Essential Statistics for Bioscientists E-Book

Mohammed Meah

• Herausgeber: John Wiley & Sons
• Kategorie: Wissenschaft und neue Technologien
• Sprache: Englisch

Seitenzahl: 309

Veröffentlichungsjahr: 2022

Beschreibung

Dive into the most common statistical tests and software packages used for scientific data analysis and interpretation 

In Essential Statistics For Bioscientists, experienced university and bioscientist Dr Mohammed Meah delivers  easy access to statistical analysis and data presentation. It is a great resource for students in the field of life and health sciences to conceptualize, analyze, and present data. This book uses three popular and commonly used statistics softwares—Microsoft Excel, Graphpad Prism, and SPSS—and offers clear, step-by-step instructions for essential data analysis and graphical/tabular display of data.   

Beginning with fundamental statistics terminology and concepts, including data types, descriptive statistics (central and spread of data), exploratory statistics (graphical display) and inferential statistics (hypothesis testing and correlation), the content gradually builds in complexity, explaining which statistical test is best suited and how to perform it.   

• A thorough introduction to basic statistical terms and building up to an advanced level of statistical application- ideal for those new to study of statistics 
• Extensive application of three popular software packages- Microsoft Excel, Graphpad Prism and SPSS 
• Numerous hands-on examples of performing data analysis using Microsoft Excel, Graphpad Prism, and SPSS 
• Considers the limitations and errors of statistical analysis 
• Essential reading for those designing and planning a research project in Biosciences 

Perfect for undergraduate students in the life and health sciences, Essential Statistics For Bioscientists will also earn a place in the libraries of anyone studying medicine, nursing, physiotherapy, pharmacy, and dentistry requiring a refresher or primer on statistical fundamentals. 

Leseprobe

Essential Statistics for Bioscientists

MOHAMMED MEAHSchool of Health, Sport & BioscienceBiological and Medical SciencesUniversity of East LondonLondonUK

This edition first published 2022

© 2022 John Wiley & Sons Ltd

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 Mohammed Meah to be identified as the author of this work has been asserted in accordance with law.

Registered Offices

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

John Wiley & Sons Ltd, The Atrium, Southern Gate, Chichester, West Sussex, PO19 8SQ, UK

Editorial Office

9600 Garsington Road, Oxford, OX4 2DQ, UK

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

While the publisher and authors have used their best efforts in preparing this work, they 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 merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives, written sales materials or promotional statements for this work. The fact that an organization, website, or product is referred to in this work as a citation and/or potential source of further information does not mean that the publisher and authors endorse the information or services the organization, website, or product may provide or recommendations it may make. 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 your situation. You should consult with a specialist where appropriate. Further, readers should be aware that websites listed in this work may have changed or disappeared between when this work was written and when it is read. Neither the publisher nor authors shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.

Library of Congress Cataloging-in-Publication Data

Names: Meah, M. S. (Mohammed Shahabu), author. Title: Essential statistics for bioscientists / Mohammed Meah. Description: Hoboken, NJ : John Wiley & Sons Ltd, 2022. | Includes bibliographical references. Identifiers: LCCN 2021060393 (print) | LCCN 2021060394 (ebook) | ISBN 9781119712008 (paperback) | ISBN 9781119712015 (pdf) | ISBN 9781119712022 (epub)Subjects: LCSH: Biometry. | Life sciences--Research--Methodology.Classification: LCC QH323.5 .M43 2022 (print) | LCC QH323.5 (ebook) | DDC 570.1/5195--dc23/eng/20220202LC record available at https://lccn.loc.gov/2021060393LC ebook record available at https://lccn.loc.gov/2021060394

Cover image: © Peter Hermes Furian/Shutterstock, Courtesy of Mohammed Meah

Cover design by Wiley

Set in 10/12pt STIXTwoText by Integra Software Services Pvt. Ltd, Pondicherry, India

Acknowledgements

I would like to thank my daughters (Maryam and Zaynah) for their constant encouragement and support during the writing of this book. I would like to thank the many project students over the years whose research project ideas have challenged me to delve into statistics. I have to say a special thank you to my mother (BR Chowdhury) and father (MM Chowdhury) for always being positive and supportive. Many thanks to my colleagues (Rane and Elizabeth) for their insightful comments and suggestions.

