Grp2-MANTHANS

7/30/2019 Grp2-MANTHANS

1/21

Chapter 3 Research Design V.E.S College of Arts, Science & Commerce

Chapter 3

Research Design

Prepared by Group 2

Manthans Research Group

Topics Page nos.

1. Impact of Problem Definition on Research Design

2. Concepts Relating to Research Design

3. Types of Research Designs.

3.1 Exploratory Research Studies

3.2 Descriptive And Diagnostic Research Studies

3.3 Hypothesis- Testing Research Studies(Experimental Studies)

4. Difference between exploratory and descriptive research

5. Basic Principles of Experimental Design

6. Formal and Informal Experimental Design

01

06

07

07

09

10

11

11

12

Group 2 -Manthans Research- Milind, Ankur, Nilesh, Tejashree, Harsha, Abhay and Sandeep 1

Members at MANTHANS include(Group 2)

Milind Doshi 10Ankur Jain 17Nilesh Jagdale 16Tejashree Patwardhan 30Harsha Punjabi 33Abhay Salitri 35Sandeep Kamath 19


2/21


1. Impact of Problem Definition on Research Design

- Research design

- Problem Definition- Components of a Problem- Impact of Problem Definition.

1.1 Research Design

A research design is the detailed blueprint used to guide a research study towardits objectives.

The process of designing a research study involves many interrelated decisions.The most significant decision is the choice of research approach, because itdetermines how the information will be obtained.

To design something also means to ensure that the pieces fit together. Theachievement of this fit among objective, research approach, and research tactics isinherently an iterative process in which earlier decisions are constantlyreconsidered in light of subsequent decisions.

1.2 Problem Definition

A problem exists when the decision-makerfaces uncertainty regarding which actionto adopt in the situation. If only one action is available (or none at all) or if there iscertainty about the outcomes of the alternatives, there really is no problem.

Defining a problem is a situation where:

1) The decision-maker has not yet determined how to exploit an opportunity or2) There are difficulties that are currently faced or are anticipated.

For instance the marketing manager may state that sales of a product have fallen by

25% because its price is too high & hence may ask the researcher to throw morelight on what is a more effective price? Actually the decline in sales may be due toany other factor or factor like poor product quality, competitors action, poorsalesmanship etc. The research dealing solely with the price may be able to solvethe problem correctly.

The existence of a disorder or a problem is the reason why the research isneeded. Once the problem is identified/disorder is located, the researcher may setthe projects objectives. The projects objectives are the specific purpose or goal ofthe research, since the objective flow from the disorder must precede the selection ofthe objectives.



3/21


1.3 Components of Problem

A problem consists of a set of specific components:

a) The decision makerand his or herobjectives;b) The environment orcontext of the problem;c) Alternative courses of action;d) A set of consequences that relate to courses of action and the occurrences

of events not under the control of the decision maker and;e) A state of doubt as to which course of action is best.

a) The decision makerand his or herobjectives;

The decision-maker may not always be represented by a single individual. Marketingdecisions may be made by a marketing group of two or more people. Moreover,some members of the group may not agree with the choice made because ofdifferences either in objectives (i.e., valued outcomes) or in their appraisal of theeffectiveness of means chosen to achieve the objectives. In other situations anindividual may be performing the role of agent for some superior or group ofsuperiors. The objectives of the decision maker provide motivation for the decision.These objectives, or goals, may range from a desire to maintain or increasecompany profits and market share to personal goals concerned with maintainingprestige and a desire to advance in the corporation.

The decision makers objectives may also be characterized by their hierarchicalnature at any given moment and their evolution over time. For example, an increasein the firms profits may come about through an increase in the firms sales, which, inturn, may be accomplished by the firms sales personnel contacting a greaternumber of new accounts per month. The goal for the salesperson may be toincrease sales contacts 10% over those made in some base period, but thisrepresents a sub goal, consistent, it is hoped, with a higher-level objective. Thedecision theorist also faces the problem of estimating changes in objective over time.

b) The environment orcontext of the problem;

Every problem exists within a context of the characteristics of the company and ofthe market-consumer tastes and preferences, level of income and rate of growth inthe market areas, the degree of competition and competitor action and reaction, andthe type and extent of governmental regulation. These environmental factors mayindividually and collectively affect the outcome of the decision made. Theresearcher must assist the manager in identifying these relevant environmentalfactors.



