VEER NARMAD SOUTH GUUJARAT UNIVERSITY, SURAT

Department of Biosciences
Ph. D. Course Work 2022-23
Assignment Submission
Ph. D Registration number: 1785
Name: Patel ShitalKumari Arvindbhai
Subject: CW 101: Advances in Research Methodology
Topic Name: Unit: 1 (1.2) Setting up Experimental Design
and Experimental Protocol
Date: 25/12/2023
Signature: _________
Setting up Experimental Design and Experimental Protocol

INTRODUCTION
Biological research includes not only observation of natural events and
their description, but also giving explanation for those events. It is also
concerned with finding solutions to problems of the natural world. In these
endeavours’ biologists adopt scientific methods. Generally, an
investigation may be a case study, a cross-sectional study or a longitudinal
study. A case study is observation and description of one or a few
occurrences of an event. The event may occur anywhere in the world,
among humans, other animals and plants, in the atmosphere, freshwater
bodies, oceans and so on. For example, the event may be a human subject
with an abnormality, which may be physical, physiological or
psychological. The natural corollary of a case study is the proposition of a
hypothesis to explain the observed event. A cross - sectional study is the
study of a defined population by observing samples collected from that
population. In such studies, the variables are not manipulated. A possible
correlation, association, etc. among the variables are investigated. Such
studies, however, do not provide any evidence to suggest that a particular
variable is the cause of a given event. A longitudinal study is an
experimental investigation. We manipulate the variables and factors in
experimental units and observe the effects. Such a study helps us to provide
tangible evidence to demonstrate the causative agent of a particular event.
Let us consider an example to understand the above basics of a
research study. Suppose a physician observes a few persons with oral
cancer and also notes that all these persons have tobacco - chewing habit.
He would report the description of the cases of occurrence of oral cancer
in these persons, and even might hypothesize that tobacco - chewing is the
causative agent. However, there is no proof that this habit actually caused
oral cancer. There may be other reports of occurrence of oral cancer in
persons who do not have this habit. This report is simply a case study. The
next step is cross - sectional study of a population of persons who have the
tobacco - chewing habit. Samples from such populations may be screened
for the occurrence of oral cancer, symptoms of oral cancer, several
biochemical parameters, etc. Simultaneously samples from a " control "
population of those who do not have this habit, may be screened for the
same factors. A comparison of the incidence of oral cancer and other
factors would certainly help to ascertain the association among the factors
studied. Establishment of correlation among various variables is also
possible.
However, such a cross - sectional study does not provide evidence the
tobacco - chewing causes oral cancer. The results simply add strength t the
hypothesis. Only a longitudinal study involving well - designed
experiments would provide substantial evidence to demonstrate that
tobacco - chewing is the cause of oral cancer. It may not be possible to
conduct the experiment c human subjects. But, animal models such as
monkeys, which can be induced into tobacco - chewing habit, can be used
(if permitted by CPCSEA).
MEANING OF RESEARCH DESIGN
The formidable problem that follows the task of defining the research
problem is the preparation of the design of the research project, popularly
known as the “research design”. Decisions regarding what, where, when,
how much, by what means concerning an inquiry or a research study
constitute a research design. “A research design is the arrangement of
conditions for collection and analysis of data in a manner that aims to
combine relevance to the research purpose with economy in procedure.”
In fact, the research design is the conceptual structure within which
research is conducted; it constitutes the blueprint for the collection,
measurement and analysis of data. As such the design includes an outline
of what the researcher will do from writing the hypothesis and its
operational implications to the final analysis of data. More explicitly, the
design decisions happen to be in respect of:
(i) What is the study about?
(ii) Why is the study being made?
(iii) Where will the study be carried out?
(iv) What type of data is required?
(v) Where can the required data be found?
(vi) What periods of time will the study include?
(vii) What will be the sample design?
(viii) What techniques of data collection will be used?
(ix) How will the data be analysed?
(x) In what style will the report be prepared?
Keeping in view the above stated design decisions, one may split the
overall research design into the following parts:
(a) the sampling design which deals with the method of selecting items to
be observed for the given study;
(b) the observational design which relates to the conditions under which
the observations are to be made;
(c) the statistical design which concerns with the question of how many
items are to be observed and how the information and data gathered are to
be analysed; and
(d) the operational design which deals with the techniques by which the
procedures specified in the sampling, statistical and observational designs
can be carried out.
