Professional Documents
Culture Documents
PDF Asking Questions in Biology A Guide To Hypothesis Testing Experimental Design and Presentation in Practical Work and Research Projects 5Th Edition Chris Barnard Ebook Full Chapter
PDF Asking Questions in Biology A Guide To Hypothesis Testing Experimental Design and Presentation in Practical Work and Research Projects 5Th Edition Chris Barnard Ebook Full Chapter
https://textbookfull.com/product/interior-design-visual-
presentation-a-guide-to-graphics-models-and-presentation-
techniques-4th-edition-mitton/
https://textbookfull.com/product/communication-research-asking-
questions-finding-answers-joann-keyton/
https://textbookfull.com/product/practical-cloud-security-a-
guide-for-secure-design-and-deployment-1st-edition-chris-dotson/
https://textbookfull.com/product/the-good-research-guide-for-
small-scale-research-projects-5th-edition-denscombe/
Introduction to Robust Estimation and Hypothesis
Testing 4th Edition Rand Wilcox
https://textbookfull.com/product/introduction-to-robust-
estimation-and-hypothesis-testing-4th-edition-rand-wilcox/
https://textbookfull.com/product/how-to-talk-to-angels-a-
practical-guide-to-asking-for-guidance-comfort-strength-first-
edition-gabriel/
https://textbookfull.com/product/philosophy-asking-questions-
seeking-answers-f3thinker/
https://textbookfull.com/product/design-a-practical-guide-to-
riba-plan-of-work-2013-stages-2-and-3-riba-stage-guide-1st-
edition-tim-bailey/
https://textbookfull.com/product/groups-a-guide-to-small-group-
work-in-healthcare-management-education-and-research-1st-edition-
glyn-elwyn/
A01_BARN5999_05_SE_FM.indd 1 10/27/16 6:45 PM
Asking Questions in Biology
Fifth Edition
Harlow, England • London • New York • Boston • San Francisco • Toronto • Sydney • Dubai • Singapore • Hong Kong
Tokyo • Seoul • Taipei • New Delhi • Cape Town • Sao Paulo • Mexico City • Madrid • Amsterdam • Munich • Paris • Milan
The rights of Chris Barnard, Francis Gilbert and Peter McGregor to be identified as authors of this work have been asserted by them in
accordance with the Copyright, Designs and Patents Act 1988.
The print publication is protected by copyright. Prior to any prohibited reproduction, storage in a retrieval system, distribution or
transmission in any form or by any means, electronic, mechanical, recording or otherwise, permission should be obtained from the
publisher or, where applicable, a licence permitting restricted copying in the United Kingdom should be obtained from the Copyright
Licensing Agency Ltd, Barnard’s Inn, 86 Fetter Lane, London EC4A 1EN.
The ePublication is protected by copyright and must not be copied, reproduced, transferred, distributed, leased, licensed or publicly
performed or used in any way except as specifically permitted in writing by the publishers, as allowed under the terms and conditions
under which it was purchased, or as strictly permitted by applicable copyright law. Any unauthorised distribution or use of this text may
be a direct infringement of the authors’ and the publisher’s rights and those responsible may be liable in law accordingly.
Pearson Education is not responsible for the content of third-party internet sites.
10 9 8 7 6 5 4 3 2 1
21 20 19 18 17
NOTE THAT ANY PAGE CROSS REFERENCES REFER TO THE PRINT EDITION
Preface vii
1 Doing science:
Where do questions come from? 1
1.1 Science as asking questions 2
1.2 Basic considerations 3
1.3 The skill of asking questions 9
1.4 Where do questions come from? 11
1.5 What this book is about 16
References 17
2 Asking questions:
The art of framing hypotheses and predictions 18
2.1 Observation 18
2.2 Exploratory analysis 23
2.3 Forming hypotheses 46
2.4 Summary 54
References 54
3 Answering questions:
What do the results say? 55
3.1 Confirmatory analysis 55
3.2 What is statistical significance? 57
3.3 Significance tests 60
3.4 Testing hypotheses 129
3.5 Testing predictions 135
3.6 Refining hypotheses and predictions 141
3.7 Summary 147
References 148
4 Presenting information:
How to communicate outcomes and conclusions 150
4.1 Presenting figures and tables 150
4.2 Presenting results in the text 163
4.3 Writing reports 164
4.4 Writing for a more general readership 179
4.5 Presenting in person: spoken papers and poster presentations 181
4.6 Plagiarism 194
4.7 Summary 195
Reference 196
Index 241
Quick test finders 249
Francis Gilbert
Peter McGregor
January 2016
Publisher’s Acknowledgements
We are grateful to the following for permission to reproduce copyright material:
Figures
Figure 2.7 and Figure 3.3 modified from Choosing and Using Statistics: A
Biologist’s Guide 3 ed., Blackwell (Dytham, C. 2010) pp. 54 and 41; Reproduced
with permission of Wiley-Blackwell in the format Educational/Instructional
Program via Copyright Clearance Center.
