The I/sI cutoffs are commonly applied at all stages of crystallographic structure solution and refinement, as it is statistically justified to consider insignificant the intensities lower than three sI. The approach reduces actual dataset to the strong, well defined reflections, yielding better refinement statistics and shorter computational time. However, the weak signals may carry some information, especially in the case of charge density studies, where high resolution and, therefore, weaker reflections are essential for the proper structure refinement, and the number of refined parameters is great.Expecting that some minute yet important effects may be lost due to the exclusion of weaker reflections, we decided to analyze the influence of applied I/sI cutoff on the series of multipole charge density refinements, conducted on a couple of datasets with sigma cutoffs of 4, 3, 2, 1 and 0 I/sI. The general quality of refinements measured by R-factors and residual density extrema improves slightly with the higher sigma cutoff applied. Interesting dependences of r(r) and Lap(r) vs I/sI cutoff are observed within the estimated errors range.