An Introduction to Statistics and Data Analysis Using Stata®

Lisa Daniels is the Hodson Trust Professor Emeritus of Economics at Washington College in Chestertown, Maryland. She specializes in development in Africa, where she worked for 10 years, beginning as a Peace Corps volunteer. During her time in Africa, she studied agricultural markets, market information systems, poverty trends, and micro- and small-scale enterprises. As part of her research on micro- and small-scale enterprises, she directed national surveys of 7,000 to 56,000 households and businesses in Bangladesh, Botswana, Kenya, Malawi, and Zimbabwe funded by the U.S. Agency for International Development. In each survey, she was responsible for the questionnaire design, sample selection, data collection and analysis, and report preparation. Her work from these surveys and other research in Africa and Asia appears in consulting reports and in peer-reviewed journals. In addition to research and fieldwork, she has taught a range of courses over the past 28 years, including a research methods course and a data analysis course that she has taught over 20 times. She has also presented her work related to teaching at more than a dozen workshops.Nicholas Minot is a Senior Research Fellow at the International Food Policy Research Institute (IFPRI) in Washington, D.C. Since joining IFPRI in 1997, he has carried out research on agricultural market reform, income diversification, spatial patterns in policy, and food price volatility in developing countries. This research often involves carrying out surveys of farmers, cooperatives, traders, and consumers to better understand changes in food marketing systems. In addition to research, he is involved in outreach and capacity-building activities, including offering short courses on the use of Stata for survey data analysis. Before joining IFPRI, he taught at the University of Illinois in Urbana–Champaign, served as a policy adviser in Zimbabwe, and analyzed survey data in Rwanda. Overall, he has worked in more than two dozen countries in Latin America, sub-Saharan Africa, North Africa, and Asia.

• Part 1: The research process and data collectionChapter 1: The research process and data collectionRead the literature and identify gaps or ways to extend the literatureExamine the theoryDevelop your research questions and hypothesesDevelop your research methodAnalyze the dataWrite the research paperChapter 2: Sampling techniquesSample designSelecting a sampleSampling weightsChapter 3: Questionnaire designStructured and semi-structure questionnairesOpen- and closed-ended questionsGeneral guidelines for questionnaire designDesigning the questionsCollecting the response dataSkip patternsEthical issuesPart 2: Describing DataChapter 4: An Introduction to StataOpening Stata and Stata WindowsWorking with existing dataEntering your own data into StataUsing log files and saving your workGetting helpSummary of commands used in chapterChapter 5: Preparing and transforming your dataChecking for outliersCreating new variablesMissing values in StataSummary of commands used in chapterChapter 6: Descriptive statisticsTypes of variable and measurementDescriptive statistics for all types of variables -- frequency tables and modesDescriptive statistics for variables measured as ordinal, interval, and ratio scales -- median and percentilesDescriptive statistics for continuous variables -- mean, variance, standard deviation, and coefficient of variationDescriptive statistics for categorical variables measured on a nominal or ordinal scale -- cross tabulationApplying sampling weightsFormatting output for use in a document (Word, Google Docs, etc.)Graphs to describe dataSummary of code used in chapterPart 3: Testing HypothesesChapter 7: The Normal distributionThe normal distribution and standard scoresSampling distributions and standard errorsExamining the theory and identifying the research question and hypothesisTesting for statistical significanceRejecting or not rejecting the null hypothesisInterpreting the resultsCentral limit theoremPresenting the resultsSummary of commands used in chapterChapter 8: Testing a hypothesis about a single meanWhen to use the one-sample t testCalculating the one-sample t testConducting a one-sample t testInterpreting the outputPresenting the resultsSummary of commands used in chapterChapter 9: Testing a hypothesis about two meansWhen to use a two independent-samples t testCalculating the t statisticConducting a t testInterpreting the outputPresenting the resultsSummary of commands used in chapterChapter 10: Analysis of varianceWhen to use one-way analysis of varianceCalculating the F ratioConducting a one-way analysis of variance testInterpreting the outputIs one mean different or are all of them different?Presenting the resultsSummary of commands used in chapterChapter 11: Cross-tabulation and the chi-squared testWhen to use the chi-squared testCalculating the chi-squared testConducting a chi-squared testInterpreting the outputPresenting the resultsSummary of commands used in chapterPart 4: Exploring relationshipsChapter 12: Linear regression analysisWhen to use a regression analysisCorrelationSimple regression analysisMultiple regression analysisPresenting the resultsSummary of commands used in chapterChapter 13: Regression DiagnosticsMeasurement errorSpecification errorMulticollinearityHeteroskedasticityEndogeneityNon-normalityPresenting the resultsSummary of commands used in chapterChapter 14: Regression analysis with categorical dependent variablesWhen to use logit or probit analysisUnderstanding the logit modelRunning logit and interpreting the resultsLogit vs probit regression modelsRegression analysis with other types of categorical dependent variablesPresenting the resultsSummary of commands used in chapterChapter 15: Writing a research paperIntroduction section of a research paperLiterature reviewData and methodsResultsDiscussionConclusions

"This book introduces statistical methods to students while, at the same time, walking them through the process by which to apply those methods to real-world problems using Stata. This is something that is severely lacking in methods texts at this time."