Lastly, I would like to thank Wiley Publishers for their encouragement and above all patience and understanding in the completion of this book.

“If it’s green or wriggles, it’s biology. If it stinks, it’s chemistry. If it doesn’t work, it’s physics or engineering. If it’s green and wiggles and stinks and still doesn’t work, it’s psychology. If it’s incomprehensible, it’s mathematics. If it puts you to sleep, it’s statistics.”

— Anonymous

List of Worked Examples of Statistical Tests

Example

Name

Page

Software

1

Example 1.1: Calculation of the mean

2

Example 1.2: Calculation of the median for an odd sample

3

Example 1.3: Calculation of the median for an even sample

4

Example 1.4: Calculation of the mode

5

Example 1.5: Calculation of the standard error of the mean (SEM)

6

Example 1.6: Calculation of the range

7

Example 1.7: Calculation of the standard deviation (SD) manually

8

Example 1.8: Calculation of the variance manually

9

Example 1.9: Calculation of the coefficient of variation (CV)

10

Example 1.10: Calculation of quartiles

11

Example 1.11: Calculation of the percentile

12

Example 1.12: Calculation of the confidence interval from sample (CI)

13

Example 1.13 calculation of the confidence interval with known mean and SD

14

Example 2.1: Drawing a dot plot manually from raw data

15

Example 2.2: Drawing a stem andleaf plot manually

16

Example 2.3: Drawing a box and whisker plot from sample data manually

17

Example 2.4: Drawing a frequency and cumulative frequency distribution curve

18

Example 2.5: Drawing a cumulative frequency distribution curve

19

Example 2.6: Use of normal distribution and Z scores to solve problems

20

Example 3.1: calculating the F test to see if the variance in two groups are equal or unequal.

21

Example 4.1: Calculation of t for a single sample

22

Example 4.2: calculation of one-way ANOVA manually using algebra

23

Example 4.3: Calculation of one-way ANOVA manually

24

Example 5.1: Calculation of difference between pairs of measurements using Sign test

25

Example 5.2: Calculation of differences between two scores using a Wilcoxon test

26

Example 5.3: Calculation of the differences in results for 2 groups using a Mann-Whitney U test

Example 5.4: Calculation of the chi-squared test

27

Example 5.5: calculation of the Spearman Rank correlation theoretically

28

Example 6.1: Putting data into an Excel spreadsheet

EXCEL

29

Example 6.2: Calculating descriptive stats using formulas ‘f

x

’

EXCEL

30

Example 6.3: Calculating descriptive stats using formulas typed directly into cells

EXCEL

31

Example 6.4: Calculating descriptive statistics using data analysis tab

EXCEL

32

Example 6.5: Plotting a bar graph using Excel

EXCEL

33

Example 6.6: Plotting multiple bar graphs using Excel

EXCEL

34

Example 6.7: Using a stacked bar to plot a Gantt chart with Microsoft Excel:

EXCEL

35

Example 6.8: Plotting a scatter plot with Microsoft Excel:

EXCEL

36

Example 6.9: Plotting a pie chart using Microsoft Excel:

EXCEL

37

Example 6.10: Plotting a frequency distribution curve with Excel

EXCEL

38

Example 6.11: How to do a paired t-test

EXCEL

39

Example 6.12: How to do an unpaired (independent) t-test and F test

EXCEL

40

Example 6.13: How to do One-way ANOVA

EXCEL

41

Example 6.14: Bonferroni Post hoc test

EXCEL

42

Example 6.15: How to do 2-way ANOVA

EXCEL

43

Example 7.1: Inputting data and replicates into Prism

Prism

44

Example 7.2: Inputting calculated means, SEMs or SD data into prism

Prism

45

Example 7.3: Calculating the mean, SD and SEM in prism.