4/21


Consider the problem of deciding whether to introduce a new consumer product.Some of the environmental factors that could affect the decision are as follows:

The types of consumers that comprise the potential market,

The size and location of the market,

The prospects for growth or contraction of the market over the planning

period,

The buying habits of consumers,

The current competition for the product,

The likelihood and timing of entry of new competitive products ,

The current and prospective competitive position with respect to price,

quality, and reputation,

The marketing and manufacturing capabilities of the company,

The situation with respect to patents, trademarks, and royalties, The situation with respect to codes, trade agreements, taxes, and

tariffs.

Although this listing is by no means exhaustive, it illustrates some of the moreimportant environmental factors that could influence the outcome of thedecision and so must be considered in the problem statement. Each problemhas a comparable set of environmental factors to be considered.

c) Alternative courses of action;

A course of action is a specification of some behavioral sequence, such as theconstruction of a new warehouse, the adoption of a new package design, orthe introduction of a new product. All courses of action involve, eitherimplicitly or explicitly, the element of time. For example, Construct awarehouse, starting next week is a different course of action from Constructa warehouse, starting next year.

Actions, of course, can be taken only in the present. A decision to stipulate aprogram of action becomes a commitment, made in the present, to follow some

behavioral pattern in the future. Courses of action may range in complexity from asingle act to be implemented immediately to a large set of related acts proceedingeither in parallel or sequentially over time. The time interval, which becomes a partof the course of action, may be highly important, since both the costs ofimplementation and the probabilities of alternative outcomes will typically vary as afunction of time.

d) Consequences ofAlternative Courses of Action;

The world of uncertainty is a familiar world for the marketer. When choosing acourse of action, a marketer can rarely be certain of the consequences, since



5/21


the choice is usually based on incomplete information about the various factorsthat influence the decisions outcome. A primary job is thus to list the possibleoutcomes of various courses of action. But these outcomes will depend on

various environmental factors.

e) A state of doubt as to which course of action is best;

To solve a problem is to select the best course of action for attaining thedecision makers objectives. A state of doubt as to which course of action isbest can arise under three main classes of conditions:

a. Certainty with respect to each course of action leading to a specificoutcome.

b. Risk with respect to each action leading to a set of possible outcomes,

each outcome occurring with a known probability.For example, if a fair coin is tossed, we may assume that over the long run theproportion of heads will approach one-half; however, on any single toss we cannotpredict whether a head or tail will appear.

c. Uncertainty with respect to outcomes, given a particular course of action.In this view of decision-making we assume that the relative frequencies of theprobabilities are not known. One version of this class of models, exemplifiedin the Bayesian approach to decision making (to be described later), assumesthat the decision maker can express various degrees of belief as to theoccurrence of alternative outcomes. Moreover, the decision maker may be

able, in many cases, to collect more information regarding the true state ofnature.

1.4 Impact of Problem Definition

A carefully formulated problem is a necessary point of departure for competentlyconducted research. There should be as clear and thorough an understanding aspossible on the part of both the researcher and the decision maker as to theprecise purposes of the research. In effect, this statement of purpose involves a

translation of the decision makers problem into a research problem and studydesign. The decision maker is faced with a problem for which he or she mustrecognize alternative courses of action, choosing among them to accomplish oneor more objectives. The research problem is to provide relevant informationconcerning recognized (or newly generated) alternative solutions to aid in thischoice. To determine what information is required, the researcher will try toidentify and understand the major elements of the problem faced by the decisionmaker.

In a very real sense, problem formulation is the heart of the research process. Assuch it represents the single most important step to be performed.



