Abstract: Taking the Journal of Information Science as an example, this paper studied the changes of characteristic parameters of generalized Lotka distribution by randomly constructing missed author information. The experimental results showed that generalized Lotka distribution was universal, while the data missing affected characteristic parameters of generalized Lotka distribution, and the influences were not entirely positive correlation. The proportion of data missing within 20% can make the affected degree of characteristic parameters lower than 5%. The statistical data should be counted as fully as possible to make the estimation of generalized Lotka distribution parameters close to the true values.