Prism

46

Example 7.4: Plotting a histogram in prism

Prism

47

Example 7.5: Producing a cumulative frequency plot using prism

Prism

48

Example 7.6: To determine the correlation coefficient using prism

Prism

49

Example 7.7: Linear regression using prism

Prism

50

Example 7.8: Students t-test using prism

Prism

51

Example 7.9: Paired t test in prism

Prism

52

Example 7.10: One-way ANOVA using prism

Prism

53

Example 7.11: 2-way ANOVA using means

Prism

54

Example 7.12: 2 Way ANOVA with raw data

Prism

55

Example 7.13: Wilcoxon non-parametric test

Prism

56

Example 7.14: Mann-Whitney non-parametric test

Prism

57

Example 8.1: Descriptive stats, graphical display of data, histograms, test of normality

SPSS

58

Example 8.2: Unpaired t test using SPSS

SPSS

59

Example 8.3: To test whether there is a difference between the means of 2 sets of paired measurements.

SPSS

60

Example 8.4: Association between data- correlation

SPSS

61

Example 8.5: Repeated Measures ANOVA

SPSS

62

Example 8.6: One-way ANOVA

SPSS

63

Example 8.7: Two-way ANOVA

SPSS

64

Example 8.8: Wilcoxon test

SPSS

65

Example 8.9: Mann-Whitney test

SPSS

66

Example 8.10: Kruskall Wallis Test (nonparametric one- way ANOVA)

SPSS

67

Example 8.11: Friedman Test (nonparametric repeated measures ANOVA)

SPSS

Introduction

“All life is an experiment. The more experiments you make, the better.”

Ralph Waldo Emerson (1803–1882) - American lecturer, philosopher and poet

The word statistics is derived from the Latin word ‘status’ – meaning political state or a government. Statistics deals with collection, organization, presentation, analysis and interpretation of data to obtain meaningful and useful information. Statistics can be split into two major areas, namely, descriptive and inferential. Descriptive statistics involves collecting, summarizing, and presenting data. Inferential statistics involves analysing sample data to draw conclusions about a population.

Statistics is an area which is often much reduced in the curriculum of undergraduate bioscience degree courses. Statistics tends to be linked to research modules. Lecturers often assume that students have a strong grasp of mathematical and statistical concepts including data analysis. However, the reality is that most students are ‘rusty’ in these areas, particularly in statistics. The most urgent need for statistics is usually for the research project which is typically in the final year of the undergraduate degree (level 6). It is unclear, during undergraduate studies, how much and when statistics should be taught. In addition, there are a variety of software packages which can be used to perform statistical analysis, and display data, not all of which can be accessed or used competently by the students. Indeed, it would be fair to say that existing software can produce extensive statistical analysis, but choosing an appropriate test and interpreting the data analysis can be challenging. It is rare to have the luxury to be able to consult a resident statistician in the Bioscience Department.

There are a variety of statistical software packages, which vary in the difficulty of use, and in what tests they can perform. An additional bonus is the ability to plot graphically, mean and individual data. The most popular software packages used currently to perform statistics and present data in graphical form are Excel (Microsoft), Prism (GraphPad) and SPSS (IBM). Microsoft Excel is a popular spreadsheet software package which is easily available, easy to use for data analysis (although types of analysis are limited), and useful to plot data graphically (limited in detail of graph). Prism is good for statistical analysis but excellent for plotting data (graphs produced are of professional standard). SPSS is the most complex, but most comprehensive statistical package. It allows a very detailed analysis of data using a wide range of tests. However, it is weak in interpreting the statistical analysis and the level of detail in plotting graphs.

A core module that most students would do is a research project. This requires them to put forward a research proposal, in which they design experiments and formulate hypotheses, collect data, analyse data, and then write a research report. From my many years of supervising undergraduates and postgraduate projects, I have observed that firstly, narrowing a project down to a specific aim and secondly, applying statistical analysis to the data obtained causes the most anxiety in students.