6/21


2. Concepts Relating to Research Design

1. Dependent and independent variableA concept, which can take on different quantitative values, is called avariable. eg: weight, height, income etc. Phenomena, which can take onquantitatively different values even in decimal points, are calledcontinuous variables. Age is continuous variable but number of children isnon continuous variable. If one variable depends on upon or is aconsequence of other variable, it is termed as a dependent variable, andthe variable that is antecedent to the dependent variable is termed asindependent variable. For example: if height depends on age, then heightis dependent variable and age is independent variable.

2. Extraneous variableIndependent variables that are not related to the purpose of the study, butmay effect the dependent variable are termed as extraneous variables.Suppose the researcher wants to test the hypothesis that there isrelationship between childrens gains in social studies achievement andtheir self-concepts. In this case self-concept is an independent variableand social studies achievement is a dependent variable. Intelligence mayas well affect the social studies achievement, but since it is not related tothe purpose of the study undertaken by the researcher, it will be termed asan extraneous variable.

3. ControlOne important characteristic of a good research design is to minimise theeffect of extraneous variable(s). The technical term control is used whenwe design the study minimising the effects of extraneous independentvariables. In experimental searches, the term control is used to refer torestrain experimental conditions.

4. Research hypothesisWhen a prediction or hypothesised relationship is to be tested by scientificmethods, is it termed as research hypothesis. It is a predictive statement

that relates an independent variable to a dependent variable.

5. Experimental and non experimental hypothesis-testing researchIn this case the purpose of research is to test a research hypothesis. It canbe of the experimental design or of the and non experimental design.Research in which the independent variable is manipulated is termedexperimental hypothesis-testing research and a research in which anindependent is not manipulated is called non experimental hypothesis-testing research.



7/21


6. Experimental and control groupIn this research when a group is exposed to usual conditions, it is termed a

control group, but when a group to exposed to some special conditions it istermed as experimental group.

7. TreatmentsThe different conditions under which experimental and control groups areput are usually referred to as treatments.

8. ExperimentThe process of examining the truth of a statistical hypothesis, relating tosome research problem, is known as an experiment. For eg: anexperiment can be conducted to examine the usefulness of a certain newly

developed drug. Experiments can be of two types viz., absoluteexperiment and comparative experiment.

9. Experimental unit(s)The pre-determined plots or the blocks, where different treatments areused, are known as experimental units. Such experimental units must beselected (defined) very carefully.

3. Types of Research Designs.

The different research designs can be categorized into research design in case of:

1. Exploratory Research Studies.2. Descriptive And Diagnostic Research Studies3. Hypothesis- Testing Research Studies (Experimental Studies)

Following are the details of different research designs:

3.1 Exploratory Research Studies

Also termed as formulative research studies.

Purpose of such studies is formulating a problem for more precise investigation. Major emphasis is on the discovery of ideas and insights.

Research design has to be flexible enough to provide opportunity for consideringdifferent aspects of a problem under study.

Inbuilt flexibility is essential.

Following are three methods in the context of research design for studies:

The survey of concerning literature

The experience survey

The analysis of insight stimulating examples.



8/21


The survey of concerning literature:

This happens to be the most simple and fruitful method of formulating the researchproblem. Hypothesis stated by earlier workers may be reviewed and their usefulnessbe evaluated as a basis for further research. In this way researcher should reviewand build upon the work already done by others, but in cases where hypothesis hasnot been formulated hi task is to review the available material for deriving therelevant hypothesis from it.

Experience Survey:

It is the survey of people who have had practical experience with the survey to be

studied. The object is to obtain insight into the relationship between variables andnew ideas relating to the research problem. For such a survey people who arecompetent and can contribute new ideas may be carefully selected as respondentsto ensure representation of different of experience. The respondents selected canbe interviewed by the investigator. An interview schedule is prepared by theresearcher for systematic questioning of informants. The interview must ensureflexibility in the sense that the respondents should be allowed to raise issues andquestions which the investigator has not previously considered. The interview maylast for few hours. Hence, it its often considered desirable to send a copy of thequestions to be discussed to the respondents well in advance. This gives anopportunity to the respondents for doing some advance thinking over various issuesinvolved so that, at the time of interview they may be able to contribute effectively.Thus, an experience survey may enable the researcher to define the problem moreconcisely and help in formulation of research hypothesis. This survey may as wellprovide information about the practical possibilities for doing different types ofresearch.

