Please help us help you and fill where relevant:
Your LimeSurvey version: 5.3.32 (build 220817)
Own server or LimeSurvey hosting: own
Survey theme/template: bootswatch_extended
==================
I configured an anonymous, token-based survey using participants table:

• Anonymized responses: On
• Enable participant-based response persistence: On
• Allow multiple responses or update responses with one access code: Off

During the pilot testing, a respondent encountered problems due to bad connection of her mobile that resulted in several interrupts during her attempts. The problem is that all four failed responses were saved under different IDs, and finally the completed response was also saved under a different ID. As a result, the data includes five responses under five different IDs by the same respondent. I think that this should not happen, since no multiple responses are allowed with the same access code and because this obviously distorts the data (for example, if you plan a missing data analysis).

Is there a way to save all interrupted responses by the same user under the same and unique ID?

Thanks

Response persistence does not work with anonymous surveys - www.limesurvey.org/manual/Participant_se...response_persistence

I would disable multiple responses and ignore the incomplete responses when analyzing the data.

Thank you for your answer. However, this makes impossible to conduct a reliable missing data analysis and/or make evidence-based decision upon multiple imputation. Since token-based response is enabled, it would make sense for LimeSurvey to record one and only one response per token regardless of failed or interrupted attempts.

baltzis wrote:

Since token-based response is enabled, it would make sense for LimeSurvey to record one and only one response per token regardless of failed or interrupted attempts.
 

Which is possible, when using the right settings.

• Anonymized responses: Off
• Enable participant-based response persistence: On
• Allow multiple responses or update responses with one access code: On (which results into updating responses cause response persistence is on)

With these settings the token is leading to one record, which the respondent can reaccess and update.
You sometimes have still some additional records cause by glitches, but you get a much cleaner responsetable.

Insisting on anonymized responses and demanding to have one record per token would result into forcing respondents to only access the survey once with their token. The token counter would then be reduced when accessing the survey and not when completing the survey. Which would result into many incompletes too.