SPSS

1
SPSS

• Recently it has gone through a name change so
  your icon on your computer may be under a
  different name (i.e. PASW- Predictive Analytics
  SoftWare).
• IBM purchased the company and said they are going
  to change the name to IBM-SPSS or SPSS An IBM
  Company

2
What is SPSS?

• Statistical software
• Statistical Package for the Social Sciences
• There are lots of others
• SAS
• STATA
• R
• SHAZAM
• Specialized software
• HLM
• LISREL
• BUGS

3
Why SPSS?

• Although SPSS is not the most powerful
  statistical software - It is the easiest to learn
  and use!!!!!!
• It is very flexible and can do most everything we
  need
• It can examine a large amount of data very
  quickly
• It is competitively priced

4
Where to start?

• Get some data
• In practice you would develop a research question
  first, a theory, hypotheses, etc. before
  collecting and analyzing data

5
Sources of data

• ICPSR - http//www.icpsr.umich.edu/icpsrweb/ICPSR/
  
• Free for universities that subscribe (UNLV does)
• For now we have a dataset on our course website
  and WebCampus
• http//faculty.unlv.edu/kfernandez/methods1.htm

6
General Social Survey

• NORC (National Opinion Research Center) located
  at the University of Chicago
• Telephone survey since 1972
• Free to anyone
• http//www.norc.org/GSSWebsite/
• It is big so use the edited 2002 version located
  on our course website

7
General Social Survey

• The dataset is huge with hundreds of survey items
  (variables).
• The dataset we are using today has been edited to
  include only 47 variables
• This is also an older dataset (sometime after
  1996).
• The NORC website has a free 2010 version if you
  are interested.
• This could be a dataset you use for your research
  design assignment

8
The Basics

• This is a one level, rectangular dataset
• What do the rows represent?
• What do the columns represent?
• This dataset could be connected to other datasets
  to create multiple-level datasets
• i.e. individual, state, country

9
First things first

• Setting the options will make life easier
• Go to the pull down menu edit
• Select Options
• Select the General tab at the top
• Under Variable List click Display Names and
  Alphabetical
• Now select the Output Labels tab
• Make sure they all say Names and Labels or Values
  and Labels

10
Lets get our feet wet

• Most tools will be under the analyze pull down
  menu
• Select Descriptive Statistics and then
  Frequencies
• Find the variable Tvhours. The list should be
  alphabetical now. Move it to the Variables box
  and click ok.
• What does this table mean?

11
Missing Data and Outliers

• Is there anything strange in any of the
  responses?
• What about missing data?
• How many respondents didnt answer the question?
• NAP Not Appropriate NA Not Answered

12
Recoding

• Variables often need to be cleaned or recoded so
  that they can be used in a statistical model
• For example find the variable income
• Run a frequency
• There is a value that if not changed will really
  mess up your analysis
• What is it and why is it a problem?
• What do we need to do?

13
Recoding

• Select the Transform pull down menu
• Click the recode into the same variable
• WARNING, WARNING
• Usually I choose the recode into different
  variable because if you make a mistake you will
  forever change that variable and wont be able to
  go back. Always have a backup of your dataset
  that is untouched so you can start over if you
  need to.

14
Recoding

• Select income
• Click Old and New Values
• In the Old Value type 24
• In the new value select System-Missing
• Click Add
• Click Continue
• Now Click OK
• Rerun the frequency for income
• Now 23 is the highest value, not refused