reading

Cover
Title Page
Copyright
About the Companion Website
Chapter 1: Introduction
1. 1.1 At Whom is This Book Aimed?
2. 1.2 At What Scale of Project is This Book Aimed?
3. 1.3 Why Might This Book be Useful for You?
4. 1.4 How to Use This Book
5. 1.5 Computer Based Statistics Packages
6. 1.6 Relevant Videos etc.
Chapter 2: Data Types
1. 2.1 What Types of Data are There and Why Does it Matter?
2. 2.2 Continuous Measured Data
3. 2.3 Ordinal Data
4. 2.4 Categorical Data
5. 2.5 Ambiguous Cases
6. 2.6 Relevant Videos etc.
Chapter 3: Presenting and Summarizing Data
1. 3.1 Continuous Measured Data
2. 3.2 Ordinal Data
3. 3.3 Categorical Data
4. 3.4 Relevant Videos etc.
5. Appendix 1: An Example of the Insensitivity of the Median When Used to Describe Data from an Ordinal Scale With a Narrow Range of Possible Values
Chapter 4: Choosing a Statistical Test
1. 4.1 Identify the Factor and Outcome
2. 4.2 Identify the Type of Data Used to Record the Relevant Factor
3. 4.3 Statistical Methods Where the Factor is Categorical
4. 4.4 Correlation and Regression with a Measured Factor
5. 4.5 Relevant Additional Material
Chapter 5: Multiple Testing
1. 5.1 What Is Multiple Testing and Why Does It Matter?
2. 5.2 What Can We Do to Avoid an Excessive Risk of False Positives?
Chapter 6: Common Issues and Pitfalls
1. 6.1 Determining Equality of Standard Deviations
2. 6.2 How Do I Know, in Advance, How Large My SD Will Be?
3. 6.3 One‐Sided Versus Two‐Sided Testing
4. 6.4 Pitfalls That Make Data Look More Meaningful Than It Really Is
5. 6.5 Discussion of Statistically Significant Results
6. 6.6 Discussion of Non‐Significant Results
7. 6.7 Describing Effect Sizes with Non‐Parametric Tests
8. 6.8 Confusing Association with a Cause and Effect Relationship
Chapter 7: Contingency Chi‐Square Test
1. 7.1 When Is the Test Appropriate?
2. 7.2 An Example
3. 7.3 Presenting the Data
4. 7.4 Data Requirements
5. 7.5 An Outline of the Test
6. 7.6 Planning Sample Sizes
7. 7.7 Carrying Out the Test
8. 7.8 Special Issues
9. 7.9 Describing the Effect Size
10. 7.10 How to Report the Analysis
11. 7.11 Confounding and Logistic Regression
12. 7.12 Larger Tables
13. 7 12.2 Reducing Tables
14. 7.13 Relevant Videos etc.
Chapter 8: Independent Samples (Two‐Sample) T‐Test
1. 8.1 When Is the Test Applied?
2. 8.3 Presenting the Data
3. 8.4 Data Requirements
4. 8.5 An Outline of the Test
5. 8.6 Planning Sample Sizes
6. 8.7 Carrying Out the Test
7. 8.8 Describing the Effect Size
8. 8.9 How to Describe the Test, the Statistical and Practical Significance of Your Findings in Your Report
9. 8.10 Relevant Videos etc.
Chapter 9: Mann–Whitney Test
1. 9.1 When Is the Test Applied?
2. 9.2 An Example
3. 9.3 Presenting the Data
4. 9.4 Data Requirements
5. 9.5 An Outline of the Test
6. 9.6 Statistical Significance
7. 9.7 Planning Sample Sizes
8. 9.8 Carrying Out the Test
9. 9.9 Describing the Effect Size
10. 9.10 How to Report the Test
11. 9.11 Relevant Videos etc.
Chapter 10: One‐Way Analysis of Variance (ANOVA) – Including Dunnett's and Tukey's Follow Up Tests
1. 10.1 When Is the Test Applied?
2. 10.2 An Example
3. 10.3 Presenting the Data
4. 10.4 Data Requirements
5. 10.5 An Outline of the Test
6. 10.6 Follow Up Tests
7. 10.7 Planning Sample Sizes
8. 10.8 Carrying Out the Test
9. 10.9 Describing the Effect Size
10. 10.10 How to Report the Test
11. 10.11 Relevant Videos etc.
