Print version ISSN 0102-0536
LUCIO, Alessandro D et al. Violation of presuppositions to the mathematical model and data transformation. Hortic. Bras. [online]. 2012, vol.30, n.3, pp.415-423. ISSN 0102-0536. http://dx.doi.org/10.1590/S0102-05362012000300010.
The adequation to the presuppositions of the mathematical model was verified and, identified data transformation for productive and morphological variables of tomato in experiments in plastic tunnel and in field, considering each harvest date and grouped harvests in two cultivation seasons. Two experiments were carried out in plastic tunnel and one in field, both in a randomized block design with three replications. We analyzed the mass, number, length and width of fruits. The tests of Tukey of non-additivity, Lilliefors, Bartlett and Run test were used to verify, respectively, the additivity of the effects of the model, normality, homogeneity and randomness of errors. In cases of violation, the data were transformed, and re-submitted to the analysis of the presuppositions of the mathematical model. The productive and morphological variables of tomato presented adequation to the presupposition of additivity of effects. Violations of the productive and morphological variables occur to the presupposition of: normality, homogeneity and randomness of errors. The grouping of all harvests for analysis provides adequation of productive and morphological variables to the presuppositions of normality and homogeneity of errors. The data transformation that provides more proportion of adequation of productive and morphological variables to the presuppositions of normality, homogeneity and randomness of errors is the square root.
Keywords : Solanum lycopersicum; grouping of harvests; additivity; normality; homogeneity; randomness.