As of now, Survey123 calculates percentages based on the overall number of records rather than on the subset of records that were actually tested, which can lead to confusion when the tested sample size is relatively small. For example, if there are 100 total records in the dataset but only 40 of those records have been tested, Survey123 still computes percentages using all 100 records as the denominator instead of using only the 40 tested samples. This means the resulting percentages do not accurately reflect the outcomes of the tested group and can be misleading for analysis and reporting, particularly in cases where testing coverage is limited or uneven. Calculating percentages from tested samples only would provide a more meaningful and interpretable representation of the results, especially when decision-making depends on understanding proportions within the actually evaluated subset rather than the full population.