Chapter 11: Kruskal–Wallis
1. 11.1 When Is the Test Applied?
2. 11.2 An Example
3. 11.3 Presenting the Data
4. 11.4 Data Requirements
5. 11.5 An Outline of the Test
6. 11.6 Planning Sample Sizes
7. 11.7 Carrying Out the Test
8. 11.8 Describing the Effect Size
9. 11.9 Determining Which Group Differs from Which Other
10. 11.10 How to Report the Test
11. 11.11 Relevant Videos etc.
Chapter 12: McNemar's Test
1. 12.1 When Is the Test Applied?
2. 12.2 An Example
3. 12.3 Presenting the Data
4. 12.4 Data Requirements
5. 12.5 An Outline of the Test
6. 12.6 Planning Sample Sizes
7. 12.7 Carrying Out the Test
8. 12.8 Describing the Effect Size
9. 12.9 How to Report the Test
10. 12.10 Relevant Videos etc.
Chapter 13: Paired T‐Test
1. 13.1 When Is the Test Applied?
2. 13.2 An Example
3. 13.3 Presenting the Data
4. 13.4 Data Requirements
5. 13.5 An Outline of the Test
6. 13.6 Planning Sample Sizes
7. 13.7 Carrying Out the Test
8. 13.8 Describing the Effect Size
9. 13.9 How to Report the Test
10. 13.10 Relevant Videos etc.
Chapter 14: Wilcoxon Signed Rank Test
1. 14.1 When Is the Test Applied?
2. 14.2 An Example
3. 14.3 Presenting the Data
4. 14.4 Data Requirements
5. 14.5 An Outline of the Test
6. 14.6 Planning Sample Sizes
7. 14.7 Carrying Out the Test
8. 14.8 Describing the Effect Size
9. 14.9 How to Report the Test
10. 14.10 Relevant Videos etc.
Chapter 15: Repeated Measures Analysis of Variance
1. 15.1 When Is the Test Applied?
2. 15.2 An Example
3. 15.3 Presenting the Data
4. 15.4 Data Requirements
5. 15.5 An Outline of the Test
6. 15.6 Planning Sample Sizes
7. 15.7 Carrying Out the Test
8. 15.8 Describing the Effect Size
9. 15.9 How to Report the Test
10. 15.10 Relevant Videos etc.
Chapter 16: Friedman Test
1. 16.1 When Is the Test Applied?
2. 16.2 An Example
3. 16.3 Presenting the Data
4. 16.4 Data Requirements
5. 16.5 An Outline of the Test
6. 16.6 Planning Sample Sizes
7. 16.7 Follow Up Tests
8. 16.8 Carrying Out the Tests
9. 16.9 Describing the Effect Size
10. 16.10 How to Report the Test
11. 16.11 Relevant Videos etc.
Chapter 17: Pearson Correlation
1. 17.1 Presenting the Data
2. 17.2 Correlation Coefficient and Statistical Significance
3. 17.3 Planning Sample Sizes
4. 17.4 Effect Size and Practical Relevance
5. 17.5 Regression
6. 17.6 How to Report the Analysis
7. 17.7 Relevant Videos etc.
Chapter 18: Spearman Correlation
1. 18.1 Presenting the Data
2. 18.2 Testing for Evidence of Inappropriate Distributions
3. 18.3 Rho and Statistical Significance
4. 18.4 An Outline of the Significance Test
5. 18.5 Planning Sample Sizes
6. 18.6 Effect Size
7. 18.7 Where Both Measures Are Ordinal
8. 18.8 How to Report Spearman Correlation Analyses
9. 18.9 Relevant Videos etc.
Chapter 19: Logistic Regression
1. 19.1 Use of Logistic Regression with Categorical Outcomes
2. 19.2 An Outline of the Significance Test
3. 19.3 Planning Sample Sizes
4. 19.4 Results of the Analysis
5. 19.5 Describing the Effect Size
6. 19.6 How to Report the Analysis
7. 19.7 Relevant Videos etc.
Chapter 20: Cronbach's Alpha
1. 20.1 Appropriate Situations for the Use of Cronbach's Alpha
2. 20.2 Inappropriate Uses of Alpha
3. 20.3 Interpretation
4. 20.4 Reverse Scoring
5. 20.5 An Example
6. 20.6 Performing and Interpreting the Analysis
7. 20.7 How to Report Cronbach's Alpha Analyses
8. 20.7 Relevant Videos etc.
Glossary
Videos
Index
End User License Agreement

