In practice, a number of approaches to data collection of fisheries statistics are used by government departments of fisheries:
(a) Occasional items of data are collected and filed as and when they become available,
(b) A system of continuous monitoring is followed to quantify aspects of a number of characteristic units in order to determine changes with time, which are presumed to represent the whole.
(c) Periodic sampling surveys (monthly, quarterly, annual, biennial or at even longer intervals) are carried out involving a significant work force and carried out for one or several short periods of time.
(d) An attempt is made to collect information from all units on a continuous basis.
These approaches all have their various drawbacks. For instance:
Approach (a) rarely allows broad conclusions on the whole sector to be made with confidence.
Approach (b) can often provide a useful indication of trends in the sector, but no reliable estimates of overall magnitude.
Approach (c) is the ideal one, but relies on availability of significant numbers of technical officers.
Approach (d) is one adopted by many fisheries departments, but is not only more expensive (even than c above), but will almost certainly result in underestimates of main quantities. One variant on the approach that may possibly be useful is however briefly described in Appendix 1.
A further desirable alternative which is rarely implemented may be suggested here, that has relevance in trained manpower shortage conditions.
(e) This is a combination of (b) and (c), whereby the schedule of visits of a few officers throughout the year, are related to the sampling frame in a statistically sound manner, so as to permit statistically sound conclusions to be made.
Here a significant work force spends extensive efforts over a relatively short period of time in getting a “snapshot” but accurate picture of the state of the fishery at regular or irregular intervals (for theory see Bazigos, 1974; Sahney, 1983).
As indicated earlier, three main components are involved here, namely:
(1) A Frame Survey - is used to make a complete list of the characteristics of the fishery as a basis for designing a statistical sampling survey programme. The principal ports, landing places, fleet size and type, and details of the operations of the fleet are documented. Transportation routes of fish and wholesale/retail transfers, transshipments, import and export routes are identified, and fishermen's communities and the main means of subsistence of fishermen and their families are described on the socio-economic side. (This component is required for all approaches to statistics gathering that hope to draw conclusions about the whole system being investigated).
(2) The fisheries catch assessment survey - is ideally carried out by a number of trained observers (statistical recorders) sampling in space and time. Statistical recorders visit a randomly selected proportion of the landing sites, markets or ports identified in the frame, and sample a present number of fishing units at pre-assigned times and days. The survey is run on successive occasions (e.g., monthly, quarterly) and generates estimates of the survey characteristics within the established reference periods.
(3) The quality check survey - is used to establish the accuracy of (2) and to de-bias the calculated estimates, if required.
It should be stressed that this type of rigorously designed survey over time is necessary if one wishes to have an estimate of the landings with measurable precision. It does, however, pose some problems for countries with a small staff in that finding personnel qualified to design and implement the survey is often rather difficult. Technically qualified personnel might also be required in the design of the surveys. Two other less rigorous approaches can however be used for data collection. A standard approach to monitoring landings, fleet size, catch composition, costs and earnings, etc., from a known portion of the whole fishery is one of these. The other, and more satisfactory one, from the point of view of statistical validity is to modify the classic sampling approach to allow for two or three officers to carry out in time, what would be accomplished by a larger team over a short period. Such an approach is described in more detail in Appendix 1.
There are a number of ways of ensuring that some data are continuously being collected for the more important key variables without the statistics officers having to collect them all in person. For these types of data-monitoring exercices however, the statistical officer needs to allocate some of his time for periodically checking the validity of the data for the sites being monitored, and will need to spend time in checking in person to see how valid the monitored data are for describing trends in areas or sites where these type of data are not available.
Listed below are a variety of sources of information where fisheries data are often collected and kept:
(a) Customs officers and harbour masters often keep records on import-export trans-shipments, and dates/times of departure and return of fishing boats to port.
(b) Departments of manpower or rural development, often keep records on employment of e.g., registered fishermen.
(c) Departments of Health and Welfare may keep records of household surveys that may include fish consumption.
(d) Fisheries Departments may keep records on vessel loans and on the number of registered fishermen.
Regular informal meetings of government officials involved directly or peripherally in fisheries (or even an inter-departmental committee to discuss fishermen's welfare) can be useful in helping to coordinate government efforts in the fisheries sector, save valuable manpower, and avoid duplication.
(e) Key points in the private sector, such as fish dealers, shipping agents, airlines (where high quality fish or shellfish may be leaving the country) can often be persuaded to record information. Here it is important that the confidentiality of such information be respected; if statistics are released (as they should be) the individual records should be combined with others to protect sources.
(f) Fishermen, auctioneers or buyers will sometimes keep records of the quantities, kinds, size categories and prices of fish being sold each day, if a small payment is made to an individual involved in this process, and if on checking by the fishery officer the result is found to be reliable.
Legislating to require buyers to record their purchases on some kind of record sheet is one frequent approach used in many countries to encourage continuous monitoring of statistics. The requirement for receiving proper data is a condition for issuing a buyer's or fish dealer's licence (other conditions being maintenance of good quality fish, etc.). A description of one such scheme that might also include industrial fishermen is given in Appendix 3. For practical reasons, it would be inappropriate to require small-scale fishermen to keep such records; an interview approach is the one to use for data collection in the artisanal sector.
Clearly, this monitoring function is one which, if applied and controlled with care, can yield a great deal of information with relatively low manpower input. The result, however, is not necessarily going to provide correct figures for total landings, without a good deal of extra work by the statistical officer, to identify gaps in the statistics, e.g., to estimate that proportion of fish not going through markets or wholesalers and to cover these off by direct observation, or quality check surveys.
In theory, it is possible to carry out a sampling survey as above, with a limited team of one or two full-time statistical officers visiting a sequence of ports on a year-round regular basis to sample a randomized number of vessels at each port. This, in fact, is probably a reasonable way of scheduling the full-time activities of a statistical officer. It is understood that the statistical officer has in his office a map of the main sampling points (a frame survey is obviously still necessary, and should be updated continuously), and that his diary and work schedule has been organized in advance of the season, assigning times and dates to visit each port/landing, site/market on a stratified random basis. (Larger ports with more landings and a higher diversity of vessels or types of catch are grouped into priority (weighted) strata, and visited more frequently than small ports, which are also visited randomly.) The officer's work schedule is to be followed with very little discretion in substitution of one port or date for another. A quality check by the supervisor, in person will also be important from time to time, to ensure that the temptation to take the easy way out, and to produce records for the more difficult sampling sites by guesswork or telepathy, is detected and dealt with severely.
This approach is probably a reasonable one that should be aimed at by small fisheries departments (see Appendix 1). It will be necessary to supplement this approach by the steady monitoring of routine data from a limited number of locations as suggested in the previous section (see also Appendix 3). A wide variety of types of data are needed.
