Survey Documentation and Analysis SDA

1
Survey Documentation and Analysis (SDA)
2
Workshop Agenda

• Overview
• What is online analysis?
• Available SDA data sets
• Statistical procedures (Frequencies, Crosstabs,
  Regression)
• Recoding, subsetting, downloading
• Teaching resources for SDA and developing
  instructional materials

3
SSRIC
Social Science Research Instructional
Councilhttp//www.ssric.org
4
The Council

• Oldest CSU discipline council
• Founded in 1972
• Representatives from CSU campuses meet three
  times per year
• Negotiates with data providers for
• access to data
• Promotes use of data analysis in
• research and teaching

5
The Council

• Annual student research conference
• at CSU Long Beach in 2008
• at CSU Sacramento in 2009
• Sponsors travel to ICPSR summer workshops in Ann
  Arbor, Michigan
• http//www.ssric.org/participate/icpsr_summer
• Works with Field Research
• Question credits to California Field Poll
• Selects faculty fellow

6
What is Online Analysis?

• Online data analysis" refers to the ability to
  perform statistical analysis using special
  Web-based software as an alternative to
  downloading data into a standalone statistical
  package on your computer.
• The software were using is called Survey
  Documentation and Analysis (SDA), which was
  developed at the University of California,
  Berkeley.

7
Alternative Statistical Packages

• You can get a complete list of available online
  statistical packages at http//statpages.org/
• Some of these include
• OpenStat
• ViSta
• Statext
• SISA

8
Advantages

• Many like SDA are free dont require a site
  license
• Only require a computer with an internet
  connection
• Some like SDA are easy to learn
• Can show students how to use some of them in 30
  minutes or less

9
Disadvantages

• Some online statistical packages (certainly not
  all) are limited in what they can do
  statistically
• Documentation is not very good for some
• Some (like SDA) can only be used with data sets
  that have already been created in a format that
  can be read by that package

10
Available SDA Data Sets
11
SDA Data Sets

• While SDA is an extremely easy statistical
  package to learn to use, its difficult to create
  SDA data sets.
• You have to purchase a SDA site license to create
  a data set and then learn how to use it.
• So we typically use SDA data sets that have been
  created for us.

12
Sources for SDA Data Sets

• SDA Archive located at UC Berkeley
  (http//sda.berkeley.edu/archive.htm)
• ICPSR Topical Archives (http//www.icpsr.org/cocoo
  n/ICPSR/all/archives.xml)
• Field data located at UC Berkeley
  (http//ucdata.berkeley.edu/data_record.php?recid
  3analyze)
• List of SDA data sets at CSU Long Beach
  (http//www.csulb.edu/library/eref/datasets.html)
  
• University of Denvers IDEA project
  (http//www.du.edu/idea/data.htm

13
SDA Archive at UC Berkeley (http//sda.berkeley.
edu/archive.htm)

• GSS Cumulative Datafile (1972-2008 2008 is a
  preliminary version).
• ANES Cumulative Datafile (1948-2000) and ANES
  datafiles for 1996, 2000, and 2004.
• Census microdata including 2000-2003 American
  Community Surveys and 1990 and 2000 U.S. 1 PUMS
  with separate files for 2000 and 1990 California
  PUMS.

14
ICPSR

• National Archive of Computerized Data on Aging
  (http//www.icpsr.umich.edu/NACDA/)
• National Archive of Criminal Justice Data
  (http//www.icpsr.umich.edu/NACJD/)
• Substance Abuse and Mental Health Data Archive
  (http//www.icpsr.umich.edu/SAMHDA/)
• International Archive of Education Data
  (http//www.icpsr.umich.edu/IAED/)

15
Field Data http//ucdata.berkeley.edu/data_record
.php?recid3analyze

• Field Polls from 1956 through 2006 are available
  as publicly-accessible SDA data sets
• More recent Field Polls are available as SPSS
  data sets (through FTP) for CSU faculty, staff,
  and students.

