1. Introduction
Normally the catches to be covered by the L-III Survey are those taken by small-scale fisheries with nonpowered boats or without any boat. Small-scale fishery units do not always sell their catch through fish markets, even when a fish market exists nearby their landing places. Thus the catch recorded at the fish market is usually smaller than the actual catch.
The L-III survey is designed so as to estimate total catch landed at all landing places other than the major landing places regardless of whether such a catch is sold through fish markets or not. Since the catch recorded at a fish market is usually underestimated, the L-III Survey is undertaken without reference to the catch record of the fish market, even when some fish markets exist in the area under survey.
2. Basic Idea of the L-III Survey
The L-III Survey is designed as a cluster sampling, in which a marine village is regarded as a cluster, as well as a sampling unit. For the selection of a marine village, “Probability Proportional to the Size of Village” (PPS) - expressed in the number of fishing households - is applied, so that naturally a marine village of larger size has a better chance to be selected in the sample. This may make for convenience for a field survey, as larger marine villages in many instances may be, located in an area which is easily accessible.
According to the theory of cluster sampling with PPS, all fishing households existing in a sample village must be interviewed by a field enumerator. However, this is not always possible due to the limitation of manpower and travel expenses available for the survey and frequent absences of fishermen who happen to be at sea at the time of interviews.
Therefore, in the L-III Survey the number of trips and total catch for each type of fishing gear are estimated by the following formula:
| Number of trips |  = N ×  |
| Total catch |  = ×  |
Where N: the number of fishing units
: Average number of trips per fishing unit
: Average catch per trip
Since the L-III Survey is conducted as a quarterly survey, the survey of N, and
refers to the facts which appeared during the course of the previous three months.
Of these three items, N is surveyed by means of a complete enumeration using Survey
Form SL-2, so that the accuracy of N will be high, is estimated as an average of the
number of trips made by some sample fishing households interviewed. Although is obtained
based on the number of trips made by such limited fishing households, the accuracy of
may be also comparatively high. This is because in the same marine village, the actual
number of trips made by each fishing household may not very much between each other.
Thus the accuracy of the estimated catch will be greatly affected by the accuracy of the
average catch per trip ().
The L-III Survey, is estimated by dividing the sum of total catch of some sample
fishing households interviewed by the sum of the number of trips made by the same.
Naturally the accuracy of is greatly affected by the accuracy of the total catch
reported by the sample fishing households interviewed. It is therefore strongly suggested
that the accuracy of is ascertained by referring to the records of fish landings per trip
which might be available at a fish market within or near the sample village.
In determining the figure of for the estimation of , it must be clearly kept in mind
that includes not only catch sold but also that consumed at home. Therefore, even when
is determined based on the record of a fish market, the part of the catch not sold to
the market must be assessed and added to the average catch estimated on the record of the
fish market.
3. Selection of Sample Fishing Households
After counting the number of fishing units for each type of fishing gear, using Survey Form SL-2, sample fishing households are selected for interview. It has been assumed that, as a standard, an enumerator may spend two full days per one sample village for each quarterly survey. In such a case, an appropriate number of sample fishing households that could be interviewed would be around 30. However, if an enumerator can afford three days or more for each quarterly survey, this sample size can be increased.
With the above standard in mind, the total number of fishing households that could be interviewed is allocated to each type of fishing gear according to the following standard:
Major fishing gear, the production of which occupies a greater part of the total catch of the sample village 5 sample fishing households.
Minor fishing gear, like castnet, portable trap without using fishing boat, etc., the production of which occupies a minor part of the total catch of the sample village 2 to 3 sample fishing households.
Once the number of sample fishing households is determined for each type of gear, a sample of fishing households is selected by means of the systematic random sampling method on Survey Form SL-2. Then, an interview is made for each of these sample fishing households. If a fisherman in the sample household is absent at the time of interview, it is allowed to shift the sample household to the next one on Survey Form SL-2.
