Marketing Research

---

Chapter 4 : Developing the research plan

4.1 Data Source

4.2 Research Approaches

The previous chapter threw some light on the steps in a marketing research process.

This chapter will discuss the stage 3 of the marketing research process with focus on stage 1 and stage 2 i.e data source and Research Approach in detail. Chapters 5 ,6 and 7 will focus on the remaining 3 stages of Step 3 of marketing research.

The five critical aspects in the research plan are

1) Data Source: Secondary and primary data

2) Research approaches: Observation, focus groups, survey, experiments.

3) Research Instruments: Questionnaire

4) Sampling Plan: Sampling unit, Sample size, and Sampling procedure

5) Contact methods: Telephone, mail, personal

 

4.1) Data Source:

There are 2 types of data sources available. They are a) Secondary and b) primary data

Secondary data: This type of data is already available.

Sources of secondary data:

Secondary data: can be collected from both internal and external sources:                                                a) internal sources: data available from sources within the organization are called internal sources. The most common sources of internal data are

1. Sales report
2. Collection report
3. Sales territory information report
4. Absenteeism report
5. Daily model wise production  report
6. Profit-loss statements
7. Balance sheets
8. Invoice
9. Inventory sheets
10. Prior research reports

             B) External sources:                                                                                 1) Census data: collected once in 10 years. This is the most authentic report on the country’s demographic profile  (age, income, sex, occupation etc.).

2) Magazines and newspapers: publish articles on various topics related to business and economy, which can be used for future reference.                                        

3) Internet:  world’s biggest storehouse of information. Using the search engine like goggle, information on any topic can be got.

4) Trade journals: journals by associations provide lots of information about the recent development, past trends and future prospects about the industry . E.g CETMA – Consumer Electronics and Television Manufacturers’ Association comes out with monthly report on the industry.

5) Research agencies: Agencies like ORG-MARG, AC NEILSEN, and IMRB conduct research on an ongoing basis on various industries. The reports generated by them are quite useful but the fees are on the higher side.

6)YELLOW PAGES/DIRECTORIES: Yellow pages in telephone directories have become an established source of information for business firms.

In India the following publications/websites provide useful information

1. Centre for Monitoring Indian Economy (CMIE)
2. Bombay Stock Exchange directory
3. Economic Times
4. Business Standard
5. Financial express
6. Business India
7. Business Today
8. Business World
9. NCAER
10. NRS IV
11. Indiainfoline.com

Advantages and Disadvantages of secondary data

Advantages:

1. Less costly as data is easily available
2. It is faster compared to primary data
3. It provides valuable insights and contextual familiarity with the subject matter.
4. It provides a base on which further information can be collected to update and use it finally.

Disadvantages:

1. Locating appropriate source and finally getting access tot eh data may be time consuming
2. The data available might be too vast and  a lot of time may be spent going through the data
3. The accuracy of secondary data  as well as its reliability would depend on the source
4. It might not be  updated and not much of use  in a dynamically changing environment

Primary Data:

Is the original data collected by the researcher. Here the researcher approaches the respondents either directly or indirectly i.e either the respondent is met personally or through mail or telephone.  Most MR involves atleast some amount of primary data search.

4.2 Research Approaches:

 Research approaches: There are four types of research approach namely--- Observation, focus groups, survey, experiments.

1. Observational Research: Fresh data can be gathered by observing the customers in real settings. This is a type of exploratory research. In this type of research, the researcher just observes the target audience. No questions are asked. The observation can be done at Malls, railway stations, supermarkets etc.

2. Focus – Group Research :

    Focus group generally consists of 6-12 consumers brought together at one place to discuss a topic of interest. It could be discusiing a new product concept, reactions to a new advertising theme for an existing product etc.

Each focus group has a moderator who steers the discussion of the group towards the objectives of the research study without the respondents being aware of it. The entire proceedings of the focus group are recorded to be reviewed later. The meeting is held in an informal surrounding. The participants are paid a small sum for attending the same. This is a type of exploratory research.

3. Survey Research: Companies undertake surveys to learn about people’s knowledge, beliefs, preferences, satisfaction, awareness, attitude etc and to measure these magnitudes in the population.

e.g Jet Airlines might want to survey—1) how many people know Jet, 2)have flown it, 3) prefer it  etc. Normally a survey research is undertaken using a questionnaire.

4. Experimental Research: The most scientifically valid research is experimental research. This calls for

• Selecting matched groups of subjects
• Subjecting them to different treatments
• Controlling extraneous variables
• Checking whether observed response differences are statistically significant.

If the extraneous effects can be controlled/eliminated, the observed effects can be related to the variations in the treatment.

e.g. If Jet might introduce in –flight phone service on one of its regular flights from Mumbai to Delhi at a price of Rs100 per minute.

On the same flight the following day, Jet could announce that the price is Rs. 75 per minute. If the plane carried the same number and type of passengers on each flight, and the day of the week made no difference then any significant difference in the number of calls made could be related to the price charged. The experimental design could be further elaborated by trying other prices, replicating the same price on some other flights and other routes.

To the extent that the design and execution of the experiment eliminate alternative hypotheses that might explain the results, the researcher and marketing managers can have confidence in the conclusions.

There are three types of Research Projects.

1. Exploratory Research: To gather preliminary data to shed light on the real nature of the problem. This is more investigative in nature.

e.g –a) Govt. Of Maharashtra may have research problem like: What are the alternate ways of solving traffic congestion in South Mumbai”

        b) Shopper’s stop: The research problem could be “ How to increase footfalls to Andheri Shopper’s Stop.

The research in this case will be highly flexible, unstructured and qualitative and the researcher begins without any preconceived notions about the conclusions.

2. Descriptive research : is basically about ascertaining certain magnitudes like

1. How many people will make a in – flight  phone call if the rate is Rs. 100 per minute
2. How many people will  drink Frooti 300ml if the price is Rs 12/-

Descriptive research accurately describes certain aspects. For e.g research undertaken to find the market share of different companies, or demographics of the people of the city.

3. Causal (experimental research) : Some research is causal – i.e. to test a cause and effect relationship.

 This research tries to show the cause and effect relationship between 2 variables.

  For e.g , If Maruti drops the price of 800 and its sales increases and all other factors remain constant, then it can be presumed that the increase in sales was due to price drop only.

However before making such inferences, reasonable proof must be available that one variable preceded the other and that there were no other causal factors that could have accounted for the relationship.

The different types of experimental designs are as follows:

1) After Only design:

                           the experimental group’s response after the experiment

E.g Measurement of advertising recall, day after recall, or measurement of increase in sales after a sales promotion scheme.

2) Before After design:         

                                        The experimental group’s reaction is first measured before and after the experimental variable is introduced to them to check out the differences in behavior. The disadvantage is that some uncontrollable variables might be responsible for the change in opinion or behavior.

E.g : A consumer is in  a department store   trying various brands of shirts.

He could be asked for his brand preferences. The sales man at the counter could be instructed to influence the choice towards a particular brand. The brand purchased by the consumer finally indicates whether the sales guys have been successful in influencing the brand choices of the consumer. A consumer’s intended purchase earlier is compared to any changes at the final purchase on account of stimuli like an advertisement, influence of dealer etc.

3) Before after with control group:                                                                                                 The design consists of having a control group that is not subjected to the variable v/s the experimental group that is subjected to the variable. The difference of their differences would give an idea of the extent of uncontrollable variables present.

