### Introduction to Quantitative Methods for Research on Governance and Public Policy

Dates: July 19-21, 2011

###### Presenter:

- Rod Alence, Associate Professor, International Relations, University of the Witwatersrand

###### Assistants:

- Xichavo Alecia Ndlovu, Wits University
- Victor Brobbey, CDD-Ghana
- Carlos Shenga, University of Cape Town

###### Hosts:

- The Ghana Center for Democratic Development (CDD-Ghana), Accra, Ghana

The aim of this three-day course is to introduce professional researchers working in the field of governance and public policy to quantitative research methods, particularly those based in academic institutions and non-commercial research organizations. The course assumes no prior training in quantitative methods nor in statistical computing. Those who have previously taken an introductory course (perhaps several years ago) will also benefit. All of the software used in the course – mainly R and GNU Emacs – is free and open-source and will be distributed to participants on USB drives.

The rationale for such a course is to begin to address the dearth of quantitative research skills among African researchers working in the fields of governance and public policy. Many researchers in these fields have studied disciplines such as political science and law, which at most African universities provide little or no practical training in quantitative methods. The course will help make participants more informed “consumers” of quantitative research, and it will equip them to be “producers” of simple analyses motivated by their own research interests.

The running example in the course is a two-sample difference-of-proportions analysis. This is arguably the simplest way of analyzing a bivariate relationship. It is broadly applicable and calls upon concepts and skills that recur in more complicated analyses – from coding and recoding, to univariate statistics and graphics, to descriptive measures of association, to statistical inference (confidence intervals and hypothesis tests). The restriction that the “grouping variable” and the “outcome variable” must both be coded dichotomously means that difference-of-proportions is rarely the best way to analyze a bivariate association. Yet it provides a useful “foundational analogy” against which more advanced concepts can be introduced.

The course combines interactive lectures with practical exercises that participants work individually or in pairs. On the first day, before introducing computer applications, participants complete a difference-of-proportions analysis “by hand,” using samples of one hundred observations each, drawn from Afrobarometer data for one country. On the second day, they learn to replicate the analysis for a full national sample using R; then they begin designing their own studies, in which they use the same approach to analyze the relationship between two variables of their choice. On the third day, they complete these analyses and give short presentations of their findings. (A course evaluation and discussion of participant interest will follow.)

###### Participants List:

- PNK Aborampah, CDD-Ghana
- Rhoda Acheampong, CDD-Ghana
- Anthony Amoah, Central University, Ghana
- Edward Ampratwum, CDD-Ghana
- Alex Antwi
- Kojo Asante, CDD-Ghana
- Maxwell A Ashon, CDD-Ghana
- Michael Asiedu
- Sherry Bempah, KNUST
- Kwesi Boateng, Ghana Anti Corruption Coalition
- Edmund Coblah
- Ernestina Dankyi
- Peter Dwuma, KNUST
- Adu Kakra, CDD-Ghana
- Nene L Kuditchar, University of Ghana
- Mary Kyei
- Daphne Lariba Nabila, Legal Resources Center
- Mark Obeng
- Joseph Ocran
- Francis Oppong, CDD-Ghana
- Sharon Praku, CDD-Ghana
- Anthony Ebow Spio
- Regina O A Tetteh, CDD-Ghana
- Abdul Mumin Yazeed
- Paul K Dzene Yidu
- Benedict Yiyugsah, CDD-Ghana

###### Schedule:

*DAY 0: PREPARATION (PRESENTER AND ASSISTANTS ONLY, 18 July)*

*DAY 1: INTRODUCTION TO QUANTITATIVE RESEARCH*

- Session 1: An introduction to quantitative research (interactive lecture)
- Thinking like a social scientist
- How quantitative and qualitative methods can complement each other
- Measuring concepts
- Samples and populations

- Session 2: Quantitative research in practice: difference of proportions (interactive lecture)
- (Re)coding data
- Tabulating data (contingency tables)
- Graphing data (mosaic plots)
- Difference of proportions as a measure of association
- Confidence intervals (“sampling error margins”)

- Session 3: “Statistics by hand” (exercise)
- Participants work in pairs to complete a structured exercise that involves a difference-of-proportions analysis. (Each pair works with a different sample of 100 observations, drawn from a national survey (Afrobarometer).)
- Discussion of the exercise.

- Session 4: Introduction to R and GNU Emacs (interactive lecture)
- Overview and demonstration
- Installation of the software on each participant’s laptop
- Using R as a calculator

*DAY 2: INTRODUCTION TO STATISTICAL COMPUTING*

- Session 1: Basic data analysis in R (interactive lecture)
- R basics
- Data and functions
- A few R “recipes” (recoding, tables, graphs)

- Session 2: “Using R” (exercise)
- Participants work in pairs on a structured exercise, applying the skills from the previous lecture to recode variables and run tables, graphs, and difference-of-proportion tests in R.

- Session 3: An extended difference-of-proportions analysis (interactive lecture)
- “Walk-through” of an R script file that runs a complete difference-of proportions analysis (very similar to the “statistics by hand” exercise, but using the full data from a national survey (Afrobarometer)).

- Session 4: Projects – conceptualization and design
- Working individually or in pairs, participants use survey data (Afrobarometer) to analyze the association between a dichotomous “grouping (independent) variable” and an dichotomous “outcome (dependent) variable” – in which the grouping variable is typically demographic or biographical and the outcome variable is typically cognitive or attitudinal.

*DAY 3: DOING QUANTITATIVE RESEARCH*

- Sessions 1 and 2: Projects (continued)
- Using the R script from the previous afternoon’s lecture as a “template,” participants run the analysis for their own projects.
- They record the findings in a simple report.

- Session 3: Presentation of results
- Presentations of 5-10 minutes on each project.
- Session 4: Taking stock
- Course summary and evaluation
- Next steps?