I have performed an ANOVA on some log-transformed data. I had planned to transform the mean differences in the log-transformed data back to the original scale and interpret them as ratios of the geometric means of the original data, as described in the top answer to this question: Interpreting test results on log-transformed data
However, some of the values of my dependent variable are 0, so I added a constant of 1 to all of the data that I transformed before I applied the log function. So now, when I transform the mean differences in the log-transformed data back to the original scale, I have the ratio of the geometric means of my data with this +1 transformation still applied. Is there any way I can obtain the ratio of the geometric means of the original data?