Sharing a house with one or more roommates is very common in college life. Roommates are generally neither friends nor family, but strangers matched to share the rent. Privacy needs of a house shared with a roommate are different from a house shared with friends or family. Moreover, the territorial behaviors that roommates can demonstrate in a shared house are distinct from the territorial behaviors of students who share rooms in resident halls. According to Altman (1975), privacy is an interpersonal boundary-control process, which paces and regulates interactions with others. The balance and equation of desired and achieved privacy can be maintained by territorial behaviors an individual perform in space. Although, starting from 80s, there is a growing literature on territorial behavior and privacy needs of college students who live in shared spaces in university residence halls (Mercer & Benjamin, 1980; Kaplan, 1982; Kastenbaum, 1984); house as a shared space for students is quite untouched phenomenon. Housing complexes close to universities generally act as a magnet for students and serve as student housing neighborhoods. Apartments are designed in a way they are expected to be shared and generally offer equal outdoor spaces. The rents of the units are parallel indicating the similarity of the housing quality. Aim of this study is to investigate the effect of privacy needs on housing preferences of college students living with roommates. The subjects of this study are 60 students who live in the student housing neighborhoods along school’s busline. All houses can be described as row houses offering either one or double story apartments. This study utilizes correlational research as methodology with the use simulation in the tactical level. A visual data set is designed and the respondents were asked to prefer among each set of visuals with the reasons behind their choice. The visuals designed for this study is as follows in order of application: First, plans and axonometric perspectives of the types of one or two story apartments, second photographs of different row houses, third photographs of different entrances to units. Plans of the housing units in the selected student housing neighborhoods were collected and categorized in multiple ways in terms of spatial choices they offered. They were redrawn to offer same utilities and formatted in the same scale and quality. To enrich the understanding of the spatial configuration perspective views were prepared. Representative types of the houses were selected from other parts of the city in order to defeat the tendency to select the pre-experienced one. The selected houses were photographed all from the same angle and later some edited in computer environment to eliminate outside effects (like cars, trees, etc.) that can influence preference. For the entrance photos similar method was followed. Respondents were told that all of them offer same rent, utilities and outdoor spaces and asked to prefer houses using the sets of visual data provided. The reasons for their selections and eliminations were asked and the answers were recorded. Each subject was separately open interviewed. Two story types were preferred more than one story apartments. General aims of selecting two story types were explained as having two different environments for different functions. Some respondents defined this as separation of the entertainment part from living part. The highly preferred type of plans were neither open nor closed but the ones that had continuity of space that gave enough privacy for each function without isolation in the spatial configuration. The results showed that there was a strong desire for home like appearance even though it was to be shared. Thus, two-story types were selected for their individual unit entrances and facade organizations. This study is significant in terms of understanding the spatial configurations that are formed by privacy requirements of individuals. Simulation as a method is preferred instead of conducting research in real life situations due to the feasibility and manageability of the study. The tool for gathering data from respondents was more controlled in terms of other influential factors that may occur in real life situations. Even though the results from a sample size of 60 people provide moderate statistical significance, findings are expected to enrich design guidelines for future projects. The methodology followed is appropriate for participatory design process.