List of Tables

Chapter 1: Introduction
1. Table 1.1 The ideal stage‐by‐stage flow of events for a research program.
Chapter 3: Presenting and Summarizing Data
1. Table 3.1 Levels of agreement among younger and older patients with the statement that “I would be happy to book future appointments electronically.” Higher values reflect stronger agreement. Medians are emboldened.
Chapter 4: Choosing a Statistical Test
1. Table 4.1 Examples of factors and outcomes.
2. Table 4.2 Identification of appropriate statistical tests for studies with a categorical factor.
3. Table 4.3 Selection of correlation or regression techniques depending on type of data used to record the factor and outcome.
Chapter 6: Common Issues and Pitfalls
1. Table 6.1 Characteristics of one and two‐sided questions and testing procedures.
2. Table 6.2 Rounding to three significant digits.
Chapter 7: Contingency Chi‐Square Test
1. Table 7.1 Examples of studies that would be analyzed by a contingency chi‐square test.
2. Table 7.2 Contingency table showing numbers and percentages with and without pneumonia among patients who do or do not receive physiotherapy and exercise in addition to normal care.
Chapter 8: Independent Samples (Two‐Sample) T‐Test
1. Table 8.1 Examples of studies that would be analyzed by an independent samples (two‐sample) t‐test.
Chapter 9: Mann–Whitney Test
1. Table 9.1 Examples of studies that would be analyzed by a Mann–Whitney test.
Chapter 10: One‐Way Analysis of Variance (ANOVA) – Including Dunnett's and Tukey's Follow Up Tests
1. Table 10.1 Examples of studies that would be analyzed by a one‐way analysis of variance.
2. Table 10.2 Necessary sample sizes per group for a one‐way analysis of variance requiring 90% power.
Chapter 11: Kruskal–Wallis
1. Table 11.1 Examples of studies that would be analyzed by a Kruskal–Wallis test.
2. Table 11.2 Number of women reporting various levels of menstrual problems with different forms of contraception.
3. Table 11.3 Descriptive statistics for levels of menstrual problems in women using different methods of contraception.
4. Table 11.4 Numbers (and percentages) of women reporting lower or higher levels of menstrual problems with different forms of contraception.
Chapter 12: McNemar's Test
1. Table 12.1 Examples of studies that would be analyzed by McNemar's test.
2. Table 12.2 Responses to the question “Have you ever smoked cannabis?” when posed as part of a questionnaire or in a face‐to‐face interview.
Chapter 13: Paired T‐Test
1. Table 13.1 Examples of studies that would be analyzed by a paired t‐test.
Chapter 14: Wilcoxon Signed Rank Test
1. Table 14.1 Examples of studies that would be analyzed by a Wilcoxon signed rank test.
2. Table 14.2 Numbers of individuals expressing various levels of satisfaction with an appointment booking system, before and after change to system. Higher scores represent greater levels of satisfaction.
Chapter 15: Repeated Measures Analysis of Variance
1. Table 15.1 Examples of studies that would be analyzed by a repeated measures analysis of variance.
2. Table 15.2 Weights of participants (kg) at the three time points, and the differences in weight comparing all possible pairs of time points (first ten participants only). Each row represents one participant.
3. Table 15.3 Appropriate sample sizes for a repeated measures analysis of variance.
4. Table 15.4 Changes in weight (kg) when comparing various pairs of time points. Means, SDs and 95% confidence intervals (Bonferroni corrected) are shown.
Chapter 16: Friedman Test
1. Table 16.1 Examples of studies that would be analyzed by a Friedman test.
2. Table 16.2 Descriptive statistics for dental hygiene scores prior to, one week after, and three months after training in tooth brushing technique.
3. Table 16.3 Descriptive statistics for the individual changes in dental hygiene scores when comparing the pre‐training period to the one week and three month time points.
Chapter 18: Spearman Correlation
1. Table 18.1 Numbers and proportions with various levels of likelihood of using the internet to research psoriasis among groups separated by level of education (1 Lowest; 4 Highest).
Chapter 20: Cronbach's Alpha
1. Table 20.1 Cronbach's Alpha values when each question was omitted one at a time.

