Large scale, organized motions play an important role in the dynamics of most turbulent flows, and hence the physical understanding as well as the prediction and control of turbulent flows require a good knowledge of these structures. For practical applications, it is necessary to have reliable detection and analysis methods for these structures. At present there is no definitive and unique answer. This paper deals with a comparative study of various methods which are in common use for extracting large scales structures. Conditional methods and correlation-based methods are tested by several research groups on a common data base. The flow considered is a turbulent plane, incompressible mixing layer.