Indian Human Development Survey: Panel data analysis using STATA

methods@manchester is an initiative funded by the Faculty of Humanities, University of Manchester. It aims to:highlight Manchester's strength in research methods in the social sciencespromote interdisciplinary and innovative methodological developmentsfoster further developments, including training, through external fundingmethods@manchester achieves these aims by:web pages that showcase the expertise in research methods within the facultypromoting and facilitatingmethods-related eventsacross the universityholding twice-yearly high profile external eventsm﻿ethods@manchester is part of Methods North West, a regional hub for research methods innovation, celebrating the collective strength, depth and breadth of methodological expertise across the participating Universities of Manchester, Liverpool, Lancaster, Lancashire and Keele.

methods@manchester

Presented by Professor Amaresh Dubey and Professor Wendy Olsen.
In the morning session, Professor Dubey introduces the Indian Human Development Survey. As a large longitudinal field survey data at two time points, 2004/05 and 2011/12, it represents all regions of a huge country. He shows how to use design weights and calculate the design coefficient DEFF. He also introduces several STATA commands, notably svyset, reshape, generate, replace, label define, merge, keep, drop and rename. Participants are given the do files to run, making their experience easy.
Participants will be exposed to merging two rounds into a panel. Professor Dubey then presents the regression modeling of an outcome in Round 2 based on structural variables measured in Round 1. The response variable is likely to be a change-factor (transitions in/out of poverty versus staying in poverty). It is likely that logit regression will be used.
The afternoon session is shared with Professor Wendy Olsen, who will introduce measures of human capabilities using microdata. The afternoon will cover regression on an index of well-being or capabilities using two competing methods---a classical scale and confirmatory factor analysis -- in Stata. One can use SEM or GSEM commands to combine a regression with a CFA, as taught elsewhere. The more advanced commands used in the afternoon include recode, replace, regress, sem, estat eqgof, and predict. Again, participants are given the do files to run. 
This workshop is of intermediate level. Some past experience with SPSS or STATA is recommended. Participants are not able to take away the whole IHDS dataset with them. Indeed participants use a masked (anonymized) dataset version in the workshop.
Anyone can register online for access to the actual IHDS data for their research. See data access URL http://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/22626?searchSource=find-analyze-home&sortBy=&q=dubey

Basement Lab

Our {communityGuidelinesLink} describe the sort of content we prohibit on Eventbrite. If you suspect an event may be in violation, you can report it to us so we can investigate.

To report other categories of prohibited or illegal content, submit {link}