List of Illustrations

Chapter 2: Data Types
1. Figure 2.1 A continuously varying measure with (a) normal, (b) skewed, and (c) bimodal distribution. In (d) the highest and lowest values (tails) from an otherwise normal distribution are missing.
2. Figure 2.2 Histogram of “Long‐tailed” data, i.e. data that includes both low and high outlying values.
3. Figure 2.3 Normal probability plots of (a) normally distributed data and (b) long‐tailed data.
Chapter 3: Presenting and Summarizing Data
1. Figure 3.1 The quartiles for a data set with ranked values ranging from 1 to 447. The quartiles, median, and interquartile range (IQR) are indicated. The mean and SD are also included.
2. Figure 3.2 Histogram of body temperatures among individuals exposed to the common cold virus.
3. Figure 3.3 Bar chart of opinions concerning changes to an appointment booking system.
Chapter 4: Choosing a Statistical Test
1. Figure 4.1 Approximately linear relationship – (a) positive correlation and (b) negative correlation.
2. Figure 4.2 A cluster of points with outlier(s).
3. Figure 4.3 Relationships that are clearly not monotonic.
4. Figure 4.4 Monotonic, but clearly non‐linear relationships – (a) positive and (b) negative.
5. Figure 4.5 Data with distinct clusters.
Chapter 6: Common Issues and Pitfalls
1. Figure 6.1 Hypothetical confidence intervals for the difference in blood pressure between actively treated and control patients. All are statistically significant (the null hypothesis figure of zero is excluded from the confidence interval), but their interpretations differ. Vertical broken lines indicate limits beyond which differences would be of practical/clinical significance.
2. Figure 6.2 Hypothetical confidence intervals for the relative risk of death for actively treated patients compared to controls. All are statistically significant (the null hypothesis figure of one is excluded from the confidence interval), but their interpretations differ. Vertical broken lines indicate limits beyond which differences would be of practical/clinical significance.
3. Figure 6.3 Hypothetical confidence intervals for the difference in blood pressure between actively treated and control patients. All are statistically non‐significant (the null hypothesis figure of zero is included by the confidence interval), but their interpretations differ. Vertical broken lines indicate limits beyond which differences would be of practical/clinical significance.
4. Figure 6.4 Hypothetical confidence intervals for the relative risk of death for actively treated patients compared to controls. All are statistically non‐significant significant (the null hypothesis figure of one is included by the confidence interval), but their interpretations differ. Vertical broken lines indicate limits beyond which differences would be of practical/clinical significance.
Chapter 7: Contingency Chi‐Square Test
1. Figure 7.1 Experimental structure where a contingency chi‐square test is appropriate.
2. Figure 7.2 Clustered bar chart showing numbers with and without pneumonia among cardiac surgery patients who do or do not receive physiotherapy and exercise in addition to normal care.
3. Figure 7.3 Stacked bar chart showing proportions with and without pneumonia among cardiac surgery patients who do or do not receive physiotherapy and exercise in addition to normal care.
Chapter 8: Independent Samples (Two‐Sample) T‐Test
1. Figure 8.1 Circumstances when an independent samples t‐test is used.
2. Figure 8.2 Histograms of visual analog scores for pain 24 hours after abdominal or laparoscopic hysterectomy. (a) Abdominal, (b) Laparoscopic.
3. Figure 8.3 Normal probability plots of pain data following (a) Abdominal or (b) Laparoscopic hysterectomy.
Chapter 9: Mann–Whitney Test
1. Figure 9.1 Circumstances when a Mann–Whitney test is used.
2. Figure 9.2 Bar charts showing numbers of individuals reporting various levels of agreement with the statement “I would prefer to get stopping smoking advice face‐to‐face rather than some other way” displayed by respondents' sex.