From what has been stated above, we can state the important features of a
research design as under:
(i) It is a plan that specifies the sources and types of information relevant
to the research problem.
(ii) It is a strategy specifying which approach will be used for gathering
and analysing the data.
(iii) It also includes the time and cost budgets since most studies are done
under these two constraints.
In brief, research design must, at least, contain—(a) a clear statement of
the research problem; (b) procedures and techniques to be used for
gathering information; (c) the population to be studied; and (d) methods to
be used in processing and analysing data.

NEED FOR RESEARCH DESIGN

Research design is needed because it facilitates the smooth sailing of the
various research operations, thereby making research as efficient as
possible yielding maximal information with minimal expenditure of effort,
time and money. Just as for better, economical and attractive construction
of a house, we need a blueprint (or what is commonly called the map of the
house) well thought out and prepared by an expert architect, similarly we
need a research design or a plan in advance of data collection and analysis
for our research project. Research design stands for advance planning of
the methods to be adopted for collecting the relevant data and the
techniques to be used in their analysis, keeping in view the objective of the
research and the availability of staff, time and money. Preparation of the
research design should be done with great care as any error in it may upset
the entire project. Research design, in fact, has a great bearing on the
reliability of the results arrived at and as such constitutes the firm
foundation of the entire edifice of the research work. Even then the need
for a well thought out research design is at times not realised by many. The
importance which this problem deserves is not given to it. As a result many
researches do not serve the purpose for which they are undertaken. In fact,
they may even give misleading conclusions. Thoughtlessness in designing
the research project may result in rendering the research exercise. It is,
therefore, imperative that an efficient and appropriate design must be
prepared before starting research operations. The design helps the
researcher to organize his ideas in a form whereby it will be possible for
him to look for flaws and inadequacies. Such a design can even be given
to others for their comments and critical evaluation. In the absence of such
a course of action, it will be difficult for the critic to provide a
comprehensive review of the proposed study.
FEATURES OF A GOOD DESIGN
A good design is often characterised by adjectives like flexible,
appropriate, efficient, economical and so on. Generally, the design which
minimises bias and maximises the reliability of the data collected and
analysed is considered a good design. The design which gives the smallest
experimental error is supposed to be the best design in many investigations.
Similarly, a design which yields maximal information and provides an
opportunity for considering many different aspects of a problem is
considered most appropriate and efficient design in respect of many
research problems. Thus, the question of good design is related to the
purpose or objective of the research problem and also with the nature of
the problem to be studied. A design may be quite suitable in one case, but
may be found wanting in one respect or the other in the context of some
other research problem. One single design cannot serve the purpose of all
types of research problems.
A research design appropriate for a particular research problem, usually
involves the consideration of the following factors:
(i) the means of obtaining information;
(ii) the availability and skills of the researcher and his staff, if any;
(iii) the objective of the problem to be studied;
(iv) the nature of the problem to be studied; and
(v) the availability of time and money for the research work.
If the research study happens to be an exploratory or a formulative one,
wherein the major emphasis is on discovery of ideas and insights, the
research design most appropriate must be flexible enough to permit the
consideration of many different aspects of a phenomenon. But when the
purpose of a study is accurate description of a situation or of an association
between variables (or in what are called the descriptive studies), accuracy
becomes a major consideration and a research design which minimises bias
and maximises the reliability of the evidence collected is considered a good
design. Studies involving the testing of a hypothesis of a causal relationship
between variables require a design which will permit inferences about
causality in addition to the minimisation of bias and maximisation of
reliability. But in practice it is the most difficult task to put a particular
study in a particular group, for a given research may have in it elements of
two or more of the functions of different studies. It is only on the basis of
its primary function that a study can be categorised either as an exploratory
or descriptive or hypothesis-testing study and accordingly the choice of a
research design may be made in case of a particular study. Besides, the
availability of time, money, skills of the research staff and the means of
obtaining the information must be given due weightage while working out
the relevant details of the research design such as experimental design,
survey design, sample design and the like.
IMPORTANT CONCEPTS RELATING TO RESEARCH DESIGN
Before describing the different research designs, it will be appropriate to
explain the various concepts relating to designs so that these may be better
and easily understood.
