The mixing of soil and binders is a very complex process. The object of the mixing process is to distribute the binder and to achieve a satisfactory mixture for the particular purpose. The requirements on homogeneity may differ for each application. The assessment and definition of mixing quality are difficult and very expensive. This paper discusses the concept of mixing quality with respect to the binder distribution, and the use of mixing indices, as a quantitative assessment. The possibility of using mixing indices with the lime–cement column method is taken up. Statistical analysis was performed on a number of samples taken from four lime–cement columns at Arboga, Sweden. The sampling methodology, the scale of scrutiny and thus the influence of the sample size are discussed. The study shows that mixing indices can be used as a quantitative measure of the mixing quality. The results indicate that the sample size has a considerable influence when the mixing quality is poor and a lesser influence when the mixing quality is relatively good. It is also shown that it is difficult to draw conclusions concerning binder distribution from only a few samples, since the distribution may vary. With more extensive knowledge, it may be possible to assess the mixing quality from only a few samples.