16
Other Sources of SDA Data Sets at ICPSR

• Voting Behavior The 2004 Election by Charles
  Prysby and Carmine Scavo (http//www.icpsr.umich.e
  du/SETUPS/)
• Investigating Community and Social Capital by
  Lori Weber (http//www.icpsr.umich.edu/ICSC/index.
  htm)

17
Statistical Procedures
18
Available Statistical Procedures

• Frequencies and crosstabulation (discussed in
  this workshop)
• Comparison of means
• Correlation matrix
• Comparison of correlations
• Multiple regression (discussed in this workshop)
• Logit/Probit regression

19
Using SDA

• Select the data set
• Look at the codebook
• Decide what statistical procedure to use
• Fill in what you want to do
• Run it

20
Data Set

• Were going to use the GSS 1972-2008 Cumulative
  Data File (2008 is preliminary data)
• http//sda.berkeley.edu/archive.htm
• Were going to use three variables
• SEX
• RELITEN
• PORNLAW

21
Frequencies

• List the variables you want to use
• ROW SEX,RELITEN,PORNLAW
• Click on Run the Table

22
(No Transcript)
23
(No Transcript)
24
Crosstabs

• Now lets use RELITEN as our independent variable
  and PORNLAW as our dependent variable to create
  two bivariate crosstabulations.
• List the variables
• ROW PORNLAW
• COLUMN RELITEN

25
Crosstabulation Continued

• Options
• Percentaging column
• Statistics
• Question text
• Color coding
• Run the Table

26
(No Transcript)
27
(No Transcript)
28
Your Turn

• Lets run two more bivariate crosstabs
• Independent variable SEX
• Dependent variables RELITEN and PORNLAW
• Go ahead and run these crosstabs

29
What Did we Discover?

• RELITEN is strongly related to PORNLAW.
• SEX is also related to both RELITEN and PORNLAW.
• Could the relationship between RELITEN and
  PORNLAW be spurious? SEX is related to both
  RELITEN and PORNLAW and could be creating the
  relationship between RELITEN and PORNLAW.
• How do we test this possibility? Lets run a
  three-variable crosstabulation with RELITEN as
  our independent variable, PORNLAW as our
  dependent variable, and SEX as our control
  variable.

30
Multivariate Crosstabulation

• List the variables
• ROW PORNLAW
• COLUMN RELITEN
• CONTROL SEX
• Options
• Percentaging column
• Statistics
• Question text
• Color coding

31
(No Transcript)
32
(No Transcript)
33
(No Transcript)
34
Spuriousness

• Was the relationship between RELITEN and PORNLAW
  spurious due to SEX?
• How do you know?
• Does that mean that the relationship can never be
  spurious?

35
Regression

• Crosstabulation is used when all the variables
  are categorical.
• What do we do when our variables are continuous
  (i.e., interval and/or ratio)?
• Regression is the answer.

36
Bivariate Regression

• Lets look at the relationship between the
  respondents socioeconomic status (SEI) and the
  amount of television one watches (TVHOURS).
• List the variables
• Dependent TVHOURS
• Independent SEI
• Options
• T-Tests
• Correlation matrix
• Color coding
• Question Text

37
(No Transcript)
38
(No Transcript)
39
Multivariate Regression

• Now lets add in another variable SEX
• But sex is not a continuous variable. How do we
  enter a variable like SEX into the regression
  analysis? Answer create a dummy variable.
• Dummy variables take on the values of 1 and 0.

40
Creating a Dummy Variable

• SEX (d1)
• SEX is the name of the variable to want to make
  into a dummy variable
• d indicates that you want to create a dummy
  variable
• 1 indicates that the value 1 will be assigned the
  value 1. All other values will be assigned the
  value 0.
• Run the table

41
(No Transcript)
42
(No Transcript)
43
Recoding, Subsetting, Downloading
44
Recoding Existing VariablesExample (from GSS
Cumulative File) ATTEND (How often Respondent
attends religious services)

