Core Quantitative Data Analysis: Statistics Assignment Question

Introduction This assessment requires you to put together a range of the skills you’ve been introduced to in CQDA to analyse and present data (with some general guidance). You should find that working through the assessment consolidates your knowledge of the skills you have picked up. Please note the importance of presentation. In addition to the content of your analysis, the grade you are awarded for each question will also depend, in part, on the clarity of your presentation. You should aim to be concise by avoiding showing data that are not really necessary, and you should, where possible, present your own tables rather than simply copying and pasting from SPSS. If you have questions about the assessment please email P.Norris@ed.ac.uk . I will answer any general questions about analysis or how to carry out SPSS procedures, but not with reference to the specific questions in the assessment. I’ll also paste an FAQ of questions I am asked on Learn. You should submit your final paper via ELMA in .pdf format. You are reminded that standard penalties will be applied to any work submitted after the deadline without a legitimate reason. Because the CQDA assignment is not a standard essay there are no defined overall word limits or penalties for exceeding such limits, but you must adhere to any instructions in individual questions which guide you as to the length of your answer. Any requests for extensions to the deadline should be submitted to the Graduate School Office. Extensions will only be considered in exceptional circumstances related to serious illness or personal circumstances (note that work pressure is not a legitimate reason). You submit your request for an extension via the following webpage http://www.sps.ed.ac.uk/gradschool/current_students/taught_msc_students/extensions This paper uses two datasets:- WDI2010exam.sav – a version of the World Development Indicators Dataset used in tutorials; the only differences are that it contains less cases (128 countries) is based on data from 2010 rather than 2011 and that it contains a wider range of variables. ESS2011exam.sav – the same UK wide European Social Survey dataset we have used in tutorials, with a couple of additional variables. These datasets are on Learn in the “Assessment 2” folder.

Question 1: Part 1: World Development Indicators Use the dataset WDI2010exam.sav. You do NOT need to weight the data. Consider the variables Region (which groups countries based on their geographical location in the world) and pop15 (Percentage of the population aged 0-14). 1.1 For each region (as defined by the variable Region) report:- i) The total number of cases ii) The mean percentage of the population aged under 15 iii) The 99% confidence interval for the mean of the percentage of the population aged under 15 iv) The median percentage of the population aged under 15 Marks will be awarded for presenting the above information in a concise manner, please do not report other statistics that are not requested. 1.2 Present suitable graphical output to show how the percentage of the population aged under 15 varies between world region. In your output, you should include a representation of the typical value in each group and an indication of how representative of the group that value is. You should also identify by name any unusual countries in each region. 1.3 In no more than 150 words report what your output for 1.1 and 1.2 shows about the relationship between geographic region and the proportion of the population aged under 15.

Question 2:  2.1 Produce a graph which summarises the relationship between a country’s GDP per capita (gdp) and its depth of hunger (hunger). The depth of hunger, or food deficit, is measured by comparing the average amount of dietary energy that undernourished people get from the foods they eat with the minimum amount of dietary energy they need to maintain body weight and undertake light activity. You should identify which variable is the independent and which is the dependent, and follow the conventions for graphically displaying the relationship between them. To aid your analysis you should calculate the appropriate coefficient of determination and present an equation which summarises the relationship. 2.2 In no more than 150 words justify your decision as to which is the dependent variable in 2.1, and describe the relationship between the two variables.

Question 3: Considering a minimum of three independent variables run a linear regression model with a country’s Deaths per million due to air or water pollution (pollution) as the dependent variable. Provide a brief report on the modelling process you undertook and the findings of your model. This should cover:- 1) For each independent variable you chose to include in your model, a hypothesis as to the relationship you expect it to have with infant mortality, giving a brief reason to support your expectation. 2) Any transformations you applied to specific variables 3) A table summarising your model 4) Any diagnostic tests you conducted and what they say about your model 5) A discussion of the substantive meaning of your results.

Question 4: Part 2: the European Social Survey 2011 (all data for the United Kingdom) Use the dataset ESS2011exam.sav. You do NOT need to weight the data. 4.1 Recode the variable agea to create a categorical variable which records each respondent’s age into one of four categories :- a) Aged up to and including 29 b) Aged 30-44 c) Aged 45-59 d) Aged 60 and over Ensure you give your new variable appropriate labelling (for both the variable itself and the values it takes). Present a frequency table for your new variable. 4.2 Using appropriate tables and statistical measures, show the relationship between your categorical variable of age (created in 4.1) and how satisfied respondents are with the job the police are doing in the country (plcjbcn). Show also whether there is any difference between men and women (using the variable gndr) in terms of the above relationship. 4.3 In no more than 150 words report what your output for 4.2 shows about the relationship between age and satisfaction with the police and whether this relationship varies between men and women.

Question 5 This question concerns the variable gvprppv, and respondents’ attitudes to whether the government should do more to prevent people falling into poverty. 1) Recode the variable gvprppv into two categories. Describe the two categories you have created and justify why you have grouped together the categories from the original variable that you have chosen to group together. 2) Identify four plausible explanatory variables, and provide a hypothesis as to their likely relationship with a person’s attitude as to whether or not the government should do more to prevent people falling into poverty. 3) Conduct a logistic regression model with your 2 category version of gvprppv (created in 1 above) as the dependent variable to test the hypotheses you suggested in 2 (above). You can recode your explanatory variables as required. 4) Briefly discuss the findings of your model. In addition to reflecting on whether your findings support your previous hypothesis, remember to reflect on issues of model fit, and model diagnostics (reporting appropriate statistical output as required).