Analysis of insight stimulating examples:

This is a fruitful method for suggesting hypothesis for research. It is particularlysuitable in areas where there is little experience to serve as a guide. It consists of the

intensive study of the selected instances of the phenomenon in which on isinterested. For this purpose the existing records may be examined the unstructuredinterviewing may take place or some other approach may be adopted. Attitude of theinvestigator, the intensity of the study and the ability of the researcher to drawtogether diverse information into a unified interpretation are the main features whichmake this method an appropriate procedure for evoking insights.

Examples for the above are:

Reactions of strangers

Reactions of marginal individuals



9/21


Study of individuals who are in a transition from one stage to another.

Reactions of individuals from different social strata.

3.2 Descriptive And Diagnostic Research Studies

Descriptive research studies are concerned with describing the characteristics ofcertain individuals or a group. E.g. studies concerning whether certain variables areassociated.Diagnostic research studies determine the frequency of with which something occursor its association with something else. E.g. studies concerned with specificpredictions, with narration of facts and characteristics concerning individual, group orsituation.The descriptive as well as diagnostic research studies share common requirements.In both the studies, the researcher must be able to define clearly, what he wants tomeasure and must find adequate methods of measuring it. The aim is to obtaincomplete and accurate information, hence, the procedure to be used must becarefully planned. It should make enough provision for protection against bias andmust maximize reliability. The design must be rigid and not flexible.

Following should be focussed:

a) Formulating the objective of the study (what is the study about and why is it beingmade.b) Designing the methods of data collection (what techniques of gathering data will

be adopted)c) Selecting the sample (how much material will be needed)d) Collecting the data (where can the required data be found and with what time

period should the data be related)e) Processing and analyzing the data.f) Reporting the findings.

Following are the steps involved in both the studies:

Step 1.Specify the objectives with sufficient precision to ensure that ht data collectedis relevant.Step 2.Select the methods by which the data are to be obtained. E.g. techniques ofcollecting the data must be devised.While designing data collection procedure, adequate safeguards against bias andunreliability must be ensured. Questions must be well examined and must beunambiguous, interests must not express their opinion.

In most studies researcher takes down samples and then wishes to makestatements about the population on the basis of the sample analyses.



10/21


The problem of designing samples should be tackled in such a form

that the samples may yield accurate information with a minimum

amount of research effort. To obtain data free from errors, it is necessary to supervise closely the

staff of field workers, as they collect and record information.

As data are collected, they should be examined for completeness,

comprehensibility, consistency and reliability.

The data collected must be processed and analysed.

This includes steps like coding the interview replies, observations, etc.;

tabulating the data; and performing several statistical computations.

The processing and analyzing procedure should be planned in detail

before actual work is started.

To avoid error in coding, the reliability of coders needs to be checked.

Similarly, the accuracy of tabulation may be checked by having a

sample of tables re-done.

Last of all comes the task of reporting the findings, i.e. communicating

the findings to others and the researcher must do it in an efficientmanner.

The layout of the report needs to be well planned so that all things

relating to the research study may be well presented in a simple andeffective style.

Thus, the research design in the case of descriptive/diagnostic studies

is a comparative design and must be prepared keeping the objective(s)of the study and the resources available.

However, it must ensure the minimization of bias and maximisation of

reliability of the evidence collected.

It can be referred to as a survey design since it takes into account all

the steps involved in a survey concerning a phenomenon to be studied.

3.3 Hypothesis- Testing Research Studies (Experimental Studies)

Hypothesis-tested research studies (experimental studies) are those

where the researcher tests the hypothesis of casual relationshipbetween variables.

Such studies require procedures that will not only reduce bias and

increase reliability, but will permit drawing inferences about casuality.

Professor R.A. Fisher begun such designs when he was working at

Rothamsted Experimental Station (Centre for Agricultural Research inEngland).

Professor Fischer found that by dividing agricultural fields or plots into

different blocks and then by conducting experiments in each of these



11/21


blocks, the information collected and inferences drawn happen to bemore reliable.

