SOCY3039/7039 ASSESSMENT TASK THE RESEARCH PAPER
Your Research Paper is the final “write-up” of your research project. It is the main
piece of written assessment for the unit and should be structured as an academic research
article. It will contain the following sections the cover sheet:
1.
2.
3.
4.
5.
Title Page
Abstract
Introduction
Literature Review
Methodology
6.
7.
8.
Results
Discussion and Conclusions
References
9. Tables and Graphs
With the exception of the abstract, the number of words for each section is provided
as a rough guide only. However, unless you have strong reasons for doing otherwise,
you should stay reasonably close to the recommended word limits for each section.
1. Title page. This is the first page of your paper. It should contain the title of your
paper and your name and student number.
2. Abstract (approximately 100 words). This should be presented on a separate
page and should briefly summarise the topic, methods, results, and conclusions
of your research. You need to present all required material concisely and clearly
to obtain high marks for this component. (2%)
3. Introduction (approximately 150 - 300 words). This section should begin on a new
page from the abstract. It should introduce and describe your research topic in
precise detail, indicate why it is important and describe the structure of your
paper. You need to clearly and precisely specify your research question, its
significance and describe the structure of your paper to obtain high marks for this
component. (3%)
4. Literature Review (approximately 500 words). This section of your paper should
review some of the relevant theoretical and/or empirical literature. In writing this
section, you should avoid simply describing one piece of research after another.
This makes for boring and uninformative reading. Instead, you should attempt to
structure the literature review thematically, by focusing on common themes in the
research literature that relate to your research topic. Common themes may
represent a common set of theoretical concepts, or arguments, a common set of
research findings, a series of contradictory findings in an area, or a common set
of research methods. You should structure your literature review around these
themes and only refer to existing literature in the context of your research
themes. You may define specific concepts in your literature review, or you may
restrict definitions to the following section. Your literature review should begin
fairly generally and become more specific as you focus in on your topic. You
should aim to write your literature review so that your research problem and
1
methodology come to appear as the most logical way to proceed to carry out
further research in your field. You do not have to call this section “Literature
Review”. Indeed it is probably better to give it a more informative and interesting
section title. You need to provide a clear coherent, well-structured description of
relevant literature to obtain high marks for this component. (5%)
5. Methodology (500 - 600 words). This section describes your data sources and
methods of analysis. You should define the data source, in terms of the sample,
population and method of survey administration (e.g. telephone or mail). You
should indicate how key variables are defined, referring if necessary to question
wording, the method of index construction, and so on. You should also describe
your analytical strategy, that is, what statistical techniques you are going to use,
and why these are appropriate. It is very important to describe your analytic
strategy, relate it back to your topic, and justify your choice of strategy and
methods. You need to provide a clear, comprehensive description and defence of
all relevant aspects of your methodology to obtain high marks for this component.
(9%)
6. Results (approximately 1,000 words). This section is the main part of your paper.
It contains the empirical evidence that you have gathered in relation to your research
problem (i.e. tables and graphs), and your written description of what this evidence
says. For the purpose of the paper, it is fine for the Results section simply to
contain written text describing what your tables and figures (graphs) show. You
can then present the tables and figures separately at the end of your paper after
the References. From the point of view of preparing your manuscript, this is much
easier than attempting to integrate tables and figures into the written text of the
document. To indicate to the reader roughly where in the text tables or figures
should occur, you can say “Table 1 about here” or “Figure 1 about here” in the
place in the text where the table or graph would appear if it was integrated into
the body of the report. (This is in addition to referring to the table or figure number
in the text describing the table or figure). Note, for a Research Paper of this size,
you will not be able to deal satisfactorily with more than about six to eight tables
and figures in total. Ten tables and graphs in would probably be the maximum
upper limit. The trick is finding and presenting the most appropriate ones for your
research topic. In this section you need to make sure you spend enough time
talking about each of your tables. For each one, you should indicate which table
you are talking about, what the table deals with (e.g. a cross-tabulation of gender
by voting intention), which figures in the table you are referring to (e.g. column
percentages, cell means), and what the numbers mean. You need to provide a
clear, comprehensive, accurate account of your results to obtain high marks for
this component. (9%)
7. Discussion and Conclusions (500 - 700 words). This section of the paper
summarises your main findings and relates your empirical results to previous
research and to your research themes. It highlights any problems or difficulties in
your empirical analysis (such as a lack of suitable data or problems of validity
with existing data), may suggest directions for further research, and concludes
the paper. This section needs to be more than just a summary and restatement of
your findings. It should link your results to the literature you began with, identify
problems and say where the research might go from here. You need to say
clearly how your findings relate to your research topic or to other research in
the area to obtain high marks for this component. You also need to provide
evidence that you have thought about possible problems/limitations of your own
research, and directions for further research. (7%)
2
8. References. This section is headed “References” and is an alphabetical listing by
author of all sources cited in the text of your paper. You should reference
correctly and consistently, and only reference those sources you have cited in the
text. You need to reference correctly to obtain high marks for this component.
