It is common to use an empirical distribution if the collected field data set is not accepted by a goodness of fitness test for a simulation study. However, a theoretical distribution is recommend to use for model building to the shortcomings of reproducing the observed data only. It is true that the possibility of passing a goodness of fit test for a set of observed data collected in the field is relatively low. In addition, determination of significance level for a theoretical distribution examination is not a simple question, either. It can be claimed that the level of significance determines whether the goodness of fit test for a observed data set is passed or not. Therefore, it is a very interesting issue that observing the effects with various levels of significance value in a simulation model. In this paper, the results of goodness of fit tests with observed data sets are presented. Also, a single line single server simulation model is constructed to examine the effect of significance level. Some selected data sets passing the Goodness of Fit test with certain levels of significance are used for this analysis.