The survey was done to check whether UK manufacturers were complying with EU rules relating to declared content.
It will also be used to verify the best methods manufacturers can use when calculating the meat content of their products.
The FSA concluded that the accuracy of its recommended method of calculation could be improved if the visual lean (VL) determination by manufacturers also improved.
For the survey the FSA collected the recipes and ingredients of 191 raw and ready-to-eat meat products from 146 different manufacturers.
In six out of 191 of the products sampled, or 3 per cent, the manufacturers overstated meat content.
A further 3 per cent did not declare how much meat the product contained.
All other products met the requirements, the FSA stated yesterday.
Manufacturers of meat products, whether prepacked or sold loose, are required to declare the percentage of they contain by species.
The FSA completed the survey to check whether meat product manufacturers are calculating and declaring meat content correctly in response to changes in labelling legislation requirements.
The FSA also used the survey to evaluate the suitability of its recommended approach that manufacturers can use when calculating meat content before the product leaves the plant.
Twenty-six local authorities collected the recipes, samples of raw materials and final products at the point of manufacture.
The survey of 191 meat products covered sausages, burgers, pies and kebabs from 146 different manufacturers.
The meat ingredients were analysed for fat, total nitrogen and hydroxyproline by a consortium of five public analysts.
A full compositional analysis was carried out on final products, where these had been collected.
The meat content was calculated using three separate methods.
Two were recipe based calculations.
The FSA method is based on the manufacturer's visual lean (VL) estimation of meat ingredients by eye and reference to typical values of fat and connective tissue.
The Liaison Centre for the Meat Processing Industry in Europe (Clitravi) method is based on analytical values of fat and connective tissue in the meat ingredients.
The third method, known as the Stubbs and More calculation, is limited to single species products and is based on analysing the meat product for nitrogen and fat and using agreed nitrogen factors to calculate the meat content.
Despite a requirement for a meat content declaration since 1984, six out of the 191 product descriptions examined did not give a quantitative meat declaration.
A further four products, or 2 per cent, quantified the meat content, but did not declare it in the correct format.
"Overall, the survey indicates that the manufacturers sampled are declaring their meat contents accurately," the FSA stated.
The statement was confirmed by comparing the manufacturer's meat declaration to the meat content determined by the Clitravi and Stubbs and More calculations, taking into account the analytical tolerances.
The results also showed that the FSA calculation of meat content accords "reasonably well" with the Clitravi calculation, which is the EU recommended method, the regulator stated.
About 70 per cent of the samples, were within 5 per cent of the Clitravi calculations, and only one sample had a higher meat content by the FSA calculation than by the Clitravi calculation, the regulator stated.
About 30 per cent of the samples had a lower meat content by the FSA calculation than the Clitravi method.
The Stubbs and More calculation, which is used by UK enforcement laboratories, also agreed "reasonably well" with the Clitravi calculation, the FSA stated.
The accuracy of the FSA calculation could therefore be improved if the visual lean (VL) determination by manufacturers also improved, the regulator concluded.
The FSA stated that it would investigate ways of helping small businesses achieve better accuracy.
Regulators have inspected manufacturers who failed to declare the meat content of their products.
Those that overdeclared the contents have revised their labels.
Many manufacturers were already aware that they were underdeclaring the meat content, and preferred to build in a margin of error given the variability of the meat ingredients, the FSA stated.