1. You have drafted a questionnaire to find out the reasons for decline in sales of a product.

Pretest your questionnaire with hypothetical data. Suggest the modifications required based on the results in the drafted questionnaire. The questionnaire to investigate the decline in sales of a product would normally cover the branding, channel of distribution, promotion, advertising, and so on. The following are the questions that one would normally go through to get the desired response. The following is the drafted questionnaire for Best Luck Ltd, a company that market XXX Cookware sets in Mauritius.

Best Luck Ltd Questionnaire Survey on customers View of the XXX Cookware 1 2

Gender Male Age Group Female

15-20 years

20-30 years

30-40 years

40-50 years

above 50 years

Monthly Household Income

Below 10,000

10,000-20,000

20,000-30,000

30,000-40,000

above 40,000

What is Your Profession?

Student

Job

Own Business

Other

Are you aware of the benefits of having a XXX Cookware at home? Yes No

Do you feel the need of having a hand held massager at home? Yes No

Do you have a XXX Cookware at home? Yes No

If yes, How much did it cost you?

Less than Rs 5,000

Rs 5,000 Rs 10,000

Above Rs 10,000

No Idea

What is its make? . 9 Up to what price will you be willing to purchase XXX Cookware?

Rs 2,000

Rs 5,000

Rs 10,000

Rs 15,000

10 What is your opinion about XXX Brand?

2

Highly Rated

Average

Poor Reputation

No Idea

11 Have you ever tried any of its products? Yes No

12 If yes, have you been satisfied with the quality of its product? Yes No

13 Have you ever seen the XXX Cookware? Yes 14 Are you aware of its selling price? Yes No No

15 Do you know about the different points of sale of the XXX Cookware? Yes No

16 Do you think that there is enough promotion and advertising for the product? Yes No

17 When purchasing an item, what are the essential qualities you look for?

Appearance

Durability

Other

18 Any other comments on the XXX Cookware?

Pretesting The pretest is a valuable indicator of the effectiveness of a questionnaire to collect data. The pretesting of questionnaire consists in selecting, approaching and interviewing a small segment in the same manner to be followed in the full scale operation and then analyzing the results in the light of the objectives of the study. Pretesting helps us to understand whether the replies provide the type of information needed or whether the respondents are misinterpreting any of the questions. In addition, results obtained in the pretest can at times suggest new ideas or hypotheses worthy of further examination. If a pretest indicates any change of importance, a further pretest may be warranted to review the questionnaire. Thus, the mere fact that the wording of a question originally misunderstood has been changed does not in itself ensure the clarity of the new form. In such cases a few interviews with the new question form are desirable. Pretests are sometimes conducted in an informal manner. The previously drafted questionnaire was pretested with the collaboration of 12 respondents, each having different academic qualifications and background, using then interviewing techniques and the following which it was noted that the questionnaire is too structured. However, on the positive side, it was noted that the questions were very clear to the respondents and no misinterpretation of question occurred. The amended version of the questionnaire is as follows:

Amended Questionnaire

Best Luck Ltd Questionnaire Survey on customers View of the XXX Cookware 1 Gender Male 2 Age Group Female

15-20 years

20-30 years

30-40 years

40-50 years

above 50 years

Monthly Household Income

Below 10,000

10,000-20,000

20,000-30,000

30,000-40,000

above 40,000

What is Your Profession? .

Are you aware of the benefits of having a XXX Cookware? Yes No

Do you feel the need of having a XXX Cookware?

Yes

No

Do you have a XXX Cookware at home? Yes No

If yes, How much did it cost you? . What is its make? .

Up to what price will you be willing to purchase XXX Cookware?

