Ascertaining the appropriate selection environment for Nile tilapia (Oreochromis niloticus) in Africa is a critical issue. Two data sets derived from two selection lines originating from a common base population were analysed in this study. The lines were selected in two different input environments, here named ‘low input’ and ‘high input’. Both data sets were combined and jointly analyzed to estimate the phenotypic and genetic parameters, with a special focus on the examination of genotype by environmental interaction. The data sets included a total of 7640 animals with phenotypic information from three discrete generations. Four different models (in terms of fixed effects) were fitted in univariate (harvest weight) and bivariate (harvest weight in each input line treated as two different traits) animal models to estimate variance and covariance components. The heritabilities estimated from the four different models by univariate analyses ranged from 0.15 to 0.41 (all with standard errors of 0.04). The genetic correlations between harvest weights expressed in the two environments, obtained from the bivariate analyses, ranged from 0.74 to 0.84 (with standard error in the range 0.15 to 0.36). We concluded that there was no significant evidence for genotype by environmental interaction for these two particular input environments.