E.g : In the experiment designed earlier , there could be customers to whom the sales man does not deliberately push a particular brand. This would constitute the control group. The difference between the purchase done by the control group and the purchase of the experimental group indicate whether uncontrollable variables had caused the change in behavior.

4) After only with control group:                                                                                                         In this case, no before measurement is made but only the after measurement is made on the experimental group. The difference between the two would indicate the effect of the experimental variable.

E.g : The experimental group may be sent privileged discount coupons for a sale , whereas the control group is not. The results of the sale would indicate whether the discount coupons are indeed useful in increasing the sales of the product.

5) The latin square design :                                                                 

                                If the effect of a single variable is to be studied over different time periods and different geographic regions , a design which takes into account possible combinations is made and the differences arising on account of either the geographic region or the time period are studied.

E.g : Suppose the owner of franchised outlets like Bata would like to know whether there is a difference in sales occurring on account of differences in three types of display used by three similar Bata Stores, situated in different locations in Bombay for a period of three  months. The design would be

Location\ Month                                1                2                3

Linking Road , Bandra                        A                B                C

Andheri                                        B                C                A

Malad                                                C                A                B

Where A, B,C are different sales displays and 1, 2,3 are different months.

If design A is instrumental in increase in sales in all 3 outlets and design “B”, “C” leads to decrease in sales, then we can conclude that the sales have increased due to design “A”  

Latin Square designs are extensively used to establish cause –effect relationships in marketing situations.

Advantages and disadvantages of

 Exploratory research : observation and focus group :

 

The advantages of focus group

1. Brings out the inner motivations of the consumers
2. Letting the researcher have a first hand experience of the consumer reactions
3. Generating new ideas

Disadvantages

1. Excessive reliance on the skills of the moderator
2. Wrong combination of consumers in focus groups , resulting in no tangible output

Observation method

Advantages

1. The respondent is observed directly so that no chances of bias on account of predisposition of any kind occurs.
2. The willingness of the respondent is not a deterrent in data collection since the respondent is not aware.
3. Categories of respondents like those in the rural areas  with whom it might be difficult to communciate as they might misinterpret the objectives of the research , lend themselves suitable to observation methods.
4. Smaller set of respondents for relatively new product concepts lend themselves more useful to observation methods.

Disadvantages

1) The particular action observed might not be the one in actual normal circumstances but due to some specific reasons on that particular day.

1. It is expensive and time consuming to set up and undertake observational studies
2. The data collected is completely dependent on the skills of the observer and the manner in which he records and interprets them.
3. Very few respondents can be observed owing to time constraints. Also the number of observers required is high.

---

Chapter 5: Research Instruments

5.1 Types of Questionnaires

5.2 Types of Questions

5.3 Scaling techniques/ scales of measurement

5.4 Steps involved in Questionnaire construction

5.5 sample questionnaires

5.6 Consumer PANELS/omnibus panels

The previous chapter covered in detail the first two stages in step 3 of marketing research plan namely—data source and research approach. This chapter will explore the third stage (of step 3) i.e. Research Instruments in greater detail

 There are mainly two types of research instruments. They are the  Questionnaire and mechanical devices

1. Questionnaire: A set of questions logically arranged presented to the respondents to answer. A sample of a questionnaire has been presented for reference.
2. Mechanical and electronic devices like galvanometer to measure interest and emotions, eye cameras to study eye movement , video camera to study physical moments and record verbal communications etc.

Lets look at the questionnaire part in greater detail

5.1 Types of Questionnaires

There are 4 types of questionnaires namely

1. Structured non disguised questionnaire
2. Structured disguised questionnaire
3. Non structured non disguised questionnaire
4. Non structured disguised questionnaire

1. Structured non disguised questionnaire

• Questions are  listed in a pre-arranged order
• Respondents are told about the purpose of collecting information

2. Structured- disguised questionnaire

• Questions are  listed in a pre-arranged order
• Respondents are not told about the purpose of conducting survey

3. Non structured non disguised  questionnaire

• Questions are not structured.
• Researcher is free to ask questions in any sequence he/she wants.

• Respondents are told about the purpose of collecting information

4. Non structured disguised questionnaire

• Questions are not structured
• Researcher is free to ask questions in any sequence he/she wants.

• Respondents are not told about the purpose of conducting survey.

 

5.2 Types of Questions:

The different types of questions that can be asked in a questionnaire

1. Closed ended questions:---

• In the closed-ended type of questions, the respondent is asked to select from a fixed list of replies.
• Respondent has to choose any one of the options given or multiple options
• This facilitates coding and helps in quantifying the answer to the questions
• Respondents don’t have to think much and answer within the options given.

2.) Open ended questions:                                                                                                                                                                 Respondents are free to answer the questions in their own words.

It does not restrict them to choose from the given alternatives as in closed-ended questions

The respondent expresses his/her thoughts in a free wheeling manner.

5.3 Scaling techniques/ scales of measurement

The different types of scales used in a questionnaire are

1. Nominal scale :  

In this scale, numbers are only used as labels, they have no numerical sanctity.         E.g To categorize male and female respondents we could say        a nominal scale of 1 for male and 2 for female.

Other examples could be to indicate categories of any variable which is not be given a numerical significance ---                                                                Religion--- Hindu—1, Muslim--- 2, Christian – 3 etc.

     Education level: H.S.C Pass--- 1, Graduate ---2, P.G--- 3 etc.

        Languages spoken: --- English--- 1, Marathi--- 2 etc.

   

2. Ordinal scale:                                                                                                 ordinal scale variables are ones, which have a meaningful order to them.

E.g. :  A typical marketing variable is ranks given to brands by respondents. These ranks are not interchangeable, as nominal scale labels are.

This is because rank 1 is higher then rank 2 and so on.

The distance between each rank is not known. Ranking simply denotes that rank 1 is higher then rank 2, rank 2 is higher then rank 3 , but by how much is unknown.

e.g Rank the following brands in the colour TV market .

The brands could be LG, SAMSUNG, VIDEOCON, SANSUI, SONY, PHILIPS . A consumer could give the following ranks

RANK                BRAND

1. LG
2. SAMSUNG
3. SONY
4. PHILIPS
5. VIDEOCON
6. SANSUI

3. Interval scale (Rating scale):                                                                 Most of the behavioral measurement scales used to measure attitudes of respondents on a scale of 1 to 5 or 1 to 7 are interval scales.

  The difference between interval and ordinal scale variables is that the distance between 1 and 2 is the same as distance between 2 and 3 and so on.

For e.g :

For brand LG for colour TV, please rate on a scale of 1 to 5 on the following features

A consumer may give the ratings as follows:---

4. Ratio scale :                                                                                 In a ratio type scale , there is a unique zero or beginning point.  Interval scale doesnot have a unique zero. Also the ratio of two values of the scale corresponds to the same ratio among the measured values.

E.g distance is a ratio scaled variable.  Starting point is zero.  2 metres is to 1 metre as 2km is to 1 km.

Some of the common ratio scaled variables are--- age, height, length, weight and income.

Other Attitude scales

1. Likert or agreement scale: A statement or series of statements with which the respondent shows the amount of agreement/disagreement.

E.g Inorbit Mall is the most attractive Mall in Mumbai                                        Strongly disagree        disagree             neither agree or disagree

     Agree                         strongly agree

2. Semantic differential scale: A scale connecting two bipolar words , where the respondent selects the point that represents his/her opinion.

