Methodology and data quality

Potato area and yield survey (Survey 3446)

Survey frame and sample selection

The target population includes all potato farms in Newfoundland and Labrador, Nova Scotia, Prince Edward Island, New Brunswick, Manitoba, Saskatchewan, and British Columbia except those on Indian Reserves and institutional farms.

The Census of Agriculture provides a list of farms and their potato area. This list is updated annually from various available sources. The list frame is stratified on the basis of Census potato area and provincial geographic boundaries.

A sample of 866 farms (67 Newfoundland and Labrador, 29 Nova Scotia, 276 in Prince Edward Island, 184 in New Brunswick, 171 in Manitoba, 93 in Saskatchewan, and 46 in British Columbia) was drawn from the list frame for the Potato Area and Yield Survey.

The yield portion of the survey is conducted in October to confirm area planted and to ask for area harvested and production. Operations that had reported no potatoes in June are not contacted a second time.

Data collection

Data collection for the seeded potato area was done by telephone interview, from June 21 to July 2, 2007 and for the yield portion between October 17 to November 12, 2007.

By the end of the collection period, about 80% of the questionnaires had been completed. The refusal rate to the survey was about 4%. Non-contact and non-response can explain the difference between the completion rate and the refusal rate. Initial sample weights are adjusted (a process called raising factor adjustment) in cases of non-response. No imputation is performed.

Estimation

The survey data collected from the Potato Area and Yield Survey are weighted in order to produce level indicators which are representative of the population. These level indicators then undergo a validation process, based on subject matter analysis and consultation with provincial specialists, before an estimate is determined.

Sampling and non-sampling errors

The statistics contained in this publication are based on a random sample of agricultural operations and, as such, are subject to sampling and non-sampling errors. The overall quality of the estimates depends on the combined effect of these two types of errors.

Sampling errors arise because estimates are derived from sample data and not the entire population. These errors depend on factors such as sample size, sampling design and the method of estimation.

Non-sampling errors are not related to sampling and may occur throughout the survey operation for various reasons like coverage, differences in interpretation of questions, incorrect information from respondents, mistakes in recording, coding or processing of data.

Data quality

Sampling error can be estimated from the sample itself by using a statistical measure called the coefficient of variation (cv). Over repeated surveys, 95 times out of 100, the difference between a sample estimate and what should have been obtained from an enumeration of all potato farming operations would be less than twice the coefficient of variation. This range of values is referred to as confidence interval. While published estimates may not exactly equal the level indicators (due to the validation and consultation process), these estimates do remain within the confidence interval of the survey level indicators.

For the 2007 area estimates, the cv's from the Potato Area and Yield Survey ranged from 3.34% to 17.35%.

For the 2007 yield estimates, the cv's ranged from .72% to 4.87%.

Farm product prices survey (Survey 3465)

Survey frame and sample selection

The target population for Prince Edward Island, New Brunswick, Quebec, Ontario, Manitoba, Saskatchewan, Alberta and British Columbia consists of all agricultural operations in these provinces which grow potatoes. The Census of Agriculture provides a list of such farms and is updated from various available sources. The list frame is stratified on the basis of the area of potatoes reported on the 2001 Census. A sample of approximately 500 farms is drawn from the list frame and is rotated as required.

Data collection

Data collection is carried out by telephone interview five business days following the 15th of every month with a reference date of the 15th. The refusal rate is usually less than 5%. No imputations are performed. Editing is done following data capture and outliers are verified and removed if necessary from the sample. A simple average is calculated on the remaining data.

Spring survey of fruits & vegetables (Survey 3407)

Survey frame and sample selection

Newfoundland & Nova Scotia - This new national annual survey contacts about 1000 fruit or vegetable growers in Newfoundland and in Nova Scotia.

The target population is all agricultural operations in Newfoundland and Nova Scotia. The Census of Agriculture provides a list of farms which is updated from various available sources.

Data collection

Data collection is carried out using a CATI application in May. No imputations are performed. Editing is done and outliers are removed. Response rates run approximately 90%. Weighting factors are adjusted to account for non-response.

United States data

American estimates are based on information furnished by crop reporters, processors, and cooperating State or Federal Agencies. The various sources include: Market News Service rail and truck shipments, Federal-State inspections, and miscellaneous data available through marketing programs.

Price estimates represent average returns to growers for all uses and for all methods of sale, including the value paid for a government diversion program. These prices are applied to the quantity sold to compute value of sales. The volume sold, and the value of sales, exclude potatoes used on own farm for seed, feed, home use, and loss.

The United States season average price is obtained by weighting State prices by quantities sold. Potato production, utilization, and value of sales in each State have been classified by seasonal group according to the period (quarter) when the largest supplies are harvested. The four seasonal groups are as follows:

A crop year is associated with the calendar year in which harvest is accomplished. Seasons are not fixed dates but are approximations because of overlapping of harvest before or after the specified season.

Generally, the marketing season closely follows harvest except for fall potatoes when marketing from storage may extend through August of the following year. Thus, the marketing year for fall potatoes is from the start of harvest through August of the following year.

Revision procedures

Area, yield, production, price or values are all subject to revision when more information becomes available. The policy is to revise the estimates when:

1. supply/disposition analysis at the end of the crop year indicates under or over-estimation of production; or
2. a new benchmark for area becomes available from the Census of Agriculture; or
3. new sources of administrative data would suggest alternatives to survey results.

The tables below contain some statistics which indicate the magnitude and direction of past revision from the preliminary to the final (following census revisions) seeded potato area and total production. The magnitude is measured by the average percent change between the preliminary and final estimates. The direction of revisions is indicated by counting the number of years that the preliminary estimate is above or below the final revised estimate.

The data indicates that the preliminary estimates of seeded area have been revised upward 9 years out of the past 10 by an average of 1.97% at the Canadian level.

The data indicates that the preliminary estimates of harvested production at the Canadian level have been revised upward by an average 1.94% over the past years. In 9 of the past 10 years, the revision has been upwards.

Conversion factors

The following metric conversion factors are used in this publication:

Area: 1 acre = 0.4047 hectare

Production: 1 hundredweight = 0.04536 tonne

To convert area from imperial to metric, provincial acreages are first converted, then the individual provinces are added for the Canadian total area in hectares. The Canadian acreage may not directly convert to the Canadian area in hectares.

The same method is used for production. Provincial hundredweights are converted to provincial tonnes, which are then added for the Canadian production in tonnes.

The Canadian yield in tonnes is the result of the total production in tonnes divided by the Canadian hectares. The yield in hundredweights/acre may not be directly convertible to the yield in tonnes/hectare.

The average farm price at the Canadian level in $/tonne is the farm value divided by the marketed production in tonnes. The price in $/cwt may not convert directly to the price in $/tonne.

Data confidentiality

Data confidentiality is ensured under the Statistics Act, which prohibits the divulging of individual or aggregated data where individuals or businesses might be identified.