This fact inspired him to develop certain experimental designs for

testing hypotheses concerning scientific investigation.

4. Difference between exploratory and descriptive research

RESEARCHDESIGN

Types of study

Exploratory of

formulative

Descriptive /

Diagnostic

Overall design Flexible design (design mustprovide opportunity forconsidering different aspectsof the problem)

Rigid design (design mustmake enough provision forprotection against and mustmaximize reliability)

(i) samplingdesign

Non- probability samplingdesign (purposive orjudgement sampling)

Probability sampling design(random sampling)

(ii) statisticaldesign

No pre-planned design foranalysis

Pre-planned design foranalysis

(iii)observationaldesign

Unstructured instruments forcollection of data

Structured or well thought outinstruments for collection ofdata

(iv) operationaldesign

No fixed design about theoperational procedure

Advanced decisions aboutoperational procedures

5. Basic Principles of Experimental Design

Professor Fisher has enumerated three principles of experimental designs:

1. the Principle of Replication;2. the Principle of Randomization; and the3. the Principle of Local Control.

According to the Principle of Replication, the experiment should be repeated morethan once. Thus, each treatment is applied in many experimental units instead ofone. By doing so the statistical accuracy of the experiments is increased. The entireexperiment can even be repeated several times for better results. Conceptually



12/21


replication does not present any difficulty, but computationally it does. It should beremembered that replication is introduced in order to increase the precision of astudy; that is to say, to increase the accuracy with which main effects and

interactions can be estimated.

The Principle of Randomization provides protection, when we conduct anexperiment, against the effects of extraneous factors by randomization. In otherwords, this principle indicates that we should design or plan the experiment in such away that the variations caused by extraneous factors can be combined under thegeneral heading of chance.

The Principle of Local Control is another important principle of experimentaldesigns. Under it the extraneous factors, the known source of variability, is made tovary deliberately over as wide a range as necessary and this needs to be done in

such a way that the variability it causes can be measured and hence eliminated fromthe experimental error. This means that we should plan the experiment in mannerthat we can perform a two-way analysis of variance, in which the total variability ofthe data is divided into three components attributed to treatments (the subject), theextraneous factors and experimental error.

6. Formal and Informal Experimental Design

Experimental design refers to the framework or structure of the experiment and assuch there are several such experimental design. Experimental design can beclassified into two broad categories. Informal experimental design and Formalexperimental design. Informal experimental design are those design that normallyuses a less sophisticated form of analysis based on differences in magnitude,whereas formal experimental design offer relatively more control and use precisestatistical procedures for analysis.Important statically designs are as follows:

1. Informal experimental design:

Before and after without control design.

After only with control design.

Before and after with control design.

2. Formal experimental design:

Completely randomized design (C.R.design).

Randomized block design (R. B. design).

Latin square design (L.S. design).

Factorial design.

The details of each of the above stated formal and informal experimental designare explained as follows.



13/21


1. Before-and-after control design : In such a design a single test groupor area is selected and the dependant variable is measured before the

introduction of the treatment. The treatment is then introduced and thedependant variable is measured again after the treatment has beenintroduced. The effect of the treatment would be equal to the level ofthe phenomena after the treatment minus the level of phenomenonbefore the treatment.The design can be represented as

______________________________________________________________Test area: Level of phenomenon Treatment Level of phenomenon

Before treatment (X) introduced after the treatment (Y)

Treatment Effect = (Y) - (X)

_______________________________________________________________________

The main difficulty of such a design is that with the passage of the time considerableextraneous variation may be there in the treatment effect.

2. After only with control design: In this design two group of arise (testarea and control area) are selected and the treatment is introduced inthe test area only. The dependant variable is then measured in boththe areas at the same time. Treatment impact is assessed bysubtracting the value of the dependant variable in the control area fromthe value in the test area. This can be exhibited in the following form

________________________________________________________________Test area: Treatment introduced Level of phenomenon after

treatment (Y)

Control area Level of phenomenon withouttreatment (Z)

Treatment effect = (Y) (Z)________________________________________________________________

The basic assumption in such a design is that the two areas are identical withrespect to their behavior towards the phenomenon considered. If the assumption isnot true, then there is then there is the possibility of extraneous variant entering intothe treatment effect. However, data can be collected in such a design without theintroduction of the problems with the passage of time. In this respect this design issuperior to before-and-after without control design.

