Food Expenditure Survey (FES)

Detailed information for 2001

The Food Expenditure Survey is a periodic survey designed to supplement the Survey of Household Spending (SHS) by providing a level of expenditure detail on food commodities not feasible in the context of the SHS methodology. Data from the Food Expenditure Survey and the Survey of Household Spending are used to update the weights used in the calculation of the Consumer Price Index.

Data release - February 21, 2003

• Questionnaire(s) and reporting guide(s)
• Description
• Data sources and methodology
• Data accuracy

Description

The Food Expenditure Survey is a periodic survey designed to supplement the Survey of Household Spending (SHS) by providing a level of expenditure detail on food commodities not feasible in the context of the SHS methodology. Data from the Food Expenditure Survey and the Survey of Household Spending are used to update the weights used in the calculation of the Consumer Price Index.

The Food Expenditure Survey is carried out in private households residing in urban and rural areas in Canada's 10 provinces as well as in the cities of Whitehorse, Yellowknife and Iqaluit.

Reference period: Calendar year

Collection period: January to December during reference period

Subjects

• Household spending and savings
• Income, pensions, spending and wealth

Data sources and methodology

Target population

The Food Expenditure Survey was carried out in private households residing in urban and rural areas in Canada's ten provinces as well as in the cities of Whitehorse, Yellowknife and Iqaluit. The survey covers about 98% of the population in the 10 provinces.

The following groups were excluded from the survey:
those living on Indian reserves and crown lands;
official representatives of foreign countries living in Canada and their families;
members of religious and other communal colonies;
members of the Canadian Armed Forces living in Military Camps;
people living in residences for senior citizens;
people living full time in institutions: for example, inmates of penal institutions and chronic care patients living in hospitals and nursing homes.

Information was not gathered from persons temporarily living away from their families (for example, students at university), because it would be gathered from their families if selected. In this way, double counting of such individuals was avoided.

Furthermore, for operational reasons, households residing in dwellings located in remote areas were excluded. These households represent less than 1% of the population described above but are difficult and expensive to contact.

Instrument design

Two instruments were used for data collection, a questionnaire and two diaries. The questionnaire collected demographic data for all members of the household as well as their eating habits. It also collected the habits of the household with regards to being away from home overnight or longer during the month preceding the survey. The diaries were designed to facilitate respondents' entering of the information while providing them with concrete information as to what was expected of them. The diary had to be appealing to the eye while enabling the collection of all the information required by all concerned parties. To attain this, the format was changed from 8½ x 14 to 8½ x 11. The questionnaire and diaries were designed by consulting employees of Prices Division, graphic designers and other employees of Dissemination Division, employees of the Agriculture and Health Canada. International experience regarding the collection of such information was also utilized. Finally, members of the section present when the last survey was conducted were consulted as well as the notes from the post-mortem. The questionnaire and diary were revised by the Questionnaire Design Resource Centre regarding plans for testing revised questionnaires. The few suggestions made by them were incorporated. The questionnaires were also submitted to Data Access and Control Services Division for approbation. The last time the questionnaire and diary were revised was for the last Food Expenditure Survey in 1996.

Sampling

This is a sample survey with a cross-sectional design.

The sample size for the 2001 Food Expenditure Survey was 9,488 dwellings.

This sample was a stratified, multi-stage sample selected from the Labour Force Survey (LFS - 3701) sampling frame. Sample selection comprised two main steps: the selection of clusters (small geographic areas) from the LFS frame and the selection of dwellings within these selected clusters. The LFS sampling frame mainly uses 1991 Census geography and 1991 population counts.

The sample was drawn for the whole year and then divided into monthly subsamples to evenly distribute data collection over the entire 2001 calendar year. This ensures that we get an accurate picture of food expenditures regardless of the season.

Data sources

Data collection for this reference period: January 1, 2001 to December 31, 2001

Responding to this survey is voluntary.

Data are collected directly from survey respondents.

The Food Expenditure Survey was conducted monthly during the 2001 calendar year. Data were collected by an interviewer by means of a personal interview using a paper questionnaire (FE2). Two weekly diaries (FE3) were also left for the respondent to complete daily for two consecutive weeks.

The questionnaire was used primarily to collect selected socio-economic characteristics (including income), as well as information on the household's purchasing habits and food expenditures if away from home during the previous month. Following the interview, respondents were asked to maintain a daily record of all food expenditures (excluding those while on a trip overnight or longer) using two one-week diaries. Respondents were asked to provide detailed descriptions of daily food purchases including type of packaging (frozen, canned, dried, other), number of units purchased, weight or volume per unit (in either metric or imperial measure), the total cost of the purchase and whether purchased from a food speciality store, convenience store, supermarket or other store. In 2001 respondents were asked to attach their grocery store receipts to the diaries so that the processing staff could identify certain food commodities or find unreported weights.

