Rumors, new ideas, and attitudes spread through groups of people via interpersonal communication. One approach to simulate these diffusion processes is to generate random or stylized networks. Normally, random, small world, and scale-free network generation algorithms are used to that end. However, the structure of the network is an important parameter for the pattern of diffusion. Therefore, the model has to fit the properties of the corresponding system. The first questions for this article are: What are the characteristics of real world interpersonal networks and which network generation algorithm fits these characteristics? Based on this examination, a new and simple to implement algorithm is introduced to generate stylized networks for modeling the diffusion of information, opinions, and messages in real world social networks by interpersonal communication. Our work also shows the connectivity of empirically collected ego networks and algorithmically generated overall networks and offers correspondences between the micro and the macro level. It is therefore possible to use micro knowledge about networks to calibrate macro networks. We also discuss scaling issues when dealing with networks with different sizes or when the simulation networks are smaller than the networks of the underlying system.