3. Before-and-after with control design : in this design two areas areselected and the dependent variable is measured in both the areas foridentical time period before the treatment .The treatment is thenintroduced into the test area only, and the dependent area is

measured both for an identical time period after the introduction of theGroup 2 -Manthans Research- Milind, Ankur, Nilesh, Tejashree, Harsha, Abhay and Sandeep 13


14/21


treatment. The treatment effect is determined by subtracting change inthe dependent variable in the control area from the change in thedependent variable in the test area. This can be shown in the following

way:________________________________________________________Time Period 1 Time Period 2

Test area: Level of phenomenon Treatment Level of phenomenonbefore treatment (X) introduced after treatment(Y)

Control area: Level of phenomenon Level of phenomenonWithout treatment without treatment

(A) (B)Treatment Effect = (Y-X) (Z-A)

__________________________________________________________________________________________

This design is superior to the other two design for the simple reason that itavoids extraneous variation resulting both from the passage of time and from non-comparability of the test and control areas. But at times due to lack of historical data,time, so it is preferred to select one of the first two informal designs stated above.

4. Completely randomized design (C.R.): involves only two principles viz., theprinciple of replication and the principle of randomization of experimentaldesigns. It is the simplest possible design and its procedure of analysis is alsoeasier. The essential characteristic of this design is that subject are randomlyare assigned to experimental treatment. For instance, if we have 10 subjectand we to test 5under treatment A and 5 under treatment B, the randomizedprocess gives the every possible group of 5 subjects selected from the groupof 10 an equal opportunity of being assigned to treatment A and treatment B.One-way analysis of variance (or one way ANOVA) is used to analyze such adesign. Even unequal application works in this design. It provides maximumnumber of degree of freedom to the error. Such a design is used whenexperimental areas happen to be homogenous. Technically, when all thevariation due to uncontrolled extraneous factors are included under the

heading of chance variation, we refer to the design of experiment asC.R.design.

The brief description on the two form of such design is explainedbelow:

i. Two-group simple randomized design: in a two-group simplerandomized design, first of all the population is defined and then from thepopulation a sample is selected randomly. Further requirement of thisdesign is that items, after being selected randomly from the population,be randomly assigned to the experimental and control groups(such



15/21


random assignmentof items of two group is called as principle ofrandomization.). Thus this design yields two groups as representative ofthe population. In the diagram form this design can be shown in this way

Two-group simple randomized design in Diagram form

Randomly RandomlySelected assigned

Since in the simple randomized design the elements constituting the sampleare randomly from the same population and randomly assigned to the experimentaland control groups, it becomes possible to draw conclusion on the basis of samplesapplicable for the population. The two group (experimental and control groups) ofsuch a design are given different treatment of the independent variable. This designof experiment is quiet common in research studies concerning behavioral sciences.The merit of such a design is that it is simple and randomizes the difference amongthe sample items. But the limitation of it is that the individual differences amongthose conducting the treatments are not eliminated, i.e., it does not control the

extraneous variable and as such the result of the experiment may not depict acorrect picture. This can be illustrated by an example. Suppose that the researcherswant to compare two groups of student who have been randomly selected andrandomly assigned. Two different treatment viz., the usual training and thespecialized training are being given to the two groups. The researchers hypothesisgreater gain for the group who receives specialized training. To determine this, hetests each group before and after the training, and compares the amount of gain forthe two groups to accept or reject his hypothesis. This is the illustration of the twogroup randomized design, wherein individual differences among students are beingrandomized. But this does not control the differential effects of the extraneous


Population Sample

Controlgroup

Experimental group

Independ

entVariable

Treatment

A

Treatment

B


16/21


independent variable (in this case, the individual difference among those conductingthe training programmes)

