Several attempts have been made in order to isolate the effect of galaxy interactions by comparing galaxy in pairs with isolated galaxies. However, different authors have proposed different ways to build these control samples (CS). By using mock galaxy catalagues built up from the Millennium Simulation, we show that the set of constrains used to define a CS might introduce biases which could affect the interpretation of results. In this analysis, we make use of the fact that the physics of interactions is not included in the semi-analytic model, so that any difference between the mock control and pair samples can be attributed to selection biases. Thus, we suggest how to build an unique and unbiased CS in order to individualize the effect of interactions. Based on this theoretical findings and using the SDSS-DR4 data, we revise some previous observational results of galaxy in pairs in order to evaluate any possible disagreement. The comparison between simulations and observations suggests an overestimated effect of the DM halo bias in semianalytical models.