10 What is your opinion about the XXX Brand? ............................. ..............................

11 Have you ever tried any product of its product line? Which one?

12 If yes, have you been satisfied with the quality of its product?

Yes

No

13 Have you ever seen the XXX Cookware? Yes 14 Are you aware of its selling price? Yes No No

15 Do you know about the different points of sale of the XXX Cookware? Yes No

16 Do you think that there is enough promotion and advertising for the product? Yes No

17 When purchasing an item, what are the essential qualities you look for? .. .. ..

18 Any other comments on the XXX Cookware?

2. What is a Semantic Differential Scale? Explain the steps in construction of the scale. When will you use this scale? The semantic differential scale (SDS) is a scaling tool which has been used frequently for measuring social attitudes, particularly in the fields of linguistics and social psychology. Typically the scale is a seven- point bipolar rating scale using adjectival opposites, although some studies have used five- and six-point scales. The SDS has been used extensively in language attitude studies as a means of measuring subjects attitudes towards various languages, dialects, accents, as well as the speakers of different varieties. Usually a tape is played of various speakers and subjects are asked to listen to the voices and make judgments about the speakers personality, looks, etc. using an SDS. For example, a subject might be asked to rate a speech sample in terms of a number of personality features. A feature such as honesty, for example, would be represented by the semantic differential scale in the following form: honest 1 2 3 4 5 6 7 dishonest There are nearly always at least eight different pairs of adjectives, and quite often there are a dozen or more. This scale has the advantage that the subjects can indicate whether they judge the speaker to be extremely honest or dishonest by marking the extremities (1 or 7, respectively), or whether they have not formed an opinion of the speakers honesty by marking position 4, a neutral position half way between the two extremes. One advantage of this technique is that it forces subjects to focus on the expected dimensions since the categories are already provided another advantage is that it is relatively easily implemented. It is not a difficult concept for lay people to understand since essentially all you are asking them to do is to decide whether they think a speaker is or isnt x, or they dont know. Nor is the procedure complex since all that is required is for the subject to make some sort of judgment and circle a number. Therefore the SDS still appears to be among the best available instruments for measuring multicomponent concepts and has been employed very successfully in attitude studies. There are two main components of a semantic differential scale: (1) The adjectives; and

(2) The layout of the scale itself. In designing a scale which can be used easily by subjects as well as providing meaningful results for the researcher, the researcher must make a number of decisions regarding both of these components. I will begin with the selection of adjectives since this is the most crucial aspect. Selection of Adjectives In selecting adjectives, the researcher should not simply think up a range of adjectives he/she thinks might sound adequate and use them straight away in a study. There are certain procedures which should be followed to maximise the relevance of traits in order to make the data meaningful. It is standard and almost essential that a pilot study be carried out in order to ascertain which measures are most appropriate for the study at hand. The first step was to conduct a pilot study to elicit adjectives from a sample group. This sample group would be played a tape of speech samples and then be encouraged to discuss freely their impressions of the speakers they had just heard. He then devised a prototype SDS using the most common adjectives elicited from these discussions. This scale was then tested using another subject group. Finally, responses were quantified and inter-correlated in order to determine whether fewer or more basic dimensions could be identified so a final SDS could be prepared to be used in the main study. While this process is a good way of ensuring meaningful adjectives, unfortunately it does not usually elicit many antonyms for the most common (and relevant) adjectives. These are left for the researcher to devise. Antonyms

A decision needs to be made regarding whether the antonyms are going to be complementary opposites (for example honestdishonest) or whether they are going to be gradable and therefore more subtle, for example (entertaining boring). At first glance using complementary antonyms seems more straight forward, and it can be a simple process for many adjectives, especially those on which a negative morpheme can be added (successful unsuccessful; honest dishonest; patientimpatient; educateduneducated; sincere insincere; kind
9

unkind; reliableunreliable). Alternatively, the negator not can precede just about any adjective (richnot rich; goodnot good; generousnot generous), but this is less satisfactory because it does not really tell you very much. For example, not rich does not necessarily mean the same as poor. While the points on the scale allow for degrees of judgement, there is no way of telling whether the subject considers notr rich to mean poor or perhaps of middle income.

For adjectives such as these gradable antonyms might be more meaningful. For example, not generous does not necessarily mean stingy, it could just mean careful or thrifty, and therefore stingy is more meaningful than not generous. Similarly, sophisticated is a less ambiguous antonym for common than uncommon (which can also mean rare). This raises the very difficult problem of which antonym to choose because it is not always clear which adjectives are more appropriate. A person who is not friendly may be a snob, or shy, or neither. Perfectionist and highly strung are two antonyms for easy-going, yet the former is relating to individuals actions, the latter to their temperament. Dishonest, two-faced, and deceitful are all antonyms for both trustworthy and sincere. It can also be seen even from this small sample that some antonyms are more appropriate than others. For example uneducated is a more appropriate antonym for educated than unskilled because being uneducated does not necessarily mean being unskilled. It can be seen that sometimes it is better to use complementary antonyms, and at other times gradable ones. However, there is no reason why a combination of complementary and gradable antonyms could not be used on the same scale. In fact it may be wiser to do so since both have their advantages and disadvantages, depending on the particular adjectives selected. Homonyms and ambiguities

It is tempting to take adjectives from someone elses study thinking that since someone else has used them then they must be okay. However, as we shall see below, there can be pitfalls in doing this. Apart from eliciting adjectives, as part of my pilot study I decided to also test out some adjectives used in previous studies. One unforeseen complication arose with the adjective pair intelligentdull. I found the feature dull was misinterpreted due to a voice quality of one
10

speaker. Because this speaker had a quiet, monotone voice, it was suspected that at least one subject considered dull to mean boring rather than unintelligent. This was apparent since the majority of subjects rated this particular speaker as being somewhat intelligent, although quite unsocial.

