We will update the Obsconst table in the Solar database (SunEntry) in January 2015. Attached is a document that describes this process.
When computing the American Relative sunspot number, k-statistics are used for each observer from at least 100 observations. These estimates will be calculated for adequate-seeing conditions as a random effect, (either G or E) and when the monthly mean sunspot numbers are high in the current solar cycle. This last requirment is especially important because if the k-factors are computed when sunspot numbers are low observers tend to 'over count'. If left unchecked this anamaly will almost certaintly result in artificially high sunspot numbers for the rest of the cycle. Something to consider during the solar minimum years to come.
If you are interested in keeping track of your own K statistic you can download William Wilson's spreadsheet found on this AAVSO page: