The Fish Freshness Meter was designed with the needs of the quality controller in mind. The term ‘quality controller’ is meant to include officials responsible for public health inspection or enforcement of mandatory regulations concerning quality, as well as persons concerned with quality control in industry and commerce. In a typical situation, a controller is required to test batches of fish that are being inspected, sold or processed as a unit, and for this a suitable sample must be taken. The meter simplifies this process of assessing a batch by sampling.
If the freshness of individual fish within a batch is measured with the meter, it will be found to vary. This fish-to-fish variation in meter readings is caused by two factors:-
(a) Variation in intrinsic freshness among individual fish.
(b) Variation of meter readings among fish of the same freshness.
Both are expressions of the variability of biological organisms. A batch of fish caught at the same time and handled and stored identically will spoil at slightly different rates because of variations in chemical constitution and bacterial activity. Hence the individual fish in the batch will have a range of freshness; a range that will tend to increase as spoilage increases.
The meter readings themselves are not direct measures of freshness, as defined in sensory terms, but are strongly associated with it. The relationship between meter readings and freshness as measured by a sensory panel will once again be different for each individual fish. For this reason it is not recommended that the meter be used to assess the freshness of single fish, except within rather wide limits.
These deviations from average behaviour tend to cancel out when the means of samples from batches are taken, and the bigger the sample the better the correlation between the mean meter reading and freshness score. It follows from sampling theory that the number of fish which ought to be sampled from a batch depends mainly on the desired precision of the mean score. It also follows that, provided the batch is large enough, the required sample size does not depend on the size of the batch. From observations made, it has been decided to use a sample size of 8 Fish as the basis for the averaging in the Torry Fish Freshness Meter.
In most marketing and distribution systems the fish are held in containers with a capacity of 50 kg. Several containers constitute a batch. In this situation fish from as many containers as possible should be sampled i.e. one from each container if the batch has eight or more, four from each container if that is all there is in the batch. If the controller or inspector wants greater confidence in the batch average, more than one set of eight fish can be sampled, and a grand average of the results calculated.
Note that the precision of an averaged result increases only as the square root of the sample size. An average of 64 readings is only twice as precise as an average of 8 readings.