E.g   Indian Airlines

Modern         _ _ _ _ _ _ _  Old- fashioned

Air hostesses

Courteous _ _ _ _ _ _ _  Rude

3) Importance scale : A scale that rates the importance of some attribute                         e.g        Airline food service to me is

Extremely important          very Important                 somewhat important

Not very important            Not at all important

4)Intention –to –buy scale :  A scale that describes the respondents intention to buy.                E.g    If an inflight telephone service was available on along flight , I would                         Definitely buy          Probably buy        Not sure        Probably not buy

                Definitely not buy.

5. Projective techniques

1. Word association : Words are presented , one at a time and respondents mention the first word which comes to their mind

E.g :--  What is the first word which comes to your mind when you hear the following :                                                                

Airlines : ____________

          Jet Airways:___________

          Air Deccan :___________

Travel :_______________

2. Sentence completion : An incomplete sentence is presented and respondents complete the sentence .

E.g When I choose my airline the most important consideration in my decision is ______________________________________.

3. Story completion : An incomplete story is presented and the respondents are asked to complete it                                                                 
4. Picture completion :  A picture of two characters is presented , each one making a statement . Respondents are asked to identify with the other and fill in the empty balloon.                        

e)  Thematic Apperception test (TAT) :                                                          A picture or a series of pictures is/are presented and the respondents are asked to make up a story about what they think is happening or may happen in the picture.

Specialised scales

1. Semantic differential scale : This scale is widely used for developing a company profile or brand profile. The profile is generally described by 4-8 pairs of bipolar adjectives, separated from each other by a scale having seven equidistant segments. Each respondent is asked to rate the company/brand on the opposing adjectives. The average score over all respondents is then obtained for all the adjectives. This gives the graph of the company/brand profile. Comparison between companies/brands can also be made using the profiles so obtained.

e.g a  study was done to obtain the profile of four major players in the real estate market. The major players are Lokhandwalla, Evershine, Hiranandani and Oberoi. The adjectives were

1. not reliable _ _ _ _ _ _ _  very  reliable
2. traditional _ _ _  _ _ _ _  modern
3. economical _ _ _ _  _ _ _ costly
4. medium quality construction _ _ _  _ _ _ _ superior quality construction

2) Thurstone scale: This is based on a logical procedure to measure the attitude of the respondent towards a certain phenomenon or behavior. E.g. we might be interested in finding about the attitude of respondents towards the use of credit cards or going on vacation, or savings and investment etc. A collection of about 100 or more statements relating to the attitude are made in general. A panel consisting of 15 to 20 judges are independently asked to sort these statements into 11 piles, from the most favorable towards the attitude to the most unfavorable.

e.g Suppose we are interested in the attitude of a certain socio-economic class of respondents towards savings and investments. The final list of statements could appear somewhat like this:

1. One should live for the present and not for the future. So savings are an absolute no.
2. There exist many attractions to spend the savings on.
3. It is better to spend savings than risk them in investments.
4. Investments are unsafe and block money
5. You earn to spend and not to invest
6. It is not possible to save these days
7. Certain fixed amount of the income should be saved and invested
8. The future is uncertain and investments are a safeguard  against it
9. Some amount of savings and investment is a must for every earning individual
10. One should endeavor to always save more so that much more can be invested.
11. All the savings should be invested for the future

A respondent agreeing with statements 8,9,11 would be deemed to have a favorable attitude towards savings and investment.

3) Constant sum scale

The importance placed by a respondent on one brand v/s the others or one attribute v/s others is measured in relative terms using the constant sum scale. The respondents are required to divide a constant sum  (usually 100) between brands or attributes as per the importance he attaches to them. The average response over all respondents would give the final score for each brand/attribute.

Suppose a researcher is interested in knowing the importance a respondent places on various attributes while taking a decision to purchase a laptop. While a rank order data would give the ranks of the various attributes, it would fail to suggest the importance of each attribute. A constant sum scale helps in finding the importance. A constant sum scale question would be framed as follows---

Divide 100 points among the attributes listed, so that the division will reflect how important each attribute is to you in your decision to buy a laptop.

A particular response from one respondent is like given below

Evaluation criteria                    Importance                                                        price                                5                                                     Processor                        30                                                                  Display quality                20                                                Memory                        25                                                        After sales support                10                                                                        Weight                        10                                                                        Total                        100

4) Paired comparison: When brands are to be compared with one another for a given attribute , they can be placed in a matrix form and the respondents can be asked to judge each pair and tick the one which he feels fares better on the attribute being considered . The proportion of respondents for each combination is then worked out. For example, the respondent may be asked to give their preferences for each pair of 4 brands of Colour TV on the product attribute -  sound quality. The four brands mean 4*3/2= 6 pairs (In general n brands means == n*(n-1)/2 pairs). The matrix would look like this

For example , if for the pair Samsung, LG, the sound quality of Samsung is better then LG , he would place a tick mark or circle in first cell of second row. If LG is better then Samsung then he would place a tick mark in row 1 cell 2.

5) Multidimensional scaling (MDS)

Consumers evaluate brand on multiple attributes simultaneously. This is done by comparing various brands for their similarities/dissimilarities. In other words, the consumer relates one brand with the other, based on certain imagined distance between the two. MDS seeks to measure this distance to determine the approximate position of each brand in the consumer’s mind. This graph of brand position is termed as perceptual map. It may be two dimensional, three dimensional or even more. The dimensions are the attributes over which the brands have been evaluated. These are to be interpreted by the researcher.

Suppose there exists 6 brands for a given product category. They can be paired two at a time as 15 pairs (6*5/2) which are distinct. The most similar pair would have the least distance of 1 and the most dissimilar pair would have the highest distance of 15. The respondents are asked to rank all the 15 pairs. The standard package for MDS  is then used to generate a map of two dimensions. MDS has wide application for segmentation, vendor evaluation , brand positioning etc.

e.g The product positioning/perceptual mapping to obtain the relative position of milk food drinks, soft drinks , fruit juices , energizing drinks , syrups/sherbets etc. resulted in the following two dimensional graph

                                                                

---

5.4 Steps involved in Questionnaire construction

.

Dddddd

---

1. Content of individual questions:

The points to be noted here are

1. Is the question required: If the information needs are satisfied by the inclusion of the question, only then it must be used.                                                 For e.g.  If the survey is about an adult product, questions pertaining to children and their ages might not be required
2. Has the respondent experienced the situation described in the questionnaire.:                                                                         The respondent should have an idea about the product/service before he compares various brands.                 For e.g. How do you rate Fun and Frolic time share as against other time shares? The respondent should atleast know what a time share concept is and also Fun and Frolic time share.
3. Is the question taxing the respondent’s memory?. Is he/she likely to remember such information?                                                                        People generally  do not remember daily events which are normal in nature without some aids to recall. The memory of  the respondent is already cluttered. So a question like  “On which page and which issue of the magazine did you see the ad? Will hardly give the correct response. The episodic memory is relatively uncluttered. So events related to important times like marriage, births , promotions, favorite sports etc. Are more likely to be remembered correctly.
4. Will the respondent part with the information voluntarily? Information on personal life, bad habits, status symbol is generally not parted with very honestly . For e.g  questions like        1) What is your age?

2. What is your income?
3. What contraceptive do you use?
4. Do you consume liquor  daily?

                Suitable rating scale / ratio scale  or non structured techniques need to be used.