1. Dependent and independent variables: A concept which can take on
different quantitative values are called a variable. As such the concepts like
weight, height, income are all examples of variables. Qualitative
phenomena (or the attributes) are also quantified on the basis of the
presence or absence of the concerning attribute(s).
2. Extraneous variable: Independent variables that are not related to the
purpose of the study, but may affect the dependent variable are termed as
extraneous variables. Suppose the researcher wants to test the hypothesis
that there is a relationship between children’s gains in social studies
achievement and their self-concepts. In this case self-concept is an
independent variable and social studies achievement is a dependent
variable. Intelligence may as well affect the social studies achievement, but
since it is not related to the purpose of the study undertaken by the
researcher, it will be termed as an extraneous variable. Whatever effect is
noticed on dependent variable as a result of extraneous variable(s) is
technically described as an ‘experimental error’. A study must always be
so designed that the effect upon the dependent variable is attributed entirely
to the independent variable(s), and not to some extraneous variable or
variables.
3. Control: One important characteristic of a good research design is to
minimise the influence or effect of extraneous variable(s). The technical
term ‘control’ is used when we design the study minimising the effects of
extraneous independent variables. In experimental researches, the term
‘control’ is used to refer to restrain experimental conditions.
4. Confounded relationship: When the dependent variable is not free from
the influence of
extraneous variable(s), the relationship between the dependent and
independent variables is said to be confounded by an extraneous
variable(s).
5. Research hypothesis: When a prediction or a hypothesised relationship
is to be tested by scientific methods, it is termed as research hypothesis.
The research hypothesis is a predictive statement that relates an
independent variable to a dependent variable. Usually a research
hypothesis must contain, at least, one independent and one dependent
variable. Predictive statements which are not to be objectively verified or
the relationships that are assumed but not to be tested, are not termed
research hypotheses.
6. Experimental and non-experimental hypothesis-testing research: When
the purpose of
research is to test a research hypothesis. it is termed as hypothesis-testing
research. It can be of the experimental design or of the non-experimental
design. Research in which the independent variable is manipulated is
termed ‘experimental hypothesis-testing research’ and a research in which
an independent variable is not manipulated is called ‘non-experimental
hypothesis-testing research’. For instance, suppose a researcher wants to
study whether intelligence affects reading ability for a group.

OBSERVATION
The basic requirement for any biological research is observation.
Observation of what has happened and is happening in nature leads to a
attempt to explain the event, i.e., answering questions such as what, how
and, if possible, why. Very often, the explanation is in the form of
hypothesis.
Examples 1. Let us consider an example. We observe large - scale mortality
of fish in a lake. Naturally, we are concerned about the happening of this
event and would like to have an explanation for it. As biologists, we would
s the quality of the water, especially for possible pollution. We analyse
samples of water from the lake for all its physio - chemical properties and
finally find out that the water in the lake has unusually very high level of
cadmium a heavy metal. We suspect that the high level of the cadmium is
the cause of death of fish. We hypothesize that cadmium at high levels is
toxic fish. This is only one of the several possible hypotheses. Several other
factors such as bacterial, viral, or any other parasitic infection of
combination of factors might have caused the mortality.
 HYPOTHESIS AND NULL–HYPOTHESIS
A hypothesis is put forth as a solution to a problem or as an explanation
for an observed event. A hypothesis is a statement making a prediction that
an event will occur under stated, specific conditions.
Examples:
1. Prolonged exposure of men to lowered oxygen pressure will cause
increased haemoglobin level in them.
This hypothesis might have arisen as a result of the observation of higher
haemoglobin levels in men who are living at high altitudes for several
years. It is common knowledge that with increase in the altitude from the
mean sea level there is decrease in the atmospheric oxygen pressure.
