In SPSS, why do the SUM() and MEAN() functions keep cases with missing values instead of dropping those observations?
Statistical functions in SPSS [e.g., SUM(),
MEAN(), SD()] perform calculations using all
available cases. SPSS will not automatically drop observations with
missing values, but instead it will exclude cases with missing values
from the calculations. SPSS will correctly estimate the mean with the
MEAN() function by using all non-missing values.
However, problems can arise when trying to exclude missing cases
and estimate results based only on observations with complete
information. For example, suppose two variables (v1 and
v2) sum to create an index variable
(v3). While v1 has 10 valid cases with no
missing values, v2 has 8 valid cases and 2 missing
values. Use the following syntax to add the two variables and create
an index, v3:
The resulting index variable v3 has 10 cases and no
missing values. When SPSS encounters a missing value in any of the
v2 cases, it ignores it and sets v3 equal to
v1. Essentially, SPSS treats the missing values of
v2 as zeroes. The results can potentially be misleading.
To ensure that v3 is equal to the sum of
v1 and v2 and that all missing cases are
dropped rather than ignored, specify the minimum number of valid cases
that SPSS should use to calculate a given function. For example, to
create an index variable v3 using only observations
without missing values, execute the following syntax:
The .2 appended to the end of the SUM
function in the above example can be any integer. Use it to indicate
the minimum number of valid cases necessary to perform a given
calculation.
For more about statistical and mathematical software, email the UITS Stat/Math Center, visit the center's web page, or phone 812-855-4724 (IUB) or 317-278-4740 (IUPUI). The center is located in Bloomington at 410 N. Park Avenue, and is open for consultation by appointment Monday-Friday 9am-5pm.
Last modified on December 20, 2011.
