This study aims to identify the factors affecting representativeness, conservatism, and overconfidence biases in a sample of Turkish institutional investors. In difference to other studies, especially the individual investments of institutional investors in capital markets are focused on. The respondents who are selected with the method of convenience sampling are asked survey questions. In the study, models are formed with the stepwise method using linear regression analysis, and survey questions that are assumed to represent the examined biases are selected as the dependent variable. The age, gender, marital status, education, whether or not having children, and experience of institutional investors are included in the analysis as the independent variables. This way, which of the independent variables that are found to be statistically significant and the subcategories of these independent variables are most influential on the biases is revealed. As a result of the analysis, in addition to determining the variables that affect a single bias, it is understood that the variable of education level affects all examined biases, and especially the effect of low education levels on the examined biases is higher. In the regression analysis, experience is identified as a significant variable affecting representativeness and conservatism biases. It is observed that especially the institutional investors who are inexperienced or less experienced interact with these two biases more in their decisions related to their individual investments. Gender is found to be significant variable on both conservatism and representativeness biases, especially it was understood that female institutional investors were more effective than men on the representativeness and conservatism bias. Furthermore, it is determined that the marital status variable and the divorced institutional investors as a subcategory of this variable are significantly influential on conservatism and overconfidence biases.