Another point to note is that the meaning of the adjectives must be clear or they are likely to be misinterpreted by the subjects. The following is the list of adjectives which they selected from prior literature: dullsharp complexsimple positivenegative smartdumb worthlessvaluable difficulteasy goodbad carefulsloppy weakstrong thinthick roughsmooth fastslow

It is difficult to imagine what is meant when a speaker is described as thin or thick, or as rough or smooth. Even the positivenegative dimension is somewhat ambiguous. Does it equate to optimisticpessimistic or as confidenthesitant? Is dullsharp referring to intelligence, or wits, or perhaps to a sharp tongue? Does complexsimple mean sophisticatedcommon or fastidiouseasy-going? If it means the latter, what is the difference between complexsimple and difficulteasy? Are the dimensions of weak strong, fast slow referring to physical or mental traits? Is smooth a positive or negative trait (where it can have connotations of sleazy or sneaky)? Clearly there are a lot of ambiguities in this scale which means that the results it would provide would be almost impossible to interpret.

11

Sometimes an adjective may be used which is not so much ambiguous but may be simply uninterpretable. Of course, another problem which may arise is that the adjectives may simply not be relevant for your particular subject group. The notion of relevance is discussed in the following section. Relevance

Relevance to subjects

Deciding which adjectives are appropriate will depend upon the subject group, the aims of the study, and all the relevant social factors. Here it becomes even clearer why it may be more prudent to take the time and elicit adjectives openendedly from a sample group and test them in a pilot study as Williams (1974) suggested. It is important to have adjectives which will be both meaningful to the study at hand and have some relevance to the particular subject group. Subject groups for the pilot study must be selected using the same criteria as the subjects to be used in the main study and the same stimulus speakers should also be used. This procedure is aimed at testing and eliminating any adjectives which may be inappropriate. Selecting groups according to the criteria to be used in the main study is important because what is relevant to one group might not be relevant to another. Relevance to study In attitude studies the adjective pairs are broken down into underlying dimensions either by factor analysis or they can be selected a priori on the basis of previous research and sociolinguistic theory . There have been numerous ways of dividing features. The more popular ones have been evaluation, potency, and activity; competence, personal integrity and social attractiveness; power and solidarity .

12

If the a priori method is chosen then care must be taken to ensure that the adjectives are truly representative of the dimensions selected. For example, the adjectives beautifulugly would be much more suited to the dimension of social or physical attractiveness than to the dimensions of power and solidarity.

Number of adjectives There seems to have been no real correlation up to now between the numbers of speakers the subjects are required to listen to and the number of traits they must decide on when marking an SDS. While it is necessary to include enough adjectives to provide the researcher with the desired information, there should not be so many that the task becomes tedious for the subjects. The number chosen should therefore partly depend upon the number of voices the subjects are required to listen to. A large number of speakers and a large number of traits would mean that the subjects are much more likely to become bored than if one of these factors were reduced. The scale would be more effective with either fewer traits or fewer speakers. The subjects would feel less fatigued, and results would be more likely to be an accurate reflection of their true opinions, resulting in more meaningful data.

Of course this does not mean that if there are only two or three speakers then 40 or 50 traits can be used! It is unlikely that so many would be useful in an attitude study anyway. A little judgement and common sense must also prevail. Layout of scale

Positive or negative polarity

While the procedure of filling in an SDS is fairly straight forward, careful thought about the arrangement of the polarity can make the task easier on the subjects. Once again, this is not necessarily a straight forward procedure. Three main areas of difficulty which can arise are
13

discussed positively or negatively valued.

below.

(1) As illustrated earlier, one weakness of SDS is that it may not always be clear if a trait is

(2) This example also illustrates the fact that adjectives can be gender- dependent. For example, an adjective such as tough may be viewed positively by males but negatively by females, regardless of the gender of the speaker, or it may be viewed negatively when judging females but not when judging males. It may also depend upon the social network of the individual. (3) Another difficulty is that some features do not necessarily have a negative/positive polarity. For example, features such as activepassive are not intrinsically positive or negative. Five- six- seven- nine-point scales