(2%)
9. Tables and Graphs. You do not need to head this section “Tables and Graphs”.
Here you present the tables and figures described in your Results section. You
should present your tables first, one per page, beginning with Table 1. Next
present your figures, beginning with Figure 1, one figure per page. You should
follow appropriate guidelines for the presentation of tables, with tables having a
table number and title. All tables should be properly labelled and sourced. Copy-pasting
output directly from STATA’s window or log files is not acceptable. You need to present
your tables and graphs appropriately to obtain high marks for this component. (3%).
If you use footnotes to make substantive comments, these should appear at the
bottom of the page on which the footnote occurs. Do not use footnotes to acknowledge
existing literature.
SOME POSSIBLE RESEARCH QUESTIONS
The following are some broad questions that you can use to develop a
specific topic for your Research Paper. Your own topic will probably need to
be more specific than these general questions.
1. What factors shape attitudes towards environmental issues?
2. What determines how people feel about the government?
3. Do individuals from different religious denominations have different attitudes towards homosexuality?
4. Who is satisfied with their job?
5. Who is “proud to be Australian”?
6. What is the effect of having children on men’s and women’s work hours?
SOME USEFUL EXAMPLES
3
The following articles follow (reasonably) closely the structure of your
Research Paper and may also be useful models for it:
Phillips, T.L. 1996. Symbolic boundaries and national identity in Australia.
British Journal of Sociology. 47: 113-134.
Tranter, B. 1999. Environmentalism in Australia: Elites and the public. Journal
of Sociology. 35: 331-350.
Western, M.C. 1999. Who thinks what about capitalism? Class consciousness
and attitudes to economic institutions. Journal of Sociology. 35: 351-370.
Healy, K., M. Haynes and A. Hampshire. 2007. Gender, social capital and location:
Understanding the interactions. International Journal of Social Welfare. 16:110-118.
Roberts, L.D. and D. Indermaur. 2007. Predicting punitive attitudes in Australia.
Psychiatry, Psychology, and Law, 14: 56-65
ASSESSMENT CRITERIA
In marking your final report, we will assign marks to each section as shown
above. You will obtain high marks for each section by including all required
material, presenting the material well and clearly (using correct grammar,
syntax, spelling, referencing style, tabular presentation) and structuring the
sections in a logical and coherent fashion. You will lose marks if your ideas
are unfocused, unclear or difficult to follow, material is badly or incoherently
structured and organised, material that should be present is missing, and/or
material is badly or incorrectly presented. In looking at each component, we are
interested primarily in your analytic ability and ability to organise and use
resources (i.e. structure material) appropriately, followed by the clarity with
which you present the relevant information. Note that if material is poorly
expressed or poorly organised, it generally indicates deficiencies in analytic
ability. Well-thought out and well-focused ideas are usually also well
structured and clearly presented. The following table defines what we mean by
analytic ability, organisation and use of resources and clarity of expression
and presentation.
Analytic Ability This criterion refers to your ability to formulate a
focused research topic, identify relevant research
issues, develop and implement a sound empirical
strategy, and situate your results within the existing
literature.
This criterion refers to the coherence and logical
clarity of the structure of your paper, and the
appropriateness of your literature and evidence.
This criterion refers to your use of correct grammar,
syntax and spelling, the clarity of your use of
English, the degree to which you acknowledge and
list references correctly and present tabular and
graphical material appropriately.
Organisation and
Use of Resources
Clarity of
Expression and
Presentation
4
NOTE THAT THE MOST IMPORTANT CRITERION FOR THE RESEARCH PAPER
RELATES TO YOUR ABILITY TO IDENTIFY A RESEARCHABLE TOPIC AND CARRY
OUT AN APPROPRIATE QUANTITATIVE ANALYSIS TO INVESTIGATE IT.
GENERAL ADVICE
1. Follow the structure for the paper described above. Each section of your
paper should cover the material relevant to that section. You may wish to
add an appendix section at the very end containing supplementary
material.
2. Try to do at least one draft of the paper, put it aside, and then come back
and edit the paper thoroughly, for spelling, grammar, presentation, and
logical coherence. All researchers edit their work and write multiple drafts.
3. If you don’t have time to edit and redraft your paper, at least make sure
you read it before you submit it. Make sure you correct grammatical and
spelling errors. Reading the paper aloud is the best way to identify
grammatical errors.
4. Pick a focused topic that concentrates on few key variables that you’re
interested in.
5. Develop a logical analytical strategy to examine your research question.
This means, identify the variables that are relevant, decide on your
methods of analysis (e.g. examining means or cross-tabulations, carrying
out a regression analysis, etc.) and carry out and write up your analyses in
a logical order.