2. Spirulina promotes growth in fish.
3. Cadmium Inhibits growth in plants.
4. Ginger has antimicrobial activity.
5. Lantana camera has cytotoxic activity.
6. Eucalyptus oil has antibacterial activity.
A hypothesis should be one that can be tested. It is very difficult to " prove
" a hypothesis beyond any doubt. All that we can do is to conduct well-
designed, controlled experiments to collect evidence in the form of valid data
to support our hypothesis. The data that we collect from the experiments are
subjected to statistical analysis, which would test a null form of our
hypothesis, the null-hypothesis. The result of statistical test of significance
calculates the probability for occurrence of the null hypothesis. If the
probability is very low, i. e., lower than the Level of Significance, we reject
the null - hypothesis. Rejection of null - hypothesis leads to the suggestion that
we have collected evidence to support our hypothesis, that is, in other words,
we cannot reject our hypothesis. But we should not forget that the " evidence
" we have obtained from an experiment using a few samples may not be
sufficient to " prove " our hypothesis beyond doubt. On the other hand, if the
statistical test gives us a probability of the occurrence of the null - hypothesis
equal to or greater than the Level of Significance, then we cannot reject the
null - hypothesis. Failure to reject null - hypothesis leads to the suggestion that
the evidence we collected from our experiment is not sufficient to support our
hypothesis. In other words, we have to reject our hypothesis, at least for the
time being, until we are able to provide sufficient evidence. A null - hypothesis
is generally a hypothesis of no difference, hence the name null - hypothesis.
When we say a hypothesis or a null - hypothesis should be testable that does
not mean those that cannot be tested are totally false. Untestable hypotheses
may be true or false. Similarly testable hypotheses may be true or false. What
we, experimental biologists, are concerned is only about testable hypotheses.
BASIC PRINCIPLES OF EXPERIMENTS
While observing what has happened or is happening in nature is the basis for
any hypothesis, experiment is the basic scientific method of testing the
hypothesis that tries to explain an event. In a way it is also observation,
observation of events (results) that occur under controlled conditions. The
purpose of such observations is to demonstrate or " prove " the functions, the
relationships, the causes, the effects, etc. of one or more factors that exist in
nature. Generally, in any biological experiment, either in the field or in the
laboratory, we try to manipulate (i.e., change by removing or adding or
altering) one or more factors in a sample obtained from a defined population,
and observe the resulting changes in specific parameters. On the basis of these
observations we infer about the population. If our inferences based on the
results of the experiment are to be valid, we must lay out all aspects of the
experiments before the commencement of the experiment, i.e., we must design
our experiment to the minutest detail. The designing varies depending on the
specific field of biology, which may be ecology, toxicology, ethology,
genetics, immunology, molecular biology and so on. It will not be possible to
go into the details of designing experiments in each field, and therefore we
shall discuss only the general aspects, which apply to all or most of the
designs.
The following are the three important basic steps we have to take at the
beginning of any biological experiment.
1. Aim We must define and record precisely the objective of the experiment
that we are proposing to carry out.
2. Plan We must write down in detail the strategy we are going to adopt in the
conduct of the experiment. The planning includes definition of the population,
sampling procedure, sample size, determination of dosage, mode of treatment,
control, randomization, methodology for the evaluation of the parameters,
collection and analysis of the data, etc.
3. Procedure We must clearly state in writing the experimental procedure, i.e.,
how we are going to perform the experiment in practice. For example, how to
collect the sample of fish, water, air, etc. should be specified. How to transport
the sample from the field to the laboratory? How to sacrifice the experimental
animals? All physical activities, operational details and requirements
(glassware, apparatus, instruments, chemicals, etc.) expected during an
experiment should be thought of in advance. Even the time, the specific time
to start and end an experiment, the time for lunch break, etc. should be decided.
Before we venture into the different experimental designs it would be useful
if we understand the meaning and usage of certain common terminology we
come across in any experimental design.
Experimental Unit and Sampling Unit
We must clearly understand the distinction between experimental unit and the
sampling unit. The experimental unit refers to the material used, such as a
laboratory animal, cage, culture plate, agriculture field, a plot in the garden, a
pot in a greenhouse, etc. A treatment is applied to an experimental unit. A
sampling unit may be the experimental unit itself or it may be a fraction of it.
For instance, in order to test the toxic effects of a heavy metal on fish, we may
have two aquaria one treated with a heavy metal and the other serving as
control. Each aquarium may have a number of fish, say, fifty. At specified
time intervals we may sample fish from each tank for various analyses such
as haematological, histological and biochemical. In this experimental set - up,
the experimental unit is the aquarium and not the fish in it. The fish are the
sampling units, a part of the experimental unit.
