Assignment 3 – Survey Analysis Exercise
Assignment 3: Survey Analysis
35 pages, including tables and optional figure(s)
Please submit your assignment as a single file. Include your ID number and page number on each page.
Overview
For your third exercise, you will conduct a quantitative analyses of survey data collected by the General
Social Survey (GSS). You must specify a research question based on two variables and test for an
association between these two variables.
All analyses can be conducted using the Survey Documentation and Analysis (SDA) website
(http://sda.berkeley.edu/sdaweb/analysis/?dataset=gss16 (http://sda.berkeley.edu/sdaweb/analysis/?
dataset=gss16) ).[1]
The key components are:
1. Introduction and research question
2. Description of data and sample
3. Description of selected measures
4. Univariate analysis, including table of frequencies and/or summary statistics
5. Multivariate analysis, including crosstabulation table and chisquare hypothesis test
6. Interpretation and conclusion
Details
Your assignment will be evaluated primarily on the inclusion and correctness of the 6 key components listed
above. I recommend that you make a separate subsection for each component. Here is additional
information for each one:
1. Introduction and research question (~1 paragraph)
2018/11/10 Assignment 3 – Survey Analysis Exercise
https://canvas.eee.uci.edu/courses/12744/assignments/256675 2/5
Begin your exercise with a brief introduction that motivates your analyses. In most cases, this should take
one paragraph. At the end of your introduction, you need to clearly state your research question. For full
credit, you must state a research question in which you clearly identify an independent and
dependent variable. I also recommend that you state a hypothesis.
You may use the research question from Exercise 1 if appropriate. If not, I strongly suggest you ask for
feedback on your research question during discussion section or office hours.
NOTE: Unlike Exercise 1 (a full research report), you do not need to include a literature review or references
to previous research. But you still need to introduce your topic and explain why it is of interest. This context
allows your reader to better understand why you made the methodological choices you did in your analysis.
2. Description of data and sample (~2 paragraphs)
Describe the source of the data and the specific sample that you have analyzed. The data have been
collected as part of the General Social Survey, so you can find information on the GSS website.
Your sample will vary based on your population of interest. This may include all respondents to the GSS
(from 1972 to 2016), or it may include only specific years or only a particular subgroup (such as only women,
or only respondents with a religious affiliation).
For full credit, you must report the sample size for your analysis.
3. Description of measures (~2 paragraphs)
Your analysis should focus on two primary variables. Describe them. For full credit, you must explain the
survey items. What was the question and what were the response categories? What concepts do these
measures represent? How well do they relate to your stated research question? How were they asked, and
what were the response categories? What level of measurement (nominal, ordinal, interval, ratio) applies to
each?
You should also describe any recoding or collapsing of variable categories you used in your analysis. I
recommend that you create simplified binary (two value) version of more complicated variables. If so,
explain your procedure. For instance, if you are considering education, you might compare “college
graduates or above” versus “less than college.” If so, you need to explain why (maybe your hypothesis is
that college experiences are the most important aspect of education) and how (maybe you recoded years of
education into 16 and above versus 15 and below).
4. Univariate analysis: frequencies and means of selected variables (~12 paragraphs and 2 tables)
Summarize your variables. For full credit, you must present table with frequencies or summary
statistics (such as mean, median, mode, and standard deviation as appropriate) of your primary variables.
You can include output from SDA directly or create your own tables. You may also include a graph for each.
2018/11/10 Assignment 3 – Survey Analysis Exercise
https://canvas.eee.uci.edu/courses/12744/assignments/256675 3/5
In text, describe any notable features of these variables, such as the most or least common category.
5. Multivariate analysis: crosstabulation and chisquare hypothesis test (~23 paragraphs and 1 table)
Analyses the association between two variables. For full credit, you must report a crosstabulation table
of your two variables. The rows and columns should be clearly labeled, and you should report column
percentages. You can include output from SDA directly or create your own table.
Your interpretation in the text must include the chisquare statistic for this table, the pvalue, and the
statistical conclusion. Explain whether the association between the two variables statistically significant
and how you know.
In addition, how do you interpret key values in the table cells? Are any unusually large or small compared to
what would be expected by chance?
6. Interpretation and conclusion (~12 pages)
Briefly summarize your findings. What do they mean? For full credit, you must explain the answer to your
research question reflected by this analysis. Do the results support or refute your original expectations?
Your interpretation should also include your own assessment of your study. What are the strengths and
limitations of these analyses, including those related to the sample, measures, and analysis? Do you results
lead to further questions for future research?
Note on Selecting your Topic and Questions
You do not need to select a topic that is related to your previous exercises or your final assignment. Indeed,
while there are hundreds of questions in the GSS for you to choose from, you are limited by what is
available. But I recommend that you look for a topic that is at least thematically related to your interests. My
hope is that this exercise will give you ideas for your final research proposal.
Grading
Your paper will be evaluated on:
1. The inclusion of all required components
1. Introduction and research question
2. Description of data and sample
3. Description of selected measures
4. Univariate analysis: frequencies and summary statistics of selected variables
5. Multivariate analysis: crosstabulation and chisquare hypothesis test
6. Interpretation and conclusion
2018/11/10 Assignment 3 – Survey Analysis Exercise
https://canvas.eee.uci.edu/courses/12744/assignments/256675 4/5
2. The quality of the components of your paper. For example, this includes:
1. Completeness: addresses aspects described in details above
2. Depth of analysis of responses and reflection on research process
3. Cohesiveness: alignment between motivation, questions, and analysis
4. Clarity in writing and style
3. Timeliness: late assignments will be penalized a letter grade per day
Resources
We will discuss examples in class and in section. Additional resources are posted on the course website.
Relevant resources online:
Survey Documentation and Analysis website, including documentation and tutorials:
http://sda.berkeley.edu/ (http://sda.berkeley.edu/)
Tutorial on SDA interface and basic tables:
Intro to the User Interface and Basic Tables Updated (https://www.youtube.com/watch?
v=rBC3wT5GE_c)
Minimize Video
Tutorial on recoding variable values:
OntheFly Recodes (https://www.youtube.com/watch?v=NOYeR82DUM4)
(https://www.youtube.com/watch?v=NOYeR82DUM4)
SDA’s GSS website: http://sda.berkeley.edu/sdaweb/analysis/?dataset=gss14nw
(http://sda.berkeley.edu/sdaweb/analysis/?dataset=gss14nw)
The General Social Survey website: http://gss.norc.org/ (http://gss.norc.org/)
Intro to the User Interface and Basic Table…
2018/11/10 Assignment 3 – Survey Analysis Exercise
https://canvas.eee.uci.edu/courses/12744/assignments/256675 5/5
You can search for variables using the Data Explorer:
https://gssdataexplorer.norc.org/variables/vfilter
(https://gssdataexplorer.norc.org/variables/vfilter)
The following sections of the textbook are most relevant:
Chapter 8, especially Box 8.2 “Writing Survey Questions”
Chapter 12, Especially sections:
Box 12.1 “Codebook Documentation” (369370)
“Inspect and Modify Data: Nominal and OrdinalScale Variables” (372375)
“Carry out Preliminary Hypothesis Testing: Nominal and OrdinalScale Variables” (379386)
“Elaboration of Contingency Tables” (392395)
[1] Note help links on the SDA website, especially links for “General” and “Recoding Variables.” There is
also a 15 minute tutorial video listed in the Resources section below.
