Hi Muthannan

I've just done some "playing" with your Death Score method vs the Reed & Muench method - I can send you the data separately if you'd like. What I did was to do the estimates from both methods by varying both the number of inoculations (25, 20, 15,12, 10 and 5 ) and the dilution series (i.e. 10-fold, 5 -fold, 3-fold and 2-fold). The results are interesting.

Your method gives values essentially identical to those from Reed & Muench if a widely-spaced dilution series, such as a 10-fold, is used for the inoculations no matter how many inoculations were used (the difference between 25 mice  inoculated for example was the same as that for 8 mice inoculated. However, the Death Score method was less comparable to Reed & Muench, and increasingly "underestimated" the R&M values as the serial dilutions were closer; but again it did not matter much how many inoculations were used in the comparisons. For 10 inoculations (mice, or tissue culture wells or whatever) with 10-fold dilutions the DS method was log 0.02 lower than R&M, for 5-fold it is log 0.30 lower, log 0.53 lower for 3-fold and log 0.70 lower for 2-folds. 

Since these deviations look fairly consistent I would think that if the Death Score method was used it would be simple to just correct for such deviations depending on what dilution series was used,  if it was important to reconcile or compare DS estimates with R&M. 

Dear Dr Scotti,

Thanks for your update. Kindly send the data you have used for the comparison. Once again, I would like to mention that my method will yield the result similar to Spearman-Karber method than RM. Regards Rama

Hi Rama

Sorry to be so long in doing this. I just did a comparison of the "death score method vs the R&M method. When I did this previously many months ago I "made up" data which I suspected may have not given a proper comparison and probably resulted in artefacts in the estimates.. What I just did was to dig out and go through my (very) old titration data and used these for calculating and comparing the results of these virus titrations from both methods. The same (actual) data, was used for both estimates; the R&M calculation was done using the spreadsheet I have posted but modified slightly so that it also included an additional estimate using your Death Score Method, i.e. same data, same time comparison with real data used. Your method and the R&M are impressively comparable. There are very minor differences in the estimates, not unexpected, and the differences are really only in the second decimal place! I don't want to necessarily post this data on research gate but would be happy to send it to you for you to look at if you have an email address I could send you the file directly. To say again, your method certainly gives as good an estimate as the Reed and Muench method and is much simpler. 

Hi Paul,

Thanks for the comparison and the update. For the benefit of the researchers, would it be possible to post the data?Thanks once gain your efforts in comparing the two methods.