PSMII_Group_Report_Assessment_25-26.pdf

1 
 
Introduction and instructions 
Time allowed – One week 
READ THESE CAREFULLY 
You will be expected to submit a single report and a single R script. Only one person 
from the group should submit both of these. If multiple members of the same group 
submit the report, only the latest submission will be considered for marking. 
This group assessment  has six questions. Each group member should pick one of 
these questions. Please make it clear in the jointly submitted document which question 
was done by which group member  by adding their student ID to the top of the 
appropriate page. 
If you left your group and are attempting the assessment of your own, please pick three 
questions out of the six  included below. At least one question will need to come from 
questions 1-3, and at least one question will need to come from questions 4-6. 
This is an ‘open book’ examination. You are free to use any written materials you find 
useful, including your own notes and annotations. You are allowed to use generative 
AI software, but you must reference it if you do so following UCL guidelines 
(https://www.ucl.ac.uk/students/exams-and-assessments/assessment-success-
guide/engaging-ai-your-education-and-assessment). You can and indeed, should 
work with other group members to complete the report . As suggested earlier, the 
introduction, methods, and conclus ions sections will need to be co -authored and 
assessed as one, whilst the methodological approaches will be graded individually. 
The introduction should provide a general introduction to the data. The methods 
section should outline the methodological approaches chosen by the group members 
as well as justifying how and why certain variables were chosen. The discussion should 
draw tog ether the results and speak to the p olicy relevance of the findings . Please 
remain consistent regarding the choice of variables as this is the only way you can 
comment on how the results for the various models compare to each other. 
You will be expected to use RStudio to complete the report and you will need to submit 
your R code which has to be replicable. Please do not copy-paste the code to the report 
as it will count towards the word count. 
 
Database for analysis 
For all six approaches, you will need to analyse a subset from a different wave of the 
Crime Survey for England and Wales (CSEW). These datasets were sanitised to make 
the analysis easier for you. This is one of the datasets we analysed so, you should be 
familiar with it, but for further details please refer to this website: 
https://www.crimesurvey.co.uk/en/index.html

2 
 
Depending on your group, you should analyse the following dataset: 
 
• Group 1, 6, 11: CSEW16-17.dta 
• Group 2, 7, 12: CSEW17-18.dta 
• Group 3, 8, 13: CSEW18-19.dta 
• Group 4, 9: CSEW19-20.dta 
• Group 5, 10: CSEW22-23.dta 
 
Questions 
Imagine that you are working for the Home Office as an analyst. You are tasked to 
provide some insights into certain types of crimes. 
As a group, please try answer: 
• What are the main characteristics and the relevance of this survey 
(Introduction)? 
• Which variables you include (d) in the analysis and why? What kind of 
methodological approach is appropriate for the dataset? (Methods) 
• What kind of real-world insights did you gain from the analysis (Discussion)? 
 
As an individual, please select one of the following questions (or 3 based on the criteria 
outlined above if you are attempting this assessment of your own): 
 
1. Is there an association between your car having been damaged (cardamag) and 
your age (age)? Use as many approaches as you deem appropriate to answer 
this question. 
2. Is there an association between  age (outcome variable), your car having been 
damaged (ca rdamag), the number of property crimes experience d by the 
individual (allcrime),  and demographic character istics? Fi t a series of linear 
regressions to answer this question. 
3. Using the final model in (2) , carry out linear regression diagnos tics, including 
further diagnostics. Interpret the findings. 
4. Is there an association between your car having been damaged (cardamag), a 
set of demographic variables (including age), and the number of property crimes 
experienced by the individual (allcrime)? Fit a series of binary logistic regression 
models to answer this question. 
5. Is there an association between the number of property crimes experienced by 
the individual (allcrime), their car having been damaged (cardamag) and a set 
of demographic  variables (including age)? Fit a series of count regression  
models to answer this question. 
6. Choose any other variable in the dataset  as your ou tcome variable . Fit an 
appropriate model  using age, cardamag, allcrime, and a  set of de mographic 
variables as the explanatory variables. Interpret the findings.