Experimental Error
Another important aspect we have to bear in mind while designing
experiments is the experimental error or residual variance. Experimental error
refers to the variations among observations made on the experimental units,
which were treated alike. For example, we may inject a specific dosage of
hormone into each of a number of mice. Though the treatment is "similar " to
all the experimental units, the response to the treatment (e.g. level of glucose
in the serum) may not be the same. The variation among the observations is
the experimental error.
There are two important sources of experimental errors. They are (a) the
natural variability among the experimental materials and (b) the lack of
consistency in or during the conduct of the experiment. The natural or inherent
variability in experimental materials may be due to genetic background, age,
sex, health, physiology, immunology, parasitaemia, etc. Absence of
uniformity in the actual conduct of the experiment may be due to errors in
recording, instrumentation, chemicals and reagents used. etc. When we
conduct experiments with living things, we should realize undergo
developmental changes that the experimental units continue to and such
changes may also cause errors.
Discrimination
Any experiment that we design should be able to provide an answer to a
specific hypothesis. It should not yield results that can explain more than one
hypothesis. In other words, experiments should be capable of discriminating
between different hypotheses.
Replication
As we are familiar, no two living things, even those that are genetically
patterns in particular tissues, even within supposedly genetically identical, are
similar. There is a large degree of variability in gene expression individuals.
In addition, differences can be expected as a consequence of time of day, age,
and physiological state such as sex and stage of the estrous cycle, and as a
response to disease.
Biological variables are highly unpredictable. Even those variables that are
considered to be constant are only regulated by some homeostatic mechanism,
to remain within a wide but acceptable range of values. Therefore, in any
experiment an inference drawn on the basis of data collected from a single
observation will be invalid. For example, to test a conduct an experiment using
one human subject. We can prepare an hypothesis that bitter - gourd lowers
blood sugar level in man, we may aqueous extract of known quantity of bitter
- gourd and administer orally the extract to a diabetic patient after measuring
the initial blood sugar level After a lapse of half-an- hour the blood sugar level
of the patient may show a lower value . Can we infer that bitter-gourd has the
potency to lower blood sugar level? Certainly not! Apart from so many other
pertinent questions that can be raised against the above " experiment ", the one
that asks " is the difference between the initial and final blood sugar levels
statistically significant " can never be answered. A sample of n = 1 is not a
sample at all. Suppose, we had a second diabetic volunteer to whom also the
bitter - gourd extract was administered. The blood sugar level at the end of the
experiment might be slightly higher than the initial level, or the two levels,
initial and final, might be the same.
Thus the response to a treatment may vary in different individual units of a
sample. This natural variance can be accounted for only if the treatment is
replicated, i.e., the same treatment is administered to different sampling units.
How many times a treatment should be replicated, i.e., what should be the size
of the sample? The answer to this question depends on the expected variance
of the data, before and after treatment. It also depends on the accuracy we
desire to have in our final prediction. Higher the accuracy, i.e., narrower range
of prediction, larger is the size of the sample. If the number of replicates is
very small, our inferences will be inconclusive and invalid. On the other hand,
too many replicates may yield diminishing returns.
Generalization
The aim of a biological investigation is to make an inference about the
population from the sample, i.e., estimation of population parameters using
sample statistics. In other words, we attempt to generalize what we observed
in a sample, very often, a small sample. We must be guarded while making
any generalization. What we observe in a few fish from laboratory stock may
not apply in toto to wild fish in a lake. Likewise, we must be very cautious
while using observations made on laboratory - bred guinea pigs, for example,
for extrapolation to other mammals, especially to humans.
Controls
All biological experiments should have appropriate controls. When a sample
unit receives a " treatment " by one of several methods, it is possible that the
effect observed might be due to factors other than the one that the treatment
contained. Therefore, it is obligatory on our part to demonstrate that the effect
is in fact due to the treatment factor and not due to any others including the
method of treatment itself (e.g. handling of the experimental animals, medium
in which the treatment factor is prepared etc.). This we can do by
incorporating a proper control in our experiments.
