Planning Primary Data Collection Good decisions require good

data. Just as researchers must carefully evaluate the quality of secondary information they obtain, they must also take great care in collecting primary data to ensure that they provide marketing decision makers with relevant, accurate, current and unbiased information. This could be qualitative research that measures a small sample of customers' views, or quantitative research that provides statistics from a large sample of consumers. Table 8.2 shows that designing a plan for primary data collection calls for a number of decisions on research approaches, contact methods, sampling plan and research instruments.

RESEARCH APPROACHES. Observational research is the gathering of primary data by observing relevant people, actions and situations. Por example:

• A food-products manufacturer sends researchers into supermarkets to find out the prices of competing brands or how much shelf space and display support retailers give its brands.

• A bank evaluates possible new branch locations by checking traffic patterns, neighbourhood conditions and the locations of competing branches.

• A maker of personal-care products pretests its ads by showing them to people and measuring eye movements, pulse rates and other physical reactions.

• A department store chain sends observers who pose as customers into its stores to cheek on store conditions and customer service.

• A museum checks the popularity of various exhibits by noting the amount of floor wear around them.

Several companies sell information collected through mechanical obserA nation. For example, Nielsen and AGB attach 'people meters' to television sets in selected homes to record who watches which programmes. They provide summaries

Table 8.3

Strengths and weaknesses of the four contact methods

MAIL

TELEPHONE

PERSONAL

INTERNET

1. Flexibility

Poor

Good

Excellent

Fair

2. Quantity of data that can be collected

Good

Fair

Excellent

Good

3. Control of interviewer effects

Excellent

Fair

Poor

Excellent

4. Control of sample

Fair

Excellent

Fair

Fair

5. .Speed of data collection

Poor

Excellent

Good

Excellent

6. Response rate

Poor

Good

Good

Poor

7. Cost

Good

Fair

Poor

Excellent

8. Sample frame

Good

Excellent

Fair

Poor

SOURCE: Adapted with permission of Ma em ill an Publishing Company torn Marketing Research: Measurement and Method, 6th edn, by Donald S. Tull and Del I. Hawkins. Copyright © 1993 by Maemillan Publishing Company.

SOURCE: Adapted with permission of Ma em ill an Publishing Company torn Marketing Research: Measurement and Method, 6th edn, by Donald S. Tull and Del I. Hawkins. Copyright © 1993 by Maemillan Publishing Company.

experimental research Thegathering of primary data by selecting matched groups of subjects, giving them different treatments, controlling related factors and checkingfor differences in group responses.

Respondents may answer survey questions even when they do not know the answer, simply in order to appear smarter or more informed than they are. Or they may try to help the interviewer by giving pleasing answers. Finally, busy people may not take the time, or they might resent the intrusion into their privacy. Careful survey design can help to minimize these problems.

Experimental research gathers causal information. Experiments involve selecting matched groups of subjects, giving them different treatments, controlling unrelated factors and checking for differences in group responses. Thus experimental research tries to explain oause-and-effect relationships. Observation and surveys can collect information in experimental research.

Before extending their product range to include fragrances, researchers at Virgin Megastores might use experiments to answer questions such as the following:

• How much will the fragrances increase Virgin's sales?

• How will the fragrances affect the sales of other menu items?

• Which advertising approach would have the greatest effect on sales of their fragrances?

• How would different prices affect the sales of the product?

• How will the product affect the stores' overall image?

For example, to test the effects of two prices, Virgin could set up a simple experiment. It could introduce fragrances at one price in one city and at another price in another city. If the cities are similar and if all other marketing efforts for the fragrances are the same, then differences in the price charged could explain the sales in the two cities. More complex experimental designs could include other variables and other locations.

CONTACT METHODS. Mail, telephone, personal interviews and the Internet, a recent development, can collect data. Table 8.3 shows the strengths and weaknesses of each of these contact methods.

Postal questionnaires have many advantages. They can collect large amounts of information at a low cost per respondent. Respondents may give more honest answers to more personal questions on a postal questionnaire than to an unknown interviewer in person or over the phone, since there is no interviewer to bias the respondent's answers.

