Exploratory research: The chief purpose of exploratory research is to reach a better understanding of the research problem. This includes helping to identify the variables which should be measured within the study. When there is little understanding of the topic it is impossible to formulate hypotheses without some exploratory studies. For example, crop residues such a straw are high in lignin (a wood-like substance) and low in nutrients. This makes them a poor animal feed since the lignin acts against digestibility and the low nutrient content means poor food value. However, if treated in a strong alkali, plus a little heat, the lignin breaks down and the nutrient content increases. A company was established to exploit this technology and did so successfully for 4 seasons. After this period sales began to slow down. Three other manufacturers had entered the market by this time. The company, Animal Feed Systems, did not know whether the whole industry had slowed down or if only their product was suffering. Nor did they know if the problem was temporary in that perhaps the market comprised of "early adopters" had been saturated but it was only a matter of time before other farmers began to buy their systems when they saw how well they worked. It was also possible that if a problem did exist it could lie in any one of a number of areas: animal populations might be declining, distributors may not be promoting the product aggressively, customers may be experiencing difficulties in getting the chemicals, and so on and on.
This is a good example of a situation where insufficient knowledge prevented the development of clear objectives, since the problem could not be articulated with any precision and therefore research of an exploratory nature was required. Such research can take the form of literature searches, informal personal interviews with distributors and users/non-users of the product and/or focus group interviews with farmers and/or distributors.
Exploratory research is intended to help researchers formulate a problem in such a way that it can be researched and suggest testable hypotheses.
Descriptive research: As the name suggests, descriptive research is concerned with describing market characteristics and/or marketing mix characteristics. Typically, a descriptive study specifies the number and size of market segments, the alternative ways in which products are currently distributed, listing and comparison of the attributes and features of competitive products, etc.
This type of study can involve the description of the extent of association between variables. For example, the researcher may observe that there is an association between the geographical location of consumers and their tendency to consume red meat. Note that the researcher is able to describe the relationship rather than explain it. Nonetheless if the relationship between the two is fairly stable this descriptive information may be sufficient for the purposes of prediction. The researcher may, for example, be able to predict how fast the per capita consumption of red meat is likely to rise over a given time period.
The principal difference between exploratory and descriptive research is that, in the case of the latter, specific research questions have been formulated before the research is undertaken. When descriptive research is conducted the researcher must already know a great deal about the research problem, perhaps because of a prior exploratory study, and is in a position to clearly define what he/she wants to measure and how to do it.
Causal research: Causal research deals with the "why" questions. That is, there are occasions when the researcher will want to know why a change in one variable brings about a change in another. If he/she can understand the causes of the effects observed then our ability to predict and control such events is increased.
In summary then there are three distinct types of marketing research study: exploratory, descriptive and causal. The purpose of each is summarised in figure 1.4. In some cases, a research programme will be of one kind or another, but in other instances these three typologies will represent phases within a single marketing research investigation.

The research design is a comprehensive master plan of the research study to be undertaken, giving a general statement of the methods to be used. The function of a research design is to ensure that requisite data in accordance with the problem at hand is collected accurately and economically. Simply stated, it is the framework, a blueprint for the research study which guides the collection and analysis of data. The research design, depending upon the needs of the researcher may be a very detailed statement or only furnish the minimum information required for planning the research project. To be effective, a research design should furnish at least the following details.

a) A statement of objectives of the study or the research output.

b) A statement of the data inputs required on the basis of which the research problem is to be solved.

c) The methods of analysis which shall be used to treat and analyse the data inputs
23

Let us try to understand the elements through an example:
A newly opened supermarket sells a broad line of merchandise ranging from provision to household appliances and kitchenware. The general manager (sales) believes that the total profits of the supermarket can be enhanced by getting people to buy in larger quantities which could be achieved by offering attractive cash discount on bigger purchases. As the other executives are doubtful about this. The hypothesis can be tested by carrying out a marketing research exercise.
1) The objective is to calculate the margin earned on sales when this discount is offered and compare it with the margin when discount is not being offered.
2) The data to be collected over a period of time

a) Sales in rupees to a selected sample of customers during the period when the discount is offered.

b) Sales in rupees to the same customer when the discount is not being offered.

c) The average order size in the two periods.

d) The average margins earned during the two periods.

e) The cost of promotional inputs regarding the discount.

3) The analysis of the data will be

a) Sales in rupees in period I minus those in period II.

b) Subtract cost of incentives.

c) Also subtract cost of promotional inputs.

