Tech Tip: Output Management System (OMS) in IBM SPSS Statistics

Output Management System

(OMS) in IBM SPSS Statistics

What is OMS?

The Output Management System (OMS) provides the ability

to automatically write selected categories of output to

different output files in different formats

Can write output files into the following formats:

• Word, Excel, PDF, SPSS (.sav) , Output file (.spv), web report

format (.spw), XML, html and text

OMS provides a means to automatically writing pivot table

output from a specific procedure (e.g. Frequencies) to files in

the above formats

2

Why use OMS ?

You might want to generate a table of summary statistics not

available through the aggregate procedure and merge them

back with the original dataset

You might want to automatically exclude selected categories

of output or specific items from the Output viewer. i.e. you

may want to exclude case processing summaries or some

other output

OMS is very powerful

3

Utilities -> OMS Control Panel

Running OMS from the menu

4

The OMS control panel is used to both start and stop the

routing of output to various destinations or to exclude types

of output from the Viewer

Can create multiple independent OMS requests at the same

time

The same output can be routed to different locations in

different formats

OMS Control Panel

5

OMS Control Panel

6

1) Specify the Output Type(s)

2) Select a Command Identifier(s)

3) Select table subtype(s)

4) Specify the format of the output file

5) Specify the destination of the output file

6) Select if you want to exclude the output from the viewer

7) Add the OMS request

To create an OMS request

7

Look at creating some output using the training file

“dietstudy.sav”

“C:\Program Files\IBM\SPSS\Statistics\*Version\Samples\English”

*Version: Your SPSS Statistics Version, e.g. 20, 21, 22,…

Look at creating some output using the training file

“dietstudy.sav”

We will create a correlation matrix between using 10 tg & wtg

variables and output this to an SPSS file. We can then

calculate the average correlation

First we create the OMS request !

OMS Example

8

If we want to work with the output from this Correlation

example say to calculate the average correlation we will

need to create an OMS request

Go to OMS Control Panel

• Select Tables under Output Type

• Select Correlations under Command Identifier

• Select Correlations in the Table Subtypes

• Select New Dataset and type corr

• Select Add request

Creating an OMS request

9

Creating an OMS request

10

Creating an OMS request

11

Each time an OMS request is

created a summary alert tells

you that the request will be

created and will remain active

until we end it or we end the

SPSS session

Creating the Output to “OMS”

Analyze -> Correlate -> Bivariate

12

In order to re-route our output to a new dataset we need to

End the OMS request

Utilities -> OMS Control Panel -> Select the request -> End-> Ok

Again you see the OMS summary but the number of requests to be

ended will be 1

A new dataset named “corr” will be opened

Routing your output to the new format

13

New Output – corr dataset

14

The new file “corr” has exactly the same structure as the

correlation matrix output

The Var1 column contains the variable labels from the rows

of the Correlation table

The variable label is repeated three times because there are

three statistics associated with each variable (Pearson

Correlation, Sig (2-tailed), and N)

We only want to look at the Pearson Correlation

We can use the Select Cases procedure to only select out

cases where Var2 = ‘Pearson Correlation’

Working with the new file

15

In order to only select out the correlation values we specify to

select cases where Var2 = ‘Pearson Correlation’

In order to work with only the correlation values we can

select “Copy selected cases to a new dataset” option

We could name this dataset “Only_correlations”

Select Cases dialog

16

Selecting out only the correlation values

17

Creating a new dataset

18

The table includes the correlation between a variable and

itself (a value of 1)

We want to exclude these correlations so we can define 1 as

a missing value

We also need to calculate the absolute value of all

correlations to calculate a meaningful average

We can do this efficiently in syntax using a Do Repeat

Command

Extra Modifications

19

Defining missing values (i.e. 1 is missing)

20

Calculating absolute values

21

Can use the Descriptives procedure to report on the average

for the correlation variables

Analyze -> Descriptive Statistics -> Descriptives

Select the Mean, Min, Max and Standard Deviation

Select only the correlation variables

Select Descending Means

Calculating the Mean Correlation

22

Final Report

23

If you want to automate the OMS facility it can be ran

through syntax

The OMS Identifier dialog can paste the command subtype

identifiers

Further OMS

24

www.presidion.com

Talk to us

[email protected] +44 (0)208 757 8820 (UK) +353 (0)1 415 0234 (IRL)

www.presidion.com/ibm-spss-technical-tips

For more Tech Tips

visit