Another decision which needs to be made is how many points of choice should be on the scale. A seven-point scale, on the other hand, has the advantages of allowing neutrality and has enough gradation to give meaningful data, yet not be too tedious. Circle numbers or tick lines This may be a relatively minor point to consider, but choices here can make life a little easier on the subjects. Whether numbers or lines are used (see Fig. 1 below) may depend upon the scales polarity. For example, if all the positive attributes are listed on one side and all the negative attributes are listed on the other, then numbers 1 and 7 are consistently associated with one polarity or the other and so the choice between numbers and lines will probably make no difference. On the other hand, if the polarity is randomised, numbers could be confusing as sometimes a 7 will refer to a positive polarity and other times a negative polarity. In this case lines might be more appropriate. Numbers: Educated 1 2 3 4 5 6 7 Uneducated Lines: Educated Uneducated

14

Conclusion It can be seen that there are many considerations at every step when devising a semantic differential scale. The decisions made can be crucial in eliciting data which is analysable, meaningful, and relevant to the study at hand. Although the choices made depend on the subjects and aims of the study, it has been shown that there are many ways to improve the quality of an SDS. Conducting pilot studies is crucial to both elicit and test adjectives. Antonyms (both gradable and complementary) should be selected according to those which are the most meaningful, appropriate, and relevant to the aims of the study and to the subjects. Homonyms and ambiguous antonyms should also be avoided. Finally, a combination of a seven-point scale, randomised polarity and lines instead of numbers on the scale also seem to be the best way of laying out the scale.

15

3.

Discuss the role of modeling in research in managerial decisions making with an

appropriate illustration. How is model validation done? A manager, whichever type of organisation he works in, very often faces situations where he has to decide/chose among two or more alternative course of action. These are called decisionmaking situations. The different types of managerial decisions have been categorized in the following manner: Routine/Repetitive/Programmable v/s Nonroutine/Nonprogrammable Decisions Operating v/s Strategic Decisions

y y

The routine/repetitive/programmable decisions are those which can be taken care of by the manager by resorting to standard operating procedures. Such decisions the manager has to take fairly often and he/she knows the information to facilitate them. Usually the decision maker has knowledge in the form of this is what you do or this is how you process for such decisionmaking situations. Examples of these decisions could be processing a loan application in a financial institution and supplier selection by materials manager in a manufacturing organization. The non-repetitive/ non-programmable/ strategic decisions are those which have a fairly longterm effect in an organization. Their characteristics are such that no routine methods, in terms of standard operating procedures, can be developed for taking care of them. This element of subjectivity/ judgements in such decision making is fairly high. Since the type of problem faced by the decision maker may vary considerably from one situation to another, the information needs and the processing required to arrive at the decision may also be quite different. The decision making process followed may consist, broadly, of some or all of the steps given below: 1. Problem Definition; 2. Identifying Objectives, criteria and goals; 3. Generation/ Enumeration of alternative courses of action;
16

4. Evaluation of alternatives; 5. Selection/choosing the best alternative; 6. Implementation of the selected alternative. All the above steps are critical in decision-making situations. However, in the fourth and fifth steps; i.e., evaluation and selection, models play a fairly important role.

Models ad Modeling Many managerial decision-making situations in organizations are quite complex. So, managers often take recourse to models to arrive at decisions.

Model The term model has several connotations. The dictionary meaning of this word is a representation of a thing. It is also defined as the body of information about a system gathered for the purpose of studying the system. It is also stated as the specification of a set of variables and their interrelationships, designed to represent some real system or process in whole or in part.

Modeling Models can be understood in terms of their structure and purpose. The purpose of modeling for managers is to help them in different decision-making. The term structure in models refers to the relationships of the different components of the model. In case of large, complex and untried problem situations the manager is wary about taking decisions based on intuitions. A wrong decision can possibly land the organization in dire straits. Here modeling comes in handy. It is possible for the manager to model the decision-making situation and try out the alternatives on it to enable him to select the best one. This can be compared to non-destructive testing in case of manufacturing organizations.

17

Presentation of Models There are different forms through which models can be presented. They are as follows: 1. Verbal or prose models. 2. Graphical/ conceptual models. 3. Mathematical models. 4. Logical flow models. Verbal Models The verbal models use every day English as the language of representation. An example of such model from the area of materials management would be as follows: The price of materials is related to the quantum of purchases for many items. As the quantum of purchase increases, the unit procurement price exhibits a decrease in a step-wise fashion. However, beyond a particular price level, no further discounts are available. Verbal models use words to represent some object or situation that exists, or could exist, in reality. Verbal models may range from a simple word presentation of scenery described in a book to a complex business decision problem (described in words and numbers). A firm's mission statement is a model of its beliefs about what business it is in and sets the stage for the firm's determination of goals and objectives. Verbal models frequently provide the scenario necessary to indicate that a problem is present and provide all the relevant and necessary information to solve the problem, make recommendations, or at least determine feasible alternatives. Even the cases presented in management textbooks are really verbal models that represent the workings of a business without having to take the student to the firm's actual premises. Oftentimes, these verbal models provide enough information to later depict this problem in mathematical form. In other words, verbal models frequently are converted into mathematical models so that an optimal, or at least functional, solution may be
18

found utilizing some mathematical technique. A look in any mathematics book, operations management book, or management science text generally provides some problems that appear in word form. The job of the student is to convert the word problem into a mathematical problem and seek a solution. Graphical Models The graphical models are more specific than verbal models. They depict the interrelationships between the different variables or parts of the model in diagrammatic or picture form. They improve exposition, facilitate discussions and guide analysis. The development of mathematical models usually follows graphical models.