6. Tell the reader what you are doing in the paper and why you are doing it.
This is extremely important. You need not only to describe your methods
of analysis, but say why they are the appropriate ones to use.
Appropriateness will depend on the kinds of variables you have (for
instance their levels of measurement), and the kinds of questions you want
to investigate.
7. Make sure you spend enough time describing your results. For a paper
like this your results section will probably only contain about 5-10 tables
or graphs, based on your quantitative analyses. You’ll probably need about
½ a page (i.e. 150 words or so) to describe each.
8. It is acceptable to stick with fairly “low-level” statistical techniques, such as bivariate
cross-tabulations, analyses of means by other independent variables, and so on, if
these enable you to investigate your topic in a satisfactory way. On the other hand,
if more advanced statistical techniques like linear regression are appropriate and you
feel confident enough to use them, this will make your analyses and the overall paper
stronger.
5
SOCY3039/SOCY7039 Applied Quantitative Research
Marking Scheme: Total marks add to 60, to be scaled to 30% of course total
Instructions: Questions are to be attempted using Stata software unless otherwise indicated. The data files are stored in Blackboard and in the course folder in the S: drive and are the same that you have used in lab sessions. Pay close attention to what you are asked to do and show all working, as part marks may be awarded for correct working. You must provide a Word document with typed answers to each question and any requested Stata output. Additionally, save your results in a Stata log file and annotate this to show the question number that corresponds to each part. Include your log file(s) as an appendix to your assignment. Please submit your assignment via Turnitin. See the ECP for further details.
1. The final item of assessment for this course requires you to either: (1) write a research paper on a topic that can be addressed using variables selected from one of the two datasets made available in the lab sessions (WVS2012 or AUSSA2011); or (2) write a research paper in which you replicate and extend the analyses reported in a previously published study. This first exercise will assist you with planning the topic of your research paper. You are required to:
a. Read the guidelines for the research paper and write down the most important criterion for producing a good research paper (in one sentence only). (1 mark)
b. Peruse the datasets WVS2012.dta and AUSSA2011.dta to identify a variable (or variables) that measures a concept that is of interest to you and that could be used as the basis for a research paper. For example, if job satisfaction is the selected topic then you would need to identify a variable(s) that contain responses to questions that ask how satisfied a person is with their life. After careful consideration of the variables available:
(i) Write down a topic that you might like to address in your research paper in no more than three sentences. (1 mark)
(ii) Identify the variable (or subset of items that could be used to create an index) that would act as your dependent variable. Write down the variable name(s) and variable label(s). (1 mark)
(iii) It is important to understand the range of values and the probability distribution of a variable before analysing it. Use Stata to produce a frequency distribution of your dependent variable or each of the items that form your index. Write down the values and value labels. (2 marks)
c. Your research topic will typically require investigation of the relationships between your variable of interest and other variables measuring factors that you suspect may be associated with your dependent variable. After further consideration of what associations your research topic might address:
(i) Write down a list of 5 variables available in your dataset that could be considered as independent variables influencing your dependent variable. (1 mark)
(ii) Identify the level of measurement for each of these independent variables (nominal, ordinal or interval/ratio). Select two of them and describe the statistical test that you would use to examine the strength of association between each of these variables and the dependent variable that you identified before. (4 marks)
2. Using the data file WVS2012.dta complete the following exercises:
a. Write down the variable names and variable labels corresponding to the 6 questions aimed at gathering respondents’ views on “income equality”, “private vs. government ownership”, “self vs. government provision”, “competition”, “hard work vs. luck”, and “achieving wealth”. What is the range of values for each of these 6 variables? (1 mark)
b. Use Stata to summarize these 6 variables and use the results to manually calculate the standard error of the mean for each variable. (3 marks)
c. Format these 6 variables by typing in Stata: format varlist %4.2f . Then, use Stata to construct 95% confidence intervals for the means of each of them. Interpret the results. (2 marks)
d. Use Stata to test the null hypothesis that people tend to have neutral feelings about income equality. Write down the null and alternative hypotheses, and the formula for the test statistic used in this test. What is the p-value associated with the test statistic? Interpret the results. (3 marks)
3. Suppose that we are interested in assessing the relationship between self-reported happiness and general health. In the AUSSA2011.dta dataset, the variable used to measure happiness is c1 and the variable used to measure health is c26. Conduct the following tasks.