The importance of research design lies in the fact that it makes a statement of what is to be done in order to achieve the research objectives and how it is to be done. It is an expression of what is expected of the research exercise in terms of results and the analytical input needed to convert data into research findings.
The research design furnishes a clear idea as to the activities that would need to be undertaken in order to achieve the research objective. It is therefore, helpful if the research design after being finalised is put in writing to enable the researcher to have a frame of reference and prevent the study from deviating.
At the stage of analysis and interpretation also, the research design helps in providing direction to the computation and interpretation process to arrive at solution and recommendations. This is however not to suggest that a design is a rigid straitjacket to which the study must always conform. The research design only represents an expectation of likely results but as the study proceeds, many unexpected results may come forth which may necessitate framing of new hypothesis or at least modifying some. The research design is only a guiding and not a limiting framework for research study.
3.3 CAUSALITY: THE BASIS OF CLASSIFICATION OF VARIOUS TYPES OF RESEARCH DESIGNS
There are four types of research designs which are broadly classified as:

1) Exploratory Designs

2) DescriptiveDesigns

3) Quasi-Experimental Designs

4) Experimental Designs

These designs will be discussed in detail very shortly. However, one point may be noted that the basis of classification of various types of research designs is causality, which is the subject matter of discussion of this section.
Let us start our discussion with an example. Suppose the sales manager of a company manufacturing table fans carried out a training programme for its salesmen operating in a state. Three months after the training programme, it was observed that there was an increase in sales in that state by 40 per cent. The sales manager claimed that the training programme was effective and therefore, the salesmen operating in other states should also undergo the same training programme. We can see that the sales manager is inferring that the training programme has resulted in a sales increase of 40 per cent. That is to say that training programme is the cause of increased sales.
The sales manager is referring to sales training as a causal variable and the sales increase as the effect variable. We may now ask a question whether this statement is valid? Has the sales training really resulted' in increase in sales? The possible answer is that we cannot say with certainty that the sales training has caused the increase in sales. There could have been other factors which might have been responsible for the increased sales. Increase in sale might have been caused by decrease in the price of the table fans, a strike at a competitor's factory, increased product penetration in the distribution channel, weather conditions, etc. Therefore, it is very essential that the sales manager should know that conditions under which proper causal statements can be made. To be able to make any causal statements, one should ensure whether the following three conditions are met:

i) We must have very strong evidence to say that there exists a strong association between an action (causal variable) and the ultimate outcome (effect variable).

ii) The other condition for the causal relationship is that the action (causal factor) must precede the observed outcome (effect variable).

iii) We must have strong evidence to say that there were no other possible factors (causal factors) which could have resulted in the observed outcome.

The first condition is that of concomitant variation. Concomitant variation is the extent to which a cause, X, and effect Y occur or vary together. In our example, the sale training programme and increase in sales would need to occur together. To be able for us to conclude causality, the condition of concomitant variation between variables in question must exist. However, it may be worth mentioning that a strong association between two variables does not constitute a proof of a causal relationship, It is quite likely that the strong association between two variables may be as a result of random variation or both the variables may be influenced by an extraneous variable.
The second condition for causal relationship is the requirement that the causal factor (action) must occur either prior or simultaneously with effect factor(outcome). However, the fact that the action precedes outcome does not establish causal relationship. It might be a mere coincidence that sales training took place prior to increase in sales of the table fans. There is also a possibility that sales training and increase in sales of table fans are strongly associated. This however, does not prove the existence of causal relationship.
The third condition for inferring causation is the absence of other possible causal factors. This means that all other possible factor influencing the outcome (in our case increase in sales of table fan) are either absent or are kept constant. It is only then we could say logically that the sales training has resulted in increase of sales of fable fans. In reality, it is impossible to find the absence of other factors or to hold some factors constant. For example, we know that the sales of table fans is influenced by weather conditions. Is it possible to keep weather conditions constant? Or can we be sure that the competitor would not change the price? The obvious answer is "No".
In a case where the outcome is completely determined by only one causal factor, we can say that causal factor is the deterministic cause of the outcome. That is the `causal factor' in this case is both necessary and sufficient condition for the occurrence of outcome. However, in a situation where the outcome is influenced by a host of causal factors, any of the causal factor is the probablistic cause of the outcome. That is to say it is a necessary but not a sufficient condition for the occurrence of outcome.
There are three possible ways to control the influence of extraneous variables. Firstly, we may physically control the extraneous variable. For example, a company trying to study the impact of two different packagings of a product on sales may control the extraneous variable like price by keeping it constant for both packaging containing the same amount and quality of good. The second way to control the effect of extraneous variables if the physical control is not possible is to randomize the assignment of treatments to test units. The third way to control the extraneous variables is through the use of experimental designs, the discussion of which -would follow in the subsequent sections. If the control of extraneous variable on the dependent variable is not possible by any one of the method, we say that experiment is confounded and such an extraneous variable is called a confounding variable

