Becoming a Social Science Researcher, DEV5005A
Quantitative research project
Assessment
Project report due: 23 January 2025, 3pm
Length: No word limit.
Submission: Electronic submission
Proportion of overall module mark: 50%
Submission will be electronic via blackboard. Students should anonymise their work, and include their student number as the heading.
Learning Outcomes
Ability to design and carry out research projects with a primary research component in human geography and development studies.
Develop skills to analyse and present data from geography and development research.
Improved understanding of different sources of secondary data, where to find these, and questions to consider about the nature and quality of data.
Quantitative research skills (e.g. handling of survey data, statistical data analysis, interpretation of quantitative findings and presentation of survey data)
Formulate social science research questions and identify the most suitable methodological approaches to answer these
Evaluate the strengths and weaknesses of different study designs
Inclusivity
In DEV we want to give everyone an equal and fair opportunity to show what they have learned. For example, if you have a health condition or specific learning disability we are very happy to provide this assessment brief and other materials in an alternative format and/or adjust the way we give feedback on your work if you feel that this would improve accessibility for you.
Inclusive steps like this at UEA are called Reasonable Adjustments . They do not make assessments easier ; the aim is to make sure that nobody is excluded.
Any student who feels that they would benefit from such reasonable adjustments is invited to contact Student Support Services via the following self referral form link https://my.uea.ac.uk/divisions/student-and-academic-services/student-services/online-referral-form. You are also welcome to speak with your advisor, who can guide you on getting this process started.
Marking Criteria
Each question has a number of marks associated with it (see below), out of a total 100 marks for this coursework. These are informed by the senate scale, but we give direct feedback per question. For example, question 1 assess Sources and Evidence and Academic Referencing . You will only get the highest marks in questions 3-4 by demonstrating good Criticality and Analysis. Presentation will be assessed throughout.
Note: You must use the below format and numbering.
Quantitative Report (100 marks total)
Identify and present your research question. Discuss some relevant literature. You should use academic references in the normal fashion. (20 marks)
Answer your question using statistics and graphs. To start, include some summary statistics showing the distribution/level of a variable of interest. (15 marks)
Include at least one test of equality, i.e. one of: (25 marks)
A one group t test
A two group t test
A one way ANOVA
A two way ANOVA
Include at least one test of association, i.e. one of (25 marks)
Correlation statistic
Regression
Chi squared
Discuss and conclude. The point of the discussion is to, briefly, explain how what you ve found differs/concurs with what others found. (15 marks)
(You MUST Include your do file as an appendix, either using screen shots or by copying and pasting.)
Advice
Data
You will use data from afrobarometer (https://afrobarometer.org/data/merged-data). We will use this dataset during the course, so you will be familiar with it.
To see an example of what can be done, see Adida, Ferree, Posner, and Lea Robinson (2016), Who s Asking? Interviewer Coethnicity Effects in African Survey Data, Comparative Political Studies, 2016, 49(12): 1630-1660 http://doi.org/10.1177/0010414016633487
This paper shows that an interesting question can be asked using this dataset. Use the paper as a guide: it starts with motivating an interesting question, reviews the relevant literature, sets out some data analysis and concludes. Your research report should do the same, albeit on a smaller scale. We will practice how to do the stata part: you just need to explain to the reader what your question is, and what the answer is.
The original data comes from afrobarometer, with details on the questionnaires used available at https://www.afrobarometer.org/surveys-and-methods/questionnaires The exact wording of a question can be found by cross referencing the question number in the dataset and the questionnaire.
There are different rounds of data available for a number of sub-Saharan countries. You can choose to focus on one time period or country, depending on your question. You must use afrobarometer data though.
The data covers about 120 questions. The easiest ways to see what is included is to a) read the pdf questionnaire for one of the countries/rounds included or b) to use the online data analysis tool. The topics included are:
Socio-Demographics & sample characteristics
Interview conditions
Economy
Democracy and politics
Institutions and leaders
Participation-Civic engagement
Government services and performance policy preferences
Citizenship and identity
Equality and gender
Taxation
Crime and security
Corruption
International relations
Country specific
Structure
Think of this as a mini report. For example:
Introduction.
Motivate the question C why should the reader care?
Summarise some previous literature. Be specific.
For the discussion of others work (q1) see Adida, Ferree, Posner, and Lea Robinson s paper for an example on how to do this. We are aware your sections will be shorter.
Summary Statistics
Explain the data you ll use, briefly.
This is a good place to include the summary statistics
Explain briefly how you ll answer the question.
Analysis
Present your data analysis.
Discuss and interpret the results.
Explain anything we should know in order to understand the results.
Discussion & Conclusion
One brief section, summarising what the data show.
List any weaknesses of the analysis.
Compare to previous work (i.e. reference the literature you discussed earlier).
Advice: technical
Note that the questionnaire shows which numbers denotes an answer was Refused or don t know . You ll need to replace these values in stata with . , using replace q=. if q==9999 or similar.
We have deliberately tried to replicate a part of Adida et al s paper in the course. Do pay attention to the technical details here.
Advice: General
What question should you ask? Well, the requirements imply you need to pick some continuous variable to be your dependent variable. In the paper they have many dependent variables, but you don t have as much time/space, so I d just pick one. For example, Adida et al could just have looked at what affects a respondents view of their tribe s economic conditions (which is just one part of figure 1).
You will then need to select one or more independent variables you care about. You could include the respondent s characteristics such as gender, age, language, education, religion, employment, country.
You will then need to think about sensible control variables, i.e. other things that might influence your dependent variable.
For this, you will need to think about what others have done.
Try to be focused: there is no hard word limit, but you may get higher marks for being focused. Better to be focused, clear and sensible.
You do not need to include your do file, but it may make things easier. If you present your data well, it should be clear whether you have successfully cleaned your data. But to remove doubt in the readers mind, showing the code in an appendix could be helpful.
Advice: Questions
There are many questions you could ask. To give you a sense of the range, any of the following could be answered using that dataset:
Does gender affect people s view of the acceptability of female leaders in Kenya?
Are interviewees in afrobarometer more likely to be coded as friendly if they are of similar age to the interviewer?
Does news consumption relate to one s view of the government in rural sub-Saharan Africa?
What is the main reason for not reporting corruption? Does this differ by age, gender and ethnicity?
Do Wolof respondents have greater trust in the president than other Senegalese?
Note the difference in questions between causal questions (e.g. does income cause people to be happy?) and descriptive questions (e.g. are richer people happier?). It is much safer to pick a descriptive question.
請加QQ：99515681 郵箱：
[email protected] WX：codinghelp