Suppose, for example, pituitary gland on the experimental fish are Suppose an
experiment is designed to find the efficacy of a plant extract which we suspect
has the potency to reduce blood - glucose level in rabbits The experimental
condition would be to administer rabbits with this extract and to find whether
the treatment had an effect on the blood glucose level of the experimental
rabbits . In order to ascertain that the effect observed in the experimental
rabbits is in fact due to the extract and not due to other factors such as the
medium in which the extract was prepared, the technique of administering the
extract, etc., we must have a control. Rabbits that are similar ( genetically and
physiologically ) to the experimental rabbits, are administered with placebo,
which has everything except the plant extract ( e.g. if the extract had been
prepared in physiological saline, the placebo would be the physiological saline
minus the plant extract it would have undergone all the steps of the extraction
process including adjustment of pH, temperature, etc. ) . Such a control is often
referred to as negative control, because of the removal of basic factor that is
being tested Comparison of the effects of the two treatments, experimental and
control would help us judge the efficacy of the plant extract in lowering b
glucose levels in rabbits. A positive control is one in which the factor being
tested is added in the placebo. Suppose, for example, we wish to know the
effect of removal of pituitary gland on the blood - glucose level of a species
of fish. The experimental fish are those that were hypophysectomized. We can
have two types of controls for such an experiment. In one control, we can have
fish that were sham operated, to eliminate the effect of the operation procedure
itself. Another control is hypophysectomized fish injected with extract of
pituitary or implanted with pituitary. Then second category is a positive
control in the sense that it has something added that is missing in the
experimental animal.
The negative and positive controls, also referred to as baseline controls, are
basically opposite of experimental units, with reference to only one specific
condition while all the other conditions are the same. Normal, healthy,
untreated animals, plants, etc. may also serve as baseline controls. Another
type of control, referred to as known standard control, is used to control the
possible experimental errors. Suppose, we are testing in the pancreas of fish.
We might be using antiserum raised in rabbits immunohistochemically the
presence of a hormone, pancreatic polypeptide, against avian pancreatic
polypeptide. It is possible we might get negative results. Our inference might
be that the pancreas of fish is lacking pancreatic polypeptide. But to ascertain
that the procedure we followed was correct and that the antiserum we obtained
was potent enough to cross - react with the polypeptide, we must have a
control test run in the pancreas of a bird and / or a mammal, which are known
for certain to give positive results.
When we talk about histochemistry or immunohistochemistry, the need for
controls is more obvious. For example, when we use a histochemical test to
localize a particular chemical substance in a tissue (e.g. carbohydrate,
proteins, lipids, nucleic acids, enzymes, hormones, etc.), we need to make sure
that the substance stained is really the substance intended to be stained by the
method we adopted. A simple control for such a situation is a negative control.
We can remove the particular chemical substance (e.g. removal of
carbohydrates by enzyme digestion) from the tissue and then apply the
histochemical test. Comparison of the experimental and control tissues would
ascertain the validity of the method applied.
Blind controls are useful in many experiments, especially when we know what
results are expected to occur. For example, in an experiment designed to test
the effectiveness of growth promoting hormone, the expected result is
 obvious. There is always the possibility of " falsifying ' data on our part if we
have a prior knowledge of the likely results. This problem can be overcome
by the " blind " design in which a person (a technician, perhaps) who is not
informed which plant received which treatment is entrusted with the task of
evaluation of the results. Thus the bias toward an " expected " result can be
avoided. Double blind controls can also be used, in which not only the person
who evaluates the results is blind as to what treatment i given to which subject,
but also the subjects are blind to the nature of the treatment they receive.
Randomization
Ensuring that every experimental unit gets equal chance of being assigned
to a treatment, in all aspects including space chance of be randomization
In other words, all treatments have an equal chance of being assigned to
each unit in the experiment. Randomization begins from the collection of
the sample and continues to the end of the experiment.
An experiment may contain two or more groups receiving different
treatments. The allotment of sampling units to the different groups should
be random. For example, we may procure 20 mice with known and similar
genetic background, for an experiment to test the effectiveness of a drug to
reduce heartbeat rate. We then divide the 20 mice into two groups each,
one for drug treatment and one for control (placebo) treatment. How are
we to select the first ten and second ten? " Catching " blindfold the mice
from the cage may not be really a random selection of a group. It can be
claimed that the mice that allow themselves to be caught earlier than the
others are in some way less active than those that do not allow themselves
to be caught. An easy way out is to assign numbers to each of the mice and
pick the numbers by lottery. The first ten numbers would form one group
and the remaining ten, the second group. Which group will receive which
treatment should also be at random, may be by the toss of a coin. Within a
group, say the experimental group, which of the ten mice will get first
injected with the drug? Is there any difference between the first one to be
injected and the last one to be injected? It is possible that the investigator
who injects the drug or the placebo gets tired with the passage of time and
therefore , the amount of test material injected , handling of the animal ,
etc. are different for those treated last when compared to than a simple
intramuscular injection , such as one involving a surgery , the those treated
first . If the experimental procedure is more complicated difference due to
elapse of time may be more pronounced. This necessitates randomization
in time. Randomization in time pertains not only to the order of treatment
 at the beginning of the experiment but also to the order of observations or
measurements to be taken during the course and at the end the experiment.