You’re out for a walk one autumn afternoon when you notice a squirrel picking up
acorns under some trees. Several things strike you about the squirrel’s behaviour.
For one thing it doesn’t seem to pick up all the acorns it comes across; a sizeable
proportion is ignored. Of those it does pick up, only some are eaten. Others are
carried up into a tree, where the squirrel disappears from view for a few minutes
before returning to the supply for more. Something else strikes you: the squirrel
doesn’t carry its acorns up the nearest tree but instead runs to one several metres
away. You begin to wonder why the squirrel behaves in this way. Several possibilities
occur to you. Although the acorns on the ground all look very similar to you, you
speculate that some might contain more food than others, or perhaps they are
easier to crack. By selecting these, the squirrel might obtain food more quickly
than by taking indiscriminately any acorn it encountered. Similarly, the fact that
it appears to carry acorns into a particular tree suggests this tree might provide a
more secure site for storing them.
While all these might be purely casual reflections, they are revealing of the way
we analyse and interpret events around us. The speculations about the squirrel’s
behaviour may seem clutched out of the air on a whim, but they are in fact
structured around some clearly identifiable assumptions, for instance that achieving
a high rate of food intake matters in some way to the squirrel and influences its
preferences, and that using the most secure storage site is more important to it than
using the most convenient site. If you wanted to pursue your curiosity further, these
assumptions would be critical to the questions you asked and the investigations
you undertook. If all this sounds very familiar to you as a science student, it
should, because, whether you intended it or not, your speculations are essentially
scientific. Science is simply formalised speculation backed up (or otherwise) by
equally formalised observation and experimentation. In its broadest sense most of
us ‘do science’ all the time.
Searchable online literature databases, like the However, the process inevitably involves some
Web of Knowledge, Google Scholar, BIOSIS compromises.
or PubMed, allow you to search for articles For example, suppose you were interested
by particular authors, or on particular topics, in steroid hormone secretion as a cause of
or according to some other category, such as immune depression in laboratory mice. You
a journal title or research organisation. An might start, seemingly reasonably, by typing
example of the kind of search fields on offer, ‘steroid hormone AND immune depression
in this case for the Web of Knowledge, is shown AND laboratory mice’ into the ‘Topic’ search
in Fig. (i). field in Fig. (i) and hitting the ‘Search’ button.
The key to using the search fields effectively Disappointingly, and rather to your surprise,
lies in the precision with which you specify this yields nothing at all – apparently nobody
your terms: too general and you will be has published anything on steroid hormones
swamped with articles that are of little or no and immune depression in mice. At the
interest; too narrow and you will wind up with other extreme, a search for ‘immune AND
only one or two and miss many important mice’ yields over 45,000 articles, a wholly
ones. To help with this, the search fields unmanageable number, of which many can
provide various means of linking terms so be seen at a glance to be irrelevant to your
that searches can be focused (the AND, OR, needs. Clearly, something between the two is
NOT options – called ‘operators’ – in Fig. (i)). what is required.
Figure (i) A screen capture from the Web of Knowledge as it appeared in 2016. Like other similar sites, it is
regularly updated, so the exact appearance of the search field screen may change.
