Childhood Diet

In 1950 (4 years), descriptive information about diet was obtained by asking the mother or carer “What did this child have for each meal yesterday?” and referring specifically to 'Breakfast', 'Dinner', 'Tea or high tea' and 'Last thing at night'.  There were few indications of amounts of food.   A further question was “Do you give this child food between meals?”  It had a “Yes”, “No” or “No answer” response, but no further information.

The derivation of food and nutrient intakes was undertaken with MRC Human and Nutrition Research (HNR) (now closed) in Cambridge. The diet records were coded using the HNR in-house program, DIDO (Diet In, Data Out) and nutrient analysis performed using the in-house suite of programs.  A preliminary scan of the text data files showed that there were many entries of non-specific “meat”, “vegetables” or “pudding”.  Representative compositions for these were created, using a 33% sample of the records to obtain frequencies where these foods had been specified.

Thus the composite “meat” was 57% beef and 43% mutton.  A composite “vegetable” was made up of 50% peas, 21% cabbage, 16% green beans and 13% carrots and a composite “salad” was 29% lettuce and 71% tomatoes.  Rice pudding, being the most common one, was assumed for unspecified “pudding”.  “Bread and butter” was frequently mentioned but margarine was rare.  However, in view of the rationing of fats, a composite was assumed, taken as butter (58%) and margarine (42%).  Tea was assumed to be taken with milk and sugar.  160 of the children ate school or nursery dinners.  From descriptions, a standard meal consisting of meat, potatoes, a vegetable followed by custard with stewed fruit or pudding was used for unspecified meals.  One third of a pint of milk was added since this was distributed free to all school or nursery attending children.

A few (less than 1%) of the dietary records showed supplements such as cod liver oil and vitamin drops.  These were not coded as it was not known how many other children might have been receiving supplements of which there was no record.


Apart from the relatively few that were reported by weight, portion sizes suitable for four year olds were allocated using the amounts of rationed foods available and Ministry of Health 1949 recommendations on feeding young children, which describes types of foods suitable for this age group and the quantities that should be given.  Since neither of these two sources separated boys and girls, no distinction could be made in the portion sizes for the two genders.

Although the nutrient content of foods in the 1989 DIDO database was satisfactory for many foods, others have significantly changed in composition since 1950.  Meat had a higher fat content in 1950 and recipes for meat dishes contained a lower proportion of meat to other ingredients.  Bread had a different composition due to the higher extraction rate of flour, and there were fewer nutrient fortifications to breakfast cereals.  More appropriate nutritive values from earlier food tables (McCance and Widdowson, 2nd and 3rd editions) were therefore used for these and other typical foods of the time.


Key publications

Prynne CJ, Paul AA, Price GM, Day KC, Hilder WS & , Wadsworth MEJ (1999): Food and nutrient intake of a national sample of 4-year-old children in 1950: comparison with the 1990s Public Health Nutr. 2, 537–547


Obtaining the standard childhood diet variables

  • You can obtain a list of the standard topic variables to use in an NSHD data sharing request by selecting the link at the end of this page.
  • The XML version of the link can be sent to the LHA as part of a data sharing request.
  • Simply use your browser's Save As facility to save the list of variables to a file on your computer.
  • You can then email the XML version to us in support of your application, or when you fill in the data access application form and it asks which variables you need, mention the 'childhood diet standard basket' along with any others.


Please note that the standard childhood diet variable basket is very limited, containing only the two original questionnaire variables. The vast majority of the childhood diet data are the derived variables, which can be found by 'searching by library' for 'NewDiet50' in Skylark.

HTML version of the standard variables to view

Download the XML for the childhood diet standard basket