Mathematical Models The mathematical models describe the relationships between the variables in terms of mathematical equations or inequalities. Most of these include clearly the objectives, the uncertainties and the variables. These models have the following advantages: They can be used for a wide variety of analysis. They can be translated into computer programs. Mathematical models are perhaps the most abstract of the four classifications. These models do not look like their real-life counterparts at all. Mathematical models are built using numbers and symbols that can be transformed into functions, equations, and formulas. They also can be used to build much more complex models such as matrices or linear programming models. The user can then solve the mathematical model (seek an optimal solution) by utilizing simple techniques such as multiplication and addition or more complex techniques such as matrix algebra or Gaussian elimination. Since mathematical models frequently are easy to manipulate, they are appropriate for use with calculators and computer programs. Mathematical models can be classified according to use (description or optimization), degree of randomness (deterministic and stochastic), and degree of specificity (specific or general). Following is a more detailed discussion of mathematical model types.
19

Types of Mathematical Models Descriptive Models Descriptive models are used to merely describe something mathematically. Common statistical models in this category include the mean, median, mode, range, and standard deviation. Consequently, these phrases are called "descriptive statistics." Balance sheets, income statements, and financial ratios also are descriptive in nature. Optimization Models Optimization models are used to find an optimal solution. The linear programming models are mathematical representations of constrained optimization problems. These models share certain common characteristics. Knowledge of these characteristics enables us to recognize problems that can be solved using linear programming.

Deterministic Model Deterministic models are those for which the value of their variables is known with certainty. In a previous example, the manager knew profit margins and constraint values with certainty. This makes the linear programming model a deterministic optimization model. Models that have values that are not known with certainty are said to be stochastic or probabilistic models. For example, a manufacturer that is having trouble deciding whether to build a large or small facility knows that the solution to this capacity problem depends upon the volume of demand that materializes. High demand would require a large facility while low demand would require a small facility. While the manufacturer has no way of knowing with certainty what demand will be, it can at least determine the
20

probability of the occurrence of each. For example, if the manufacturer estimates that the probability of the occurrence of high demand is 70 percent and the occurrence of low demand is 30 percent, it can use this information along with the monetary value (expected payoff) of each situation to construct mathematical models such as payoff matrices or decision trees to find an optimal decision This type of model can be said to be a stochastic optimization model. Some models can even be very similar with the degree of randomness being the key differentiator. For example, in project management techniques, program evaluation and review technique (PERT) and the critical path method (CPM) are very similar except that CPM is used whenever the required time to complete the activities is known and PERT is used whenever the required activity times are not known but can be estimated. CPM is considered to be deterministic while PERT generally is said to be probabilistic. Once the activity times are established, the two techniques are virtually the same throughout the remainder of the problem's completion. Logical Flow Models The logical flow models are a special class of diagrammatic models. Here, the model is expressed in form of symbols which are usually used in computer programming and software development. These models are very useful for situations which require multiple decision points and alternative paths. These models, once one is familiar with the symbols used, are fairly easy to follow.

Benefits of Model Use The goal of modeling use is to adequately portray realistic phenomenon. Once developed properly, a great deal can be learned about the real-life counterpart by manipulating a model's variables and observing the results.

21

Real-world decisions involve an overwhelming amount of detail, much of which may be irrelevant for a particular problem or decision. Models allow the user to eliminate the unimportant details so that the user can concentrate on the relevant decision variables that are present in a situation. This increases the opportunity to fully understand the problem and its solution. In his book, Operations Management, William J. Stevenson lists nine benefits of models: Models generally are easy to use and less expensive than dealing with the actual situation. Models require users to organize and sometimes quantify information and, in the process, often indicate areas where additional information is needed. Models provide a systematic approach to problem solving. Models increase understanding of the problem. Models enable managers to analyze "what if" questions. Models require users to be very specific about objectives. Models serve as a consistent tool for evaluation. Models enable users to bring the power of mathematics to bear on a problem. Models provide a standardized format for analyzing a problem.

