2007 Survey of the Recreational Cod Fishery of Newfoundland and Labrador

Appendix A - Survey Methodology and Data Quality

This section outlines the full methodology used in the 2007 Survey of the Recreational Cod Fishery of Newfoundland and Labrador.

In the absence of a licence to fish for cod and, with it, any associated reporting mechanisms to follow up after the fishing season, a household survey of residents of Newfoundland and Labrador had to be undertaken. This was the only possible way of successfully assessing the numbers of anglers, the days fished, cod caught, etc. The procedures carried out to ensure a successful survey are discussed below.

The survey fieldwork was carried out by Bristol Omnifacts Research, a major research firm, located in St. John's, NL.

Sampling

As explained in the Survey Overview section, a random sample of households was interviewed in each of Statistics Canada's Census Divisions in the province (Divisions 10 and 11, which comprise Labrador were treated as a single Census Division). For the ease of reporting results from the survey, these Census Divisions were grouped into six specific geographic areas (see Figure 1).

Data Collection, Processing, Coding and Estimation

The survey was launched on October 18, 2007, within a week and a half of the closing date of the recreational cod fishing season. Press Releases and news items before and during the season included references to the post-season survey whenever possible, so anglers in the province were fully aware that the survey would be undertaken. The survey was conducted by telephone, with the interview lasting between three minutes for households in which no one had fished recreationally for cod and five minutes if anglers in the household had participated in the 2007 recreational cod fishery.

While most coding was provided on the documents, answers relating to where anglers had fished for cod were post-coded into bays and coastal areas, based on responses provided by respondents. Many bays around the province carry the same name, despite being located in diverse areas. Hence, for post-coded bays, entries had to be assessed to ensure that the specific bay identified was the most likely area fished by respondents. A combination of the bay name and the telephone number prefix of the respondent were used in combination to determine cases where there might have been miscoding based on bay name alone. Only a few number of responses had to be corrected, based on an unlikely location for fishing activity when compared to the residence location. In all instances, a bay with the same name was located near the residence location. For example, there are three bays called Bonne Bay: one on the west coast, one on the south coast and a third at the tip of the Northern Peninsula. The Bonne Bay referred to in the post-coded bays was the one on the west coast of the province. While possible, it was considered highly unlikely that an angler living in the Burin or Hermitage Districts would be fishing for cod on the west coast of the province.

A major requirement in the editing and coding phase was verification that the information provided by anglers was both reasonable and sound for the numbers of anglers in the household and for the information related to the recreational cod fishery. This included assessing the number of fish caught, the number of cod caught compared to days fished, the number of individuals, etc. There were no responses that were changed due to unexpected or unreasonable data provided during the interviews. As well, no responses were deemed suspect as being out of scope for the fishery. For example, the total number of days the fishery was open was 35, but the maximum number of days fished was reported at 20. Similarly, anglers could have harvested a maximum of 175 cod during the season and no angler reported exceeding this number. It should be noted that in those instances where anglers reported more than 5 fish per day being harvested, no attempt was made to "make the catch legal" (i.e., 5 fish). These responses were accepted at face value in light of the assurances anglers were given guaranteeing anonymity in the survey.

The required programming to produce survey estimates consisted of:

• systematic editing procedures;
• substitution algorithms for missing data (only age in this survey);
• data output programs covering all data collected; and
• special output programs for publication purposes.

Weighting Procedures

Weighted estimates were produced to ensure that the raw survey data reflected estimates for the population of Newfoundland and Labrador. The multi-step process for weighting the results of the survey is as follows:

• as shown in Table A.1, calculate household weights for the 10 Census Divisions used in the survey by dividing the number of households reported for each Census Division by the number of households in the sample

Source: Statistics Canada.

• the household weights were applied to the information collected on individuals during the interviews

• the estimated numbers of males and females, under 16 years of age, in total and the number who fished as well as the number of males and females over 16 years of age, in total and the number who fished, were calculated

• the total number of males and females were compared to the estimated population information provided by Statistics Canada from the 2006 Census of Canada, with the purpose of adjusting the estimated numbers of anglers, based on the rates of activity found in each group in each Census Division (Table A.2)

Source: Statistics Canada.