However, posta! questionnaires also have disadvantages. They are not very flexible: they require simple and clearly worded questions; all respondents answer the same questions in a fixed order; and the researcher cannot adapt the questionnaire based on earlier answers. Mail surveys usually take longer to complete and the response rate — the number of people returning completed questionnaires -is often very low. Finally, the researcher often has little control over the postal questionnaire sample. Even with a good mailing list, it is often hard to control 'K'hoat the mailing address fills out thequestionnaire.IK

Telephone intervie<wing is the best method for gathering information quickly and it provides greater flexibility than postal questionnaires. Interviewers can explain questions that arc not understood. Depending on the respondent's answers, they ean skip some questions or probe further on others. Telephone interviewing also allows greater sample control. Interviewers can ask to speak to respondents with the desired characteristics, or even by name. Response rates tend to be higher than with postal questionnaires.'1'

However, telephone interviewing also has drawbacks. The cost per respondent is higher than with postal questionnaires and people may not want to discuss personal questions with an interviewer. Using interviewers increases flexibility, hut also introduces interviewer bias. The way interviewers talk, small differences in how they ask questions and other differences may affect respondents' answers. Finally, different interviewers may interpret and record responses differently, and under time pressure some interviewers might even cheat by recording answers without asking questions.

Persona/ interviewing, takes two forms - individual and group interviewing. Individual interviewing involves talking with people in their homes or offices, in the street, or in shopping malls. The interviewer must gain their co-operation and the time involved can range from a few minutes to several hours. Sometimes people get a small payment in return for their time.

Group interviewing consists of inviting six to ten people to gather for a few hours with a trained moderator to talk about a product, service or organization. Tim moderator needs objectivity, knowledge of the subject and industry, and some understanding of group and consumer behaviour. The participants are normally paid a small sum for attending. The meeting is usually in a pleasant place and refreshments are served to foster an informal setting. The moderator starts with broad questions before moving to more specific issues, and encourages easy-going discussion, hoping that group interactions will bring out aetual feelings and thoughts. At the same time, the moderator 'focuses' the discussion — hence the name focus-group interviewing. The comments arc recorded by written notes or on videotapes for study later. Focus-group interviewing has become one of the key marketing research tools for gaining insight into consumer thoughts and feelings.20

Personal interviewing is quite flexible and can collect large amounts of Information. Trained interviewers can hold a respondent's attention for a long time and ean explain difficult questions. They can guide interviews, explore issues and probe as the situation requires. Personal interviews can utilize any type of questionnaire. Interviewers can show subjects actual products, advertisements or packages, and observe reactions and behaviour. In most cases, personal interviews can be conducted fairly quickly.

The main drawbacks of personal interviewing are costs and sampling problems. Personal interviews may cost three to four times as much as telephone interviews. Group interview studies usually employ small sample sizes to keep focus group A small sample of epical consumers under the direction of a group leailerTO/IOelicits their reaction to a stimulus .swc.'/i as an ad or product concept.

time and costs down, and it may he hard to generalize from the results. Because interviewers have more freedom in personal interviews, the problem of interviewer bias is greater.

Which contact method is best depends on what information the researcher wants and on the number and types of respondents needed. Advances in computers and communications have had an impact on methods of obtaining information. For example, most research firms now do Computer Assisted Telephone Interviewing (CATI). Professional interviewers eall respondents, often using phone numbers drawn at random. When the respondent answers, the interviewer reads a set of questions from a video screen and types the respondent's answers directly into the computer. Although this procedure requires a large investment in computer equipment and interviewer training, it eliminates data editing and coding, reduces errors and saves time. Other research firms set up terminals in shopping centres - respondents sit down at a terminal, read questions from a screen and type their answers into the computer.21

Internet data collection is still in its infancy. But as the use of the World Wide Web and online services widens, online research is becoming a quick, easy and inexpensive method.22 Online researchers recognize that Web surfers are not representative of the population. Online users tend to be better educated, more affluent and younger than the average consumer, and a higher proportion are male. These are important consumers to companies offering products and services online. They are also some of the hardest to reach when conducting a research study. Online surveys and chat sessions (or online focus groups) often prove effective in reaching elusive groups, such as teen, single, affluent and well-educated audiences.

