# Statistical Research Techniques in International Development

Module Code:
15PECC052
Credits:
15
FHEQ Level:
7
Taught in:
Term 2

Statistical Research Techniques in International Development is a module designed specifically to address the distinctive needs of research in international development. It not only covers standard topics in econometrics and quantitative research skills, but also gives practical training in using survey data, understanding sampling techniques and qualitative statistical methods. This module also forms part of requirements of doctoral training in Economics and International Development.

The aims of Statistical Research Techniques in International Development are threefold:

1. to introduce students to statistical inference and a range of statistical tests;
2. to encourage clear and coherent expression of statistical results;
3. and to promote critical reading of statistics in economics and international development.

The topics covered will include sampling methods, survey design, comparison of means, analysis of contingency tables, correlation and regression. Open access computers at SOAS are available for students to complete the course work.

#### Prerequisites

Students may take Econometrics as an alternative to Statistical Research Techniques in International Development, if it is more suitable for their research interests, subject to academic approval.

#### Objectives and learning outcomes of the module

The learning outcomes are as follows:

• Understand the characteristics of probabilistic and non-probabilistic sampling methods.
• Define and describe the properties of simple random sampling, stratified random sampling and multi-stage sampling.
• Identify the main sources of non-observational (sampling, non-response, frame selection) error in relation to sample surveys.
• Identify the main sources of observational error (questionnaire design, interviewer bias, interviewee bias) in relation to sample surveys.
• Describe a variable, both numerically and graphically, with reference to its level of measurement.
• Distinguish between marginal, joint and conditional probabilities in relation to a contingency table.
• Use a range of measures of association related to a contingency table including conditional distributions, odds ratios, standardised residuals, Pearson’s Chi-Squared test and Cramer’s V.
• Create confidence intervals and hypothesis tests for a number of sample statistics.
• Test for the equality of two sample means.
• Test for the equality of three or more means using one-way ANOVA.
• Test for the equality of variance between two sample variances.
• Be able to interpret a range of correlation coefficients for nominal, ordinal and interval data.
• Understand the main assumptions of the classical linear regression model.
• Generate and interpret the results of simple bivariate and multivariate regression.
• Interpret the standard inferential statistics for simple bivariate and multivariate regression.
• Understand the purpose and consequence of different functional forms in estimating regression equations.
• Use dummy variables as independent variables in a regression.

#### Method of assessment

Assessment weighting: Exam 100%.