*2.1. Subjects and Study Design*

The pilot study was carried out January–June, 2013. Twelve Somali- and Iraqi-born mothers with children aged 10–21 months living in Oslo and Akershus counties, Norway, were recruited through several methods: the Norwegian National Population Register, open kindergartens and by using the snowball method. Inclusion criteria were the mothers' country of birth and the child being approximately 12 months old, born in Norway and with no serious health problem or disease requiring a special diet. The mothers received a bilingual information letter and provided written consent. Respondents received a shop voucher after completing two recalls. The study was approved by the Norwegian Regional Committees for Medical and Health Research Ethics.

To measure the dietary intake among toddlers, a structured 24-h recall method was used [15]. In the 24-h recall method, the mothers were interviewed twice, usually 1–2 weeks apart, by trained field workers about the exact food and beverage intake of their child during the preceding 24 h. If other caretakers were involved, the mother was asked to obtain information about the child's food consumption while under their care. A researcher (C.M., M.B.A., or N.K.G.) was present and observed the interviews. In addition to the food and beverage intake of the child, information about the performance of the interviews and methods was collected through an evaluation form.

#### *2.2. The 24-H Recall Method*

#### 2.2.1. Picture Library for Food Identification

To help the mothers and the field workers identify the correct foods given to the child, a library with pictures of food items commonly eaten by children in Norway was developed. A list of food items to be included in the library was made based on knowledge about Norwegian children's dietary intake. In addition, food items identified as eaten by Somali and Iraqi children through an informal qualitative prestudy and food items suggested by the field workers were also included. The foods were photographed in supermarkets and independent shops owned by immigrants in Oslo, with permission from the owners/managers. A Canon Ixus 860 IS digital camera was used, and the pictures were edited in iPhoto on a MacBook Pro (Apple Inc., Cupertino, CA, USA).

The library contained pictures of a wide selection of industrially produced baby foods for children aged 8–15 months, as well as other foods and beverages. Before the pilot study, the field workers suggested adding Nido milk powder, different types of meat, cheese and biscuits. The "bread scale", a Norwegian labeling scheme for fiber and wholemeal content of bread, was also included in the library.

To ease retrieval of pictures during the interviews, the 336 unique pictures were categorized in 19 different folders on an iPad (Table 1). The folders contained between 4 and 50 pictures, with the largest number of pictures in the folder "fruits and vegetables". Because certain foods may be categorized differently by different people, some pictures were placed in more than one folder. For instance, smoothies appeared both in the folders "snacks" and "juice and nectar". Thus, the final library contained 405 pictures in 19 folders.


**Table 1.** Number of pictures in each folder of the picture library.

#### 2.2.2. Photographic Booklet and Measuring Equipment for Portion Size Estimation

This study used a photographic booklet for portion size estimation developed for the Spedkost and Småbarnskost surveys [8–10]. It included 17 color photograph series of selected food items representing different, usually four, portion sizes appropriate for toddlers ranging from small (A) to large (D), with up to six different portion sizes for baby cereal. The mother used the booklet as a tool to identify portion sizes eaten by her child. The field workers also brought a kitchen scale and three measuring cups (in deciliters and milliliters) to weigh or measure foods or volume in tableware from the respondents' homes whenever possible. Frequently, both methods were used to compare the results.