Random replication design (IN A DIAGRAM FORM)

Random Selection Random Selection

Random Randomassignment assignment

Group 1 E Group 2 E Group 3 E Group 4 E

Group 5 CGroup 6 CGroup 7 C E= Experimental groupGroup 8 C C= Control group

Treatment BTreatment A


Population(Available for

study)

SAMPLE(To be

studied)

Population(Available toconduct

SAMPLE(To conducttreatments)

Independent variable or causal variable


17/21


(ii) Random replication design:The limitation of the two-group randomized design is usually eliminated withinthe random replications design. In the illustration just cited above, the teacherdifferences on the dependent variable were ignored, i.e., the extraneousvariable was not controlled. But in a random replications design, the effect ofsuch differences are minimised (or reduced) by providing a number ofrepetitions for each treatment. Each repetition is technically called areplication. Random replication design serves two purposes viz., it providescontrols for the differential effects of the extraneous independent variables and

secondly, it randomizes any individual differences among those conducting thetreatments.

From the diagram it is clear that there are two populations in the replicationdesign. The sample is taken randomly from the population available for studyand is randomly assigned to, say, four experimental and four control groups.Similarly, sample is taken randomly from the population available to conductexperiments (because of eight groups eight such individuals be selected) andthe eight individuals so selected should be randomly assigned to the eightgroups. Generally, equal number of items are put in each group so that the sizeof the group is not likely to affect the results of the study. Variables relating to

both population characteristics are assumed to be randomly distributed amongthe two groups. Thus, this random replication design is, in fact an extension ofthe two-group simple randomized design.

5. Randomized block design (R.B design) is an improvement over theC.R. design. In the R.B design the principle of local content can be applied alongwith the other two principles of experimental designs. In the R.B. design, subjectsare first divided into groups, known as blocks, such that within each group thesubjects are relatively homogeneous in respect to some selected variable. Thevariable selected for grouping the subjects is one that is believed to be related to themeasures to be obtained in respect of the dependent variable. The number ofsubjects in a given block would be equal to the number of treatments and onesubject in each block would be randomly assigned to each treatment. In general,blocks are levels at which we hold the extraneous factor fixed, so that its contributionto the total variability of data can be measured. The main feature of the R.B design isthat in this each treatment appears the same number of times in each block. The R.Bdesign is analysed by the two-way analysis of variance (two-way ANOVA) technique.

Let us illustrate the R.B. design with the help of an example. Suppose four differentforms of a standardized test in statistics were given to each of five students (selected



18/21


one from each of the five I.Q. blocks) and following are the scores which theyobtained.

Very low IQ Low IQ Average IQ High IQ Very High IQ

Student A Student B Student C Student D Student E

Form 1

82 67 57 71 73

Form 2

90 68 54 70 81

Form 3

86 73 51 69 84

Form 4

93 77 60 65 71

If each student separately randomized the order in which he or she took the fourtests (by using random numbers or some similar device), we refer to the design ofthis experiment as a R.B. design. The purpose of this randomization is to take careof such possible extraneous factors (say as fatigue) or perhaps the experiencegained from repeatedly taking the test.

6. Latin squares design (L.S design) is an experimental design very frequentlyused in agricultural research. The conditions under which agricultural investigationsare carried out are different from those in other studies for nature plays an importantrole in agriculture. For instance, an experiment has to be made through which theeffects of five different varieties of fertilizers on the yield of a certain crop, say wheat,is to be judged. In such a case the varying fertility of the soil in different blocks inwhich the experiment has to be performed must be taken into consideration;otherwise the results obtained may not be very dependable because the outputhappens to be the effect not only of fertilizers, but it may also be the effect of fertilityof soil. Similarly, there may be the impact of varying seeds on the yield. To overcome

such difficulties, the L.S design is used when there are two major extraneous factorssuch as the varying soil fertility and varying seeds.

The merit of this experimental design is that it enables differences in fertilitygradients in the field to be eliminated in comparison to the effects of differentvarieties of fertilizers on the yield of the crop. But this design suffers from onelimitation, and it is that although each row and each column represents equally allfertilizer varieties, there may be considerable difference in the row and columnmeans both up and across the field. This, in other words, means that in L.S. designwe must assume that there is no interaction between treatments and blockingfactors.