5. Can a single question be fragmented into small  multiple questions for better understanding ?

             For e.g , Do you think MTV has eroded our cultural values and has an influence on the teenagers in having no respect for their elders.? The question can be split up as : Do you feel MTV has                                                                                        1) eroded our cultural values

        2) influenced disrespectful attitude of teenagers towards elders                        3) both

6. Can a diagram/card/actual prototype of the product be used instead of a written question?                                                                                 Certain  questioning aids like cards, actual advertisements, small packages of the product in question or a video cassette about the problem in question can be used to break the monotony of verbal questions.

2. Types of questions :                                                                                There are two types of questions                                                         a) Open Questions : Questions where a respondent is free to express his opinion or ideas and not required to necessarily pick out from any alternatives which are given are termed as open questions. E.g

1. Name the domestic microwave oven brands that you are aware of.
2. Do you use yellow pages? , why

   Advantages of Open questions

1. They serve as a good introductory or preliminary round of questions. The respondent can get to talking  about hie opinion and also feel important  and knowledgeable.
2. They are less prone to result in prejudiced answers, influenced by choices listed. For e.g , if top –of-mind awareness of brands is wanted , a listing of various brands should not be given.
3. Also  the open-ended responses might result in certain responses which might have not come through jn the exploratory survey

 

Disadvantages

1. They are often time consuming
2. The interviewer might not record responses to the open questions correctly, resulting in interviewer bias.
3. The responses obtained might be so varied , it might not be possible to arrive at any conclusion. For e.g. Out of 100 respondents asked to name the brands that they are aware of , atleast 80 of them name different regional brands.
4. The coding and editing task of summarizing information from open ended questions is often very laborious and not worth the effort.

Closed questions

1. Multiple –choice questions : Questions in which a fixed number of choices are made available to the respondent  as answers are termed as multiple choice questions. The respondent is generally asked to make exactly on choice . In case the number of alternatives or choices  chosen are  more, then he is asked to rank the choices in order of importance.

E.g

1) How did you decide  to buy  the  washing machine                                        a) On your own b) Advised by a friend  c) Advertisements                d) Gift     e) Demonstration of the product.

2) What is your concept of orange juice?                                                        A) Nutritive b) low calorie c) should easily mix d) natural

3) What is a reasonable price you would be ready to pay for an updated version of the yellow pages.                                                                                A) below Rs 50         b) Between Rs. 50-100  C) Between Rs.100-150                 d) between Rs.150-200

Advantages :

1. They are faster to administer
2. They lend themselves to analyses using various statistical techniques
3. The coding , editing  and tabulation process is simplified to a large extent.
4. They are not prone to interviewer bias.

Disadvantages

1. If the number of choices given is too large, the respondents  might get confused.
2. If there are too many such questions , the questionnaire becomes very monotonous.
3. If one of the choices given is don’t know or can’t say , the respondent might conveniently mark that to save any bother.

2. Dichotomous questions This is a form of multiple- choice questions , but the choice is limited to just two alternatives., Yes or No. When very confirmed opinions or events that have occurred in the past have to be obtained, it is advisable to provide just two options. Care should however be taken  to see that there does not exist a very large grey area between yes and no. In that case  a multiple choice separating the two extremes will be more desirable to use like

Definitely no                   May be not           probably yes               definitely yes

E.g of dichotomous question

1. Do you own a washing machine

        Yes          No

Advantages of Dichotomous questions

1. They are easy to administer
2. Their analysis is simple
3. They are not prone to interview bias

Disadvantages

1. The choice between just two alternatives might be too narrow
2. Use of dichotomous questions might deliberately force the respondent for ayes or no , giving incorrect results
3. It cannot be used to measure opinions or attitudes or usage rate as there will always be more then two alternatives in such cases.

3) Wording of the questions                                                                        a) Avoid leading questions : Questions for which the answer is obvious should be avoided

2. Would you like a Rs 5/- off on your favourite brand of tea?
3. Do you like your children to be healthy?

B) Decide whether objective or subjective questions should be used

 Whether it is the opinion of the respondent which is desired or a general thinking of the people.

 E.g : Do you feel you have a better choice and easy accessibility of products/services after using  the yellow pages?

 V/s

Does the use of yellow pages give a better choice and easy accessibility of products/services?

3.  Avoid  generalised questions                                                                        Questions using words like usually, frequently , in general, most often are to be avoided as they mean different things to different people . For e.g   Do you drink liquor frequently?                        For one respondent twice a week might constitute frequent and for another twice a day might be frequent. It is better to describe the frequency in detail in the form of a multiple- choice question.

4.   Sequence of Questions :                                                                        a) Do the opening questions win the interest of the respondent?                        The opening question should arouse the curiosity of the respondent  as well as draw his interest towards the subject matter of the questionnaire.

For e.g. Which are the brands of cola drinks you are aware of and which is your favorite brand. Personal questions are definitely avoided as openers like what is your age, income etc.

B) Are difficult questions ensconced in the body of the questionnaire.?                        Questions pertaining to attitudes are usually put in the middle when  the respondent has been aroused sufficiently about the subject . Towards the end of the questionnaire , a fatigue effect may set in and the respondent might not give correct responses.

4. Are questions on demographic information at the end?                                        A respondent might feel invasion of privacy if a barrage of personal questions is put to him right in the beginning. Also he may be put off and not answer all the other questions properly. On the other hand , towards the end he ahs already  established a rapport with the interviewer or with the questionnaire  )in case of mail survey) and might not hesitate to give information of personal nature also.

5. Layout of questionnaire:                                                                         a) Quality of paper and reproduction: It is very important that the questionnaire is perceived as friendly. The quality of paper used, the type of printing  and reproduction substantiate the authenticity , seriousness and clear mindedness of the researcher.. A cluttered questionnaire with bad reproduction would leave a bad impression and result in a poor or incomplete response.

B) clarity for analysis and administration : The number and order of questions should be in sequence  so that there is an ease of administration . None of the questions should be left out by the interviewer for want of clarity. The choice of various alternatives should be indicated clearly. Instructions for filling up of the questions also should be spelt out in detail for every question.

    C) Precoding : the final responses would be tabulated and subjected to computer analysis. Each and every response possible needs to be coded. The codes are indicated on the questionnaire itself to facilitate the final tabulation and analysis.

6) Pretesting of questionnaire: The prototype of the questionnaire is administered on a small sample of respondents. A number of factors are checked out at this stage.

• Whether the respondent understands all the questions
• The time taken to admini9ster the entire questionnaire
• Whether certain words in the questionnaire need explanation
• Whether the respondent interprets the question correctly
• Does the opening question generate interest
• Does the fatigue effect creep in very early in the questionnaire
• Is the respondent embarrassed about certain questions
• Does he require to think a lot to answer the question
• Are there many open questions taking a lot of time
• Are there any unnecessary questions
• Should some questions need to be added

The deficiencies in the questionnaire are ironed out in the pretest stage. The questionnaire designed is normally administered on a pilot sample.

7) Final draft of questionnaire

After the pretest stage , certain questions may be eliminated altogether, certain  multiple choice questions might have fewer or more categories, certain dichotomous questions may be changed to multiple choice questions , the sequence may be changed, wordings may be changed, certain questions mat be added, certain open ended questions might become close- ended and vice versa

The final instrument so obtained is the QUESTIONNAIRE to be used for the survey.

5.5 sample questionnaire.

 

Let’s presume that LG wants to do a survey to find what consumers feel about LG  as a brand of Colour TV . Also LG wants to know which parameters consumer looks for before  buying an Colour TV. LG also wants to check whether consumers are brand loyal.

