Laser speckle contrast imaging is a technique used for imaging blood flow without scanning. Though several studies have attempted to combine spatial and temporal statistics of laser speckle images for reducing image noise as well as preserving acceptable spatiotemporal resolution, the statistical accuracy of these spatiotemporal methods has not been thoroughly compared. Through numerical simulation and animal experiments, this study investigates the changes in the mean speckle contrast values and the relative noise of the speckle contrast images computed by these methods with various numbers of frames and spatial windows. The simulation results show that the maximum relative error of the mean speckle contrast computed by the spatiotemporal laser speckle contrast analysis (STLASCA) method, in which the speckle contrast images are computed by analyzing the 3-D spatiotemporal speckle image cube, is approximately 5%, while it is higher than 13% for other methods. Changes in the mean speckle contrast values and the relative noise computed by these methods for animal experiment data are consistent with the simulation results. Our results demonstrate that STLASCA achieves more accurate speckle contrast, and suggest that STLASCA most effectively utilizes the number of pixels, thus achieving maximized speckle contrast, and thereby maximizing the variation of the laser speckle contrast image.