated to callers. Families generally excluded from calling were withdrawn and opted-out families, known phone problems, medical exclusions and special cases.

In the Perception study, callers attempted to contact 412 families. Because calling capacity was limited for this study, families were only called if one twin had finished but the other twin had not, or if a twin had started but failed to complete the battery. As in the questionnaire study, families were given the opportunity to finish under their own steam, and reminders were sent, before calling was started.

Callers were TEDS employees. In some cases, callers were able to prompt families to return the paper questionnaires that they had been sent; in other cases, callers collected the questionnaire data directly by telephoning, recording them onto fresh copies of the questionnaires.

Twin rewards

No rewards were offered for the completion and return of contact forms and questionnaires. However, both vouchers and entries into prize draws were offered as rewards in the Perception, Bricks, Kings Challenge, Navigation and Fashion web studies.

On completion of the web battery (in each of these studies), each twin was sent a £10 electronic Flexecode voucher. The voucher was sent by email, using the email address recorded by the twin in the consent form. Flexecodes had two advantages over the Love2Shop and iTunes vouchers used in the 16 year study: firstly, they could be redeemed with a range of retailers, hence twins could choose for themselves where to spend them; and secondly, sending the vouchers by email avoided postage costs.

There were two prize draws for the Perception study: one for the pilot study (jointly with participants in another pilot, for the Music study), and one for the main study. The Bricks and Fashion studies each involved just one prize draw. In the Kings Challenge study, there were two prize draws for the two waves of invited twins respectively. In the Navigation study, there were four prize draws for successive waves of invited twins In every prize draw, each twin was given a single entry for completion of the web battery. A single twin was selected at random as the winner in each draw. For each twin drawn as a winner, the co-twin was given the same prize provided that s/he had also completed the battery. If the co-twin had not completed the battery, another singleton twin was selected for the second prize.

Data entry

General data entry issues (for all studies including the 18 Year) are described in a separate page.

Data from the twin contact study are purely administrative, consisting of twin addresses, telephone numbers and email addresses. These data have therefore been recorded as updates to the TEDS administrative database, and have not formed part of the dataset. These contact details were entered manually, by TEDS office staff. Updates to contact details in the questionnaire consent form were handled in the same way.

Other administrative data, both for the twin contact study and the questionnaire study, included mailing dates, return dates, reminder dates and other contacts with families. These also were entered manually in the TEDS admin database.

The questionnaire data were entered by manual keying into an Access database, by staff in the TEDS office. Manual keying was essential because of the complexity of some of the data - data entry often required extensive interpretation and coding by the data entry staff. For example, qualifications or further education choices were sometimes recorded ambiguously, and key details were sometimes omitted by twins. Details of qualification types, and how they were entered, are described on the 18 year exam results page.

Data from the Perception, Bricks, Kings Challenge, Navigation and Fashion web studies were effectively entered by the twins themselves. As they answered items on their computers, their responses were recorded on the TEDS studies web server. The web server was programmed to produce, when required, a file containing all the web activity data. The data file is a plain text file, with tab-delimited variables, containing one row of data for each twin who completed the test. The data file was copied from the web server when needed for construction of the dataset.

The Access database into which questionnaire data have been entered contains all the cleaned raw data from the 18 Year questionnaire study, and is the direct source of data for the analysis dataset. The data entry process incorporates data integrity rules that ensure clean data are entered. For example, only valid twin IDs may be entered, and may not be duplicated; only valid categorical responses may be entered; and consistency is enforced in questions with multiple parts. For a discussion of general data cleaning issues in the TEDS data, see the data cleaning page.