A questionnaire can be made on the following pattern

Q.1 Which brand of colour TV do you own ?_____________

Q.2  When did you  buy your colour TV ?____________

Q.3 Which other brands you considered before buying your brand? __________

Q.4 If you were to buy a TV today which brand will you buy and why ?

Q.5  Given below are  a list of  9 features/parameters which consumers normally look for in a colour TV before buying. Rank the same in order of importance . Rank 1 being the most important , rank 9 being the least important.

Features                            Rank

Price

Aesthetics

Sound quality

Picture quality

Country of origin

Brand name

Colour

Unique facilities

Sales promotion schemes

Q.6 Compare the following 3 brands on the features mentioned above . Each feature is to be given a rating of 1 to 5  for each brand . For e.g , For the feature Price , rate the brands seperately on a scale of 1 to 5

Q.7 Among the colour TV brands which advertisement  you like the most and why?

Q.8  As per you which brand of colour TV is the most advanced in terms of technology ?

Q.9 What is the first word which comes to your mind when you hear the following    Colour TV : ____________

LG         :___________

Samsung :___________

Onida :___________________

        Entertainment :_______________

 Some more questions can be added as per the interviewer’s requirement.

 

5.6 Consumer PANELS/omnibus panels

Difference between panel and survey:

In a survey, a fresh sample is selected

In a panel research, the same panel is used again and again for the collection of information.

A panel  comprises of respondents who may be

House wives  ----(for getting information on household products or channel viewing habits etc.)

Households  

Executives --  (for getting information on clothes etc)

Basically information regarding purchases, frequency of purchases, brand preferences can be collected on a regular basis.

The panel maintain their consumption  in a diary provided by the company.

Definition:  “Panel consists of persons, households or business firms who report their purchase activities at periodic intervals and who are selected based on a combination of their willingness and  representativeness.

In a consumer panel, there is a permanent sample of respondents.

Advantages of consumer panel

1. Continuous supply of information: information is provided on an ongoing basis. Any kind of similarity or change in consumption a pattern ca be easily gauged.
2. Longer interviews are possible : as the panel members are met on an ongoing basis , the panel members can be interviewed for a longer time ---- ( the panel members know the interviewer)
3. Economical method: as information is collected from same consumers on a regular basis, it’s comparatively cheaper.
4. Time saving :
5. Reliable information

Disadvantages

1. Biased information : Over a period of time , the panel members may give biased answers
2. Absence of representative character: the small panel can never represent the entire class of consumers. Drawing conclusions based on panel readings can lead to wrong judgements.
3. Panel members drop out : over a period of time , some panel members drop out. In such a scenario, identifying new consumers with similar characteristics is not easy.
4. Limited cooperation: initially the panel members are very cooperative. As time goes by, their cooperation level/ enthusiasm  falls.

---

Types of consumer panels

1. Purchase panel : is useful to study the purchasing habits and trends of consumers The selected consumers are asked to record their purchases in the diaries provided to them.

E.g of purchase panel --- housewives---- to record purchase behavior of items like washing powder, toothpaste, soaps etc.

2)  Audience panel : Many channels want to know hoe their programs are fairing . An audience panel records the viewership/listening patterns  (TV/RADIO). The panel records on a diary or automatically (people meter) the channels they watched in the previous week.        This helps the companies in media planning.

3. Product testing Panel: before launching a product, its tested among the product testing panel. Their views on features, packaging are recorded. This information can be used for future product upgradations.
4. Dealer panel: companies collect information from certain dealers on a regular basis. Information like competitor product and pricing etc are recorded.
5. Retail audit panel: panel of retail outlets whoa re willing to give information about sales and stock on  a regular basis (normally monthly) of all brands in their respective shops/showrooms .

This helps the researcher draw useful inferences . This data is useful in finding out who are the leading brands (market shares), which brand is doing well in which state etc.        ORG MARG does retail audit on an ongoing basis and they publish reports. The companies subscribe to this report. For e.g   For Colour TV/Washing machines the cost is Rs 3,00,000 per year. 12 monthly copies are provided.

Chapter 6   SAMPLING PLAN :

6.1  Census v/s sample

6.2  The preparation of the sampling plan

  6.3 SAMPLING TECHNIQUES

6.3.1  Probability sampling techniques

6.3.2 Non probability sampling techniques

6.4 Types of Errors in Marketing Research

The previous chapter covered in detail stage 3 of step 3 i.e questionnaire design. This chapter will take a close look at stage 4 of step 3 of marketing research namely “SAMPLING PLAN”.

If we understand what sampling plan is all about,  it is important to understand the difference between census and sample

6.1  Census v/s sample

It is possible to do census if the population size is small. Foe e.g , if there are 200 solar coooker owners in a town, it may be possible to meet all of them., If their addresses were available or could be obtained.

In some cases like survey of disributors or dealers ,or even industrial buyers , it may make sense to do a census. Particularly if opinions or buying behaviour of respondents in a small Population are likely to be widely divergent.

But in most cases , if populations are reasonably large or very large , it makes little sense to do a census

On major reason is the cost and other reason it may take too long. Data may arrive too late for decision making.

For the above mentioned reasons , most researches are done on the basis of sampling i.e only some respondents from the population are surveyed.

6.2 The preparation of the sampling plan calls for the following decisions :-

1. Sampling unit: This answers the question----“ Who is to be surveyed” The researcher must define the target population that will be sampled. The sample unit can be housewives, teenagers, executives, business people etc. The sample unit depends on the product/service. For e.g in the Jet airways survey the sample unit would be business travellers, vacation travellers of both gender.
2. Sample size: This answers the question ---“ How many people should be surveyed”.  Large samples give more reliable results then small samples. But the cost may not justify a large sample.  Generally a sample size which is 1-2% of the target population is good enough to generate fairly accurate results.
3. Sampling procedures: This answers ” How should the respondents be chosen”. The two types are a) Probability sampling and b) Non probability sample.

  Some more  terminology’s in Sampling

Sample element: This is the unit about which information is sought by the marketing researcher for further analysis and action The most common sampling element in marketing research is a human respondent who could be a consumer, a dealer, a person exposed to an advertisement, company, household etc.

Population: This is not the entire population of a given geographic area, but the predefined set of potential respondents (elements) in a geographical area for e.g , a population may be defined s  “ all mothers who buy branded baby food in Mumbai”. Or all teenagers who watch MTV in Bangalore/India “

Sampling frame: is a subset of the defined target population from which a realistic sample is selected. For e.g , We may use a telephone directory of Mumbai as a sampling frame to represent the target population defined as “ adult residents of Mumbai” . Obviously there would be number of elements (people) who fit our population definition , but don’t figure in the telephone directory. Similarly some who have moved out of Mumbai , would still be listed.

Thus a sampling frame is usually a practical listing of the population , or  a definition of the elements or areas which can be used for sampling exercise.

Sampling unit :  If individual respondents form the sample elements  and if we select some individuals in a single step , the sampling unit is also the element. But in most research, there is a multistage selection.

For e.g, We may first select specific areas or blocks in a city. These form the first stage sampling units. Next, we may select specific streets within a block or area and these are called second stage sampling unit. Then we may select apartments or houses---- third stage sampling units. At the last stage we reach the individual sampling element --- the respondent.

6.3 SAMPLING TECHNIQUES

There are two methods of sampling. They are

1. Probability sampling and
2. Non-probability sampling 

6.3.1  Probability sampling techniques:

These are techniques where each sampling unit (households or individuals) have an equal  or known  chance of being included in the sample.