When appropriate, online research offers marketers two distinct advantages over traditional surveys and focus groups: speed and cost-effectiveness. Online researchers can field quantitative studies and fill response quotas in only a matter of days. Online focus groups require some advance scheduling, but results are practically instantaneous. Research on the Internet is also relatively inexpensive. Participants ean dial in for a focus group from anywhere in the world, eliminating travel, lodging and facility costs, making online ehats cheaper than traditional focus groups. And for surveys, the Internet eliminates most of the postage, phone, labour and printing costs associated with other survey approaches. Moreover, sample size has little influence on costs. There is little difference between 10 and 10,000 on the Web. There is also no difference in the speed and cost of conducting an international survey rather than a domestic one.

However, using the Internet to conduct marketing research does have some drawbacks. The method shares a problem with postal surveys: knowing who's in the sample. Trying to draw conclusions from a 'self-selected' sample of online users, those who clicked through to a questionnaire or accidentally landed in a chat room, can be troublesome. Online research is not right for every company or product. For example, mass marketers who need to survey a representative cross-section of the population will find online research methodologies less useful, since most low-income consumers do not have online access.

Rye contact and body language are two direct, personal interactions of traditional focus-group research that are lost online. And while researchers can offer seasoned moderators, the Internet format - running, typed commentary and online 'emoticons' (punctuation marks that express emotion, such as :-) to signify happiness or :-o for surprise) - greatly restrict respondent expressiveness. Similarly, technology limits researchers' capability to show visual cues to research subjects. But just as it hinders the two-way assessment of visual cues, Web research can actually permit some participants the anonymity necessary to elicit an unguarded response.

Table 8.4

Data collection methods: expendí itUre toi \iyL)

quantitative

and qualitative

(% of

SURVEY (%)

TELEPHONE (%)

INTERVIEWS (%)

Belgium

l

15

Bl

1Q

8

Denmark

1B

24

G1

B

21

Finland

19

38

Gl

n.a.

n.a.

France

n.a.

n.a.

n.a.

n.a.

n.a.

Germany

B

18

BB

B

10

Greece

is

3

82 h

1Q

3

Ireland

1

2

l2

2G

3

Italy

4

27

44

1Q

10

Luxembourg

-

10

BB

B

-

Netherlands

G1

18

G4

4

9

Norway

1Q

20

BQ

1Q

5

Portugal

12

9

BB

l

3

Spain

G

16

BG

1G

3

Sweden

2G

39

2G

4

5

Switzerland

8

27

4B

n.a.

n.a.

Turkey

B

4

BQ

1B

5

United Kingdom

8

15

Bl

1Q

-

NOTES: * Less than 0.5 per cent; h including panel turnover.

Source: Report o/ESOMAR Working Party on .1991 Market Statistics (1992)

NOTES: * Less than 0.5 per cent; h including panel turnover.

Source: Report o/ESOMAR Working Party on .1991 Market Statistics (1992)

To overcome such sample and response problems, NPD and many other firms that offer online services construct panels of qualified Web regulars to respond to surveys and participate in online focus groups. NPD's panel consists of 15,000 consumers recruited online and verified by telephone; Greenfield Online picks users from its own database, then calls them periodically to verify that they are who they say they arc. Another online research firm, Research Connections, recruits in advance by telephone, taking time to help new users connect to the Internet, if necessary.

Some researchers arc wildly optimistic about the prospects for marketing research on the Internet; others are more cautions. One expert predicts that in the next few years, 50 per cent of all research will be done on the Internet, although others are more sceptical.