Having taught bioscience students for more than 25 years, I am clear that more help, guidance and resources should be made available to students in using statistics and displaying data. This book is intended for all undergraduate students at levels four (year 1), five (year 2) and six (year 3), studying the biological sciences (biomedical science, medical physiology, pharmacology, pharmaceutical science, human biology, biochemistry, microbiology, and biotechnology). Although most examples are drawn from the biological sciences, the statistical methods and tests covered in the book are applicable and useful for (i) students in other disciplines in medical and health subjects, including medicine, physiotherapy, podiatry, nursing, pharmacy, dentistry, and sports science, (ii) postgraduate research, and (iii) a quick refresher for those who are rusty on statistics and using statistics software.

The book starts from a basic level and builds in complexity, allowing readers to dip into the area they are more familiar with. It does not assume any prior knowledge of the area. The book layout is as follows:

Chapter 1

introduces common terms used in statistics

Chapter 2

shows an overview of how to display data

Chapter 3

considers statistical significance and choosing inferential tests

Chapter 4

gives background to some common parametric tests

Chapter 5

gives background to some common nonparametric tests

Chapter 6

explains how to use Microsoft Excel with examples

Chapter 7

explains how to use GraphPad Prism with examples

Chapter 8

explains how to use IBM SPSS with examples

Chapter 9

briefly considers misinterpretations/errors of statistics in analysis.

The appendices have sections on common formulas and symbols, deciding on sample size, historical milestones in statistics, background to Prism, answers to sample problems, and reference tables of critical values for statistical tests.

This book is not comprehensive in its coverage (the focus is on the most commonly used statistical tests in biomedical science) as that would have increased the size and complexity of the book. I have tried to keep the mathematical input to a minimum; however, there are areas such as analysis of variance where this was unavoidable. For those who want more depth and detail in maths and statistics, suggestions for further reading are provided. This book does not cover qualitative analysis (e.g. interviewee responses, social context, interactions with people).

This book:

Introduces statistical terms and analysis from the basics to a more advanced level.

Shows clear step by step use of three common software used in analysing data and producing graphs.

Uses examples of common statistical tests.

Does not describe areas such as enzyme and substrate reactions (e.g. Scatchard plots), or non-linear curve fitting or multiple regression.

Helps in deciding the factors to consider for study designs.

Helps in choosing appropriate tests to analyse data and to display data.

The reader should be able to answer the following questions from the use of this book.

What:

Is your study design?

Sample size is appropriate?

Are the descriptive statistics appropriate for the sample data?

Is the difference between standard deviation and standard error of the mean?

Is a confidence interval?

Is a normal distribution?

Is significance and how do you test for it?

Is the interquartile range?

Is the difference between parametric and non-parametric tests?

How do you:

Graphically describe your data?

Plot a frequency distribution plot?

Check if the sample data is normally distributed?

Decide on which statistical test to use?

Do a paired (related groups) t test?

Do an unpaired (independent groups) t test?

Do a non-parametric test (Wilcoxon)?

Do a non-parametric test (Mann–Whitney)?

Do a 1-way ANOVA test?

Do a repeated measures 1-way ANOVA?

Do a 2-way ANOVA test?

Do a correlation test?

Use Microsoft Excel software to do statistical analysis?

Use SPSS software to do statistical analysis?

Use Prism software to do statistical analysis?

I hope the readers and users of this book will (a) get a better understanding of statistical concepts and (b) be able to use the software packages with more confidence and thereby aid them in their degree studies.

Mohammed Meah BSc MSc PhD FHEA

Senior Lecturer in Physiology

Course Leader for Medical Physiology and Human Biology

University of East London,

London

“It would be so nice if something made sense for a change.”

