Potato Area and Yield Survey

Detailed information for 2007

The purpose of this survey is to obtain information to produce national and provincial level, estimates of potato production.

Data release - July 20, 2007

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

Description

The survey produces provincial estimates of area of potatoes planted, harvested, average yield and production. These data are used by agricultural industry analysts and producers to make production and marketing decisions and by government analysts to monitor the potato or agriculture industry or develop agricultural policies in Canada. The data are used in the calculation of farm income estimates and flow to the Canadian System of National Accounts. Further, the data are used in the calculation of net farm income projections, produced by Agriculture and Agri-Food Canada in co-operation with Statistics Canada and the provinces.

Reference period: Refers to the annual production.

Collection period: June and November

Subjects

• Agriculture and food (formerly Agriculture)
• Crops and horticulture

Data sources and methodology

Target population

The universe includes all potato producers in Canada although data are collected directly only in Newfoundland and Labrador, Prince Edward Island, New Brunswick, Nova Scotia, Manitoba, Saskatchewan and British Columbia; other provinces report through administrative sources. Potato producers on Indian Reserves and Institutional farms are excluded from this survey.

Instrument design

The original development of the questions for the telephone survey was based on other surveys including the Census of Agriculture. The Census of Agriculture questionnaire was tested using focus groups and pilot surveys.

Sampling

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

The population consists of all potato farms in Newfoundland and Labrador, Nova Scotia, New Brunswick, Prince Edward Island, Manitoba, Saskatchewan and British Columbia (other provinces report through administrative sources). The frame includes all potato producers, excluding indian reserves and institutional farms, who reported more than $1,000 on the Census of Agriculture, supplemented by known new operations. Stratification and allocation are by size of operation. The sample is selected using a stratified simple random sampling method.

The sample is stratified by size of operation with about 270 producers contacted twice and 520 producers contacted once.

Data sources

Responding to this survey is mandatory.

Data are collected directly from survey respondents and extracted from administrative files.

The survey information is collected by telephone interview from Agriculture Division's Truro (Nova Scotia) Office of Statistics Canada. Farm operators are contacted directly by phone and, in cases where the operator is difficult to contact, there are multiple follow-up telephone calls that ultimately result in a high contact rate.

Information is collected in June for Prince Edward Island and New Brunswick and in November for Newfoundland and Labrador, Prince Edward Island, New Brunswick, Nova Scotia, Manitoba, Saskatchewan and British Columbia. Information from other provinces is obtained through administrative sources.

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

Error detection

Responses are evaluated during the contact to ensure that there are no significant inconsistencies with historical patterns and the current year's growing conditions. The top contributors to the unweighted and weighted estimates are reviewed and verified.

Imputation

Total non-response (e.g. refusals and no contacts) is accounted for by weighting adjustments to each stratum. Some missing information is imputed manually during the edit process based on information contained in the questionnaire, previous reports, the Census of Agriculture or industry trends.

Estimation

The response values for sampled units are multiplied by a sampling weight in order to estimate for the entire surveyed population. The sampling weight is calculated using a number of factors, including the probability of the unit being selected in the sample. Raising factor (weight) adjustment is used in the estimation process to account for respondents who could not be contacted or who refused the survey.

Once the weights have been calculated for each record, any level of required estimates may be obtained using domain estimation (i.e. provincial, agricultural region, etc.).

Quality evaluation

The survey results are evaluated through comparisons to previous estimates and other sources when available. Norms and current growing conditions are used as a guide when evaluating the data or comparing to other data sets. A primary tool in the evaluation and final determination of the data involves supply-demand analysis. Production estimates are evaluated against disappearance information, administrative information and the Census of Agriculture.

Disclosure control

Statistics Canada is prohibited by law from releasing any information it collects which could identify any person, business, or organization, unless consent has been given by the respondent or as permitted by the Statistics Act. 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.

Revisions and seasonal adjustment

Once every five years the published potato area estimates are aligned with the results of the Census of Agriculture. Any adjustments made to the data during the Census year are then smoothed in over the historical five-year period between the Censuses. The impact of the revisions is normally less than 5%, however, the impact can be higher.

Data accuracy

Sampling and Non-sampling Errors

A sample survey is subject to two major types of errors; sampling and non-sampling.

1. Sampling Errors

Sampling errors are an inherent risk of sample surveys. They result from the difference between the value of a variable if it is randomly sampled and its value if a census is taken (or the average of all possible random values). These errors are present because observations are made only on a sample and not on the entire population. The sampling error depends on factors such as the size of the sample, variability in the population, sampling design and method of estimation. For example, for a given sample size, the sampling error will depend on the stratification procedure employed, allocation of the sample, choice of the sampling units and method of selection. (Further, even for the same sampling design, we can make different calculations to arrive at the most efficient estimation procedure.) The most important feature of probability sampling is that the sampling error can be measured from the sample itself.

Sampling error can be measured by the standard error (or standard deviation) of the estimate. The coefficient of variation (CV) is the estimated standard error percentage of the survey estimate. Estimates with smaller CVs are more reliable than estimates with larger CVs. Generally any estimate with a C.V. value under 5% is considered to be of excellent quality.

C.V.'s can range from a low of 3% in larger producing provinces to a high of 18% in provinces that with much less potato area.

2. Non-sampling Errors

Non-sampling errors result from a systematic flaw in the structure of the data-collection procedure or design of any or all variables examined. They can include coverage error, which can result from incomplete listings from which the sample is selected. Data response error can come from a flaw in the questionnaire design or questions, plus the inability or unwillingness of the respondent to provide correct information. Non-response error comes from respondents' refusal to answer or who reply too late to be included in the estimates. Another source of non-sampling errors are processing errors, which may occur at various stages of processing such as coding, data entry, verification, editing, weighting, and tabulation.

Measures such as response rates are used as indicators of the possible extent of non-sampling errors. The response rate ranges from 65 to 85%. The wide range in response rates results from years when producers were not finished planting or harvesting by the end of the collection period. We might not have been able to contact the producers, or they might not have been willing to give their information without having their numbers finalized.