_{1}), live weights

_{ }in 30th (X

_{2}), 45th(X

_{3}), 60th (X

_{4})

^{ }and 75th

^{ }(Y) days) of 18 Hamdani breed single-male lambs born in early March of 2001. According to results of MRA, although all independent variables including in model explained approximately 92% of variation in dependent variable, Y, the effect of only independent variable X

_{4 }on dependent variable Y was significant (p<0.01). With respect to residual analysis, it could be said that the assumptions of normal distribution and homogeneity of error terms in MRA were provided. As the value of Durbin-Watson statistics equaled to 2.31, there was not a sequent correlation among error terms, that is, the assumption that error terms independent from each other was ensured. Considered the leverage and influence diagnostics calculating for observations of sample data set, only two observations (2nd and 16th observations) of all observations-both outliers and potential effective (influence) observations- should be carefully examined. It could be concluded that diagnostics would be an important statistics for researchers because they could give an idea about whether the basic assumptions would be provided for reliability of MRA, data set and goodness of fit.]]>