Randomization is also necessary wherever there is variation due to space,
for example, when we arrange experimental cages, aquarium tanks, pots,
etc. in the laboratory, their position may make a lot of difference with
reference to light, aeration and like factors. A pot placed close to a away
from the window. Unless we have the facility to maintain window might
receive more light and air than those that are kept artificially a uniform
lighting, aeration, humidity, temperature, etc., variations are bound to arise
due to space and influence the result of the experiment.
Randomization is more important especially in field experiments mainly
because of the heterogeneity of soil. While soils in adjacent areas are more
homogeneous, those far apart are more heterogeneous. Randomization is
becoming more and more important and complex in the field of molecular
biology where microarray experiments are routinely conducted to ascertain
the functions of genes in relation to different environmental factors.
Measurement
Very often, the results of an experiment require to be measured using some
kind of scale, which may be an ordinary metric scale, a digital balance or
a sophisticated spectrophotometer. Whatever instrument we use, we must
be concerned with the accuracy and precision of our measurements.
Accuracy is closeness of measured value to the true value. Ironically, the
true value of a measurement is never known and, therefore, we would never
be able to say whether our measurement is accurate or not. For example,
suppose we measure the length of a fish as 4.3 cm. This value is obtained
by using a scale that could give values up to a tenth of a centimetre. If we
use a still finer scale, one with Vernier facility, we might be able to get
measurements up to a hundredth or even a thousandth of a centimetre. But
how can we be certain that the value we obtained is accurate? Suppose we
know the true value of the length of the fish as 4.25 cm. Any measurement
that is close (e.g. 4.3 cm) to this true value can be described more accurate
when compared to a value that is far away (e.g. 4. 0 cm or 45 cm).
Precision is the reproducibility of a measurement or numerical result. suppose,
we measure using a digital balance the weight of a fish and g result as 10.6 g. We
weigh the same fish again in the same balance and g a result as 10.5 g. A third
measurement may give a result of 10.9 g O results are not precise. And suppose
we use another balance and our result are something like 10.6 g, 10.5 g, and 10.5
g. Now we can say our result are more precise than the ones obtained using the
first balance. We ca describe the second balance as a precision instrument because
it yields reproducible measurement values. Precision is practical and more
valuable than accuracy.
A Few Common Experimental Designs
All experimental designs require a thorough understanding of the bas principles
of statistics, especially those of inferential statistics. At the end of an experiment
we collect data from the sample, analyse it, and infer about the population from
which the sample was obtained. If we know beforehand what statistical test
should be applied to the data, we c design the experiment accordingly. That is
why it is suggested to experimental biologists to think about the statistical tests
to be used before the start of the experiment. Sir R. A. Fisher (1890-1962), whose
contribution to Design of Experiment in Biology is well known, said in his
Presidential address to the First Indian Statistical Congress in 1938 " to consult
the statistician after an experiment is finished is often mere the experiment died
of ". to ask him to conduct a post mortem examination. He can perhaps say what
the following examples of experimental designs are based on statistical
principles. However, we are not going to consider the detailed statistic
applications. procedures of each design but only the principle behind and the
possible One - group design A random sample is collected from the population
and its statistics (e.g. mean) is compared with the population parameter (e.g. μ).
For example, a sample of persons having the habit of consuming alcohol may be
collected and examined for their blood cholesterol level. It may be tested whether
the mean blood cholesterol level of this sample s significantly different from
mean population level.
Bibliography
➢ Gurumani N, Research methodology for biological sciences, New Delhi:
C. Janarthanan; 2020. 147-160
➢ C.R. Kothari, Research methodology methods and techniques, New Delhi:
New age international (p) limited, publishers; 2004. 31-52