Submit an empty/incomplete survey

Can the interviewer submit/send an incomplete or empty survey?
Our survey needs to be fulfill, otherwise is not going to work. Also, I tried to create validation conditions for each question using the ccondition “variablename != null” but is not working.
Thanks in advance, S


we will need a detailed explanation of what the “is not going to work” and “is not working” mean.

For the first one I can only guess. For the second, perhaps see here:

You can probably attract more answers from the peer users if you introduce yourself, your survey, and particular settings.

Best, Sergiy Radyakin

Oh sorry, let me explain better.
I have a complex survey with 147 questions but to test how Survey Solution works with incomplete answers or incorrect answers from the interviewers I created another to be answered as a sample mode: just to test

This simple survey on demo has only 2 questions:

  1. Question: Protected area name
    -> type of question: text
    -> variable name: apname
    -> validation condition 1: apname!=null
    -> error message: Empty data
  2. Question: Area (hectares)
    -> type of question: numeric
    -> variable name: area
    -> validation condition 1: area!=null
    -> error message: resposta vazia
    -> validation condition 2: area>0
    -> error message: area must be bigger than zero. Please, check your answer.

After importing the survey on demo, I tested in three ways:
1st) The interviewer respond correctly to the questions and submit
2nd) The interviewer did not respond to any question and submit
3rd) The interviewer respond correctly to one question and incorrectly to another question and submit

For the 2nd situation, Survey Solution shows the warning “Answer was not saved - Empty value cannot be saved” under each question. I pressed “Complete Interview” and it shows 2 unanswered. I pressed “complete” and was submitted.

For the 3rd situation, Survey Solution shows the warning “Answer is invalid Area must be bigger than zero. Please, check your answer”. I pressed “Complete Interview” and it shows 2 answered and 1 error(s). I pressed “complete” and was submitted.

So, because of these situations I was wondering if there is any way of “blocking” the submission to avoid having incorrect and/or empty data.

Thanks in advance, S


I think it is important to separate

  1. SUBMITTING an answer to a question;

  2. COMPLETING an interview, and sending (submitting) it to the server;

  3. have an interview FILLED OUT completely.

  4. every question type may accept or not accept some values. For example, date-type question will not let you enter the 32 of July of 2011. Even if you try, it won’t let you do this.

  5. every interview assignment will be eventually completed (will have the status Completed). Whether the questions are answered or not, whether they are answered correctly or not. Once the interviewer can’t get better information they must complete. It is a reality of life: the respondent may say she is tired and terminate the interview. And she doesn’t care whether it will be possible for you to do analysis later. Interview finished, done, and hence “Completed”.

  6. If you managed to get the answers to all of your questions your interview is completely FILLED OUT, congratulations. With real interviews of 300 questions and more this is almost never the case. On the contrary, it may raise suspicion that the interviewers are entering the imagined numbers (when the respondent refused to answer) just to avoid being punished. And this is what they almost certainly will resort to if you add a condition like you want to add. So instead of the true missing value, you will have values imputed (not inputted, but imputed) by the interviewers, which are not necessarily good at this.

My question was actually about YOUR procedures. You wrote “Our survey needs to be fulfill, otherwise is not going to work”. Why? Why do you wish to discard THE WHOLE INTERVIEW IF ONLY ONE VALUE IS MISSING? <== this is effectively what you are requesting. And if there is a reason that I am not seeing, why not just discard this assignment later, when you export the data? Using Stata’s terminology:
drop if (apname=="##N/A##") | area==(-9999999)

Best, Sergiy


thank you for your explanation.
My team and I understand your point of view. However, we think it will be a great improvement that Survey Solution had an option to establish the question as “required” or “not required”.
Our survey (with 147 questions) has been design to guarantee the maximum number of answers. For this, we are including options as “Not important”, “No data available” and “Not applicable”.
It is true that we can “discard” or “not discard” the survey but it depends on the type and importance of data we want to collect. In our case, 99% of the survey are questions related to the status of the protected area. If the interviewer “forget” to answer to the question “Please, identify the protected area” we will have to discard the whole interview even when 99% of the survey is answered correctly.

Note that, having an empty answer is not the same of not knowing, not having data and/or not applicable to the case.


Dear Susana,

I know nothing about your survey, but if 99% of the questions are related to the status of the protected area, place them in a section or a sub-section, and place a condition on it:
This will lock out those 99% of question until the interviewers answer the status question.

Note that historically, we had this option “mandatory question”. As a result, users marked all questions as mandatory. We removed it intentionally to prevent deadlocks.

Best, Sergiy