There is no one best contact method to use. The one chosen depends on the information needs, cost, speed and other issues. Table 8.4 shows the traditional data collection methods used across Europe. Rational reasons may account for only part of the variation shown. Face-to-facc interviews account for half the number, hut these figures are particularly high in southern Europe and the United Kingdom. The low penetration of telephones in some of these countries may be an influence, but it may also reflect cultures who like socializing. Ireland's high use of group discussions may show that land's love of conversation. The Scandinavians' use of telephone interviews is partly explained by their being large countries with small populations. Their mobile phone use also shows their telephone orientation:

sample

A segment of the population selected/or marketing research to represent the population as a "whole.

10 per cent are mobile phone users compared with 6 per cent in the United States and about 3 per cent in Germany and Japan.23 In some countries, postal surveys do not work because of low literacy, but another reason is the unwillingness of people to respond. Research agencies and managers also have preferred methods, so they will also exert some personal influence on the choice of method. The relatively low use of the Internet in Europe means that Internet data collection in Europe will lag behind the United States. In addition, the large differences in penetration of the Internet across Europe mean that its use will not be uniform.

Increasing consumer resentment has become a major problem for the research industry. This resentment has led to lower survey response rates in recent years -one study found that 38 per cent of consumers now refuse to be interviewed in an average survey, up dramatically from a decade ago. Another study found that 59 per cent of consumers had refused to give information to a company because they thought it was not really needed or too personal, up from 42 per cent just five years earlier.21 The research industry is considering several options lor responding to this problem. One is to educate consumers about the benefits of marketing research and to distinguish it from telephone selling and database building. Another option is to provide a freephone number that people can call to verify that a survey is legitimate. The industry has also considered adopting broad standards, such as Europe's International Code of Marketing and Social Research Practice. This code outlines researchers' responsibilities to respondents and to the general public. For example, it says that researchers should make their names and addresses available to participants, and it bans companies from representing activities like database compilation or sales and promotional pitches as research.

SAMPLING PLANS. Marketing researchers usually draw conclusions about large groups of consumers by studying a small sample of the total consumer population. A sample is a segment of the population selected to represent the population as a whole. Ideally, the sample should be representative, so that the researcher ean make accurate estimates of the thoughts and behaviours of the larger population.

Designing the sample calls for three decisions. First, to/io is to be surveyed (what sampling unit)? The answer to this question is not always obvious. For example, to study the decision-making process for a family car purchase, should the researcher interview the husband, wife, other family members or all of these? The responses obtained from different family members vary, so the researcher must determine the information needed and from whom.23

Second, hv*us many people are to be surveyed (what sample sise)? Large samples give more reliable results than small samples. However, it is not necessary to sample the entire target market or even a large portion to get reliable results. If well chosen, samples of less than 1 per cent of a population can often give good reliability.

Third, how are the people in the sample to be chosen (what sampling procedure)? Table 8.5 describes different kinds of sample. Using probability samples, each population member has a known chance of being included in the sample, and researchers can calculate confidence limits for sampling error. But when probability sampling costs too much or takes too long, marketing researchers often take non-probability samples, even though their sampling error is not measurable. These varied ways of drawing samples have different costs and time limitations, as well as different accuracy and statistical properties. Which method is best depends on the needs of the research project.

RESEARCH INSTRUMENTS. In collecting primary data, marketing researchers have a choice of two main research instruments: the questionnaire tmd mechanical de-vices.

Table 8.5

Types of sampling

Probability sample

Simple random sample

Stratified random sample Cluster (area) sample

Non-probability sample Convenience sample

Judgement sample

Quota sample

Every member of the population has a known and equal chance of selection.

The population is divided into mutually exclusive groups (such as age groups), and random samples are drawn from each group. The population is divided into mutually exclusive groups (such as blocks), and the researcher draws a sample of the groups to interview.

The researcher selects the easiest population members from which to obtain information.

The researcher uses his or her judgement to select population members who are good prospects for accurate information. The researcher finds and interviews a prescribed number of people in each of several categories.

