TEDS Data Dictionary

4 Year Study

Contents of this page:

Introduction

The 4 Year study data were collected by means of booklets:

  • Parent booklets.
    Parent-reported data, relating mainly to the family (but with some twin-specific items).
  • Twin booklets.
    Parent-reported data specific to each twin, including various language and behaviour measures. Included the "Parca" parent-administered cognitive tests.

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

The 4 Year data were collected between 1998 and 2000 from all TEDS families (twins born between 1994 and 1996). Data collection was timed to coincide as closely as possible with the twins' 4th birthdays.

The sample

The 4 Year sample included all three birth cohorts (1994, 1995 and 1996). Families were generally excluded from the 4 Year sample if they had not returned the 1st Contact booklet. In addition, in the 1994 cohort only, families were excluded from the 4 Year sample if they had returned neither the 3 year nor the 2 year booklets. However, there were exceptions to these rules.

The 4 Year booklets were sent to roughly 12500 of the 16810 families in the original TEDS sample from ONS. Hence there were around 4300 families that were not sent the 4 Year booklets. Roughly 2000 of these had withdrawn from TEDS or were known address problems; the remainder were mostly families that had not returned the 1st Contact booklet, or families in the 1994 cohort that had returned neither the 2 year nor the 3 year booklets.

The data returns for the 4 Year study are summarised in a separate page. There are further pages comparing samples and returns for different TEDS studies.

Data collection

Each family in the sample was sent three booklets: a parent booklet plus two copies of the child booklet (pdfs). The parent booklet included a consent form on the first page. The measures used in the booklets are described in detail on another page. The measures were entirely parent-reported, although the child booklet included cognitive tests that were administered by the parents on the children.

The booklets were designed to be completed when the twins were precisely 4 years old; they were therefore sent to families at or just before the twins' 4th birthdays. This involved regular mailings of booklets (generally once per month) between December 1997 and December 2000.

Regular reminders were sent to families who did not return the booklets promptly. Up to 8 reminders were sent, over a period of up to 11 months after the original booklets were sent to each family.

Data entry

General data entry issues are described in another page. In the 4 Year study, data entry was handled externally by NOP Numbers, a commercial company. The data were returned in pre-formatted Excel workbooks containing multiple worksheets. It is not clear whether the data were entered directly into Excel, or whether the data were exported into Excel after data entry into some other software system.

Data entry staff at NOP carried out basic coding of the raw data by converting tick boxes to numeric code values - see 4 Year raw data item coding for the parent and child booklets (pdfs). The Drawing task items in the child booklet were scored by TEDS staff before the booklets were sent to NOP. The Drawing items (with the exception of the new "Draw a Man" task) were the same at age 4 as at age 3; the coding rules for these items are fully described in 3 Year Drawing coding (pdf).

The inside front cover of the 4 Year parent booklet asks for contact details of a relative or friend of the family; and the first page of the booklet is a consent form, asking for family name and address details. The verbatim text data from these two pages was entered and does exist in the raw electronic data (in one of the worksheets of each Excel file). However, this does not form part of the dataset, and was only entered so it could be used for administrative purposes. At the time of the 4 Year study, these data were used to update the TEDS admin database.

In the main body of the parent booklet there are some items where a free text response is invited. However, the verbatim text was only entered for a few of these items (sibling names on page 1, and descriptions of health problems on page 11). For the other items, the verbatim text was not entered and hence does not exist in the electronic raw data. The parent booklet raw data coding (pdf) clearly shows which text items were taken up. The child booklet contains only one question inviting a free text response, but the verbatim text was not taken up for this item.

The Excel workbooks containing the original raw data, as returned by NOP Numbers, still exist in most cases and have been retained; however a large portion of the original raw data for the 1994 cohort is missing (at least one workbook has been lost). All of the raw data, including the data from the "missing" workbooks, was copied, cleaned and aggregated into a single Access database file. This Access database is now treated as the master copy or source of all 4 Year data to be used in analysis. All the raw data files (Excel and Access) are described in full in 4 Year raw data files. The ways in which the raw data have been cleaned are described in another page.