• ATTEND0 Never1 Less than once a year2 Once a
  year
• 3 Several times a year
• 4 Once a month
• 5 2 to 3 times a month
• 6 Nearly Every Wk
• 7 Every week
• 8 More than once a week
• 9 DK/NA (Missing)

• ATTENDR
• 1 Seldom (0 to 3)
• 2 Sometimes (4 to 5)
• 3 Often (6 to 8)
• 9 Missing (9)

45
(No Transcript)
46
(No Transcript)
47
(No Transcript)
48
(No Transcript)
49
Your Turn

• Recode AGE into the following categories
• 1 18-29
• 2 30-64
• 3 65 and older
• Obtain FREQUENCIES for the result

50
For More Information, See

• http//sda.berkeley.edu/HELPDOCS/helpnewv.htmreco
  de

51
Compute a New Variable Example (from GSS
Cumulative File) Alienation Index

• Create measure of ALIENATION from these variables
  asked in 1978 only (all coded as 1agree,
  2disagree, other missing data)
• ALIENAT1 PEOPLE RUNNING COUNTRY DONT CARE
• ALIENAT2 RICH GET RICHER, POOR POORER
• ALIENAT3 WHAT YOU THINK DOESNT COUNT
• ALIENAT4 YOU'RE LEFT OUT OF THINGS
• ALIENAT5 POWERFUL PEOPLE TAKE ADVANTAGE OF YOU
• ALIENAT6 PEOPLE IN WASH D.C. ARE OUT OF TOUCH

52
(No Transcript)
53
(No Transcript)
54
(No Transcript)
55
Your Turn

• Create an index of parental education (MAEDUC
  PAEDUC)/2

56
For More Information, See

• http//sda.berkeley.edu/HELPDOCS/helpnewv.htmcomp
  ute

57
Subsetting and Downloading

• Example create and download a subset of the GSS
  cumulative file, selecting only cases from 2008,
  all Case Identification variables and some
  Personal and Family Information variables
  (MARITAL, AGEWED, DIVORCE, WIDOWED).
• At end of each intermediate step, click on
  Continue button.

58
(No Transcript)
59
(No Transcript)
60
(No Transcript)
61
(No Transcript)
62
SPSS Syntax File
63
Creating an SPSS system file

• Run SPSS (syntax) file against data (ASCII) file.
• For more information, see
• http//www.ssric.org/data/icpsr_direct (scroll
  down)
• http//www.ssric.org/data/icpsr_direct (scroll to
  Syntax Files)
• http//www.icpsr.com/cocoon/ICPSR/FAQ/0062.xml
• http//web.pdx.edu/stipakb/download/Data/SDA_data
  _to_SPSS.pdf (portions outdated)

64
File Directory
65
(No Transcript)
66
(No Transcript)
67
Your Turn

• Subset and download your own custom GSS SPSS
  system file.

68
Sample Instructional ApplicationsCrosstabs With
a Control Variable
69
Example 1

• GSS Cumulative File (selecting 2002 and 2004
  only)
• Crosstab Voting in 2000 election (VOTE00) by
  computer usage (COMPUSE).
• Repeat, but with a control for respondents
  education level (DEGREE).

70
Example 2

• ANES 2004 Study
• Instructors note In addition to using this
  example in teaching use of control variables, I
  also use it in teaching about reactivity in
  interviewing.
• Run frequency distribution for V5205 (Working
  mother can have warm relationship with kids).
• Crosstab V5205 with V1109a (Respondent gender).
  Weight by Post-election weight
• Repeat, but use V4103 (Interviewer gender) as
  independent variable
• Run frequency distribution for V4103
• Repeat 1 with a control for V4103
• Repeat 2 with a control for V1109a

71
Teaching Resources for SDAand Developing
Instructional Materials
72
ICPSR Web-Based Instructional Materialshttp//www
.icpsr.umich.edu/ICPSR/training/index.htmlinstruc
tional
73
Investigating Community Social
Capitalhttp//www.icpsr.umich.edu/ICSC/index.html

74
Voting Behavior the 2004 Electionhttp//www.icps
r.umich.edu/SETUPS/index.html