In this paper, considering a regression setup with the outcome following a normal distribution, the presence of missing observations has been assessed by assuming that they are missing at random, i.e., the setup involves MAR conditions (Rubin [1]). Here, a logistic regression scheme is employed to model the missingness probabilities, along with the formulation of the joint probability distribution of the observed data and missingness indicators. Moreover, assuming that the parameters are distinct, it is established that for likelihood-based inference the missingness approach is quite ignorable. By virtue of the EM algorithm, a maximum likelihood estimation process is developed. A simulation study is performed to compare the bias and variance in the settings of MAR and MNAR contexts, to validate the theoretical developments. Such findings indeed strengthen the robustness of the whole inference procedure as well as determine the potential risks when the missingness mechanism deviates from this framework.