A selected sample of births in 9 of the 36 municipalities in the Groningen clay area (about half the province of Groningen) and a double sample for the city of Groningen were taken. For every municipality the first 120 births were chosen from 1 August 1811, 1 January 1830, 1 January 1850 and 1 January 1870 onwards (5,280 persons). 1.Which source forms the basis for the sample? Civil birth certificates: the first 120 births from 1 August 1811, 1 January 1830, 1850, 1870 onwards were transcribed on a form: name child, name parents, occupation parents, signature father, date of birth, date of certificate and number, place of birth (municipality and village). Stillborns reported in the civil birth registration were not taken into account. The 1811 cohort starts 1 August, because at this date the civil registration started in the province of Groningen, before that date you have to search in baptism registers which give far less information (for instance no occupation). A problem with the 1811 cohort is also that the surnames were often nog correct, as a lot of Dutch people (especially in the countryside) only chose a surname in the years around 1811, and they also did not always use them. Actually, this was a major problem with the construction of the 1811 cohort, resulting in quite a lot of people leaving no trace except for the birth certificate. For some villages the baptism records were inspected, as they sometimes give other extra information on names. For the sample persons marriage and death certificates in the civil registration were searched for individually in the province of Groningen. At that time this was rather time consuming as every municipality had a ten year index (making searching easier, but not that easy), and overall indices were missing (at the moment excerpts of the whole civil registration of the province of Groningen have been digitalised: allegroningers.nl, and the same is the case for large parts of the rest of the Netherlands, for instance: wiewaswie.nl. This makes searching for lost people much easier than at the moment the database was generally constructed) The relevant information in these certificates was transcribed and put on a form. Also the sample persons and their family were looked for in the available censuses (usually 1830 and 1840) in the municipalities of birth. After 1850 for most of the municipalities a dynamic Population Register is available, which was renewed every 10 year (but not always, some Population Registers even stretched over 40 years: for instance 1860-1900). The information and changes in this information in the Population Register were also taken over on the form. Especially the names an birth dates of children of sample persons were taken over from the Population Register, and only controlled using the birth and death registration (especially to find the registration of stillborns and mortality of babies after a few days that was missing in the Population Register). Sample persons were followed after migration in the Population Register of the municipality of settlement (usually in Groningen, for the rural cohorts 1830, 1850 and 1870 also in the rest of the Netherlands). Migration to the USA (not always reported in the Population Registers) was also controlled using the available departure registers of the province of Groningen from 1846 onwards (which were, however, also not complete). 2. Sampling units: Individual sample persons were followed through their life. Nine rural municipalities were chosen for the selection procedure scattered rather evenly across the region. Zuidhorn and Hoogkerk in Westerkwartier district (west), Leens, Uithuizen and Bedum in Hunsingo district (northwest), Appingedam and Stedum in Fivelingo district (northeast), Beerta and Winschoten in Oldambt district (east). However, both rural centres (Appingedam and Winschoten) were selected. A double sample was taken for the large city of Groningen: 240 3. Variables used for selection: All children born irrespective of gender. 4. Selection method: Clustered (see before).