Statistics for Linguistics with R: A Practical Introduction

✍ Scribed by Stefan Th. Gries

• Publisher: De Gruyter Mouton
• Year: 2013
• Tongue: English
• Leaves: 374
• Edition: 2nd rev. ed.
• Category: Library

✦ Synopsis

This book is the revised and extended second edition of Statistics for Linguistics with R. The volume is an introduction to statistics for linguists using the open source software R. It is aimed at students and instructors/professors with little or no statistical background and is written in a non-technical and reader-friendly/accessible style.

It first introduces in detail the overall logic underlying quantitative studies: exploration, hypothesis formulation and operationalization, and the notion and meaning of significance tests. It then introduces some basics of the software R relevant to statistical data analysis. A chapter on descriptive statistics explains how summary statistics for frequencies, averages, and correlations are generated with R and how they are graphically represented best. A chapter on analytical statistics explains how statistical tests are performed in R on the basis of many different linguistic case studies: For nearly every single example, it is explained what the structure of the test looks like, how hypotheses are formulated, explored, and tested for statistical significance, how the results are graphically represented, and how one would summarize them in a paper/article. A chapter on selected multifactorial methods introduces how more complex research designs can be studied: methods for the study of multifactorial frequency data, correlations, tests for means, and binary response data are discussed and exemplified step-by-step. Also, the exploratory approach of hierarchical cluster analysis is illustrated in detail.

The book comes with many exercises, boxes with short think breaks and warnings, recommendations for further study, and answer keys as well as a statistics for linguists newsgroup on the companion website.

Just like the first edition, it is aimed at students, faculty, and researchers with little or no statistical background in statistics or the open source programming language R. It avoids mathematical jargon and discusses the logic and structure of quantitative studies and introduces descriptive statistics as well as a range of monofactorial statistical tests for frequencies, distributions, means, dispersions, and correlations. The comprehensive revision includes new small sections on programming topics that facilitate statistical analysis, the addition of a variety of statistical functions readers can apply to their own data, a revision of overview sections on statistical tests and regression modeling, a complete rewrite of the chapter on multifactorial approaches, which now contains sections on linear regression, binary and ordinal logistic regression, multinomial and Poisson regression, and repeated-measures ANOVA, and a new visual tool to identify the right statistical test for a given problem and data set. The amount of code available from the companion website has doubled in size, providing much supplementary material on statistical tests and advanced plotting.

• Preserves the critically appraised features of the first edition: think breaks, recommendations for further study as well as many exercises and complete answer keys for all exercises from the companion website and newsgroup
• Requires no prior background in statistics and presupposes only elementary linguistic knowledge
• Explicit step-by-step and identically organized explanations of all methods and their results (in particular in the new regressions chapter)
• Suitable both for adoption in one- or two-semester courses or for self-study
  

