The following working examples refer to the practice dataset from the practice questionnaire .

1. Compute

1.1 Transform -> Compute

1.2 We want to construct a new variable to examine whether TV is a major source of advertisement received by our respondents.

• we have asked 5 types of products
• the relevant questions are 2, 6, 10, 14, 18
• the coding is 1 (chosen) and 0 (not chosen)
• it is different from questions 4, 8, 12, 16, 20
• *** to add the "values" in questions 2, 6, 10, 14, 18 is meaningful ONLY if the coding is either 1 or 0

compute tvad = v203 + v603 + v1003 + v1403 + v1803.
execute.

• the new variable will have range of values from 0 to 5

1.5 After we have constructed 6 new variables, we can compare which source is the most popular, and which is the least, using the Frequencies command to examine the distributions.
 

2. Count

2.1 Transform -> Count

2.2 We want to construct a new variable to examine how important TV is the source of advertisment referenced by our respondents.

• unlike 1.2, it is meaningless to use Compute , since the relevant questions (4, 8, 12, 16, 20) are coded as 0 (if not chosen), 1 (ranked first), or 2 (ranked second)
• Count has a different logic of computation from Compute
• Count will not add up the "values" of the variables
• it will tally (i.e. counting the number of occurrence) the existence of "values" across the variables as specified in the command
• we do not mind if TV is ranked first or second, so codes 1 and 2 will fit our criteria

• the relevant variables (not questions) are v403, v803, v1203, v1603, v2003

• the new variable will have range of values from 0 to 5

2.5 After we have constructed 6 new variables, we can compare which source is the most important, and which is the least, use the Frequencies command to examine the distributions.
 

3. Recode - Another Ways to creat new variable

3.1 In the above examples, you may or may not notice that Compute and Count are able to create new variables.

3.2 There are other commands in SPSS have the same function of creating new variables.  Among others, a commonly used procedure is Recode ... Into Different Variables .

3.3 In its simple operation, Recode is used to:

• combine response categories to facilitate crosstabulation analysis
• assign undesirable or illegitimate answers to the number corresponding to the missing value defined
• in either case, the variable under treatment CANNOT be recovered to its original value, but all labels and missing value declaration will remain unchanged.

3.5 Since Recode ... Into Different Variables creates a new variable, labels and missing value have to be defined again.

3.6 Either simple Recode or Recode ... Into Different Variables does not require to tranform all original values.  But the results are different.

3.7 Simple Recode

• keeps the untouched values as the original one and also the value labels will remain the same
• keeps the missing value
• the undesirable consequence is that if the value used to represent the combined categories is not existing originally, you have to label it
• on the contrary, if the value assigned to the combined answers corresponds to any original value, the label will follow the original one which may not truely reflect the meaning of the combined categories.  You have to re-label it when necessary.

• will NOT copy untouched values in the original variable to the newly created variable
• all untouched values in the new variables will be treated as "system missing value"
• to deal with this, you are require to "recode" any untouched answer to its own value
• variable label, values labels and user-defined missing value are empty, that means you have to define them again

• Into Same Variables (i.e. replaces the values of original variables)
• Into Different Variables (i.e. new variable created and the original variable remains unchanged, but missing values in new variable to be redefined)
• Recode -> If ... : clean up undesirable or illegitimate answers

4. To record procedures of variable transformation

4.1 Keep track of what you have done to the variables.

4.2 To apply the same procedures when new cases are added.

4.3 In all procedures, click    button instead of OK.
4.4 The SPSS Syntax Editor will pop-up, the commands of the corresponding tranformation are shown.

4.5 Highlight the necessary commands and also the "EXECUTE."

4.6 Press the Run Current button to complete the transformation.

4.7 If the result of the tranformation (and other procedures) meet your request, click the SPSS Syntax Editor window and save the syntax file, the extension of the file will be .sps .

4.8 At later time to re-use the commands, File -> Open -> Syntax
 

5. Basic Data Analysis

5.1 Univariate analysis: Analyze ->

• Descriptive Statistics -> Frequencies
• Descriptive Statistics -> Descriptives
• revision on basic inferential statistics

• Descriptive Statistics -> Crosstabs
• Correlate -> Bivariate
• Compare Means -> Independent-Samples T Test