Lewis Carroll (1832–1898), English novelist. From the book Alice in Wonderland

CHAPTER 1 Basic Statistics

“The word ‘statistic’ is derived from the Latin status, which, in the middle ages, had become to mean ‘state’ in the political sense. ‘Statistics’, therefore, originally denoted inquiries into the condition of a state.”

— Wynnard Hooper (1854–1935) - English author

Expected Learning Outcomes

Explain the common terms used in statistics.

Describe, interpret and calculate descriptive statistics.

Distinguish the differences between confidence interval, standard error and standard deviation.

Outline common study designs.

Describe the parts of a research proposal.

Data

Without data there would be no statistics!

Data is Information that can be analysed, interpreted and presented statistically.

Data can take many forms, digital data, personal data, sample data, laboratory experimental data, field data, population data.

In statistics we categorize these into numerical and non-numerical data.

What is Statistics?

The earliest use of statistics came from rulers and governments, who wanted information (data), such as the number of people, resources (e.g. food, gold, land) in order to set taxes, fund infrastructure (building projects), raise and maintain armies, and go to war (Appendix 1). To make accurate decisions, ideally you would want to collect all the information or data available about a defined group or category. This is called the population (the entire group of individuals or observations).

The modern equivalent of this is called a census or survey of the population (usually every 10 years) of a country, which collects information such as the total number of people, ethnicity, age, and gender. The data obtained is called the population data. Some examples include, those with a disease or condition (e.g. diabetes or hypertension), smoking, animals, or plants.

However, it is not practical or possible to get the population data most of the time, so we take a random sample which can be representative of the population. The sample is a defined group of individuals or observations such as smoking habits taken from an identified and specific population. The sample should be representative of the population and is chosen by setting inclusion and exclusion criteria. These criteria define the characteristics or features of the sample. For example, inclusion criteria could be healthy, males, aged 20–30; exclusion criteria participants (or subjects) not suitable for selection might be smokers, not on any medication or have a medical condition.

“By a small sample, we may judge of the whole piece.”

Miguel de Cervantes (1547–1616), from his novel “Don Quixote”; Spanish novelist, poet and playwright

Statistics can also be defined as: a branch of mathematics which involves data collection, data presentation, data analysis, and interpretation of data which comes from a population or sample of the population. In statistics, we usually take data from population samples. A population sample consists of a certain proportion/percentage of the total population determined by the researcher. The bigger and more representative the sample size, the more valid would be the results of data analysis. For example, we take height measurements (an example of data collection) of 50 male and 50 female level 5 bioscience students. This sample of 100 students would be taken from a total population of 645 level 5 bioscience students, to determine the average (mean) height of both male and female students. We would then see whether this average height is representative of the average height of all the level 5 bioscience students.

Types of Statistical Methods

In statistics, we analyse sample data to make predictions and generalisations about the population data. The sample data collected is described according to the amount of data (total data), centre of data (average, middle and most commonly occurring) and spread of data (e.g. lowest value and highest value). These are described numerically and called descriptive statistics. The statistical terms used in descriptive statistics (e.g. mean, mode, median, standard deviations) are described below.

The statistical terms used in descriptive statistics using a variety of graphs (Chapter 2), this is called exploratory statistics. Further analysis to look for differences between sample data and population data, differences between two or more groups of sample data, or looking for associations or relationships between sample groups is called inferential statistics. The types of inferential statistical tests used will be determined by data size, data distribution, data type and number of groups of data. Inferential tests can be classified into two types: parametric or tests which are based on known probability distributions of the population (Chapters 3 and 4), and non-parametric which do not follow a distribution (Chapters 3 and 5). A distribution in statistics describes the possible values and likely occurrences of these values (e.g. in tossing a coin, getting a head or tail, and the likelihood of getting heads or tails with further attempts) in experiments (Chapter 2).

Types of Data

Data can be observations or measurements. Data can be classified into quantitative (numerical – numbers) or qualitative (categorical – non-numerical).

