Logistic Regression

Prerequisites

Bookings for this module open on THURSDAY, 11 DECEMBER at 10:00 am
For more information see: http://www.ssrmc.group.cam.ac.uk/ssrmc-modules/core/making/windows

This module is part of the Social Science Research Methods Centre training programme which is a shared platform for providing research students with a broad range of quantitative and qualitative research methods skills that are relevant across the social sciences.

Often researchers deal with outcomes that come in the form of ‘yes’ or ‘no’ responses to questions, or where respondents only have two options to choose from. Similarly, there are occasions where researchers are using unordered (e.g. red, yellow, green) or ordered categories (low, medium, high) as response variables. This module will teach you about how to analyse these different types of data. This will include (a) interpreting outputs, (b) conducting diagnostic tests, (c) calculating effect sizes and (d) making predictions.

• Mphil and PhD students from participating departments taking the Social Science Research Methods Centre training programme as part of their research degree

• a working knowledge of statistical methods up to the level of multiple linear regression
• familiarity with interpreting regression output from statistical programmes
• a University Information Services (Computing) Desktop Services password (http://www.ucs.cam.ac.uk/linkpages/newcomers)
• access to CamTools

Number of sessions: 4

• Session 1: Binary I - Probability, odds & logit
• Session 2: Binary II - Logistic regression in SPSS
• Session 3: Binary III - Diagnostics
• Session 4: Binary IV - Categorical and ordinal

To teach students how to analyse different types of data using SPSS

• to deal with outcomes that form "yes" or "no" responses to questions and unordered or ordered response variable

Presentations, demonstrations and practicals

STATA on MCS

One written exercise [optional, dependent upon department]

• Field, A. Discovering Statistics Using SPSS (3rd ed.) London:Sage.
• Tarling, R (2009) Statistical Modelling for Social Researchers: Principles and Practice. London: Routledge.
• Long, J.S. (1997) Regression models for categorical and limited dependent variables London: Sage. (Advanced)

In addition, there is a chapter on ‘Nonlinear estimation’ in the Electronic Statistics Textbook (EST): http://www.statsoft.com/textbook/stathome.html

• To gain maximum benefits from the course it is important that students do not see this course in isolation from the other MPhil courses or research training they are taking.
• Responsibility lies with each student to consider the potential for their own research using methods common in fields of the social sciences that may seem remote. Ideally this task will be facilitated by integration of the SSRMC with discipline-specific courses in their departments and through reading and discussion.

Four sessions of two hours each.

Once a week for four weeks.