As the fishery sample census was the mainstay of the 1973 fishery census, the present section specifically deals with its methodology followed in detail.
3.1.1 Design of Sample Census
The 1971 population census of Indonesia established a number of census blocks throughout the entire area of the country. During the course of complete enumeration, which is known as a listing, the identification of both agricultural households and fishing households was carried out. This made it possible to prepare a complete list of census blocks which involved any number of agriculture or fishing households. The CBS sample census for both agriculture and fishery used such a list as a sampling frame. However, the distinction between marine fishing households and brackishwater culture households was not clear in the complete enumeration of the 1971 population census. Therefore, the identification of census blocks with brackishwater culture was made mainly by referring to the census maps prepared for the 1971 population census, which indicate the location of brackishwater ponds.
Basically, both the selection of a sample and census enumeration which were followed by the CBS sample census were very simple. At first, for each province a certain number of census blocks were randomly selected. Then, within a sample census block selected, all fishing establishment/fishing operator's households and fishing labourer's households were interviewed with a predetermined questionnaire.
With regard to sample size, at first the total number of sample blocks throughout the three islands was determined on the base of the amount of census budget available. Then this sample size was allocated to marine fishery and brackishwater culture separately. Thereafter, it was further allocated to each province so that, in principle, at least 30 sample census blocks may be allocated to each province.
Sample sizes allocated to each province and the sampling ratio are seen in Tables la and 2a of Appendix 2. It may be noted from these tables that the sampling ratio for brackishwater culture was extremely high as compared with that of the marine fishery. During the planning stage of a fishery census the number of fishing boats for the marine fishery was known roughly, whereas, in the case of brackishwater culture, the number of fish culture households was not known. As a result, the population size as well as the sample size for brackishwater culture was arbitrarily estimated.
As may be noted from the allocation of sample size, initially the fishery census intended to provide data at provincial level and above. However, as the sampling ratio for brackishwater culture was extremely high, data processing for this subsector of a fishery was carried out at district level and above.
The complete enumeration of the 1973 agriculture census revealed that the total number of fishing labourer's households was unexpectedly high as compared with that of fishing operator's households. However, this fact was not known during the planning stage of a fishery census. As a result, throughout the three islands the number of sample fishing labourer's households become slightly higher than that of sample fishing operator's households (see Table 1b of Appendix 2). This resulted in a considerable waste of census funds given to less important survey objects.
In the CBS sample census for the marine fishery, fishing establishments/households with powered boats were poorly covered due to the absence of stratification. As a result, a DGF additional fishery census had to be undertaken. This resulted in a tremendous delay for the data analysis as well as data processing of the fishery census. As a matter of fact, this was the most serious shortcoming in the design of the CBS sample censuses.
3.1.2 Complete Enumeration
In the 1973 agriculture census the identification of agriculture holdings and fishing households was done for all census blocks regardless of whether they were in a rural or urban area. Thus, statistically, the complete enumeration of the 1973 agriculture census was the most ideal. For the complete enumeration, Form A was used (Appendix 2).
As seen in Form A, for every census block all households/establishments were listed at first regardless of whether they were involved in agriculture/fishery or not. Then, agriculture households and fishing households were identified one by one. Thereafter, with a view to secure a sampling frame as well as a population size for every subsector of a fishery, fishing households so identified were further classified into the following classes:
Fishing Operator's household
Marine fishery1
Inland fishery:
brackishwater culture
freshwater culture
cage culture
paddy field culture
open water fishery
Fishing labourer's household (marine fishery only)
Because the majority of census enumerators were not familiar with the fishery sectors, many simple mistakes in identification occurred during the field operations.
1 This household was further classified according to size of boat used.
The tabulation on the number of fishing households, based on Form A, was done by the head office of the CBS, limited at first to the marine fishery and brackishwater culture. Such a tabulation was done first at village level, and then figures at village level were aggregated to those at subdistrict, district and provincial level. An example of such a tabulation can be seen in Appendix 4.
When the outcome of the above tabulation was carefully reviewed, a number of simple errors were found. The nature of errors commonly found was:
Marine fishing household
Inland districts with no marine fishery households were often treated as marine districts;
Marine districts or marine subdistricts were sometimes missing;
Villages which were purely inland villages were sometimes treated as marine fishing villages, and
Fishing households with nonpowered boats were often misidentified as fishing households with powered boats.
Brackishwater fish culture households (BWFC)
Districts which do not have any BWFC were often treated as districts with BWFC;
A district or subdistrict which apparently has BWFC was sometimes missing.