The questionnaire is by far the most common instrument. Broadly speaking, a questionnaire consists of a set of questions presented to a respondent for his or her answers. The questionnaire is very flexible - there are many ways to ask questions. Questionnaires need to be developed carefully and tested before their large-scale use. A carelessly prepared questionnaire usually contains several errors (see Table 8.6).

In preparing a questionnaire, the marketing researcher must decide what questions to ask, the form of the questions, the wording of the questions and the ordering of the questions. Questionnaires frequently leave out questions that need answering, hut include questions that cannot be answered, will not be answered, or need not be answered. Each question should be checked to see that it contributes to the research objectives.

The/orm of the question can influence the response. Marketing researchers distinguish between closed-cud and open-end questions. Closed-end questions include all the possible answers, and subjects make choices among them. Part A of Table 8,7 shows the most common forms of closed-end questions as they might appear in an SAS survey of airline users. Open-end questions allow respondents to answer in their own words. The most common forms are shown in part R of Table 8.7. Open-end questions often reveal more than closed-end questions because respondents are not limited in their answers. Open-end questions are especially useful in exploratory research in which the researcher is trying to find out what people think, but not measuring how many people think in a certain way, Closed-end questions, on the other hand, provide answers that are easier to interpret and tabulate.

Researchers should also use care in the wording of questions. They should use simple, direct, unbiased wording. The questions should he pretested before use. The ordering of questions is also important. The first question should create interest if possible. Ask difficult or personal questions last, so that respondents do not become defensive. The questions should be in a logical order.

closed-end questions Questions that include all the possible answers and allow subjects to make c-hoices amtmg them.

open-end questions Questions that allow respondents to answer in their own words.

Table 8.6

A 'questionable questionnaire'1

Suppose that an adventure holiday director had prepared the following questionnaire to use in interviewing the parents of prospective campers. How would you assess each question?

1. What is your income to the nearest hundred pounds?

People don't usually know their income to the nearest hundred pounds nor do they want to reveal their income that closely. Moreover, a researcher should never open a questionnaire with such a personal question.

2. Arc you a strong or a weak supporter of overnight camping for your children? What do 'strong'and 'weak'mean?

3. Do your children behave themselves well on adventure holidays? Yes"( ) No( )

'Behave' is a relative term. Furthermore, are 'yes' and 'no' the best response options for this question? Besides, will people want to answer f/ii.s? Why ask the question in the first place?

4. How many adventure holiday operators mailed literature to you last April? This April? Wlio can remember this?

5. What are the most salient and determinant attributes in your evaluation of adventure holidays?

What are 'salient' and 'determinant' attributes? Don't use big words on me!

6. Do you think it is right to deprive your child of the opportunity to grow into a mature person through the experience of adventure holidays?

A loaded question. Given the bias, how can any parent answer 'yes'?

Although questionnaires are the most common research instrument, mechanical instruments are also used. We discussed two mechanical instruments - people meters and supermarket scanners - earlier in the ehapter. Another group of mechanical devices measures subjects' physical responses. For example, a galvanometer (!ie detector) measures the strength of interest or emotions aroused by a subject's exposure to different stimuli: for instance, an ad or picture. The galvanometer detects the minute degree of sweating that accompanies emotional arousal. The taehistoscupe flashes an ad to a subject at an exposure range from less than one-hundredth of a second to several seconds. After each exposure, the respondents describe everything they recall. Eye cameras study respondents' eye movements to determine at what points their eyes focus first and how long they linger on a given item.2'1

Online Survey Champion

Online Survey Champion

There are people all over the world trying to find ways to make money online. From stay at home moms looking to make a few extra dollars to college students and entrepreneurs, the allure of making your own hours and working from home or from the local coffee shop is very appealing.

Get My Free Ebook


Responses

  • arvid anderson
    How to plan for the collection of primary datay?
    7 years ago
  • mellisa
    What are the two main types of research instruments used to collect primary data?
    7 years ago
  • outi
    What are primary data for marketing decisions?
    7 years ago
  • ERIC
    What are the strengths and weaknesses of the internet for primary data collection?
    6 years ago

Post a comment