Chapter 10: One‐Way Analysis of Variance (ANOVA) – Including Dunnett's and Tukey's Follow Up Tests
1. Figure 10.1 Structure of an experiment or survey where a one‐way analysis of variance would be appropriate.
2. Figure 10.2 Histograms of changes in blood cholesterol levels (mmol/l) over the three month study period for participants (a) making no dietary change, (b) using a cholesterol lowering spread in place of their normal spread, and (c) adding 35 g of nuts per day to their diet.
Chapter 11: Kruskal–Wallis
1. Figure 11.1 The structure of a study suitable for analysis by a Kruskal–Wallis test.
2. Figure 11.2 Stacked bar chart showing percentages of individuals reporting various levels of menstrual problems displayed by contraceptive type. Higher grades indicate greater problems.
Chapter 12: McNemar's Test
1. Figure 12.1 Structure of a study where a McNemar's test would be used.
Chapter 13: Paired T‐Test
1. Figure 13.1 Structure of a study where a paired t‐test would be used.
2. Figure 13.2 Dot plot showing changes in systolic blood pressure (mmHg) after one month's use of a combined oral contraceptive.
3. Figure 13.3 A ladder plot of systolic blood pressures (mmHg) before and after one month's use of a combined oral contraceptive.
4. Figure 13.4 Normal probability plot of the individual changes in systolic blood pressure used to test for any evidence of a long‐tailed distribution.
Chapter 14: Wilcoxon Signed Rank Test
1. Figure 14.1 Structure of a study where a Wilcoxon signed rank test would be used.
2. Figure 14.2 Proportions of participants expressing various levels of satisfaction with the appointment booking system before and after introduction of the new system. Higher scores represent higher levels of satisfaction.
3. Figure 14.3 Bar chart of individual changes in satisfaction scores following introduction of the new appointment booking system. Negative and positive figures represent reduced or increased satisfaction, respectively.
Chapter 15: Repeated Measures Analysis of Variance
1. Figure 15.1 Structure of a study where a repeated measures analysis of variance would be used.
2. Figure 15.2 Ladder plot showing individual weights (kg) for participants at the three time points in the study (first ten individuals only).
3. Figure 15.3 Weights (kg) of participants at the three time points in the study (mean ± SD). (Data for all 106 participants are described.)
Chapter 16: Friedman Test
1. Figure 16.1 Structure of a study where a Friedman test would be used.
2. Figure 16.2 Stacked bar chart showing dental hygiene scores prior to, one week after, and three months after training in tooth brushing technique.
3. Figure 16.3 Bar chart showing individual changes in dental hygiene scores one week after and three months after training in tooth brushing technique.
4. Figure 16.4 Mean dental hygiene scores (± SD) at the three time points in the study.
Chapter 17: Pearson Correlation
1. Figure 17.1 FEV1 values (liters) among residents versus number of years employment in a local cotton mill.
2. Figure 17.2 Possible relationships between FEV1 and period of employment in a cotton factory.
Chapter 18: Spearman Correlation
1. Figure 18.1 Proportions of new mothers in four age bands who provided varioussatisfaction ratings for the maternity service (1 = Strongly dissatisfied; 2 = Dissatisfied; 3 = Neutral; 4 = Generally satisfied; 5 = Very satisfied).
2. Figure 18.2 Data relationships that would not be suitable for Spearman correlation.
3. Figure 18.3 Proportions with various levels of willingness to research psoriasis on the Internet among four groups separated according to educational level (1 Lowest; 4 Highest).
Chapter 19: Logistic Regression
1. Figure 19.1 (a) A strong relationship between a binary outcome and a continuously varying measured factor, which is likely to be statistically significant. (b) A weaker relationship that is less likely to be significant.
2. Figure 19.2 Percentage of patients developing a new case of gallstones during the two years of the study. Patients have been divided into four groups based on dosage of drug received.