EXPLORATORY RESEARCH DESIGN
Ideally all marketing research projects must start with an exploratory research as this helps in providing a sharper focus of the situation and a clearer definition of the problem at hand. The exploratory research design, as the name suggests, involves getting a feel of the situation and emphasises a discovery of ideas and possible insights that may help in identifying areas of further rigorous study. For example a food product manufacturer, wanting in introduce a breakfast cereal may be in knowing the desirable attributes of such a product before really defining the product concept. The main objective of the exploratory research is to fine tune the broad problem into specific problem statement and generate possible hypotheses. It therefore, gives useful direction for farther research. The exploratory studies are mainly used for:

1) Providing information to enable a more precise problem definition or hypothesis formulation.

2) Establishing research priorities.

3) Giving the researcher a feel of the problem situation and familiarising him with the problem.

4) Collecting information about possible problems in carrying out research, using specific collection tools and specific techniques for analysis.

Since exploratory studies are not conclusive studies, the design of the study is highly flexible and informal. However, rarely ever does formal design exist in case. of exploratory studies. Structured and/or standardised questionnaires are replaced by judgement and intuitive inference drawing on the basis of collected data. Convenience sampling rather than probability sampling characterises exploratory designs. The generally used methods in exploratory research are:

a) Survey of existing literature

b) Survey of experienced individuals

c) Analysis of selected case situations.

3.5 DESCRIPTIVE RESEARCH DESIGN
Primarily in use for preliminary studies, this type of designs are employed to facilitate description and inference building about population parameters and the relationship among two or more variables. Description or inference could be' quantitative or qualitative in nature. Descriptive designs only describe the phenomenon under study attempting to establish a relationship between factors. The data collected may relate to the demographically or the behavioural variables of respondents under study or some situational variables. For example, descriptive research design could be suited to measure the various attributes of successful sales people, or evaluate a training programme or a retailing situation.
The design could be used to study how customers behaved when a new sales promotion programme was introduced. It does not, however, determine the extent of association between the different variables i.e. the income and age of people as associated with response to the sales promotion. Descriptive design can however, be used to draw inferences about the possible relationships between variables.
As descriptive designs are aimed only at providing accurate descriptions of variables relevant to the problem under consideration, they are generally used for preliminary and explorative studies. Very often however, the decision makers choose to accept descriptive data which would permit inference drawing about causality between variables. They may not want to or may not be able to afford experimental studies in terms of time involved and as such descriptive design may at times be used for conclusive studies also.
Descriptive designs are a very frequently used, perhaps the most commonly used category of research design. In short descriptive research can be used for the following purposes.

a) To describe the characteristics of certain groups of interest to the marketer e.g. users of the product, potential users, non users, possible receivers of promotional communication by the company and so on.

b) To estimate the proportion of people in a given population who behave in a certain way for example the proportion of consumers who are prone to deals.

c) To make specific predictions for specified future periods.

d) To develop inferences whether certain variables are associated, for example income and shopping place preference.

In comparison to an exploratory research study, the descriptive study is more formal and less flexible. As the descriptive design in directed at collecting qualitative and quantitative data to enable, description of variables, it involves formulation of more specific hypothesis and testing them through statistical inference of discipline designs may include case research designs, longitudinal and cross sectional designs as well as focus group studies. Of these we shall discuss the panel, cross-sectional and focus group designs.
Panel Research Design
Also known as longitudnal research, the panel design involves the continual or periodic information collection from a fixed panel or sample of respondents. The elements of this panels may be retail outlets, deal9rs, consumers or just individuals. A panel is expected to be constant over time and adjustments are made to provide for dropouts and representativeness. A continual measurement of variables relating to the elements of the panel provides a very good idea of their response patterns and may provide clues for future behaviour .The longitudal analysis used involves repeated measurements of the same variables to facilitate a variety of inferences to be drawn about the behaviour of the elements of the panel. Recently the omnibus panel has also been devised where though a fixed sample of elements is maintained but the information collected from the sample members may vary over time or vary between individuals. For example, at one time information on attitudes with respect to a proposed packaging change may be