#
Analyzing Social Science Data
50 Key Problems in Data Analysis

- David de Vaus - University of Queensland, Australia

**Other Titles in:**

Data Collection & Analysis | Quantitative Methods | Research Methods & Evaluation (General)

**Analyzing Social Science Data**guides students in: problems with the initial data; problems with the initial variables; how to handle too much data; how to generalize; problems of analyzing single variables; problems examining bivariate relationships; and problems examining multivariate relationships

The book is a *tour de force* in making data analysis manageable and rewarding for today's undergraduate studying research methods.

`I'm full of admiration for this book. Once again, David de Vaus has come up with a superb book that is well written and organized and which will be a boon to a wide range of students. He has taken a vast array of problems that users of quantitative data analysis procedures are likely to encounter. The selection of issues and problems ... reflects the experience of a true practitioner with a grasp of his field and of the intricacies of the research process. The selection of issues clearly derives also from experience of teaching students how to do research and analyse data....A large number of practitioners will want the book. I was surprised at how much I learned from this. This will be a vital book for the bookshelves of practitioners of the craft of quantitative data analysis' - *Alan Bryman, Professor of Social Research, Loughborough University*

*
PART ONE: HOW TO PREPARE DATA FOR ANALYSIS
How to Code Data
How to Code Questions with Multiple Answers
Can the Respondent's Answers be Relied on?
How to Check that the Right Thing is Being Measured
TWO: HOW TO PREPARE VARIABLES FOR ANALYSIS
How to Deal with Variables with Lots of Categories
How to Identify and Change the Level of Measurement of Variables
How to Deal with Questions that Fail to Identify Real Differences Between Cases
How to Rearrange the Categories of a Variable
What to do with Gaps in the Data
What to do with People who 'Don't Know', 'Have no Opinion' or 'Can't Decide'
How to Tell if the Distribution is Normal
How to Tell if the Relationship is Linear
How to Tell if Outlier Cases are a Problem
What to do if the Required Variable is not Available
How to Compare Apples with Oranges
*

PART THREE: HOW TO REDUCE THE AMOUNT OF DATA TO ANALYSE
How to Work Out Which Variables to Use
How to Combine Information from a Set of Variables into a Single Measure
How to Build a Good Likert Scale
How to Build a Scale Using Factor Analysis
PART FOUR: HOW AND WHEN TO GENERALISE
What Does it Mean to Generalize?
How to Judge the Extent and Effect of Sample Bias
How to Weight Samples to Adjust for Bias
What are Tests of Significance?
Should Tests of Significance be Used?
What Factors Affect Significance Levels?
Is the Sample Large Enough to Achieve Statistical Significance?
Should Confidence Intervals be Used?
PART FIVE: HOW TO ANALYSE A SINGLE VARIABLE
How to Use Tables Effectively to Display the Distribution of a Single Variable
How to Use Graphs for Single Variables
Which Summary Statistics to Use to Describe a Single Variable
Which Statistics to Use to Generalise about a Single Variable
PART SIX: HOW TO ANALYSE TWO VARIABLES
How and When to Use Crosstabulations
Which Graph to Use
How to Narrow down the Choice When Selecting Summary Statistics
How to Interpret a Correlation Coefficient
Which Correlation?
How much Impact Does a Variable Have?
How to Tell if Groups are Different
Which Test of Significance?
How are Confidence Intervals used in Bivariate Analysis?
PART SEVEN: HOW TO CARRY OUT MULTIVARIATE ANALYSIS
Understanding Bivariate Relationships

Using Conditional Tables as a Method of Elaboration Analysis
Using Conditional Correlations for Elaboration Analysis
Using Partial Tables as a Method of Elaboration Analysis
Using Partial Correlations for Elaboration Analysis
What Type of Data are Needed for Multiple Regression?
How to do a Multiple Regression
How to Use Non-interval Variables in Multiple Regression
What Does the Multiple Regression Output Mean?
What Other Multivariate Methods are Availabe?
November 19, 2016

May 10, 2016

May 10, 2016

March 9, 2015

September 2, 2013

June 4, 2012

December 20, 2011

October 12, 2011

September 21, 2011

March 25, 2010
Comparing Scores on Different Variables |

The Logic of Elaboration Analysis |

I really like de Vaus problem-based approach to quantitative data analysis. Rather than having to read an entire book from A to Z, de Vaus compiled a list of 50 problems in data analysis that deemed him important or ubiqious enough. I personally can relate to most of these problems and I think that most researcher will come across those as most of them are fundamental. Writing a method book is always a compromise between depths and breadths. The great thing of its structure is that it can be read both from beginning to end as well as from problem to problem because of the crossreferences. That's why I've would liked a more extensive coverage of some of the problems, but that's of cause subjective. Also, it would be great if the book could be updated in terms of newer controversies, i.e. which measurement of effect size should be used. Multivariate methods would be a great addition too.

**Education , Universität Ulm**

Very useful guide for students looking to complete research in social science, with a number of concepts being transferable to other areas.

**HE Equine, Hartpury College**

recommended as a "rein check" for those students who are looking to embark on a social science project with limited prior experience.

**HE Equine, Hartpury College**

Good book which deals with those tricky situations which are not always tackeled in similar books.

**Social and General, Cork Institute of Technology**

The problem-based approach to data analysis is a key strength of this book.

**Cardiff School of Social Sciences, Cardiff University**

Analysing Social Science data: 50 Key Problems in Data Analysis, covers areas of data analysis which many research methods books do not cover in as much detail as this text does. The book consists of 50 short chapters which are group in to seven parts such as: Part One: How to Prepare Data for Analysis; Part Two: How to Prepare Variables for Analysis; Part Three: How to Reduce the Amount of Data to Analyse; Part Four: How and When to Generalise; Part Five: How to Analyse a Single Variable; Part Six: How to Analyse Two Variable; and Part Seven: How to Carry out Multivariate Analysis. This book is important for anyone undertaking a quantitative research project as the text covers succinctly how to code data, how to assess the reliability of research participants answers, how to judge question validity, and tackles problems with measuring the ‘mean average’ as well as preparing data for use with SPSS social software for data analysis. This book is essential reading for Masters and PhD students, as well as, researchers designing and implementing survey research in the social sciences.

**School of Computing, University of the West of Scotland**

This is a must book for research practitioners in almost all fields. Since the book was introduced in 2002, it has been widely recommended by the research community.

I recommend this book to my students who have little background in statistical methods. This book deals with the core assumptions about statistical analyses, common errors and problems and ways to address them.

The topical nature of the book serves as a user friendly guide for researchers to probe a specific issue. Highly recommended.

**Education , Higher College of Technology**

Enjoyable and easy to understand. It gives some really great suggestions on how to avoid many of the pitfalls in Data Analysis. Well worth keeping as a reference point.

**Psychology , Glyndwr University**

We really liked the format of this book and feel that 'outing' key problems and providing a clear rationale for how to deal with them is very relevant to students new to researching and using data. We will also use this book for a Year 2 module called Researching Education (100 students)

**Education , Roehampton University**

This text will be adopted for the next academic year and the foreseeable future. The book is very well written and at a level that students on non mathematics/statistics courses can understand. Many of the issues that this text addresses are the ones that I encounter on a day-to-day basis when students consult me with their data analysis problems. For example, dealing with outliers, missing data and non-normally distributed distributions. I particularly like the way the chapter headings are stated as questions and that each chapter is relatively short. These features should reduce student anxiety, which is known to be an important issue for most students learning statistics.

**Dept. Sport, Health & Exercise Science, Hull University**