• the numbers of males and females 16 years of age and over who fished were then adjusted for each Census Division to ensure that the total number of males and females 16 years of age and over matched Census data (Table A.3)

• a similar approach for population and anglers under 16 years of age was completed

• for anglers 16 years of age and over, respondent weights were calculated in order to develop population estimates for their angling activity (Table A.4)

• the above weights were used to produce population estimates of effort and harvest in the 2007 recreational cod fishery as well as estimates of wolffish interceptions in the fishery

• a slightly different approach for anglers under 16 years of age was required since the information calculated using household weights resulted in total numbers of males and females. This required that estimates of effort and harvest in the 2007 recreational cod fishery be derived by calculating days fished per angler and fish caught per angler by sex and then applying these averages to the estimated number of male and female anglers under 16 years of age

Explanation:

In Census Division 1, the original estimates for males under 16 years of age were: 24,944 males, with 2,341 fishing 8,262 days and catching 35,322 cod

However, from Census data, we know that the number of males under 16 years of age in Census Division 1 was 21,651, thereby requiring that the number of anglers be reduced to 2,028. From the original estimates for these anglers, they fished about 3.5 days each and caught 15.1 cod each. Hence, for this Census Division, the estimates of days fished and cod caught by males under 16 years of age were calculated as 7,158 days and 30,598 cod, respectively.

Tables A.5a/b and Tables A.6a/b provide detailed estimates and adjustments for male and female anglers under 16 years of age.

No adjustments were made to reflect the minor change of approximately 1,000 individuals in the population of Newfoundland in 2007.

¹ based on the 2006 Census of Canada
² proportional allocation based on initial estimates
³ based on averages found in initial estimates

The adjusted estimates found above were used to allocate activity and harvest to residence areas and to NAFO areas of effort and catch for male anglers under 16 years of age.

¹ based on the 2006 Census of Canada
² proportional allocation based on initial estimates
³ based on averages found in initial estimates

As was done for under-age male anglers, the adjusted estimates found above were used to allocate activity and harvest to residence areas and to NAFO areas of effort and catch for female anglers under 16 years of age.

Data Quality

Sampling Error

Information on reliability is presented in terms of the coefficient of variation of the mean, at one standard deviation. Because exact estimates of variance cannot be calculated, the coefficient of variation of the mean is used as a proxy for the assessing the variability of the data. Table A.7 presents the statistical reliability of key variables for various presentations of the data.

Source: Statistics Canada.

Non-Sampling Error

During the data collection phase, efforts were made to reduce the occurrence of non-sampling errors in the survey. These efforts included a complete verification of the reported data, validity and consistency edits.

One source of non-sampling error that is not considered to play a significant role in this survey is errors introduced due to recall issues related to effort or catch. The time between fishing activity and the survey was no more than three months, limiting any short-term recall regarding the angling activity in 2007.

Data Limitations

Despite all efforts to improve data accuracy, the survey results have limitations. Understanding these limitations will help the reader make informed decisions before conducting further research and analysis using the estimates in this report:

1. 1. Estimates of cod fishing effort and harvest reported for age groups by either residence area or by areas fished will have reduced reliability
2. 2. Data tables presented by bay or coastal area are for information purposes only as the survey was not designed to produce data at this level. Reliability levels are presented in Table A.8.

Data comparability

The 2007 survey results are relatively comparable with the results for cod fishing by residents of Newfoundland and Labrador from the 2000 Survey of Recreational Fishing in Canada (SRF) and the 2001 Log Book Recall Assessment for Newfoundland and Labrador. Although not directly comparable, data on cod fishing in NAFO sub-division 3Ps in 2005, from the 2005 Survey of Recreational Fishing in Canada, and the 2007 results are reasonably close. These comparisons are provided in Table A.9 below.

¹ Number of fish harvested is based on an average weight of 2.4 kilograms divided into the total weight estimate derived in 2001. Information on the specific number of cod caught was not available.
² total weight was calculated using an average weight of 2.4 kilograms per cod. This estimate is considered to be conservative.