SPSS Syntax Export
ATLAS.ti can export your coded data in form of an SPSS syntax file as plain ASCII text file. This allows you to modify the syntax before running it in SPSS.
When transforming your data into numbers, codes and code groups become variables and all quotations become cases. Additional variables like document numbers, document groups and document type indicating where a quotation belongs to and document groups allow you to aggregate your data in SPSS.
Creating a Syntax File
From the Import & Export tab select SPSS Job.
The SPSS Job Generation Window opens. You have the following options:
Use separate file: When checked, the data matrix is written to a separate file. This is mandatory if the size of the matrix exceeds a certain size. SPSS cannot handle large data sets within a syntax file. For regular size projects, leave this option unchecked.
Specify the name of the data set. This name is used as the file name and as the FILE reference from the DATA LIST section. You only need to enter a name here if you generate separate data files.
Create a fresh data file during next run can be unchecked if the data has not changed since it was last created. This may save some processing time.
Include quotation's author. Check if you want to export an additional variable that indicates the author for each quotation. This is for example useful if you want to use SPSS to calculate inter-coder reliability.
The value separator is % by default, but can be changed to something else. See Naming Convention For Scaled Codes.
Create task section. Enable this option if you want templates for procedures included at the end of the syntax file.
Create SAVE OUTFILE instruction. Enable this option if you want SPSS to save the data as *.sav file after running the syntax file.
Specify all desired properties and click Create.
You can now run the file in SPSS, or make some changes to the syntax file before you run it.
The Syntax File Explained
The output of the SPSS generator is a complete SPSS syntax file containing variable definitions, optionally the data matrix and some default jobs statements:
VARIABLE LABELS are taken from code and code group names.
VALUE LABELS for variables created from codes are:
Yes (1) - code is assigned
No (0) - code is not assigned
Every case (quotation) is described by the codes that are attached it to, the document,the position in the document (start and end position), author, media type and date of creation.
Start and End Positions
There are two variables for the start position (SY and SX) and two variables for the end position (EY and EX) for a quotation. Depending on the media type, they are used for different start and end coordinates.
|Data Type||SY (start)||SX (start)||EY (end)||EX (end)|
|Text (rich text)||paragraph||Column character, quotation based||paragraph||Column character, quotation based|
|page||character count, page based||page||character count, page based|
|Image||SX and SY indicate the position of the upper left hand corner of the image||EX and EY indicate the position of the lower right hand corner of the image|
The date of creation is shown in the unix timestamp format. You can convert it into human-readable time by using a converter, for instance: https://www.epochconverter.com/
VARIABLE LABELS D 'Document'.
VARIABLE LABELS QU 'Q-Index'.
VARIABLE LABELS SY 'Start y-Pos'.
VARIABLE LABELS SX 'Start x-Pos'.
VARIABLE LABELS EY 'End y-Pos'.
VARIABLE LABELS EX 'End x-Pos'.
VARIABLE LABELS TI 'Creation Date'.
VARIABLE LABELS C1 'Audio'.
VARIABLE LABELS C2 'just a name'.
VARIABLE LABELS C3 'Evaluation'.
VARIABLE LABELS C4 'Buddha'.
VARIABLE LABELS C5 'Code name with more than 40 characters'.
VARIABLE LABELS C6 'Fish'.
VARIABLE LABELS C7 'Geo Code'.
VARIABLE LABELS C8 'Octopus'.
VALUE LABELS C1 to C2 1 'YES' 0 'NO'.
VALUE LABELS C3 0 'NO' 1 'bad' 2 'good' 3 'not so good'.
VALUE LABELS C4 to C8 1 'YES' 0 'NO'.
VARIABLE LEVEL C3 (ORDINAL).
Document Type Declaration.
IF (D = 5) MediaType = 1.
IF (D = 4) MediaType = 2.
IF (D = 6) MediaType = 3.
IF (D = 1) MediaType = 4.
IF (D = 3) MediaType = 5.
IF (D = 2) MediaType = 6.
- 1 'text'
- 2 'PDF'
- 3 'graphic'
- 4 'audio'
- 5 'video'
- 6 'geo'
Treatment of Code Groups
Code groups in SPSS jobs count how many of the codes in the code group have been assigned to the quotation.
Code groups are named: CG1, CG2 and so on.
Calculation: COMPUTE CG1 = C10 + C11 + C12 + C13 + C14 + C15.
VARIABLE LABELS CG1 'name of code group'.
Scaled codes are ignored in the computation of code groups variables. See Export for further Statistical Analysis.
Treatment of Document Groups
For each case (= quotation) there is a document and document group variable. If the document a quotation is part of also belongs to a document group, the value for the document group is 1.
Calculation: COMPUTE DG1 = 0. IF (D = 5 or D = 4) DG1 = 1.
VARIABLE LABELS DG1 'DG_Document group name'.