Some errors were corrected either by DGF or by C S. In many instances the data tabulated by the CBS were sent to Provincial Fisheries Services for further correction. Thus, it took more than a year until all doubts were settled.
After finishing the complete enumeration of the 1973 agriculture census, all the completed Forms A were forwarded to the head office of the CBS through its provincial branches. However, no tabulation was done either at district level or at provincial level. If such a tabulation was done at district level and the results were shown to District Fishery Service officials, such simple errors could easily have been eliminated.
CBS tabulation based on Form A was done at first for marine fishery and brackishwater culture only, as a sampling frame for these two subsectors of a fishery was urgently needed for the design and implementation of a new national production survey. However, as a sampling frame obtained for these two subsectors caused such unexpected work as mentioned above, the construction of another frame for the remaining subsectors of the fishery, i.e, fresh water culture, open water fishery, etc., based on Form A was abandoned.
3.1.3 Sample Census
For the marine fishery with the use of Form H1 a complete enumeration of all establishments and households was carried out for the entire area of sample census blocks. During the course of such an enumeration all fishing establishment/household and fishing labourer's households were identified. Then, with the use of Form H2 an interview was made with all survey objects so identified.
For brackishwater culture (BWFC) a slightly different procedure was followed. At first, with the use of Form H3 all names of owners or operators of fish ponds which are located within the area of a sample census block were identified regardless of whether they have their houses within the area of the selected census block or not. Then, an interview was made with all these BWFC households using Form H4. This means that an interview took place sometimes outside the area of the sample census block. As mentioned earlier, census blocks with BWFC were identified by means of area approach by referring to a census map. In BWFC areas, not all owners or operators live within the area of census blocks so identified. This was the reason why the above procedure was followed.
As mentioned earlier, the DGF sample census was undertaken primarily to remedy poor coverage of the CBS sample census, especially for fishing establishments/households with powered boats. However, the DGF sample census also aimed at stabilizing the annual survey on powered fishing boat statistics with the use of a complete list of powered fishing boats, which would become available during the course of its complete enumeration.
3.2.1 Design of Sample Census
In the DGF sample census a typical sample census methodology was used. It comprised a complete enumeration, selection of a sample of surveyed units and a sample census. Therefore, at first a complete list of fishing establishments/households with powered boats as well as of powered fishing boats was prepared with the use of Form B (see Appendix 3). Then, on the basis of such a list, a sample of fishing establishments was chosen, and a sample census took place for the selected sample.
3.2.2 Complete Enumeration
The type of fishing gear employed by powered fishing boats normally requires a fishing license. Therefore, Form B, which was a complete list of fishing establishments/households with powered boats, was prepared, based chiefly on the record of fishing licenses, which are kept by Provincial Fisheries Services. For coverage purposes, Form B also covered powered fishing boats without fishing licenses.
An inboard powered boat is mobile for a good distance due to its cruising capacity and often changes its fishing base, even in the same year. This may cause a double counting of the same powered boat over time. Such a case may happen particularly for trawlers. It was, therefore, strictly instructed that Form B was to be completed for the owners or operators of powered fishing boats who have their own residence or office in the surveyed areas.
It was emphasized to the Provincial Fisheries Services to make Form B as complete as possible. However, after finishing the DGF sample census, it was noted that very imcomplete coverage occurred for trawlers of Aceh, the Malacca Strait coast of North Sumatra, Jakarta and the south coast of Central Java. If such population sizes are used for estimation purposes, a great underestimate will occur of the catch taken by the fishing establishments/ households with powered boats. Therefore, the population sizes obtained through Form B were amended by referring to records of the Harbour Master's Office and other data sources. Again, this additional work became another factor in delaying the data processing and data analysis of the marine fishery census.
3.2.3 Selection of a Sample
The CBS sample census for the marine fishery aimed at providing data at provincial level. Therefore, the selection of a sample was made at the Provincial Fisheries Service, based on the completed Form B.
During the planning stage of the DGF sample census, there was a plan to collect cost and earning data of fishing operations by type of fishing gear and, further, by the size of fishery management within that fishing gear. Therefore, attempt was made to take five or more of the sample fishing establishments/households from each size of management of a certain gear. However, as powered fishing boats are unevenly distributed among provinces, in certain provinces where the number of powered boats was few, all fishing establishments/households with powered boats were taken as a sample.
3.2.4 Sample Census
It was assumed that data processing as well as data analysis would be done by pooling all questionnaires recovered from both the CBS sample census and the DGF sample census. Therefore, the same questionnaire as used for the CBS sample census was used for the DGF sample census.