The other major distinguishing feature of probability sampling methods is that are unbiased.

The scheme of selection of units from the target population is pre-specified and then the sample is selected according to the scheme and not according to any biases or preferences of the researcher.

The major types of PST are

1. Simple Random sampling
2. Systematic sampling
3. Stratified random sampling
4. Cluster sampling
5. Area Sampling
6. Multi stage or combination sampling

1. Simple Random sampling

In this  method , each unit in the population has an equal chance of being included in the sample. If there are N units in the sampling frame then the chance that a unit will be selected for the sample is 1/N.

E.g we wish to estimate the average income level of 100 employees in a  company. Due to time constraint the interviewer can ask say only 5 people randomly selected from the 100 employees.

If we wish to use random sampling method. We could make a list of 100 employees. Then a number could be allotted to each employee.

 We could then write these 100 numbers on small pieces of paper, one number on each paper. Shuffling these pieces of folded paper, we pick up at random 5 papers. These 5 employees are then asked the questionnaire.

2. Systematic Sampling :                                                                                                         In this method , the first unit is selected at random from the sampling frame. Other units are then selected at a regular interval depending upon the size of the sampling frame, so that every 5th unit or 10th unit or 100th unit can be selected. If the population size is denoted by N, and sample size by n, then N/n is the interval denoted by k.

    E.g We have a population of 300 students for some research. We need  a sample of 15 out of these. The sampling fraction or interval 300/15=20. This means 1 out of every 20 students will be selected on an average. We divide the list into 20 parts. Out of the first 20 students, we choose any one at random . Let us say we choose student number 7. Thereafter, we choose student numbers 7+20, 7+20+20, 7+20+20+20 and so on in a systematic sampling plan.

Therefore the selected students will be numbers 7, 27, 47, 67, 87, 107…………,, 297. All these 15 students will comprise our total sample for the study.  

3. Stratified sampling :                                                                                                 The entire population is divided into a number of mutually exclusive and collectively exhaustive strata. A simple random sample is then drawn from each strata such that the size of the sample drawn from each strata is proportional to the size of the strata. The stratification of strata is such that variance between strata is high and within strata is low.

E.g A consumer population may be divided into age brackets --- <25, 25-40, 41-50, >51 . Then a sample is taken from each of the strata defined .

4. Cluster sampling :                                                                                                 The entire population is first divided into clusters using geographical areas, city blocks, membership of some groups such as a club, social organization etc. As dimensions. A sampling frame of clusters is first constructed. Clusters are then chosen at random. Thereafter, either  all units in the chosen clusters are studied or a simple  random sample from each cluster is chosen.
5. Area sampling :                                                                                                 Clustering the sample by geographical  area is termed as area sampling. The population is divided using dimensions like states, towns, districts, cities etc. Each of these constitutes a cluster. A sample of clusters is then taken and either fully studied or random sampling is carried out from each chosen area

6.3.2 Non probability sampling techniques

There are 4 major non-probability sampling techniques .

These are

Quota sampling

Judgement sampling

Convenience sampling

Snowball sampling

1) Quota sampling : The entire population is divided into a number of mutually exclusive and collectively exhaustive strata. A quota or number of respondents for each strata is decided and is selected by the field worker .

2) judgement sampling :                                                                         this method relies on the judgement of the researcher as to who should be in the  sample. It obviously suffers from a researcher bias.

3) convenience samplng :                                                                                 this is employed many times in pretesting of questionnaires. It involves picking up any available set of respondents for the researcher to use.                                                 For e.g , students could be used as a sample by a marketing researcher who lives in a college town. The students may or may not be representative of the target population for the study, for the product being researched.

Other examples of convenience sampling includes on the street interviews, or from employees of one office block or factory.

Another example of convenience sampling is the one by tv reporters who catch any person passing by and interview him on the street.

4) snowball sampling :                                                                         this technique is used when the population being sought is a small one and chances of finding them by traditional methods are low.                                                 For e.g , to find owners of mercedes benz cars in a city , we may go to one or two , and ask them if they know anyone else who owns mercedes.                                                                                         Or to find golf /tennis players. One can ask each of the 5 golf/tennis players if they know one or two others.

One respondent being used to generate names of others is called snowballing and it can be done again on the second set of respondents. It couls also be called “networking” to find respondents.

As a general rule, this method is useful for niche markets .

Convenience Sampling :

Advantages

It is very easy to operationalize the data collection and is often resorted to when quick decisions are to be made and resources are limited.

The cost required to contact the units chosen is very less.

Disadvantages:

The result would differ from one researcher to another.

 The level of accuracy of results is rather low.

Judgement Sampling :

Advantages:

It is very easy to operationalize the data collection and is resorted to when quick decisions are to be made and resources are limited.

The cost to carry out field work is not high

The element of bias is less then that in convenience sampling

Disadvantages:.

The bias of judges is difficult to avoid

Errors in sampling are not possible to estimate so the accuracy levels of results is rather low.

The time required is more as compared to convenience sampling.

Quota sampling :

Advantages

The favourable features of this method are that the bias on account of sampling is least.

 The validity of results can also estimated and checked .

 The planning and execution is fairly easy and with results having reasonable accuracy

This method is normally used in descriptive studies.

Disadvantages:

The investigator bias cannot be eliminated completely as certain units in each class may be neglected completely.

The time required for selection is fairly high since the quota has to be completed.

The cost is higher as compared to other methods of non –probability sampling

Simple Random sampling (srs) :

Advantages :

The population parameters can be estimated with known accuracy.

Disadvantages:

I t is very time consuming

It requires strict control in the field so that the selected unit is the one from which data is collected.

The cost of arriving at the final sample is high as compared to non probability sampling .

Systematic sampling (sys):

Advantages :

The population parameters can be estimated with known accuracy.

The time required is less the SRS as it is simple to execute

Disadvantages:

There might be a clustering of certain units in the population resulting in their being completely left out in the sample.

The cost of arriving at the final sample and contacting the individual units thereafter is high.

Stratified sampling:

Advantages :

The population parameters can be estimated with known accuracy.

The time required is  fairly less.

The cost is not very high in comparison to SRS and SYS as sample size required is lower as compared to SRS.

Disadvantages:

A complete and exhaustive list of sampling units  is  required which at times may not be available.

The other disadvantage is the non-availability of dimensions over which the population could be completely stratified.

Cluster sampling

Advantages:

The population parameters can be estimated with known accuracy.

The time required is fairly less since the sampling units would be close to each other geographically.

The cost is also not very high.

It is most useful for test marketing situations.

Disadvantages:

If the units in the cluster are homogeneous , the efficiency of the estimate  will be reduced.

6.4 Types of Errors in Marketing Research.

Any research has an error margin associated with it. There are two major types of errors associated with a research study.

These are called:

1. Sampling errors or Random errors
2. Non-sampling or Human error.

Sampling Error: This is the error which ocurs due to the selection of some units and non-selection of other units into the sample. It is controllable if the selection of sample is done in a random, unbiased way. If a probability sampling technique is used, it is possible to control this error. In general, This error reduces as sample size increases.

Non-Sampling error: This is the effect of various errors in doing the study, by the interviewer, by the data entry operator or the researcher himself.

Handling a large quantity of data is not an easy job and errors may creep in at any stage of the research.

The data entry operator may interchange the column of “yes” and “no” responses while entering or compiling data. The interviewer may cheat by not filling up the questionnaire in the field and instead fudge the data.