The reason the first search turned up phrase are recognised, in this case, say, ‘immune
nothing is not, of course, because nobody depression’, ‘depressed immunity’ or ‘depressed
has published anything on the topic, but immune response’. Running the search again
because the search term was restricted to in this form yields around 450 articles, much
a very specific combination of phrases. It better than zero or 45,000, but with quite
could well be that people have published on a lot of them still redundant. If the search
the effects of steroid hormones on immune is specified a little more tightly as ‘immun*
depression in mice but didn’t use the precise SAME depress* AND mice AND hormone’,
phrases selected. For instance, they may have however, it turns up around 40 articles, and all
reported ‘depressed immune responsiveness’ much more on target.
or ‘depressed immunity’, rather than ‘immune All the searchable databases use these kinds
depression’, and referred to specific hormones, of approaches for refining searches, some
such as testosterone or cortisol, rather than the very intuitive, some less so. One thing you
generic term ‘steroid’. There are various ways will quickly notice, though, is that exactly the
of catering for this. In the Web of Knowledge, same search can turn up a different number
the form ‘immun* SAME depress* AND mice’ and selection of articles depending on which
in the ‘Topic’ search field allows the system to database you are using – BIOSIS, for example,
search for any term beginning with ‘immun’ manages to find something under the initial
or ‘depress’, such as ‘immune’, ‘immunity’, over-specific search that drew a blank on
‘immunocompetence’, ‘depression’, ‘depressed’ the Web of Knowledge. For this reason, it is
and so on, thus picking up all the variants. The good practice to run searches on a selection
term ‘SAME’ ensures similar combinations of of databases.
Google, can often turn up information from the professional literature too, but
just as often you’re likely to get information from unregulated personal websites,
or other sources of uncertain provenance. Taking received wisdom at face value
can be a dangerous business – something even seasoned researchers can continue
to discover, the famous geneticist and biostatistician R®. A. Fisher among them.
In the early 1960s, Fisher and other leading authorities at the time were
greatly impressed by an apparent relationship between duodenal ulcer and certain
rhesus and MN blood groups. Much intellectual energy was expended trying to
explain the relationship. A sceptic, however, mentioned the debate to one of his
blood-group technicians. The technician, for years at the sharp end of blood-
group analysis, resolved the issue on the spot. The relationship was an artefact
of blood transfusion! Patients with ulcers had received transfusions because of
haemorrhage. As a result, they had temporarily picked up rhesus and MN antigens
from their donors. When patients who had not been given transfusions were tested,
the relationship disappeared (Clarke, 1990).
Where at all feasible, therefore, testing assumptions yourself and making up
your own mind about the facts available to you is a good idea. Indeed, science
is often characterised as systematic scepticism – a demand for evidence for every
assertion. It is impossible to draw up a definitive list of what it is an investigator
needs to know as essential background; biology is too diverse a subject, and every
investigation is to some extent unique in its factual requirements. Nevertheless, it
is useful to indicate the kinds of information that are likely to be important. Some
examples might be as follows:
Question Can the material of interest be studied usefully under laboratory conditions or will
unavoidable constraints or manipulations so affect it that any conclusions will have
only dubious relevance to its normal state or functions?
For instance, can mating preferences in guppies usefully be studied in a small plastic
aquarium, or will the inevitable restriction on movement and the impoverished
environment compromise normal courtship activity?
Or, if nutrient transfer within a plant can be monitored only with the aid of
a vital dye, will normal function be maintained in the dyed state or will the dye
interfere subtly with the processes of interest?
Question Is the material at the appropriate stage of life history or development for the desired
investigation?
There would, for instance, be little point in carrying out vaginal smears on female
mice to establish stages of the oestrous cycle if some females were less than 28 days
of age. Such mice may well not have begun cycling.
Likewise, it would be fruitless to monitor the faeces of infected mice for the eggs
of a nematode worm until a sufficient number of days have passed after infection
for the worms to have matured.
Question Will the act of recording from the material affect its performance?
Question Does the investigation make demands on the material that it is not capable of meeting?
By the same token, assumptions about the relationship between the staining
characteristics of cells in histological sections and their physiological properties
might need verifying before concluding anything about the distribution of
physiological processes within an organ.
The list could go on for a long time, but these examples are basic questions of
practicality. They are not very interesting in themselves but they, and others like
them, need to be addressed before interesting questions can be asked. Failure to
consider them will almost inevitably result in wasted time and materials.
Of course, even at this level, the investigator will usually have the questions
ultimately to be addressed – the whole point of the investigation – in mind, and
these will naturally influence initial considerations. Before we develop this further,
however, there is one further, and increasingly prominent, issue we must address,
and that is the ethics of working with biological material.
