ECON7030 Descriptive Research
Task
In an empirical project the research question is answered by formulating a hypothesis and testing this hypothesis using appropriate data and relevant statistical techniques. In this assignment, you will focus on data & descriptive research. Accordingly, you will:
1. Finalize your research hypothesis.
2. Describe the data relevant for testing your research hypothesis.
3. Provide a brief outline of your empirical model and a description of relevant variables.
4. Clean and process the data. Construct the relevant variables.
5. Create a table of summary statistics (or figure) following your OP.
6. Discuss the results reported in the table (or figure) reporting summary statistics.
Resources
Refer to Canvas module ‘Project resources’ for resources on data mining/descriptive research: Value of descriptive research.
Formatting requirements
− Use 12 pt font and 1.15-line spacing, normal margin.
− Maths/Equations: The word equivalent for displayed math and other equations is 16 words per row for single-column equations. Two-column equations count as 32 words per row.
Submission Instructions
− Submission deadline: 20 April, 11:59 pm.
− Students should submit their assignment by uploading a PDF via Canvas assignment. This will be processed with Turnitin (similarity detecting software).
− Assignments submitted after the deadline will be subject to Faculty of Arts and Social Sciences late submission guidelines.
− Simple Extensions and special consideration can be requested through the official F system: the university webpages contain information for students about applying for a special consideration and simple extensions
Structure of the report: Descriptive research
Title page
Write the title of OP and mention your SID, Semester #, Year.
Section 1. A clear statement of the hypothesis
Clearly state your research question and hypothesis.
Section 2. A clear and concise description of data/Sample
In describing the data, follow YOUR OP. At the least, you must state the following:
ü Data source (name of the survey) (e.g., IPUMSCPS, HILDA)
ü Survey Location (e.g., Nigeria)
ü Data source (name of the survey) [e.g., DHS].
ü Sampling strategy (e.g., describe relevant sampling strategy/survey design – available in the data sites)
ü What information was collected in the survey? (e.g., DHS provides representative data on population, health, HIV, and nutrition at the household level/& individual level (for women and men aged 15-49 years old, and for children aged 0-14 years).
ü Year of study (e.g., “I … used cross section data for 2008 for Nigeria.” Or I used panel data from 20 waves of HILDA – 2002-2020).
ü Which sample are you using? (I am using DHS Women’s sample).
ü What is your unit of analysis? (Individual ?Household? Country?) [e.g., I obtained individual-level anthropometric and sociodemographic data for women of reproductive age: 15-49 years old].
ü Sampling size (e.g., The Nigerian component contains responses from 7,000–35,000 women in each year. I used a sample of 10,746 women)
ü Study design: cross section, panel, repeated cross section, time series (e.g., study design for my research is repeated cross section).
Note: Clearly indicate if your replication data is different from/or similar to OP data.
Section 3. outline your proposed statistical model.
i. Briefly discuss your proposed empirical model. For example: Following OP (cite), I estimate linear probability models to assess the association between years of schooling and the probability of being overweight/obese. Estimated coefficients of OLS may be biased due to endogeneity of schooling with resect to overweight (following form unobserved factors influencing both schooling and overweight status).I will discuss these additional empirical challenges in the final project report.
[Note: at this stage just provide an outline of the basic model which you propose to estimate). Additional econometric complications/challenges will be discussed in A4and A5].
Section 4. Daat Mining, Variable construction and description of key variables.
i. How did you clean and process the data to construct the relevant variables.
Describe how you cleaned the raw data to construct the outcome variable and the key explanatory variable(s).
ii. Clearly describe the relevant variables of your proposed model.
Key outcome variable. For example: The key outcome variable (dependent variable) of interest in my proposed model is whether an individual is overweight/or obese.
How did you construct the variable? “I divided body weight by height squared to calculate each individual’s body mass index (BMI) and transformed this into a dichotomous indicator of whether an individual had a BMI above pre-established overweight and obese thresholds set by the World Health Organization (WHO 2015) based on risks to cardiovascular health: 25 kg/m2 (overweight or obese) and 30 kg/m2 (obese).”
Key explanatory variable. For example: the key explanatory variable is ‘expected years of schooling’ measured by number of years of schooling completed by an individual at the time of the survey (based on DHS variable v133).
Key control variables. For example: “I control for individual level characteristics such as age, income, occupation; family background characteristics, and dummy variables for regions in Nigeria.” For variable descriptions see the table of summary statistics reported in the next section.
SECTION 4. Report summary statistics of variables
You will replicate a figure or the table of summary statistics from your OP.
· If you are directly using OP data (same sample/setting/time-period), then simply reproduce the table from OP.
If you are using a different sample /different setting/or time-period), then report the summary statistics using your sample.
An example from (Barlow, 2021):
OR
You must:
- Clearly label tables and figures, adding appropriate notes to make it clear what variables are included and how they are constructed.
- Present tables and figures professionally. Use relevant Stata codes (e.g., esttab / outreg / putexcel commands and/or graph export commands for figures) to produce tables/figures. DONOT copy-paste Stata output directly into the body of the assignment.
- Figures, tables, and graphs should be informative and self-explanatory.
SECTION 5. Discussion of results: Descriptive statistics
Discuss the descriptive statistics reported in the above table or figure. The discussion must refer to the broader literature where relevant. An example following (Barlow, 2021).
“Table 1 summarizes their demographic and anthropometric characteristics. On average, 33.6% of the sample was overweight or obese across all survey years. This number is slightly higher than the proportion reported in previous analyses of DHS data, which likely reflects the slightly older age of my sample (Kampala and Stranges 2014; Neupane et al. 2015). ….Figure 5 plots schooling disparities in overweight/obesity; years of schooling are grouped according to the schooling levels associated with different years of schooling. The figure shows that the probability of being overweight/obese was lowest among those with the least years of schooling (0 years of schooling/no education) and highest among those with the most years of schooling (>12 years of schooling/postsecondary education). The proportion of respondents who were overweight or obese increased over time, from 1.9% in 2003 to 12.1% in 2013….”
Referencing
− The selected papers must be from relevant economics journals.
− You may include references to articles from fields other than economics as long as they are relevant, and are sourced from high-quality journals (Demography, Social Indicator Research, Social Science & Medicine to name a few).
− In addition to journal papers, you may include book chapters, government reports etc. However, pay attention to quality of these resources and cite them appropriately.
− Stay restricted to resources published during the last 10 years (that is, avoid older resource unless it is a seminal work).
Note:
− Use author-date, Harvard style references. Please consult: https://libguides.library.usyd.edu.au/ld.php?content_id=50827391 .
− Any source that you cite in the main text must be included in your reference list.
− Conversely, do not include in your reference list the sources that you have not cited in the main text.
− Pay attention to the quality of the references.
Appendix (optional)
You may OPTIONALLY include additional materials (questionnaires, extra information on description of variables, data, model). Typically, these are the materials which are not crucial to your conclusion !
Reference:
Barlow, P, 2021, ‘The Effect of Schooling on Women’s Overweight and Obesity: A Natural Experiment in Nigeria’, Demography, vol. 58, no. 2, pp. 685–710, doi: 10.1215/00703370-8990202.