Is it possible to convert a data table in "R" to an octave scale (as done, for example, in the MVSP multivariate stats program)? I work with tables of word or category frequencies across a number of texts or text segments, e.g.: Token sect_1 sect_2 sect_3 sect_4 sect_5 sect_6 sect_7 sect_8 sect_9 sect_10 sect_11 sect_12 sect_13 sect_14 sect_15 sect_16 sect_17 sect_18 sect_19 sect_20 sect_21 sect_22 sect_23 advance 0 0 0 0 0 1 0 0 0 0 4 0 0 0 2 0 0 0 0 0 0 0 0 aed 0 1 3 0 0 1 0 0 0 0 4 0 0 0 0 4 2 3 0 0 0 1 1 agree 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 antibiotics 0 0 0 0 0 0 0 0 0 0 0 3 1 0 0 0 0 0 0 1 0 0 0 However, the texts/segments are typically of different lengths and the analysis program doesn't calculate proportional frequencies. (NB: It also doesn't select *all* words in the texts, so it is not possible to calculate true percentages "after the fact".) What I want to do is to transform the data before calculating distances and carrying out clustering or multidimensional scaling, so that the differences in text/segment size don't (heavily) bias the results. Many thanks, Andrew Wilson