A hypothesis, then, can be little more than an intuitive feeling about how
s omething works, or how changes in one factor will relate to changes in another,
or about any aspect of the material of interest. However vague it may be, though,
it is formative in the purpose and design of investigations because these set out
to test it. If at the end of the day the results of the investigation are at odds with
the hypothesis, the hypothesis may be rejected and a new one put in its place. As
we shall see later, hypotheses are never proven, merely rejected if data from tests so
dictate, or retained for the time being for possible rejection after further tests.
Hypotheses, too, can be so broad or imprecise that they are difficult to reject.
In general the more specific, mutually exclusive hypotheses that can be formulated
to account for an observation the better. In our chaffinch example, the first
hypothesis was that the display was sexual. Another might be that it reflected
aggressive defence of food. Yet another that it was an anti-predator display. These
three hypotheses give rise to very different predictions about the behaviour and it
is thus, in principle, easy to distinguish between them. As we have already seen,
however, distinguishing between the ‘sexual’ and ‘aggressive’ hypotheses may need
more careful consideration than we first expect. Straw man hypotheses are another
common problem. Unless some effort has gone into understanding the material,
there is a risk of setting up hypotheses that are completely inappropriate. Thus,
suggesting that our displaying chaffinch was demonstrating its freedom from avian
malaria would make little sense in an area where malaria was not endemic. We shall
look at the development of hypotheses and their predictions in more detail later on.
As we have already intimated, questions do not spring out of a vacuum. They are
triggered by something. They may arise from a number of sources.
1.4.1 Curiosity
Questions arise naturally when thinking about any kind of problem. Simple
curiosity about how something works or why one group of organisms differs in
some way from another group can give rise to useful questions from which testable
hypotheses and their predictions can be derived. There is nothing wrong with
‘armchair theorising’ and ‘thought experiments’ as long as, where possible, they
are put to the test. Sitting in the bath and wondering about how migratory birds
manage to navigate, for example, could suggest roles for various environmental cues
like the sun, stars and topographical features. This in turn could lead to hypotheses
about how they are used and predictions about the effects of removing or altering
them. By the time the water was cold, some useful preliminary experiments might
even have been devised.
that the paper is a review, then it should be a good source of general information,
but it is unlikely to contain detailed information on experimental details.
2. Do the results bear out the title and detail in the Abstract?
Skim-read the Results, particularly looking at the figures or tables because only
the main results are shown in these formats; look for p-values and details of the
statistical tests that produced them.
In our example paper, the three figures show the key results (large males
more likely to win, and to initiate fights; escalated encounters more likely when
opponents are of similar size) and the figure captions include p-values and details
of the statistical tests that produced them.
If the paper has passed the two assessments of quality/relevance in Step 3 (as
our example paper does), the next step depends on whether you are looking for
general or detailed information.
We’ve said a little about how science works and how the kind of question-asking on
which it is based can arise. We now need to look at each part of the process in detail,
because while each may seem straightforward in principle, some knotty problems
can arise when science is put into practice. In what follows, we shall see how to:
• frame hypotheses and predictions from preliminary source material;
• design experiments and observations to test predictions;
• analyse the results of tests to see whether they are consistent with our original
hypothesis; and
• present the results and conclusions of tests so that they are clear and informative.
The discussion deals with these aspects in order so that the book can be read
straight through or dipped into for particular points. A summary at the end of each
chapter highlights the important take-home messages, and the self-test questions at
the end show what you should be able to tackle after reading the book.
Remember, the book is about asking and answering questions in biology – it is
not a biology textbook or a statistics manual, and none of the points it makes are
restricted to the examples that illustrate them. At every stage you should be asking
yourself how what it says might apply in other biological contexts, especially if you
have an interest in investigating them!
We suggest you get hold of a statistical package in order to follow the analyses.
The one we use in detail in the book is R®. You can download the installation file
of R® free from www.r-project.org, and follow the simple instructions to install it
on your computer. This book will enable you to run analyses in R®; a simple and
very clear guide to learning more about R® is available in Zuur et al. (2009).