Model Construction The accuracy of the results of the model analysis is dependent upon how well the model represents reality. The closer the model is to its actual counterpart, the more accurate the conclusions drawn and the predictions made about the object of attention. Hence, the model user must strive for the most accurate representation possible. Model users also must be careful to identify the decision variable values that provide the best output for the model. This is referred to as the model's optimal solution. However, the model user also must be careful not to include
22

irrelevant variables that may cloud the picture and cause inaccurate conclusions or force the model user to spend an unnecessary amount of time in analysis. In their book Operations Management: Concepts, Methods, and Strategies, Mark Vonderembse and Gregory White present a step-by-step process for successfully building a useful model: Define the problem, decision, situation, or scenario and the factors that influence it. Select criteria to guide the decision, and establish objectives. A perfect example of this is the use of heuristics in assembly-line balancing to guide the decision and the criteria of maximizing efficiency/minimizing idle time as an objective. Formulate a model that helps management to understand the relationships between the influential factors and the objectives the firm is trying to achieve. Collect relevant data while trying to avoid the incorporation of superfluous information into the model. Identify and evaluate alternatives. Once again, the example of assembly-line balancing is appropriate. The user can manipulate the model by changing the heuristics and comparing the final results, ultimately finding an optimal solution through trial-and-error. However, the production of alternatives may not be necessary if the model in use initially finds an optimal solution. Select the best alternative Implement the alternative or reevaluate If the user is not familiar with models and their use, he would be wise to study the variety of models that are available for use and seek to understand their purpose and how each is used to generate results. Additionally, the user would be well served to learn how the individual model's results are interpreted and used, and what assumptions and limitations apply to each.

Advantages and Disadvantages of using Models

23

Models provide the most effective means developed for predicting performance. It is hard to conceive a prediction system that is not finally a model. To construct a model of a real process or system, careful consideration of the system elements that must be abstracted is required. This in itself usually is a profitable activity, for it develops insights into the problem. When building a model, we are immediately struck with the magnitude of our ignorance. What do we really know? Where are the gaps in available data? It is often impractical or impossible to manipulate the real world system in order to determine the effect of certain variables. Business systems are typical, for in order to use the business system itself as a laboratory could be disastrous and very costly. The dangers in using predictive models lie in the possibility of oversimplifying problems to keep models in workable form. The decision maker may place too much faith in a seemingly rigorous and complete analysis. It is important for the model user to realize that model development and model solution are not completely separable. While the most accurate representation possible may seem desirable, the user still must be able to find a solution to the modeled problem. Model users need to remember that they are attempting to simplify complex problems so that they may be analyzed easily, quickly, and inexpensively without actually having to perform the task. Also desirable is a model that allows the user to manipulate the variables so that "what if" questions can be answered. Models come in many varieties and forms, ranging from the simple and crude to the elegant and exotic. Whatever category they are in, all models share the distinction of being simplifications of more complex realities that should, with proper use, result in a useful decision-making aid. Models are important and widely used in management. Marketing managers utilize the product life cycle model to facilitate understanding of the phases of product life. Accounting managers use ratios, such as the current ratio and the quick ratio, to quickly grasp the ability of an organization to pay its bills in the short term. Information systems managers have flow diagrams to depict the logic needed to develop a computerized order-entry system. Financial managers use net present value and internal rate of return in analyzing investment alternatives. Operations managers have precedence diagrams, decision trees, lot sizing models, material requirements planning, assembly-line balancing, and a host of other models they can use to make better decisions. Organizational performance is a result of the decisions that management makes. Models make these decisions easier to understand and often can lead to an optimal choice.
24

Role of Modeling in Research in Managerial Decision-Making: An Illustration The example we will consider here is the case of a co-operative state level milk and milk products marketing federation. The federation has a number of district level dairies affiliated to it, each having capacity to process raw milk and convert it into a number of milk products like cheese, butter, milk powders, ghee, etc The diagrammatic model of the processes in this setup is depicted in Figure 1(see next page). The typical problems faced by the managers in such organizations are that: (a) the amount of milk procurement by the individual district dairies is uncertain, (b) there are limited processing capacities for different products, and (c) the product demands are uncertain and show large fluctuations across seasons, months and even weekdays. The type of decisions which have to be made in such a set up can be viewed as a combination of short/intermediate term and long term ones. The short term decisions are typically product-mix decisions like deciding: (1) where to produce which product and (2) when to produce it. The profitability of the organization depends to a great extent on the ability of the management to make these decisions optimally. The long term decisions relate to (1) the capacity creation decisions such as which type of new capacity to create, when, and at which location(s) and (2) which new products to go in for. Needless to say, this is rather complex decision-making situation and intuitive or experience based decisions may be way off from the optimal ones. Modeling of the decision-making process and the interrelationships here can prove very useful. In absence of a large integrated model, a researcher could attempt to model different subsystems in this set up. For instance, time series forecasting based models could prove useful for taking care of the milk procurement subsystem; for the product demand forecasting one could take recourse, again, to time series or regression based models; and for product-mix decisions one could develop Linear Programming based models.

