Analysing differences between two age determination methods by tests of symmetry

A common problem in fisheries science is the comparison of two methods for obtaining ages of individual animals. Often, indices of precision are computed for this purpose. We believe such indices are inappropriate both as measures of precision and for comparative purposes because they do not properly account for age effects or take the experimental design into consideration. We suggest that if the overall level of agreement is low between two ageing methods used on the same sample of fish, then one can use a test of symmetry to look for evidence of systematic disagreement. A χ 2 test is used to determine if the number of fish assigned age i from method 1 and age j from method 2 differs significantly from the number of fish assigned age j from method 1 and age i from method 2. Such a test can also be used to determine the range of nominal ages over which two methods appear to give comparable results.