Leg-coverings are now being introduced, and the last princess of the
blood royal I saw added to her comfort, though she destroyed the
poetry of her appearance, by a tightly-fitting pair of black cloth “pants”
with a gold stripe! This garment will doubtless soon become general.
In ancient days the Persian ladies always wore them, as may be
seen by the pictures in the South Kensington Museum. In those times
the two embroidered legs, now so fashionable as Persian embroideries
(“naksh”), occupied a girl from childhood to marriage in their making;
they are all sewing in elaborate patterns of great beauty, worked on
muslin, in silk.
The outdoor costume of the Persian women is quite another thing;
enveloped in a huge blue sheet, with a yard of linen as a veil,
perforated for two inches square with minute holes, the feet thrust into
two huge bags of coloured stuff, a wife is perfectly unrecognisable,
even by her husband, when out of doors. The dress of all is the same;
save in quality or costliness, the effect is similar. And yet with such a
hideous disguise, a Persian coquette will manage to let the curious
know if she have a good face and eye, by lifting her veil in a sly and
half-timid way. The only thing I know exceeding in folly the chimney-pot
hat, is the outdoor dress of the Persian woman. Expensive, ugly,
uncomfortable, hot in summer, cold in winter, words fail to express its
numerous disadvantages; it has one positive quality—as a disguise it is
perfect, and its use favours the intrigues rife in the country.
As for the children, they are always when infants swaddled: when
they can walk they are dressed as little men and women, and with the
dress they often, nay generally, ape the manners; a Persian child of the
upper class being a master of etiquette, an adept at flattery, and a
mirror of politeness. It is a strange custom with the Persian ladies to
dress little girls as boys, and little boys as girls, till they reach seven or
eight years; this is often done for fun, or on account of some vow,
oftener to avert the evil eye.
Persian women are very fond of their children, and pet them greatly.
The love of the Persian for his mother is very great; he never leaves her
to starve, and her wishes are laws to him, even when he is an old man,
and she an aged crone. The mother is always the most important
member of the household, and the grandmother is treated with
veneration. Mothers-in-law are not laughed at or looked down on in
Persia; their presence is coveted by their sons-in-law, who look on them
as the guardians of the virtue of their wives. The uncle, too, is a much
nearer tie than with us, that is to say, the paternal uncle: while men look
on their first cousins on the father’s side as their most natural wives.
Possibly this is because their female cousins are the only women they
have any opportunity of knowing anything of personally. Black slaves
and men-nurses, or “lallahs,” are much respected and generally
retained in a household, while the “dyah,” or wet nurse, is looked on as
a second mother, and usually provided for for life.
Persians are very kind to their servants, and try to make their people
look on them as second fathers; a master will be often addressed by a
servant as his father, and the servant will protect his master’s property
as he would his own, or even more jealously.
A servant is invariably spoken to as “butcha” (child). The servants
expect that their master will always take their part, and never allow
them to be wronged; if he does not do so, he cannot obtain a good
class domestic, while if he sticks to the man, he never leaves him.
The slaves in Persia have what Americans call “a good time;” well
fed, well clothed, treated as spoiled children, given the lightest work,
and often given in marriage to a favourite son, or taken as a “segah,” or
concubine, by the master himself (and respectable Persians only take a
“segah” for ninety-nine years, which is equivalent to a permanent
marriage), slaves have the certainty of comfort and a well-cared-for old
age. They are always looked on as confidential servants, are entrusted
with large sums of money, and the conduct of the most important
affairs; and seldom abuse their trust.
The greatest punishment to an untrustworthy slave is to give him his
liberty and let him earn his living. They vary in colour and value: the
“Habashi” or Abyssinian is the most valued; the Souhāli or Somāli, next
in blackness, is next in price; the Bombassi, or coal-black negro of the
interior, being of much less price, and usually only used as a cook. The
prices of slaves in Shiraz are, a good Habashi girl of twelve to fourteen,
forty pounds; a good Somāli same age, half as much; while a Bombassi
is to be got for fourteen pounds, being chosen merely for physical
strength. They are never sold, save on importation, though at times
they are given away. Strange as it may appear, to the mind of any one
who has lived in Persia, slavery in that country to the African is an
unmixed good. Of course the getting to Persia, and the being caught, is
another thing. But I have never seen a Persian unkind to his own horse