Quantitative Data

This numerical data is split into continuous (lots of data collected in very small steps), and discrete (number of specific events occurring). Discrete data take specific number values (e.g. number of pregnancies, number of vaccinations) and give less information. Continuous data (e.g. height, weight, cholesterol, cell counts, concentration of substances in fluids, decimals, percentages, ratios) – very small steps, give more information; this type of data are used mainly in parametric tests (Chapter 4).

Qualitative Data

This non-numerical data can be split into discontinuous data, such as whole numbers, ranks, scales, gender, colours, species, classes, position in a race, and blood groups. This is also called categorical data, which can also be divided into nominal data (data which cannot be ranked in size, e.g. blood groups, gender, ethnicity) and ordinal data (data which can be ranked, e.g. position in a race, anxiety scale, pain scale, intensity of exercise). This type is predominantly used in non-parametric tests (Chapter 5).

Both quantitative and qualitative data can be expressed by the term variable. A variable is a specific factor, property, or characteristic of a population or a sample. It is the name given to the data that is collected, e.g. height, weight, gender, colour, and size.

Collection of Data

Data can be collected from observations (e.g. epidemiological study), surveys (e.g. questionnaires and interviews) or experiments (e.g. clinical trials, pilot study).

Framingham Heart Study (1948)

This was a famous medically important long term epidemiological study, initiated by the National Institute of Health (NIH). It collected a large amount of data on the epidemiology and risk factors of cardiovascular disease in 5209 adults in 1948. They identified hypertension (elevated blood pressure), high cholesterol (fat) levels, and cigarette smoking, as major risk factors for cardiovascular disease (Mahmood et al. (2014)).

Clinical Trial

A clinical trial compares the effect of one treatment with another. It involves patients, healthy people, or both in four stages (Phase 1 to Phase 4).

The rapid increase in infections and mortality caused by the coronavirus (Covid-19) worldwide in 2019 led to the need to develop effective vaccines. The two most popular vaccines in the UK are the Oxford–AstraZeneca and the Pfizer Biontech. Both of these underwent clinical trials after ethical approval. Each trial involved two groups, in which half the volunteers were given the vaccine and the other half were given a placebo. The groups were randomly selected and were matched (e.g. for age and gender). Volunteers did not know whether they were receiving the vaccine or the placebo, nor did the researchers know (double blind). Prior to using human volunteers, the vaccines were tested on animals.

The Oxford–AstraZeneca trial sampled 23,848 people across the UK, Brazil and South Africa between April and November 2020.

The Pfizer Biontech trial sampled 46,331 people from 153 sites around the world between July and December 2020.

Surveys

Ask a series of questions where the answers are subjective (based on personal opinion). These can be non-numerical or numerical by adding a number scale to the responses

e.g. (i) anxiety or pain – could be classified into none, mild, moderate, or severe

(ii) Likert Scale – agree, disagree, neither agree or disagree, strongly agree, e.g. module evaluation or describing a product you have bought

(iii) Visual Analogue Scale – an increasing number scale (e.g. 1 is low and 10 is high). An example of this is the Borg Rate of Perceived Exertion which scales the intensity of exercise.

Observations, hypotheses, theories

Before the 20th century, science was based on induction from observations from which theories were formed (e.g. Newton’s laws of motion in 1664). Karl Popper, a philosopher in 1935, suggested that theories and laws should be put first as a hypothesis, which would then be tested by experimental hypothesis falsified by observations and experiments (e.g. Darwin’s theory of evolution in 1859).

Experiments

An experiment or study gathers data from a sample (assumed to be drawn randomly from the population). Experiments test a hypothesis (or prediction) of something happening. For example, suppose you wanted to know if there was a difference between an old treatment and a new treatment for a disease condition. Two predictions which we can make about this are: (i) there will be no difference between the old treatment and the new treatment (called the null hypothesis) on the disease condition, or (ii) there will be a difference between the old and new treatments (called the alternative hypothesis) on the disease condition.