19/21


(7) Factorial designs: Factorial designs are used in experiments where the effects

of varying more than one factor are to be determined. They are specially important inseveral economic and social phenomena where usually a large number of factorsaffect a particular problem. Factorial designs can be of two types: (i) simple factorialdesigns and (ii) complex factorial designs.

(i) Simple factorial designs: In case of simple factorial designs, we considerthe effects of varying two factors on the dependent variable, but when anexperiment is done with more than two factors, we use complex factorialdesigns. Simple factorial design is also termed as a two-factor-factorialdesign, whereas complex factorial design is known as multi-factor-factorial design. Simple factorial design may either be a 2 x 2 simple

factorial design, or it may be, say 3 x 4 or 5 x 3 or the like type of simplefactorial design.

Illustration : ( 4* 3 simple factorial design)

The 4*3 simplex factorial design will usually include four treatments of theexperimental variable and three levels of the control variable. Graphically it may takefollowing form:

Experimental Variable

CONTROLVARIABLE

TREATMENTA

TREATMENTB

TREATMENTC

TREATMENTD

Level 1 Cell 1 Cell 4 Cell 7 Cell 10



This model of a simplex factorial design includes four treatments viz. A, B, C and Dof the experimental variable and three levels viz (I), (II), and (III) of the control

variable and has 12 different cells as shown above. This shows that a 2 * 2 simplefactorial design can be generalized to any number of treatments and levels. In sucha design the means for the columns provide the researcher with an estimate of themain effects for the levels. Such a design also enables the researcher to determinethe interaction between treatments and levels

(ii) Complex factorial designs: experiments with more than two factors at atime involve the use of complex factorial designs. A design whichconsiders three or more independent variables simultaneously is called acomplex factorial design. In case of three actors with one experimentalvariable having two levels, the design used will be termed 2 * 2 * 2 *



20/21


comple factorial design which will contain a total of eight cells as shownbelow:


Treatment A Treatment B

ControlVarable 2

Level I

ControlVariable 2

Level II

ControlVariable 2

level I

ControlVariable 2

Level II

Level I

Control

Variable 1 Level II

Cell 1 Cell 3 Cell 5 Cell 7

Cell 2 Cell 4 Cell 6 Cell 8

To obtain the first order interaction say, for EV * CV 1in the above stated design, theresearcher must necessarily ignore control variable 2 for which purpose he may

develop 2 * 2 design from the 2 * 2* 2 design by combining the data of the relevantcells of the latter design as has been shown below:


Treatment A Treatment B

Control Level 1 Cells 1, 3 Cells 5, 7

Level 2 Cells 2, 4 Cells 6, 8

Similarly, the researcher can determine other first order interactions analysis of thefirst order interaction, here essentially a simple factorial analysis as only twovariables are considered at a time and the remaining on eis ignored. But the analysisof the second interaction would not ignore one of the three independent variables inthe case of a 2 * 2* 2 design. The analysis would be termed as a complex factorialanalysis.

Factorial analysis are used mainly because of two advantages :



21/21


1. They provide equivalent accuracy ( as happens in the case of experimentswith only one factor) with less labour and as much are a source of economy.Using factorial designs, we can determine the main effects of two ( in simple

factorial design ) or more ( in case of complex factorial design ) factors ( insimple factorial design ) or more ( in case of complex factorial design) factors( or variables in one single experiment.

2. They permit various other comparisons of interest. For example, they giveinformation about such effects which cannot be obtained by treating onesingle factor at a time. The determination of interaction effects is possible incase of factorial design.

There are several research designs and the researcher must decide in advance ofcollection and analysis of data as to which design which prove to be moreappropriate for his research project. One must give due weight to various points suchas type of universe and its nature, objective of his study, source list or samplingframe desired standard of accuracy and the like when taking decision in respect ofthe design for his research project.


Grp2-MANTHANS

Documents

Transcript of Grp2-MANTHANS