Polarimetric Synthetic Aperture Radars (PolSAR) have been used to detect oil slicks at sea surface. The numerous platforms available, acquiring data in different formats and configurations, pose as a challenge to understand which is the better format and statistical modeling to improve the oil detection. To contribute with this issue, a combination of different data formats in single look complex, intensity and amplitude, with full and dual polarimetric channels, were evaluated considering adequate statistical modeling to classify each data type. The better results were obtained by the full and dual-pol matrices, however when the HV channel is excluded the accuracy levels are damaged. Therefore, it is better use the data in intensity or amplitude preserving the HV channel, than use a polarimetric data without this channel. The classifier demonstrated potential to detect the three types of oils released, being more effective in detecting biogenic oils rather than mineral oils. The uncertainty levels increase from the center to the border of the mineral oil slicks, indicating the presence of transition regions, possibly related to different weathering mechanisms. Future studies should be done including more SAR images, with known occurrences and field data to investigate properly the trade-offs related with each data format to discriminate different oil types.