correct_data deleting some validation log entries

[rostal]

@collinschwantes
The validation log is filtered to only get the relevant rows. This includes filtering out blanks in the new_value. This step of filtering should be removed as for example the participant had selected Other, but then the reviewer found that their response was actually an answer they could have selected. So they set the new_val of the other explain question to "" and select the correct choice from among the options.

Please change:

changes <- dplyr::pull(
    dplyr::summarise(
       dplyr::filter(validation_log,
                            is_valid == "FALSE" | is_valid == "F", 
                            !is.na(field), 
                            field != "", 
                            !is.na(entry),
                            entry != "",
                            new_val != ""), 
          n = dplyr::n()),
          n)

to:

changes <- dplyr::pull(
   dplyr::summarise(
      dplyr::filter(validation_log,
                           is_valid == "FALSE" | is_valid == "F", 
                           !is.na(field), 
                           field != "", 
                           !is.na(entry),
                           entry != ""),  # dropped new value = "" as this is sometimes a valid response.
          n = dplyr::n()),
         n)