Dealing with relational data is an important and challenging topic not only in life science, as in the characterization of dynamics of living systems, but also in psychology and social sciences, as in the examination of interpersonal behavior. In this context, stochastic process models are often used for the mathematical description of system dynamics and transition processes. In this observational study, interpersonal attachment of 96 participants (74 females and 22 males, mean age of 23.94 ± 4.16 years) measured by a digital application was represented and analyzed as Discrete Time Markov Chains (DTMCs) after the participants formed 48 dyads. For each participant, 100 DTMCs were formed based on a single measurement by continuously increasing the interval between two states and tested for Markov property as well as for homogeneity with their counterpart within a dyad. The results show, firstly, that the majority of the DTMCs satisfy the Markov property and can therefore actually be interpreted as DTMCs and, secondly, that the DTMCs of a dyad can be both homogeneous and heterogeneous, with the middle intervals appearing to be the most suitable in the two tests. Nonetheless, the results should be further investigated by either optimizing the methodology itself or supporting it with additional measures of interpersonal behavior, e.g., by observing interpersonal interactions or by eye-tracking measurements.