BF994: Assignment Overview and Template
Guide to this template
You should submit your assignment in this template. Please submit the .Rmd file and the PDF you generate
using it.
This demonstrates that you have coded everything correctly.
You can create headings like this
Or like this for sector tier headings etc
You can run commands inside this environment:
library(haven)
Outside of these environments you can type freely.
Structure of the report
You should download the data set that you were assigned, and work through the elements of the report
described here.
Please apply your judgement and common sense in curating the material that you submit.
Additional parameters of the report
The report should be no more than 12 pages when the PDF is compiled. You can compile this file from
within R Studio using [Ctrl+shift+K] or clicking the ‘Knit’ button above. 12 Pages is the MAXIMUM, but
good reports are often shorter.
You should submit the PDF that this document creates as well as this .Rmd file through MyPlace before the
submission deadline.
There will be a draft submission upload link so that you can put the PDF through Turnitin prior to your
final submission. This is to provide you with an opportunity to catch any accidental plagiarism.
You should consult the “Policy and Procedure for Late Submission of Coursework” document on MyPlace if
you are not able to submit this report on time.
You can spellcheck this document through the ‘edit’ menu. This is recommended.
Part 1 of the report
In this part of the report you are required to do two things.
First, to provide some background information to motivate your study. For instance, why is the question at
hand important You will want to support your discussion with some links to the academic literature in the
area.
1
Second, you should describe and explore the assigned dataset.This should include explaining where the data
come from (i.e. the details of the survey from which the data are drawn), as well as some plots and descriptive
statistics to illustrate your mastery of the methods discussed in the class. Good empirical papers often
motivate the paper using descriptive statistics to set the scene and motivate the subsequent empirical choices.
Finally it is up to your good judgement to decide what the descriptive analysis of the dependent variable
should incorporate.
Part 2 of the report
In this part of the report you should address yourself to using regression models to try to answer the question
associated with your dataset below.
The goal of the analysis is to produce a model which you will use to explain the variation of a variable of
interest.
You should decide on what model to estimate (you are free to consult the literature on these questions too),
and undertake a range of robustness tests.
The emphasis here is on explaining why you are doing what you are doing and interpreting the output of the
different routines you run on the data.
A good structure will involve setting out the rational for and estimating a baseline model, interpreting its
results, and exploring its robustness. You may then want to estimate a slightly different model (e.g. with
another variable added) and explore the effect of this change on the model results.
Some additional comments:
In general, there is not a perfect/unique model. Several competing models could be equally good at explaining
the data. It is up to your good judgement to decide what variables are included to explain the dependent
variable. You should make sure that you explain why you are including particular variables though! Similarly
it is up to your good judgement to decide what the model(s) should look like.
Part 3 of the report
In this final part of the report you should look at the model that you have estimated in Part 2, and discuss
the extent to which you feel that you are able to claim to show a causal relationship between your covariates
and the dependent variable. You should justify your answer.
This should include an explanation and discussion of omitted variable bias, and conditioning on observables.
You may also note in this part of the report any alternative approaches that you think might be useful for
causal inference and which might overcome the shortcomings of the model that you selected in Part 2.
This part of the report should be 2-3 pages in length, and will not include any additional empirical material.
Assignment to a project
You have been assigned to a project denoted by a letter A – G. The question for your project, the data you
need and the dependent variable in the dataset are set out in this section.
Question for the report
Project A – What determines a person’s exposure to crime
Project B – What are the determinants of the size of a main residence
Project C – What are the determinants of drinking behaviour
Project D – What are the determinants of job satisfaction
2
Project E – What determines a person’s exposure to crime
Project F – What are the determinants of the size of a main residence
Project G – What are the determinants of drinking behaviour
Data for the report
Project A – Crime Survey for England and Wales, 2013-2014: Teaching Dataset https://discover.ukdataservi
ce.ac.uk/catalogue/ sn=7911&type=Data%20catalogue
Project B – English Housing Survey, 2012-2013: Household Data Teaching Dataset https://discover.ukdatas
ervice.ac.uk/catalogue/ sn=7853&type=Data%20catalogue
Project C – Health Survey for England, 2011: Teaching Dataset https://discover.ukdataservice.ac.uk/catalog
ue/ sn=7402&type=Data%20catalogue
Project D – Workplace Employment Relations Survey, 2004: Teaching Dataset https://discover.ukdataservice
.ac.uk/catalogue/ sn=5897&type=Data%20catalogue
Project E – Crime Survey for England and Wales, 2011-2012: Teaching Dataset https://beta.ukdataservice.a
c.uk/datacatalogue/studies/study id=7401
Project F – English Housing Survey, 2008-2009: Household Data Teaching Dataset https://beta.ukdataservi
ce.ac.uk/datacatalogue/studies/study id=6949
Project G – Health Survey for England, 2002: Teaching Dataset https://beta.ukdataservice.ac.uk/datacatal
ogue/studies/study id=5033
Dependent variable
Project A – bcsvictim
Project B – NBedsX
Project C – totalwu
Project D – a8g
Project E – bcsvictim
Project F – nbedsx
Project G – drating
3