Or the respondent may say onr thing, but another may be recorded by mistake. These errors are usually a function of the sample size. I.e The larger the sample size, the larger the non-sampling error.

Total Error: This is the total of sampling and non-sampling error. The sampling error can be estimated in the case of probability samples but not in the case of non-probabilty samples. Non sampling errors can be controlled through hiring better field workers, qualified data entry persons and good control procedures throughout the project.

---

Chapter 7 : Modes of communication (contact method)

7.1 Personal interviews (survey)

7.2 Mail Survey

7.3 Telephone survey

This chapter will deal with the last stage of step 3 of marketing research namely The contact  Method.

The three main types of contact method are

1. Personal interviews , 2) Mail survey and 3) telephone

Email as a tool for contact method is now being increasingly used by many companies.

7.1 Personal interviews (survey)

When the time available for research is large the personal interview is used.In the personal method , an interviewer is personally required to visit the respondent with the set of questions that are to be asked.

Advantages

1. The accuracy obtained is very high as the right persons are contacted and if there is difficulty in understanding certain questions the interviewer can take care of it.
2. If the interviewer feels that the respondent is not furnishing the correct facts, by observing the respondent, the interviewer can make his own interpretations.
3. When the literacy levels are low or when the respondent would find it difficult to fill up the questionnaire  on his own , this method is the best alternative available.
4. When the questions require spontaneuous answers , this is the best method

Disadvantages

1. The cost involved is very high since it requires field interviewers == who have to be paid. Also if a person is not available he may have to be contacted again and again.
2. The number of respondents that can be contacted is not very high, as the time taken to contact the respondents and the time spent on interview itself is very high.
3. The interviewer may have the tendency to contact some other person similar to respondent to complete his quota of respondents. This may affect the accuracy of results, thus necessitating a tight control on fieldwork.
4. There is a high chance of interviewer bias on account of recordinhg incorrect responses.

7.2 Mail Survey

Survey done through mail/post office.

Merits of mail survey

• Economical: The cost includes cost of postage (pre paid envelopes are sent along with the survey) and the printed questionnaire.
• Wide geographical coverage :
• Interviewer’s bias is eliminated
• Convenience to respondents: -- as thee is no pestering by interviewer and there is no pressure to fill up on the spot. The respondent can take their own time and this allows them to think before answering.
• Respondents and supervisors not required.—major expenses saved
• Family views can be ascertained.

Demerits of mail survey

• Respondents not replying at all or not replying on time.
• Exhaustive and correct list of mailing address of respondents is required. The accuracy is debatable
• Additional on the spot linked/thought questions cannot be asked as in personal interview.
• Not suitable when quick reply is required.
• Non verbal (facial) expressions cannot be noted.

7.3 Telephone survey

Telephone is the medium through which the information from respondents is collected.

A brief interview of the respondent is taken on telephone.

Merits of telephone survey

• Economical
• Quick response
• Busy people prefer telephonic interviews
• As the interviews are short, more people can be contacted in a day.
• The interviewer’s time and money is saved considerably.

Demerits

• Questionnaire has to be short and sweet.
• Limited information is given by respondents
• Non – verbal responses cannot be seen and analyzed.
• Difficulty in checking the authenticity of the respondent.

Chapter 8 : Collecting , Analyzing   data

8.1 Collecting the information: The fieldwork

8.2 Analyzing the information

8.3 Hypothesis testing

8.4 Z-Test

8.5 t-test

8.6 The Chi-square test for cross tabulations

8.7 ANOVA

 8.1 Collecting the information: The fieldwork

a) Designing the form: For getting the right information, the researcher has to prepare the form, which contains questions to be asked to respondents. The form should be designed in such a way that the information can be collected with speed and accuracy. The form of the questionnaire depends upon the nature of information sought, the kind of respondents and data collection methods.

b) Field work: The researcher has to appoint well-trained people to collect the information form samples selected for research. They must be properly trained, directed and motivated. Usually the people selected for collecting information have the following characteristics

  1) well mannered

2) Cultured

3) They know the importance of time management

4) Pleasing personality

5) High degree of patience

6) Have requisite language fluency

Once the required number of people are selected, a formal training in terms of what is the requirement, product knowledge is imparted to the selected people. They then collect the information from the respondents.

8.2 Analyzing the information

a) Editing : It is done in two stages . The first stage is the field editing which is done to detect the glaring omissions and inaccuracies, immediately after collection because the interviewers have fresh memory about the lapses and wrong statements. The second stage editing is office editing to evaluate completed forms. This is done by complete scrutiny of the questionnaire.

b) Coding : It consists of assigning symbols and numerical to each answer. It’s a technical procedure for categorizing the data.. It transforms the raw data into symbols and numerical. For e.g. Religion--- Hindu—1, Muslim--- 2, Christian – 3 etc.

      

3. Tabulation: It is the process of arrangement of data in rows and columns to identify what is the number of cases in each category. There are two types of tabulations namely-

Simple and cross tabulation

Simple tabulation : involving a single variable constitutes univariate analysis.. In simple tabulation every question is treated separately. For every question, the number of responses in each category of answers is counted. E.g If the sample size is 500 , and all 500 have answered the question, the simple tabulation of a respondent’s gender may look like the following

The simple tabulation for another question on the questionnaire may look like this

1. regular users of  Cinthol soap = 200
2. Non USERS = 300

Cross Tabulation: this is the second stage after simple tabulation is done.

A cross tabulation can be done by combining any two of the questions and tabulating the data together. This is a 2 variable cross tabulation.

This is bivariate analysis. An example could be cross tabulation between the brand preference for brands of tea and region to which respondent belongs.

e.g

The interpretation for North is out of 50 people 25 drink Brooke Bond. Also all India number of people who drink Brooke Bond is 80.

d) Data analysis: The tabulated data has to be analyzed. Appropriate techniques of analysis should be utilized to analyze the data.  

Analysis of data

Analysis of data is the process by which data is converted into useful information.

The three types of analysis are

1. Univariate , involving a single variable at a time
2. Bivariate, involving two variables at a time,
3. Multivariate , involving three or more variables simultaneously

For univariate and bivariate analysis of metric data (interval scale) the ‘t’ or ‘z’  test is used.

The two types of ‘t’ tests are

1. The independent sample ‘t’ test
2. The paired sample‘t’ test.

8.3 Hypothesis testing

steps in hypothesis testing

1. Formulate a hypothesis
2. Set up  suitable significance level
3. Choose a test driterion
4. Compute

5) make decisions

step 1 : formulate a hypothesis :Suppose as marketers of a brand of jeans, we wanted to find out whether a set of customers in Delhi and a set of customers in Mumbai thought of a particular brand in the same way or not.

Suppose we conducted a small survey in both cities and got ratings on an interval scale. We now want to do a statistical test to find out if the two sets of ratings are “significantly different from each other or not. We have to now set a level of “statistical significance” and select a suitable  test. We also need to specify a null hypothesis.

The null hypothesis represents a statement to be used to perform a statistical test to prove or to disprove the statement.

 E.g --- The null hypothesis for the‘t’ test would be “There is no significant difference in the ratings given  by customers from Mumbai and Delhi”.

The alternate hypotehsis is “there is significant difference in the ratings  given by customers in mumbai and delhi”.

Step 2) Set up a suitable significance level.