25

Figure 1: A diagrammatic model of the interrelationships in a Mix Marketing Federation

Model Validation When a model of a decision making situation is ready a final question about its validity should always be raised. The modeler should check whether the model represents the real life situation and is of use to the decision maker.

26

A number of criteria have been proposed for model validation. The ones which are considered important for managerial applications are face validity, statistical validity and use validity. In face validity, among other things, we are concerned about the validity of the model structure. One attempts to find whether the model does things which are consistent with managerial experience and intuition. This improves the likelihood of the model actually being used. In statistical validity we try to evaluate the quality of relationships being used in the model. The use validity criteria may vary with the intended use of the model. For instance, for descriptive models one would place emphasis on face validity and goodness of fit.

27

4. Write short note on the following: a. Factor loading and factor analysis Factor analysis Factor analysis is used to uncover the latent structure (dimensions) of a set of variables. It reduces attribute space from a larger number of variables to a smaller number of factors. In many scientific fields, particularly behavioral and social sciences, variables such as intelligence or leadership quality cannot be measured directly. Such variables, called latent variables, can be measured by other quantifiable variables, which reflect the underlying variables of interest. Factor analysis attempts to explain the correlations between the observations in terms of the underlying factors, which are not directly observable. Factor analysis closely resembles principal components analysis. Both techniques use linear combinations of variables to explain sets of observations on many variables. In principal components, the intrinsic interest is in the observed variables. The combination of variables is primarily a tool for simplifying the interpretation of the observed variables. In factor analysis, the intrinsic interest is in the underlying factors, the observed variables are relatively of little interest. Linear combinations are formed to derive the factors.

Factor loadings The factor loadings are the correlation coefficients between the variables and factors. Factor loadings are the basis for imputing a label to different factors. Analogous to Pearson's r, the squared factor loading is the percentage of variance in the variable, explained by a factor. The sum of the squared factor loadings for all factors for a given variable is the variance in that variable accounted for by all the factors, and this is called the communality. In complete principal components analysis, with no factors dropped, communality is equal to 1.0, or 100% of the variance of the given variable. The factor analysis model does not extract all the variance; it extracts only that proportion of variance, which is due to the common factors and shared by several items. The proportion of
28

variance of a particular item that is due to common factors (shared with other items) is called communality. The proportion of variance that is unique to each item is then the respective item's total variance minus the communality. Various rotation strategies are proposed in the literature: Varimax, Oblimin, Quartimin, but the most common rotation strategy is the Varimax rotation. The goal of these rotation strategies is to obtain a clear pattern of loadings, i .e., the factors are somehow clearly marked by high loadings for some variables and low loadings for other variables. This general pattern is called Simple Structure. Varimax rotation seeks to maximize the variances of the squared normalized factor loadings across variables for each factor. This is equivalent to maximizing the variances in the columns of the matrix of the squared normalized factor loadings. : b. Different types of experimental design Before an experiment is performed, the question of experimental design must be addressed. Experimental design refers to the manner in which the experiment will be set up, specifically the way the treatments were administered to subjects. Treatments will be defined as quantitatively or qualitatively different levels of experience. For example, in an experiment on the effects of caffeine, the treatment levels might be exposure to different amounts of caffeine, from 0 to .0375 mg. In a very simple experiment there are two levels of treatment; none, called the control condition, and some, called the experimental condition. The type of analysis or hypothesis test used is dependent upon the type of experimental design employed. The two basic types of experimental designs are crossed and nested. Crossed designs In a crossed design each subject sees each level of the treatment conditions. In a very simple experiment, such as one that studies the effects of caffeine on alertness, each subject would be exposed to both a caffeine condition and a no caffeine condition. For example, using the members of a statistics class as subjects, the experiment might be conducted as follows. On the first day of the experiment, the class is divided in half with one half of the class getting coffee with caffeine and the other half getting coffee without caffeine. A measure of alertness is taken for each individual, such as the number of yawns during the class period. On the second day the
29

conditions are reversed; that is, the individuals who received coffee with caffeine are now given coffee without and vice-versa. The size of the effect will be the difference of alertness on the days with and without caffeine. The distinguishing feature of crossed designs is that each individual will have more than one score. The effect occurs within each subject, thus these designs are sometimes referred to as within subjects designs. Crossed designs have two advantages. One, they generally require fewer subjects, because each subject is used a number of times in the experiment. Two, they are more likely to result in a significant effect, given the effects are real. Crossed designs also have disadvantages. One, the experimenter must be concerned about carryover effects. For example, individuals not used to caffeine may still feel the effects of caffeine on the second day, even though they did not receive the drug. Two, the first measurements taken may influence the second. For example, if the measurement of interest was score on a statistics test, taking the test once may influence performance the second time the test is taken. Three, the assumptions necessary when more than two treatment levels are employed in a crossed design may be restrictive.

