The following elements can be part of a numeric expression:

• Constants: (numbers) are specified conventionally, for instance 1, 30.9 -0.045, but you can also write 25.3E-3 using exponential notation.
• Operators
• Parentheses (): Used to indicate the hierarchy or operations and to make complex expressions more readable.
• Functions

There are two situations where the result of the evaluation of an expression is undefined and SPSS returns a SYSMIS value:

• A value of a variable is missing (user missing or SYSMIS)
• Domain error: the operation is not defined, e.g. division by zero, a log or a square root of a negative number.

If an operation implies a missing value and 0, the result is not always SYMIS, for instance [mv is a missing value] 0*mv, 0/mv will produce 0 (as do 0**mv or MOD(0,mv).

arg may be a simple argument or a numeric expression. The result

Statistical functions compute for each case a summary statistic across all variables specified as argument to the function (varlis).

• <varlis> is a variable list, e.g. MEAN(SYMP1,SYMP3,SYMP8,SYMP9,SYMP10), but also MEAN(ATT1 TO ATT20) using TO to indicate a sequence of variables.
• The result is based on the valid values across all variables. If all values are missing the result is SYSMIS. SD, VARIANCE and CFVAR require at least two valid observations, the other functions 1.
• You can control the minimum valid values for a function using the MEAN.n() notation, where .n indicates the minimum number of valid values required to compute the statistic; if the number of valid values is less than .n, SYSMIS is returned.

See also the SYSMIS() and MISSING() logical functions explained here with logical expressions

SPSS has a large number of specialized functions, useful when working with time data (time and date oriented functions and lags), string and conversion functions, as well as functions dealing with random variables and (statistical) distributions. The table below shows just a few, common ones as an example.