Having formulated the hypothesis, the next step is to test its validity at a certain level of significance. The confidence with which a null hypothesis is rejected or accepted depends upon the significnace level used for the purpose. A significance level of say 5% means that in the long run, the risk of making the wrong decision is 5%. The researcher is likely to be wrong in accepting a false hypothesis or in rejecting a true hypothesis in 5 out of 100 occasions.

step 3: Select Test criterion :                                                                        z test is used when sample size is greater then 30.                                  t test is used when sample size is less then 30.

step 4 compute : using necessary formulaes.

step 5 Make decisions

8.4 Z-Test

E,.g

A company manufacturing automobile tyres finds that tyre life is normally distributed with a mean of 40000 kms and standard deviation of 3000 kms. It is believed that a change in the production process will result in a better product and the company has developed a new tyre. A sample of 64 new tyres has been selected. The company has found that the mean life of these new tyres is 41200 kms. Can it be concluded that the new tyres is significantly better then the old one?

In a problem of this type , we are interested in testing whether or not there has been an increase in the mean life of tyres. In other words, we would like to test whether the mean life of new tyres has increased beyond 40000 kms

The various steps in testing the hypothesis are as follows

1. Null hypothesis : Ho:u= 40000 kms
2. Alternative hypothesis : H1:u > 40000 kms

2)The significance level is taken as 0.05. That is 5 out of every 100 occasions, there is a risk of being wrong in accepting or rejecting the hypothesis.

3. The test criterion is Z test (as sample is greater then 30)
4. Computations

             Interval estimate of the mean

X-ZOx= lower point

X+Zox= upper point

Where X= is the mean of the sample

Z= represents the number of standard errors for the specified confidence level

Ox= size of standard error.

Ox= O/n

        = 3000/8= 375

Z=   ( X-u)/375

  = (41200-40000)/375

= 3.2

5. decision – At 0.05 level of significance, the critical value of Z= 1.64

As the computed value of Z=3.2  falls in the rejection region, the null hypothesis is rejected. I.e the alternative hypothesis that u>40000 kms is accepted.

We conclude that the new tyre is significantly better then the old one.

8.5   t-test

Hypothesis testing when sample <30 === “t –test”

The Z test is based on the assumption that the sampling distribution of the mean is a normal distribution . This is applicable when the sample is large i.e 30 or more. When a sample is less then 30 , the assumption of normal distribution doesn’t hold good and hence t-test has to be used.

e.g

A Company manufacturing ice-cream sells it in 500 gms packs.

Periodically, a sample is taken to check whether, on the average, each pack contains 500 gms. A sample of 16 packs is taken and the sample mean is found to be 460 gms. And the estimated standard deviation 0 = 40 gms. Does the sample mean differ significantly form the intended weigh of 500 gms.

1. Null hypothesis = there is no difference between the sample mean and the population mean. Thus Ho= u=500 gms and we have a two tail test. The t statistic is

Ox= O/n

        = 40/4= 10

t=   ( X-u)/10

  = (460-500)/10

= -4

This t has a t distribution with 16-1 = 15 degrees of freedom. Assuming a significance level of 0.05 , the value of t from table is 2.131

As the calculated value of t is within the rejection region , the null hypothesis is rejected.. We conclude that the sample mean differs significantly from the population mean of 500 gms. The company should bring the production process under control.

Types of Errors in Hypothesis testing

When a hypothesis is tested , there are 4 possibilities

1. The hypothesis is true but our test leads to its rejection
2. The hypothesis is false but our test leads to its acceptance
3. The hypothesis is true and our test leads to its acceptance
4. The hypothesis is false and our test leads to its rejection

Of these 4  possibilities , the first two lead to an erroneous decision. The first possibility leads to Type 1 error, and the second possibility leads to Type 11 error.

8.6   The Chi-square test for cross tabulations

In the case of cross tabulations featuring two variables, a test of significance called the Chi-square test can be used to test if the two variables are statistically associated with each other significantly. The user who is analyzing the data on the computer and using a statistical package , can request a Chi –squared test along with any cross –tabulations. Commands such as CROSSTABS on most statistical packages have the option of doing a chi-squared test.

e.g Let us assume that we have conducted a consumer survey for a brand of detergent. One of the questions dealt with income category of the respondent. Another asked the respondent to rate his purchase intention These 2 variables are listed below. Both variables are coded. Income codes and their relative incomes are as follows

Code                        Income in Rs.

1. Less then 5000
2. 5001 –10000
3. 10001-20000
4. >20000

Purchase intention codes are as follows :

Code                        Explanation

1. None-no intention to buy
2. Low-low intention to buy
3. High- high intention
4. Very high- very high intention
5. Certain – certain to buy

These two variables were cross-tabulated from a sample of 20 respondents. The output of cross tabulation is as follows

Is there a significant association between respondent income and purchase intention.

After analysis using the computer , we get the following answer

Significance level= 0.09690

Degree of freedom =12

Value =18.6667

The interpretation is as follows: the Chi-square test is showing a significant association between these two variables at a 90 % confidence level (10%significance level)

In a Chi-square test , for a 90% confidence level, if the significance level is greater than or equal to 0.1 , it signifies that there is no association between the two variables in the cross tabulation and if the significance level is less then 0.1 , then it signifies that there is a significant relationship  between the selected variables

This leads to the conclusion that  the price of detergent is important in its purchase.

8.7  ANOVA

ANOVA stands for analysis of variance , the generic name given to a set of techniques for studying  the cause and effect of one or more factors on a single dependent variable. It is also used to test differences among means of independent samples.

The null hypothesis is Ho:u1=u2=u3 …..

The alternative hypothesis is H1: all the means are not equal.

Then following steps are involved in the analysis of variance

1. Calculate variance between the samples
2. Calculate variance within the samples
3. Calculate the F ratio by the following formula

F= Variance between the samples/variance within the samples

4. Compare the value of F as arrived in step number 3 with the critical value of F such as 5% significance level for the applicable degrees of freedom

When the calculated value of F is less then the table value of F , the difference in sample means is not significant and the null hypothesis is accepted. In contrast , when the calculated value of F is more then the critical value of F , the difference in sample means is regarded as significant and the null hypothesis is rejected.

e.g Suppose a manufacturer of a breakfast food is interested to know the effectiveness of 3 different types of packaging . He puts each kind of packaged breakfast food into five different stores. He finds that during a given week the number of packages sold were as follows.

Packaging 1: 25, 28, 21, 30, 26

Packaging 2: 27,25,25,33,30

Packaging 3: 22,29,26,20,23

The mean sales for these three packaging are packaging 1= 26, packaging 2= 28and packaging 3= 24. What the manufacturer would like to know is whether the differences among the means are significant.

Uses of ANOVA

ANOVA is used when the independent variables are of nominal scale and dependent variable is metric (continuous) the independent variables may be different levels of prices or different pack sizes or different product colors and the effect (dependent variable) could be the sales of the product. ANOVA is used in experimental designs.

---

Chapter 9 : Marketing research report

A marketing research report normally contains the following topics, which are self-explanatory.

1. Title page
2. Table of contents
3. Executive summary
4. Introduction
5. Statement of objectives
6. Methodology

1. research design
2. Data collection method
3. Sampling
4. Field work
5. Analysis and interpretation

7.   Limitations

8.   Findings

9.    Conclusions and recommendations

10. Appendix

1. Copy of questionnaires
2. Details of samples
3. Tables not included in findings
4. Bibliography

---

Bibliography

Marketing Research

1. Marketing Research: Sangeeta Agarwal

2. Marketing Research: Rajendra Nargundkar

3. Marketing Research: Boyd and Stasch

4. Marketing Research: Green, Tull and Hawkins