In experiments, two groups of variables need to be identified. One group is the dependent variables, which are the outcome measures of the experiment (e.g. cell size, rate of growth, and cholesterol levels). The other group is the independent variables, which are the variables being manipulated in the experiment (e.g. type of treatment (drug, surgery), type of activity (exercise, training) and type of nutrient (protein, vitamin)).

Research Proposal

The experiment could be part of a research project (a short- or long-term study), and can be done in the laboratory or field (natural environment). In starting a research project or applying for research funding, it is common to write a research proposal. The proposal describes the subject area of the research and has various parts which answer the following questions:

What is the research area of interest and justification for doing the research?

What is the aim of the research?

How will the aim be investigated?

What are the expected outcomes?

Parts of a Research Proposal

“The formulation of the problem is often more essential than its solution, which may be merely a matter of mathematical or experimental skill.”

-Albert Einstein (1879–1955) - Physicist

Background literature search

: to see what is known, what is unknown, and whether this study has been done before. This would involve doing literature searches, using key words from the title of the project and

databases

(e.g. in bioscience the most popular are PubMed, Science Direct, Google Scholar), reading primary sources (e.g. journal papers) and secondary sources (e.g. textbooks). It is very important to identify what is unknown or controversial, or where there is a gap of knowledge from previous studies, an observation, or initial investigation (pilot study). This would then be the basis of the rationale (reason) for doing the study. For example, suppose you were interested in doing a research project on exercise training and haemoglobin levels. You would do a literature search with your key words being exercise, training, and haemoglobin. Your search may produce a large number of studies, which you will then need to filter, e.g. human studies, type of exercise, duration of training, gender, year of publication. You then need to identify gaps in knowledge or controversy, i.e. levels of haemoglobin between gender, ethnicity, types, and duration of exercise training.

A clear aim

: typically, this would be a question you want to answer based on the unknown or controversy identified in the background literature search. Ideally, you would want to state a narrow aim (say in the form of a question). For our example, suppose you found in the literature a controversy about the duration of exercise affecting the levels of haemoglobin. Some studies showed there was a difference after three weeks and others showed that the difference was only found after six weeks. You could have an aim such as: Is there a difference in haemoglobin levels after three and six weeks of aerobic exercise training?

Hypothesis

: a prediction of outcomes. This is stated as two parts:

Null hypothesis: there is no difference (in the mean variables) between the control and experimental conditions. For our example, the null hypothesis would be ‘there is no difference in mean haemoglobin levels after three and six weeks of training’.

Alternative hypothesis: there is a difference (in the mean variables) between the control and experimental conditions: ‘there is a difference in mean haemoglobin levels after three and six weeks of training’.

Methodology

: methods used to investigate the aim. There are four major parts in methods. An outline of the subjects (human or animal or material), equipment/techniques used, protocol (procedures or steps of the experiment), and analysis of data (statistical tests to be used). These parts would incorporate the following:

Health and safety issues (for the researchers, participants, environment).

Ethical considerations (e.g. using humans or animals for data collection in the study/experiment).

How large should the sample size be?

How many replicates (repeat measurements)?

What are the dependent and independent variables? Independent variables are those being manipulated in experiments and dependent variables are those we measure to see the effect of the manipulation.

How do you exclude confounding variables (factors that could adversely affect an experiment or other data collection procedure)?

Control and experimental conditions (what is the baseline or reference and the interventions?)

Analysis of results (descriptive and inferential statistics).

Precautions (to ensure you collect relevant study data).

Example of Biomedical experiment methodology

Consider our example: an experiment investigating the effect of exercise training on haemoglobin levels in a group of healthy adults

Health and safety issues (for the researchers, participants, environment): taking blood, preventing infections, laboratory safety, disposal of used materials, e.g. needles, equipment safety will depend on type of equipment used (e.g. cycle ergometer, treadmill), injury during training, first aider availability.

Ethical considerations (e.g. using humans or animals for data collection in the study/experiment): ethics application to Ethics Committee for approval.