✦ Table of Contents

Preface
Chapter 1. Some fundamentals of empirical research
1. Introduction
2. On the relevance of quantitative methods in linguistics
3. The design and the logic of quantitative studies
3.1. Scouting
3.2. Hypotheses and operationalization
3.2.1. Scientific hypotheses in text form
3.2.2. Operationalizing your variables
3.2.3. Scientific hypotheses in statistical/mathematical form
3.3. Data collection and storage
3.4. The decision
3.4.1. One-tailed p-values from discrete probability distributions
3.4.2. Two-tailed p-values from discrete probability distributions
3.4.3. Extension: continuous probability distributions
4. The design of a factorial experiment: introduction
5. The design of a factorial experiment: another example
Chapter 2. Fundamentals of R
1. Introduction and installation
2. Functions and arguments
3. Vectors
3.1. Generating vectors
3.2. Loading and saving vectors
3.3. Editing vectors
4. Factors
4.1. Generating factors
4.2. Loading and saving factors
4.3. Editing factors
5. Data frames
5.1. Generating data frames
5.2. Loading and saving data frames
5.3. Editing data frame s
6. Some programming: conditionals and loops
6.1. Conditional expressions
6.2. Loops
7. Writing your own little functions
Chapter 3. Descriptive statistics
1. Univariate statistics
1.1. Frequency data
1.1.1. Scatterplots and line plots
1.1.2. Pie charts
1.1.3. Bar plots
1.1.4. Pareto-charts
1.1.5. Histograms
1.1.6 Empirical cumulative distributions
1.2. Measures of central tendency
1.2.1. The mode
1.2.2. The median
1.2.3. The arithmetic mean
1.2.4. The geometric mean
1.3. Measures of dispersion
1.3.1. Relative entropy
1.3.2. The range
1.3.3. Quantiles and quartiles
1.3.4. The average deviation
1.3.5. The standard deviation/variance
1.3.6. The variation coefficient
1.3.7. Summary functions
1.3.8. The standard error
1.4. Centering and standardization (z-scores)
1.5. Confidence intervals
1.5.1. Confidence intervals of arithmetic means
1.5.2. Confidence intervals of percentages
2. Bivariate statistics
2.1. Frequencies and crosstabulation
2.1.1. Bar plots and mosaic plots
2.1.2. Spineplots
2.1.3. Line plots
2.2. Means
2.2.1. Boxplots
2.2.2. Interaction plots
2.3. Coefficients of correlation and linear regression
Chapter 4. Analytical statistics
1. Distributions and frequencies
1.1. Distribution fitting
1.1.1. One dep. variable (ratio-scaled)
1.1.2. One dep. variable (nominal/categorical)
1.2. Tests for differences/independence
1.2.1. One dep. variable (ordinal/interval/ratio scaled) and one indep. variable (nominal) (indep. samples)
1.2.2. One dep. variable (nom./cat.) and one indep. variable (nom./cat.) (indep.samples)
1.2.3. One dep. variable (nom./cat.) (dep. samples)
2. Dispersions
2.1. Goodness-of-fit test for one dep. variable (ratio-scaled)
2.2. One dep. variable (ratio-scaled) and one indep. variable (nom.)
3. Means
3.1. Goodness-of-fit tests
3.1.1. One dep. variable (ratio-scaled)
3.1.2. One dep. variable (ordinal)
3.2. Tests for differences/independence
3.2.1. One dep. variable (ratio-scaled) and one indep. variable (nom.) (indep. samples)
3.2.2. One dep. variable (ratio-scaled) and one indep. variable (nom.) (dep. samples)
3.2.3. One dep. variable (ordinal) and one indep. variable (nom.) (indep. samples)
3.2.4. One dep. variable (ordinal) and one indep. variable (nom.) (dep. samples)
4. Coefficients of correlation and linear regression
4.1. The significance of the product-moment correlation
4.2. The significance of Kendall’s Tau
4.3. Correlation and causality
Chapter 5. Selected multifactorial and multivariate methods
1. The notions of interaction and model (selection)
1.1. Interactions
1.2. Model (selection)
1.2.1. Formulating the first model
1.2.2. Selecting a final model
2. Linear models
2.1. A linear model with a binary predictor
2.2. A linear model with a categorical predictor
2.3. A linear model with a numeric predictor
2.4. A linear model with two categorical predictors
2.5. A linear model with a categorical and a numeric predictor
2.6. A linear model with two numeric predictors
2.7. A linear model selection process with multiple predictors
3. Binary logistic regression models
3.1. A logistic regression with a binary predictor
3.2. A logistic regression with a categorical predictor
3.3. A logistic regression with a numeric predictor
3.4. A logistic regression with two categorical predictors
3.5. A logistic regression with a categorical and a numeric predictor
3.6. A logistic regression with two numeric predictors
4. Other regression models
4.1. An ordinal logistic regression with a categorical and a numeric predictor
4.2. A multinomial regression with a categorical and a numeric predictor
4.3. A Poisson/count regression with a categorical and a numeric predictor
5. Repeated measurements: a primer
5.1. One independent variable nested into subjects/items
5.2. Two independent variables nested into subjects/items
5.3. Two independent variables, one between, one within subjects/items
5.4. Mixed-effects / multi-level models
6. Hierarchical agglomerative cluster analysis
Chapter 6. Epilog
References