a. Produce frequency tables for the two variables of interest. List the values and value labels for each variable. Recode “can’t choose” responses as missing values where relevant. (1 mark)
b. What percentage of people in the sample is fairly unhappy with their lives or worse in our sample? What percentage of people rates their health as good or better? Would you expect these two variables to be related in any way? If so, explain how and why. (2 marks)
c. Recode the variable c1 into a new variable called happy with the value 1 indicating that a person is happy with their life (i.e. fairly, very, or completely happy) and the value 0 indicating that the person is not happy (i.e. neither happy nor unhappy, fairly, very or completely unhappy). Label the new happy variable “Dummy variable denoting being happy with one’s life” and give its values new, meaningful labels. Produce the necessary frequency distributions to check whether the new variable has been created and labelled correctly. (1 mark)
d. Recode the variable c26 into a new variable called healthy with the value 1 indicating that a person is healthy (i.e. self-assessed health is good, very good or excellent) and the value 0 indicating that the person is not healthy (i.e. self-assessed health is fair or poor). Label the new healthy variable “Dummy variable denoting being healthy” and give its values new, meaningful labels. Produce the necessary frequency distributions to check whether the new variable has been created and labelled correctly. (1 mark)
e. To examine the association between happiness and health, produce a cross-tabulation between happy and healthy. Which is the explanatory variable and which is the outcome or dependent variable? Why? Request appropriate percentages for your table. Describe the pattern in the table. (2 marks)
f. Test for an association between happiness and health using a chi-squared test. Produce the relevant table showing the chi-squared statistic. Write down the chi-squared statistic and the p-value for this test. What do you conclude about the relationship between happiness and health? Does this result agree with what you hypothesised in part 3b? (2 marks)
4. Use variables c16a, c16b, c16c, c16d and c16e in the AUSSA2011.dta dataset to construct an index variable called doctors where high values reflect high trust in doctors. Create a do-file to record and save your Stata commands used in the tasks (a) to (i) below. Include comments throughout the file. Remember to save your results by opening a log file at the beginning of your do-file, and closing the log file at the end of the do-file. Submit both your do-file and the answers to the questions below.
a. What concept are you measuring by creating this index? Give 2 reasons why you would combine responses from these 5 variables to measure it. (2 marks)
b. Produce frequency distributions for each variable in the index. Write down the values and corresponding value labels for each variable. Recode “can’t choose” responses as missing values for all 5 variables using the recode or replace commands, overwriting the existing variables. (2 marks)
c. Do the response codes need to be reversed for any of the variables? Why? Reverse as necessary creating new variables and giving these meaningful variable and value labels. (2 marks)
d. Produce a correlation matrix for the variables based on pairwise deletion. Which pair of variables is most closely associated? How can you tell? (2 marks)
e. Generate the Cronbach’s alpha coefficient with item statistics for the variables going into the index. Use casewise deletion of missing values. Do the results suggest that the items will form a reliable index? Explain why. (2 marks)
f. Based on the output from 4e, would the index be more reliable if any variable was not included? How can you tell? Proceed without deleting any variable. (2 marks)
g. Create a variable called missing_responses that counts the amount of missing data each respondent has on the variables going into the doctors index. Produce a frequency distribution of missing_responses. For how many people are data missing on more than three items? (2 marks)
h. Create the index called doctors that is the average of each respondent’s scores on the 5 input variables only if a respondent has valid data on at least 3 items. Produce summary statistics for the doctors index. What are its average and standard deviation? (2 marks)
i. Using the doctors index that you have created, generate a second index variable called doctors2 where the potential range is 0 to 10. Use visual means to demonstrate that the index doctors2 is roughly normally distributed. (2 marks)
5. The table below shows the percentage of Australian men who were smokers in 2007-08 by age category.
Variable
Age mid-point (X) 20 30 40 50 60 70 80
% smokers (Y) 18.1 32.3 28.5 24.5 13.0 9.9 5.1
Source: Adapted from ABS 2009 Australian Social Trends
a. Enter the data for variables X (age mid-point) and Y (% smokers) into Stata as two variables using the Data Editor. Rename these as age and smokers, respectively. (2 marks)
b. Produce a scatterplot of smokers against age. Include the regression line of best fit. Describe the form and direction of the association between the percentage of men who smoke and age. (2 marks)
c. Ask Stata to compute the correlation coefficient for the association between smokers and age. Explain what this tells you about the strength of association between these two variables. Check whether this relationship is statistically significant at the 99% level adding the relevant command option. (2 marks)
d. There is an outlier in the data. At what age does this outlier occur? Provide a plausible explanation for why it is present. Remove the outlier from the data and compute the correlation coefficient again. In what way has the strength of association between the 2 variables changed? (2 marks)
e. The prediction equation for this regression is:
(i) Identify the value for the slope and explain what this tells you about the relationship between the percentage of men who smoke and age. (2 marks)
(ii) Calculate the predicted percentage of men who smoke at age 55. (2 marks)
f. In 3 sentences or less explain how the information gained from this assessment of the relationship between the percentage of men who smoke in Australia and age may inform Government policy aimed at promoting healthier lifestyles. (2 marks)