When you create question with type multiple selection not support for spss syntax file

User Support Form Building
,
1

In spss variable view label and values not present for Checkbox type questions.

How can we fix it?

2

Could you kindly simplify your issue so that we are able to understand and help you out if it’s possible with KoBoToolbox.

3

Hi,

This is a known issue and it is one of the downfalls of the syntax. Unfortunately there is no clear approach to how you can incorporate the syntax for the multiple select variables. I have always had to hardwire the syntax for my needs on the SPSS variable.

Stephane

1 Like
4

I have created a example form like below

do1
do1701×572 9.3 KB

My survey tab

do2
do2950×175 7.58 KB

My choices tab
do3

When I run the spss syntax file, It will show some error

do4
do41281×659 54.9 KB

My data view

do6
do61156×210 27.7 KB

Here Q3 column take two values because I have checked both.
Q31, Q32 column take options separately.

My variable view

do5
do5921×353 37.9 KB

Row 6, 7 values None
1My need is values should be checked, unchecked

5

Hi @bamini,

Could you mind providing us your username and project name through a private message. I would like to have a closer look at your case.

7

Hi @bamini,

I did not face any issue. I simply did the following:

  • Downloaded the data in XLS (legacy) format
  • Downloaded the SPSS Labels
  • Opened SPSS software
  • Imported XLS (legacy) in SPSS
  • Opened SPSS Labels
8

Hello @bamini , @Kal_Lam and @stephanealoo,
I think, you manually adapted some names in Excel, but not also in the SPSS Label file (like _id >> @_id), so you get log warnings for not mapping.

Our experience is to better use normal XLS (Header …) and SPSS labels export, NOT legacy (or CSV).

As Kal_Lam, we don’t encounter any problems with select_one labels:
VALUE LABELS
SexHHM
‘1’ ‘Female’
‘2’ ‘Male’

The only adaption needed is to rename internal KoBo variables with leading _ (underline) and replace the slash in the (single) columns for multiple select:
VARIABLE LABELS
WhyNotSchool ‘What is or was the Main Reason for NOT attending School?’
/WhyNotSchool_A ‘What is or was the Main Reason for NOT attending School? :: No money for school costs’
/WhyNotSchool_B ‘What is or was the Main Reason for NOT attending School? :: Supporting the family’
/WhyNotSchool_C ‘What is or was the Main Reason for NOT attending School? :: Work outside the family’

We do these textual adaptions in Excel and the SPSS label file (with simple replace function).

HINT: Pay attention:

  1. The SPSS Labels export does NOT filter (even using the new KoBo data export filter options). So, you may get more SPSS VAR labels than variables filtered/exported, resulting in a Warning on Label import (as in your example.)

  2. You may have some Labels longer than in your real (imported) data usage, like 98 for “Don’t know”. As SPSS will (first) automatically determine the variable format based on the data for ex. 1 = Yes, 2 = No (and no Don’t know case), you may want to enhance/adapt the FORMAT (You get a Warning for this choice label - see below), but only if you want to have the complete Label set imported, even the one’s not used for data (might be interesting for analysis nevertheless).
    "Warning …: The (ADD) VALUE LABELS command specifies the labeling of a value which is
    inconsistent with the type and/or the format of the variables to be labeled.
    The value and the applicable label will be ignored. …
    ‘98’ ‘Don’t know’.

  3. We generally avoid exporting with Groupname-Prefix, as you will get again the / issue. (I even think, SPSS will automatically replace / with blank for Variable names. But not for the Labels.)

Kind regards

9

Dear @stephane,
The problem is the “/” generated by KoBo for the multiple (single) items and the _ (underline) at the beginning of some internal variables. (Maybe in future the KoBo export might be optimised?)