For meals and snacks in restaurants, the type of meal (breakfast, lunch, dinner, between-meals food) was requested. Respondents were also asked to record the number of meals and total cost, and to indicate whether the meal had been purchased from a table-service restaurant, fast-food restaurant, cafeteria or other restaurant.

At the end of each one-week recording period, interviewers were required to return to the respondent's home to pick-up and review the previous week's diary for completeness and accuracy. At the same time, information was collected on items which might have influenced the previous week's purchasing pattern and level of expenditure such as household member absences during the week, number of meals served to guests or other non-household members, free meals received, food received as gifts or home-produced, food from hunting or fishing..

View the Questionnaire(s) and reporting guide(s).

Error detection

Processing consisted of:

1. Reviewing the questionnaires for missing information, for clarity and for errors.
2. Coding food commodities on the diaries according to approximately 210 different food codes. Coders also entered missing weights or volumes based on the price declared and the quantity purchased of the product.
3. Checking basic ranges for data fields during data entry.
4. Passing data through an editing system that included checking for missing data and verifying consistency between selected data fields.
5. Reviewing and correcting extreme data values.
6. Imputing missing values using the nearest neighbour method or averages (only for expenditures on meals). Detailed food codes were assigned from donor records having similar characteristics (province, quarter of the year, similar costs, etc).
7. Tabulating expenditures using a PC/client server-based system. This system provides tools (database querying, searching, and viewing capabilities) for spotting systematic errors.

To maximize accuracy during the processing of the survey data, strict procedures for quality control were followed. Editing was carried out in stages. Preliminary edits, mainly for completeness, were done by the interviewers in the field.

Imputation

Missing values were imputed using the nearest neighbour method or averages (only for expenditures on meals).

Estimation

The estimation of population characteristics from a sample survey is based on the premise that each sampled unit represents a certain number of units in the population. A basic survey weight was attached to each record in the sample to reflect this representation.

These basic weights were adjusted for non-response by calendar quarter in each of 13 metropolitan areas, 18 additional geographic areas, and three high-income strata. The 18 additional geographical areas consist of regional groupings of the remaining metropolitan areas and urban and rural areas which are based on census definitions but do not necessarily correspond exactly. For definitions of metropolitan area, urban and rural, refer to the 1996 Census Dictionary, catalogue number 92-351-XPE.

To increase the reliability of the estimates, weights were also adjusted to ensure that estimates based on relevant characteristics of the population would respect population totals from sources other than the survey. These totals are based on demographic projections as of January 19, 2002 using the 1996 Census of Population (adjusted for net undercoverage).

Controls for six age groups combined with totals for one-person households, two-person households and more than two-person households are used at the regional level. There are also totals for the number of single-parent families and couples with never-married children that are also used at the regional level.

Controls for two age groups are used in 13 metropolitan areas plus Whitehorse. In addition to these two age groups, in Whitehorse the following controls are also used: counts of one-person households, two-person households and households with more than two persons. For Yellowknife and Iqaluit, only the total number of households is used.

All weekly records from a household received the same weight. Most households (95%) have two records - one for each weekly diary completed. However, 252 households completed only one diary and therefore have only one record.

Quality evaluation

Data are compared to the results from previous years and to other data sources: the Census of Population (record number 3901), administrative and other Statistics Canada surveys.

Disclosure control

Statistics Canada is prohibited by law from releasing any data which would divulge information obtained under the Statistics Act that relates to any identifiable person, business or organization without the prior knowledge or the consent in writing of that person, business or organization. Various confidentiality rules are applied to all data that are released or published to prevent the publication or disclosure of any information deemed confidential. If necessary, data are suppressed to prevent direct or residual disclosure of identifiable data.

The respondent's name and address are removed from the questionnaire before it is archived.

The name and address do not appear on the survey database.

Approval by Statistics Canada's Microdata Release Committee is required before a microdata file may be released to the public. This committee ensures that confidentiality requirements are met as per Statistics Canada policy.

Revisions and seasonal adjustment

This methodology does not apply to this survey.

Data accuracy

Response rate for 2001 = 71.3%

Coefficients of variation for summary-level variables for Canada and the regions are supplied in the publication 62-554.