From what I know, you should use the question code or group code instead the SGQA code. However, reading here www.limesurvey.org/manual/Expression_Manager#Key_Definitions it seems like the SGQA could still be used.

But I am not quite sure what you are trying, because it looks like you check if a question code equal to it's QID and the same for the group. This doesn't make much sense to me...

Read through www.limesurvey.org/manual/Expression_Manager which helps quite good to understand how the EM works. If you would like to reference to the Question ID instead of the question code (that you can define yourself), then you must follow the rules of the SGQA code. You can't just reference the QID directly.

Thank you holch.

I have having a lot of difficulty understanding the explanations of expression manager in the manual - so I am asking here.

Your link to SGQA code gives a 404 error.

It sounds like you answer my question in the final line: cannot reference QID directly.

In that case, I'm not able to figure out what else - if anything - can be referenced directly.

I see that some Names[ID] can be referenced, but is that the only thing?

On the Expression Manager manual page there is an example of the SGQA code. It is quite simple.
S = Survey (Refers to the survey id)
G = Group (Refers to the group id)
Q = Question (refers to the question id)
A = Answer (refers to the answer id)

Here the manual page for the SGQA identifier: www.limesurvey.org/manual/SGQA_identifier

Often it helps to search with "limesurvey xxxxx" in Google, which seems to work better than the actual search in the manual.

On the Expression Manager manual page there is an example of the SGQA code. It is quite simple.
S = Survey (Refers to the survey id)
G = Group (Refers to the group id)
Q = Question (refers to the question id)
A = Answer (refers to the answer id)

Here the manual page for the SGQA identifier: www.limesurvey.org/manual/SGQA_identifier

Often it helps to search with "limesurvey xxxxx" in Google, which seems to work better than the actual search in the manual.

Try also to understand the examples on the Expression Manager page. You can learn a lot from them and how Expression manager works.

I am finding the information on that page too complex (and in the manual generally). The examples on relevance relate more to inserting previous answers than to the mechanics of writing a simple instruction to branch.

I am unable to evaluate the correctness of an expression, and don't see all the rules about syntax. So I'm just guessing at it again and again until something changes.

Whilst it's clear there are many powerful tools to aid development, there is a steep learning curve/high bar to understanding their proper use. There are many ways to misunderstand the information!

Definitely, Expression Manager is very complex and thus it is not learned in 5 minutes. This is advanced survey development, which will also require a little advanced learning and some time investment. There are also more pages.

However, it is impossible to explain the basics of EM in the forum. Here we can only help on specific problems and questions.

Okay.
I appreciate the limitations of the forum.

So far I have put in about 120 hours over 6 weeks trying to figure out how Lime Survey works.
Some of that development has been straightforward to figure out. Other bits not so much. I am hoping another week will get me to a point where I can activate a 30-question survey.

120h is quite a bit, you must have quite a complicated survey.

We always try to help, but it is always better to have a specific problem. I don't really understand your problem described above.

It's not a complicated survey - it's probably about as simple as they come. Probably too simple to have considered as an example in the manual. It's just a bunch of forward jumps. But it's laborious typing it out in order to test the expressions. So I've spent about 90 of those hours just monkeying with the coding.

The expression example I put up earlier was meant to describe a jump from either of two answers, in the form of ((Question1 AND Answer 1) OR (Question 3 AND Answer 1)). The survey looks like it will have 3 or 4 extraordinarily long expressions of this sort, where any of over 20 questions will jump to the same point later in the survey. I expect it to look like ((Question1 AND Answer 1) OR (Question 3 AND Answer 1) OR (Question 7 AND Answer 1) OR ... (Question 19 AND Answer 1) OR (Question 25 AND Answer 1)).

But I have given my questions the same variable names as the question groups they are in. This may be why some of the jumps fail. Group 1 is called Access, Question 1 is also called Access. When I named them I did not understand the limitation about naming variables for use by expression manager. I still don't know anything beyond that one fact.

So I need to go back and rename all of the questions. Or find some other way of specifying the target in EM.

So I asked if I could use another identifier like QID. That didn't work, so I will go back and rename the questions. THen see if there are other jump failures from as-yet-unknown reasons.

Here's a puzzler.

On a survey logic display, the order of group labels are displaced part way through the survey.

Initially, Group rows are Followed by Question rows and Answer rows.
At one point, the Group Name shifts, and shows up after the Answer rows.
The GID does not shift.

In short, only the label is displaced.

Here's a simplified diagram of the shift.




It looks as though the move of Question1 into Group 0 has caused the Group 1 label to displace onto the Group 2 row.

Is that right?

I suppose the answer is that the Reordering Tool doesn't know that a Group requires a Question. It allows me to move any question from one group to another, including the last remaining question.