【SOCI2252】Data Analysis

Using the datasets available on Ultra (under Assessments -> RMA Data Analysis Summative) answer all the questions in the three sections below using appropriate statistical methods and analysis.  Your work should be no more than 2,000 words in length (+10% max), excluding tables and figures.

Each section is based around a different dataset.  You should address each section separately and there is no expectation that you somehow draw connections between the different sections.  You do not need to write a general introduction or conclusion to the assessment but you may want to for each individual section.

In marking your work, the following key factors will be considered:

• Have appropriate statistical methods been used to address the question?
• Have the results of the analyses been interpreted accurately?
• Have the results been used to produce a thoughtful and thorough discussion?
• Is the work presented in a clear, accessible and professional manner?

You are not being assessed on your ability to define statistical tests or levels of measurement.  So focus your work on presenting and discussing the analyses you have undertaken.  Further, while your discussion may benefit from some reading around the themes of each section, there is no expectation that you do this and the information provided in each section should be sufficient to complete the assessment to a high standard.

In answering the questions, you should make appropriate use of tables, graphs and diagrams to present your results.  You should be mindful to include only relevant output from SPSS and are encouraged to reformat tables and graphs to ensure they are focused and accessible.  Not doing this, by including unnecessary and unformatted SPSS output, is also a good way of missing marks.

Finally, please note that the levels of measurement for all of the variables in the datasets have been set as nominal, regardless of what their levels of measurement actually are.  As such, before running your analyses, you will need to work out what level of measurement each variable is.  You do not need to explain this in your assessment but you do need to do it correctly to be able to use the right statistical methods.  It may be a good idea to update the datasets with the correct levels of measurement to avoid forgetting.

The three sections of the assessment are detailed on the following pages.

The Russell Group is a collection of British universities that have come to represent the prestigious end of UK higher education.  Director General of the Russell Group, Wendy Piatt, asserts that:

“Our students work with world-class experts, use first-rate libraries and facilities, are part of a highly motivated and talented peer group and often engage with cutting-edge research.

Graduate recruiters rank ten Russell Group universities in the top 30 universities worldwide, and Russell Group graduates typically receive a 10% salary ‘top-up’ over others. Why? Because the combination of teaching and research excellence creates the ideal learning environment which produces ‘work-ready’ graduates” (cited in Sharp, 2019)

The focus of this first question is to test Piatt’s claims regarding the outstanding educational environment of Russell Group institutions.

---

The RMA UK HEI dataset on Ultra contains data from a sample of UK higher education institutions (HEIs) taken from the 2023 National Student Survey (NSS):

• InstitutionType: this defines whether the case is a Russell Group University, another “old” university, or a “new” university founded since 1992
• TeachingScore: the proportion of students at the university who agreed that they were satisfied with the quality of teaching on their degree programme in the 2023 NSS
• FeedbackScore: the proportion of students at the university who agreed that they were satisfied with the quality of feedback on their work in the 2023 NSS
• LearningOpportunities: the proportion of students at the university who agreed they were satisfied with the opportunities for learning and skill development on their programme in the 2023 NSS

The Universities of Oxford and Cambridge – both members of the Russell Group – have been excluded from the data as they are significant outliers in terms of student intake and economic resources (see Boliver, 2015) and therefore may have an inappropriate influence on the results.

Questions:

1. Using appropriate statistics, analyse the association between InstitutionType and TeachingScore.  What do your results indicate about the quality of teaching and learning provided by different institutional groups?
2. Using appropriate statistics, analyse the association between InstitutionType and FeedbackScore.  What do your results indicate about the quality of feedback provided to students by different institutional groups?
3. Using appropriate statistics, analyse the association between InstitutionType and Learning-Opportunities.  What do your results indicate about the learning opportunities provided to students by different institutional groups?
4. Considering your answers to questions (a), (b) and (c) above, can you conclude that Russell Group universities offer educational “excellence” in excess of other British higher education institutions?

Drug policy is a contentious political issue in the UK.  On the one hand, the British government remains firmly committed to the criminalisation of narcotic substances (see for example: BBC, 2018), reflecting wider public attitudes that are mixed regarding decriminalisation of “soft” drugs like cannabis but largely opposed to the decriminalisation of “harder” drugs like cocaine and heroin (YouGov, 2018).  On the other hand, underpinned by the argument that incarceration is ineffective for deterring illicit drug use, several police areas in England and Wales have shifted towards education and support for minor drugs possession offences rather than using legal sanctions (see for example Durham Constabulary’s Drug Arrest Referral Scheme; Durham Constabulary, 2019).

Among other things, opposition to decriminalisation of narcotic substances is underpinned by the perceived social harms that result from drug use that go beyond individual health and wellbeing.  Goldstein’s (1985) seminal work, for example, argues that drugs result in violent crime owing to: (1) the effects they have on people taking them; (2) the need for money to purchase drugs; and (3) the systems which emerge around the sale and supply of drugs (e.g. criminal gangs).  However, more recent scholarship has challenged Goldstein’s conclusion, pointing to the specificity of the context his model was developed in and the paucity of empirical supporting it (see for example MacCoun et al., 2003).  Further, so-called “harm reductionists” argue that the social harms arising from drug use are more effectively mitigated by decriminalisation and treatment, including state-funded “substitution” therapy where drug users are provided with substitute substances without charge to help them break their addiction and offset the need to resort to crime to buy drugs (see for example: Erickson et al., 1997).

---

The RMA Drugs and Crime dataset on Ultra contains data collected from a random survey of people arrested by the police in England and Wales between 2003 and 2006.  There are three variables:

• Drug_Crimes: the proportion of crimes the respondent says they commit to get drugs.
• Arrest_Reason: the type of crime the respondent was arrested for at the time of the survey.
• Prison_History: whether the respondent has previously spent time in prison.

While this will not be true in every case, for the sake of simplicity, assume that all respondents had actually committed the offences they were arrested or went to prison for.  This means that any respondent who has previously been to prison (according to the Prison_History variable) can be considered a repeat offender.

Questions:

1. Using appropriate statistics, describe the distribution of the Drug_Crimes variable.  To what extent is drug use a significant characteristic of criminal offending in England and Wales?
2. Using appropriate statistics, analyse the association between Drug_Crimes and Arrest_Reason.  What do the results suggest about the social harms that arise from drug use?
3. Using appropriate statistics, analyse the association between Drug_Crimes and Prison_History.  Do the results imply that going to prison effectively deters drug use?
4. In view of your answers to (a), (b) and (c) above, what do you conclude about the debate above, i.e. the decriminalisation of currently illicit drugs?