For my study I have to conduct an online questionnaire between two groups. The content of questionnaires is same. But there is only one difference, namely reading text in questionnaire consists of two parts (A and . The task (challenge) is that two reading parts should change their order according to the order of participants. For example, first participant gets texts in order A and B (normal). The second participant should get the texts in order B and A (reverse). The third one - in normal order (A and . The forth – in reverse order (B and A) and so on.
Lime survey offers randomizing function. But this function shows the texts in random order. It means two parts are not shown one after another, but randomly, e.g. 4 times in normal order A and 2 times in reverse order B. At the result I could have unequal participants. Is there someone here who can help me to find a better solution to this problem?

Yes, with the random order you can have an unequal number of participants. However, with a bigger number of respondets this should be fairly equal distribution.

But I don't see a way to easily guarantee an equal representation. Why?

You could read out the Saved ID of the started interview which is consecutive and when it is even show one order or when it is uneven show another order. But what happens when more people with and even Saved ID drop out than others? Right. Same problem.

Maybe you could write a script that reads out what the last participant that completed has seen and then show the oposite to the next respondent. But what when more than one are in the questionnaire at the same time? Correct, they will also get the same order and if they all finish --> Uneven number...

Binafsha wrote: At the result I could have unequal participants.

That demand pops up more often. Always in the light of universities and there DOE guide lines.

If N is very low the randomization of Limesurvey could be a problem. But since online surveys tend to have a few surveys which are not valid (not finished, double entries because browser crashed etc) you end up with a few unused or unusable responses. You cannot control that in a online survey. So why invest time to control things you cannot control?