Choosing the right investment option by a fund manager or analyst is the first step that contributes to the overall performance of any portfolio of assets. The decision making process is complicated. Markowitz portfolio theory (1952, 1959) laid the theoretical foundations for asset selection and management. However the decision maker is influenced by parameters outside the realm of financial theory and mathematical models (French and French 1997; French 2001). The actual behavior of decision makers can deviate from this normative model. This can be due to the problem solving behavior of the human brain. Human problem solving theory began with the work of Newell and Simon (1972) and Simon (1978). They argue that the human memory is characterized by limitations in terms of processing capacities (Newell and Simon 1972). Given the amount of data the decision maker has to analyze, the process of asset selection is complicated and difficult. Besides the volume of data, the information items may provide information relating to the same aspect of the asset making some of the data set redundant. Besides that, some of information contained in the data set might provide contradictory signals about the performance or characteristics of the asset. Thus the information set available to a decision maker is large, multi-channeled (different data providing different information) and multi-dimensional (for example real estate assets have information pertaining to legal aspects, financial aspects, physical aspects etc.). The limitations in the decision maker’s processing capabilities and the characteristics of the information cues make the asset selection process exceedingly difficult. French (2001) in a study of fund managers from U.K finds that asset allocation uses two sets of hard information during the process, namely historic data and current market perceptions. The study also finds differences between exposure levels of the funds dictated by theory (as per portfolio theory) and actual decisions made by companies (true asset allocations of funds). Gallimore, Gray and Hansz (2000) find medium-sized and small companies’ investment decision making does not follow any normative model due to the diverse nature of property markets in the United Kingdom. Past literature in the field of decision making finds that an expert’s decision making behavior differs from that of a novice. (Bedard and Mock (1992), Bouwman (1984) and Jacoby et al. (1984, 1985, 1986, 1987)). The primary purpose of this study is to understand the impact of experience on the decision making behavior of investors and see if their behavior differs from that of inexperienced individuals. In a controlled experiment design, two groups of subjects are tested. One group is composed of experienced subjects (experts) represented by real estate professionals such as acquisition analysts, fund/portfolio managers or real estate investors (experienced individuals investing either their own money or a client’s money in real estate). The other group tested is composed of students, who are inexperienced subjects (novices). Both groups are asked to choose between two investment cases in two different cities. The two options offered are both class A office properties, institutional grade. Fifteen sets of data are given for each investment option. Data for the cases is sourced from investment management companies, involved in managing funds on behalf of institutional clients. Using a process tracing technique, each subject’s behavior is observed and recorded while making the investment choice. These observations will give us insight into the actual (descriptive) behavior of experienced real estate professionals and inexperienced novices. It will help in isolating the impact of experience on the decision making behavior of real estate investors. This study finds mixed evidence relating to the difference in the behavior of novices and experts. On the five aspects that the two groups are tested, evidence that their behavior differs in three has been uncovered. They are search pattern, number of steps and time on task. However, for the other two aspects, sequencing and cue utilization, no difference was found.