Nested designs In a nested design, each subject receives one, and only one, treatment condition. The critical difference in the simple experiment described above would be that the experiment would be performed on a single day, with half the individuals receiving coffee with caffeine and half without receiving caffeine. The size of effect in this case is determined by comparing average alertness between the two groups. The major distinguishing feature of nested designs is that each subject has a single score. The effect, if any, occurs between groups of subjects and thus the name between subjects is given to these designs.

30

The relative advantages and disadvantages of nested designs are opposite those of crossed designs. First, carry over effects are not a problem, as individuals are measured only once. Second, the number of subjects needed to discover effects is greater than with crossed designs. Some treatments by their nature are nested. The effect of sex, for example, is necessarily nested. One is either a male or a female, but not both. Current religious preference is another example. Treatment conditions which rely on a pre-existing condition are sometimes called demographic or blocking factors and are usually preferable. Often one compares against a standard or traditional treatment that acts as baseline. Randomization There is an extensive body of mathematical theory that explores the consequences of making the allocation of units to treatments by means of some random mechanism such as tables of random numbers, or the use of randomization devices such as playing cards or dice. Provided the sample size is adequate, the risks associated with random allocation (such as failing to obtain a representative sample in a survey, or having a serious imbalance in a key characteristic between a treatment group and a control group) are calculable and hence can be managed down to an acceptable level. Random does not mean haphazard, and great care must be taken that appropriate random methods are used. Replication Measurements are usually subject to variation, both between repeated measurements and between replicated items or processes. Multiple measurements of replicated items are necessary so the variation can be estimated. Blocking Blocking is the arrangement of experimental units into groups (blocks) that are similar to one another. Blocking reduces known but irrelevant sources of variation between units and thus allows greater precision in the estimation of the source of variation under study. Orthogonality Orthogonality concerns the forms of comparison (contrasts) that can be legitimately and efficiently carried out. Contrasts can be represented by vectors and sets of orthogonal contrasts are uncorrelated and independently distributed if the data are normal. Because of this
31

independence, each orthogonal treatment provides different information to the others. If there are T treatments and T 1 orthogonal contrasts, all the information that can be captured from the experiment is obtainable from the set of contrasts. Factorial experiments

Use of factorial experiments instead of the one-factor-at-a-time method. These are efficient at evaluating the effects and possible interactions of several factors (independent variables). Analysis of the design of experiments was built on the foundation of the analysis of variance , a collection of models in which the observed variance is partitioned into components due to different factors which are estimated and/or tested.

32

Q5. A sample of 48 tools produced by a machine shows the following sequence of good (G) and defective (D) tools:

G G G G G G D D G G G G G G G G G G D D D D G G G G G G D G G G G G G G G G D D G G G G G D G G Test the randomness at the 0.05 significance level. A sample of 48 tools produced by a machine shows the following sequence of good (G) and defective (D) tools: G G G G G G D D G G G G G G G G G G D D D D G G G G G G D G G G G G G G G G D D G G G G G D G G Test the randomness at the 0.05 significance level. H0: The production pattern of tools by the machine, good and defective, is random. H1: The production pattern of tools by the machine, good and defective, is not random. n1 = 38 n2 = 10 r = 13 Since n1 is larger than 20, the normal distribution will be used as an approximation to the sampling distribution of r, with Mean = E (r) = 2n1n2+ 1 n1+n2
33

= 2 x 38 x 10 + 1 38+10 = 16.83 And Standard Deviation = S.D. (r) = [2n1n2 (2n1n2-n1-n2) / (n1+n2)2(n1+n2-1)]1/2 = {2x38x10x[(2x38x10)-38-10] / (38+10)2(38+10-1)}1/2 = {[760 x (760-48)] / (2304)(47)}1/2 = (541120 / 108288)1/2 = (4.99)1/2 = 2.24

Therefore, Z = r - E(r) S.D. (r)

34

13 16.83 2.24

= 1.7098 = 1.71

Conclusion As = 0.05, the critical value of Z is 1.96. As the calculated value of Z is less than the critical

value, the null hypothesis is accepted. Thus, the production pattern of tools by the machine, good and defective, is random.

35