Analysis of Questionnaire Data with R
Chapman and Hall/CRC – 2011 – 280 pages
While theoretical statistics relies primarily on mathematics and hypothetical situations, statistical practice is a translation of a question formulated by a researcher into a series of variables linked by a statistical tool. As with written material, there are almost always differences between the meaning of the original text and translated text. Additionally, many versions can be suggested, each with their advantages and disadvantages.
Analysis of Questionnaire Data with R translates certain classic research questions into statistical formulations. As indicated in the title, the syntax of these statistical formulations is based on the well-known R language, chosen for its popularity, simplicity, and power of its structure. Although syntax is vital, understanding the semantics is the real challenge of any good translation. In this book, the semantics of theoretical-to-practical translation emerges progressively from examples and experience, and occasionally from mathematical considerations.
Sometimes the interpretation of a result is not clear, and there is no statistical tool really suited to the question at hand. Sometimes data sets contain errors, inconsistencies between answers, or missing data. More often, available statistical tools are not formally appropriate for the given situation, making it difficult to assess to what extent this slight inadequacy affects the interpretation of results. Analysis of Questionnaire Data with R tackles these and other common challenges in the practice of statistics.
"… useful for readers wishing to transfer knowledge of survey analysis and its application in other statistical packages to R. Insights in how a practitioner can use R to analyze one particular survey are very helpful and can be readily applied to one’s own work. … this text would be handy to have on my bookshelf to refer to when conducting survey analyses. … a good book to have …"
—Gregory E. Gilbert, The American Statistician, November 2014
"… excellently written and documented. The text covers many of the real-life concerns that arise when analyzing questionnaire data … . I recommend the book to any researchers and post-graduates embarking upon questionnaire design and analysis for the first time, especially in the field of social sciences."
—International Statistical Review, 80, 2012
"[T]he book is nicely compact, well organized, and for the reader who is already familiar with R, sampling, and survey methodology, it is quite easy to jump from section to section and read through them quickly. … I have found myself already referring to portions of the text as I consider various survey analyses, and I have recommended at least portions of it to students and colleagues. … an interesting and well-written book … ."
—Ronald D. Fricker, Jr., Journal of Statistical Software, Vol. 46, January 2012
Principles of Analysis
The Mental Health in Prison (MHP) Study
If You Are a Complete R Beginner
Description of Responses
Description using "Summary Statistics"
Summary Statistics in Subgroups
Evolution of a Numerical Variable across Time (Temperature Diagram)
Description of Relationships between Variables
Relative Risks and Odds-Ratios
Principal Component Analysis
A Spherical Representation of a Correlation Matrix
Focused Principal Component Analysis
Confidence Intervals and Statistical Tests of Hypothesis
Confidence Interval of a Proportion
Confidence Interval of a Mean
Confidence Interval of a Relative Risk or an Odds-Ratio
Statistical Tests of Hypothesis: Comparison of Two Percentages
Statistical Tests of Hypothesis: Comparison of Two Means
Statistical Tests of Hypothesis: The Correlation Coefficient
Statistical Tests of Hypothesis: More than Two Groups
Sample Size Requirements: The Survey Perspective
Sample Size Requirement: The Inferential Perspective
Introduction to Linear, Logistic, Poisson, and Other Regression Models
Linear Regression Models for Quantitative Outcomes
Logistic Regression for Binary Outcome
Logistic Regression for a Categorical Outcome with More than Two Levels
Logistic Regression for an Ordered Outcome
Regression Models for an Outcome Resulting from a Count
About Statistical Modelling
Coding Numerical Predictors
Coding Categorical Predictors
Assessing the Relative Importance of Predictors
Dealing with Missing Data
Random Effects and Multilevel Modelling
Principles for the Validation of a Composite Score
Item Analysis (1): Distribution
Item Analysis (2): The Multi trait Multi-method Approach to Confirm a Subscale Structure
Assessing the Unidimensionality of a Set of Items
Factor Analysis to Explore the Structure of a Set of Items
Measurement Error (1): Internal Consistency and the Cronbach Alpha
Measurement Error (2): Inter-rater Reliability
8 Introduction to Structural Equation Modelling
Linear Regression as a Particular Instance of Structural
Factor Analysis as a Particular Instance of Structural
Structural Equation Modelling in Practice
Introduction to Data Manipulation using R
Importing and Exporting Datasets
Manipulation of Datasets
Manipulation of Variables
Appendix A: The Analysis of Questionnaire Data using R: Memory Card
Importation Exportation of Datasets
Manipulation of Datasets
Manipulation of Variables
Validation of a Composite Score
After studying mathematics and getting his Ph.D. in biostatistics, the author graduated as a child and adolescent psychiatrist. He is now professor in biostatistics in Paris-Sud University, head of a master in public health and of the research lab "public health and mental health".