This edition first published 2018

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, except as permitted by law. Advice on how to obtain permission to reuse material from this title is available at http://www.wiley.com/go/permissions.

The right of Adam Mackridge and Philip Rowe to be identified as the authors of this work has been asserted in accordance with law.

Registered Office

John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, USA

Editorial Office

111 River Street, Hoboken, NJ 07030, USA

For details of our global editorial offices, customer services, and more information about Wiley products visit us at www.wiley.com.

Wiley also publishes its books in a variety of electronic formats and by print-on-demand. Some content that appears in standard print versions of this book may not be available in other formats.

Limit of Liability/Disclaimer of Warranty

The publisher and the authors make no representations or warranties with respect to the accuracy or completeness of the contents of this work and specifically disclaim all warranties; including without limitation any implied warranties of fitness for a particular purpose. This work is sold with the understanding that the publisher is not engaged in rendering professional services. The advice and strategies contained herein may not be suitable for every situation. In view of on-going research, equipment modifications, changes in governmental regulations, and the constant flow of information relating to the use of experimental reagents, equipment, and devices, the reader is urged to review and evaluate the information provided in the package insert or instructions for each chemical, piece of equipment, reagent, or device for, among other things, any changes in the instructions or indication of usage and for added warnings and precautions. The fact that an organization or website is referred to in this work as a citation and/or potential source of further information does not mean that the author or the publisher endorses the information the organization or website may provide or recommendations it may make. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this works was written and when it is read. No warranty may be created or extended by any promotional statements for this work. Neither the publisher nor the author shall be liable for any damages arising here from.

Library of Congress Cataloguing-in-Publication Data

Names: Mackridge, Adam (Adam John), 1979- author. | Rowe, Philip, author.

Title: A practical guide to statistics for health research / by Adam Mackridge, Philip Rowe.

Description: Hoboken, NJ : Wiley, 2018. | Includes bibliographical references and index. |

Identifiers: LCCN 2017055955 (print) | LCCN 2018001635 (ebook) | ISBN 9781119383598 (pdf) | ISBN 9781119383611 (epub) | ISBN 9781119383574 (hardback)

Subjects: LCSH: Medical statistics. | Medicine–Research–Statistical methods. | BISAC: MEDICAL / Epidemiology. | MATHEMATICS / Probability & Statistics / General.

Classification: LCC R853.S7 (ebook) | LCC R853.S7 M33 2018 (print) | DDC 610.2/1–dc23

LC record available at https://lccn.loc.gov/2017055955

Cover Design: Wiley

A Practical Approach to Using Statistics in Health Research

From Planning to Reporting

About the Companion Website