As part of your assessment for BUS005/135 you will submit a written project worth 40% of your
final mark. This document sets out the main aspects of this project.
The aim of the project is to integrate the learning from the entire module in the service of
conducting your own exploration of the determinants of pay gaps. To conduct this study, you
must use the extract of the Labour Force Survey as your underlying dataset that is provided for
your on QMPlus.
You will decide what source pay gaps you will examine (this is your ‘research question’), and
then conduct an appropriate data analysis and writeup.
You are encouraged to start working on your project early in the semester. It is extremely hard to
do well on this project at the last minute, as it requires some contemplation. It is as least as much
about ‘thinking’ as about ‘doing’, especially since you will have easy access to all the Stata code
you will need, through the labs and homework assignments (and associated solutions). You are
being assessed chiefly on your thoughtfulness in drawing meaning from numbers.
Although this is an individual assignment, you are more than welcome to work with a fellow
student on specific parts. Specifically, you may work together on writing Stata syntax and
thinking through the analytical choices you make. However, your written report must be
entirely yours.
Feel free to come to office hours to get help on your project, or to ask for support from the Quant
Skills Tutor. In the scheduled lab session in Week 12, lab teachers will run a ‘surgery’ where you
can get targeted assistance.
Deliverables
Using data available on the course QMPlus page and following the instructions below, you are
expected to submit a 2,000-word research report (not including appendices).
The due date for the report and the peer review is March 29th, 2019.
Submit your report via QMPlus.
BUS005/135
QUANTITATIVE METHODS
Step-by-Step Instructions
- Your first task is to choose which research question you will work on. You can start making
this choice as soon as you like. By the lab in week 5, you will be asked to do some basic
quantitative work based on this choice, so ideally you should decide before then.
o How to decide?
§ Skim through the following articles.
• Hills, J. (2010) An Anatomy of Economic Inequality in the UK An
Anatomy of Economic Inequality in the UK - Report of the
National Equality Panel. LSE STICERD Research Paper No.
CASE report 60
• Longhi S., and Platt, L. (2008) Pay Gaps Across Equalities Areas:
An analysis of pay gaps and pay penalties by sex, ethnicity,
religion, disability, sexual orientation and age using the Labour
Force Survey. Institute for Social and Economic Research,
Research Report 9. University of Essex.
o Based on these articles, identify a source of pay gaps in the UK other than gender
that interests you. A possible, but by no means exhaustive list would include
location, disability, ethnicity etc.
o Be sure that there are variables in our LFS dataset that allow you to explore your
topic (and if there are not, choose a topic that does have relevant LFS variables). - Use Google Scholar or the QMUL library website to identify 3 additional journal articles or
academic working papers that examine the pay gap of your choice. These must be from peerreviewed sources – not journalistic ones. - Use these three studies (plus relevant parts of Longhi & Platt, and Hills) to motivate your
own empirical inquiry into the pay gap of your choice. Using the Labour Force Survey data,
conduct research and write a report that does the following
o Determine whether there are significant differences in pay according to the factor
of your choosing
o Consider if these differences survive the inclusion of individual-level control
variables. Make sure you are explaining why the control variables make sense.
o Diagnose your regression results, in terms of any assumptions that might or might
not be upheld.
o Discuss the relationship between your findings that those of the previous literature
that is motivating your analysis. Why might they differ?
o Consider the confidence with which we might speak of the relationship you
describe being causal. What might limit our confidence, and what strengthens it?
BUS005/135
QUANTITATIVE METHODS
Report Structure and Tips
Your assignment should be structured like a mini-scientific article with the following major
sections: - Introduction
- Prior knowledge
- Data and Methods
- Results
- Conclusion
- References
- Tables and Figures
1) Introduction
The introduction section introduces the problem or question to be studied. It usually
briefly tells us the state of knowledge on the issue, and why and for whom the topic is an
important one.
2) Prior knowledge
In this section you will succinctly synthesize what the studies you examined tell us about
your particular question or hypotheses.
Components: - Theory: What does the work you read say in terms of a theory of model telling
you why/how your independent variable of interest shapes your dependent
variable? - Empirics: Write carefully about what the studies you read (plus other work they
cite in their own literature review sections) show in terms evidence regarding your
question or questions related to it.
3) Data and Methods
The purpose of this section is to primarily address two questions: 1) How was the data
collected or generated? 2) How was it analysed? Since there are many different ways that
research can be done, you must provide some justification for the choices that you make.
Components: - Identify, describe, and justify the main study variables (dependent, independent,
and control) that you have selected for the quantitative analysis. Justification for
these typically comes from your wider reading, should be cited, and should
include a logic. One additional reason to conduct your literature review is to
identify possible control variables, based on what other researchers have used in
their work.
BUS005/135
QUANTITATIVE METHODS - If you recoded any variables, please describe how you did so and why.
- Describe the type of statistical analysis you did, and why it is appropriate.
4) Results
The purpose of the results section is to present your key results in a logical sequence.
There should be little interpretation here – just a statement of what the results are, as
opposed to what they mean.
Place all relevant tables and figures to which you refer here in the final section of the
report.
Tables and figures (and any notes and titles for those tables and figures) do not count
towards the word limit.
Components: - Describe the characteristics of your study sample.
• Refer to a table of descriptive statistics that includes a summary of your
dependent, independent, and control variables. - Using some form of descriptive statistics, describe the relationship between
the dependent and independent variables in your analysis. - Describe the results of a bivariate and multivariate regression analysis,
including the F-statistics, regression coefficients, and R-squared.
• Refer to a table that shows both bivariate and multivariate regression
results. Again – place the table itself in the final section of the report.
Tips for a good Results section
• Be selective. The point of this section is to provide statistical results that address
your research question – not to present as many figures and tables as you can.
o As a rule, do not include any figures or tables that you do not describe in
words in the body of your report.
o This is especially important since the word limit is tight. You need to be
judicious about what matters and what does not.
• Make sure the numbers in your tables/figures and your text are the same.
• Make sure that the results you present directly relate to your research question.
• It is very useful to carefully read (and to loosely mimic) the way the studies you
read phrase their results sections. I DO NOT MEAN PLAGIARIZE.
• Also remember that there are examples in the lab outlines and in my slides on
how to talk about your findings.
5) Conclusion
The purpose of the conclusion is to interpret and describe the significance of your
findings in light of what was already known about the research problem being