CHAPTER FIVE DESCRIPTIVE AND CAUSAL RESEARCH DESIGNS LEARNING OBJECTIVES (PPT slides 5-2) 1. Explain the purpose and advantages of survey research designs. 2. Describe the types of survey methods. 3. Discuss the factors influencing the choice of survey methods. 4. Explain experiments and the types of variables used in causal designs. 5. Define test marketing and evaluate its usefulness in marketing research. KEY TERMS AND CONCEPTS 1. Ability to participate 2. Causal research 3. Computer-Assisted Telephone Interview (CATI) 4. Control variables 5. Dependent variables 6. Drop-off survey 7. Experiment 8. External validity 9. Extraneous variables 10. Field experiments 11. Generalizable 12. Incidence rate 13. Independent variables 14. In-home interview 15. Internal validity 16. Knowledge level 17. Laboratory (lab) experiments 18. Mall-intercept interview 19. Mail panel survey 20. Mail surveys 21. Nonresponse error 22. Online surveys 23. Person-administered survey 24. Propensity scoring 25. Respondent errors 26. Response error 27. Self-administered survey 28. Survey research methods 29. Telephone interviews 30. Test marketing 31. Topic sensitivity 32. Validity 33. Variable 34. Willingness to participate 35. Wireless phone survey CHAPTER SUMMARY BY LEARNING OBJECTIVES Explain the purpose and advantages of survey research designs. The main advantages of using descriptive survey research designs to collect primary data from respondents are large sample sizes are possible; generalizability of results; ability to distinguish small differences between diverse sampled groups; ease of administering; and the ability to identify and measure factors that are not directly measurable (such as customer satisfaction). In contrast, disadvantages of descriptive survey research designs include the difficulty of developing accurate survey instruments, inaccuracy in construct definition and scale measurement, and limits to the depth of the data that can be collected. Describe the types of survey methods. Survey methods are generally divided into three generic types. One is the person-administered survey, in which there is significant face-to-face interaction between the interviewer and the respondent. The second is the telephone-administered survey. In these surveys the telephone is used to conduct the question-and-answer exchanges. Computers are used in many ways in telephone interviews, especially in data recording and telephone-number selection. The third type is the self-administered survey. In these surveys, there is little, if any, actual face-to-face contact between the researcher and prospective respondent. The respondent reads the questions and records his or her answers. Online surveys are the most frequent method of data collection, with almost 60 percent of all data collection being completed with online surveys. Discuss the factors influencing the choice of survey methods. There are three major factors affecting the choice of survey method: situational characteristics, task characteristics, and respondent characteristics. With situational factors, consideration must be given to elements such as available resources, completion time frame, and data quality requirements. Also, the researcher must consider the overall task requirements and ask questions such as “How difficult are the tasks?” “What stimuli (e.g., ads or products) will be needed to evoke responses?” “How much information is needed from the respondent?” and “To what extent do the questions deal with sensitive topics?” Finally, researchers must consider the diversity of the prospective respondents, their likely incidence rate, and the degree of survey participation. Maximizing the quantity and quality of data collected while minimizing the cost and time of the survey generally requires the researcher to make trade-offs. Explain experiments and the types of variables used in causal designs. Experiments enable marketing researchers to control the research situation so that causal relationships among the variables can be examined. In a typical experiment the independent variable is manipulated (changed) and its effect on another variable (dependent variable) is measured and evaluated. During the experiment, the researcher attempts to eliminate or control all other variables that might impact the relationship being measured. After the manipulation, the researcher measures the dependent variable to see if it has changed. If it has, the researcher concludes that the change in the dependent variable is caused by the manipulation of the independent variable. To conduct causal research, the researcher must understand the four types of variables in experimental designs (independent, dependent, extraneous, control) as well as the key role random selection and assignment of test subjects to experimental conditions. Theory is important in experimental design because researchers must conceptualize as clearly as possible the roles of the four types of variables. The most important goal of any experiment is to determine which relationships exist among different variables (independent, dependent). Functional (cause-effect) relationships require measurement of systematic change in one variable as another variable changes. Define test marketing and evaluate its usefulness in marketing research. Test markets are a specific type of field experiment commonly conducted in natural field settings. Data gathered from test markets provide both researchers and practitioners with invaluable information concerning customers’ attitudes, preferences, purchasing habits/patterns, and demographic profiles. This information can be useful in predicting new product or service acceptance levels and advertising and image effectiveness, as well as in evaluating current marketing mix strategies. CHAPTER OUTLINE Opening VIGNETTE: Magnum Hotel’s Loyalty Program The opening vignette in this chapter describes the loyalty program used by the Magnum Hotel to encourage guest loyalty. Management needs to evaluate whether the strategy is working and provides a competitive advantage. No data other than a list of members and the cost of running the program were available. They decided to use qualitative research with in-depth interviews and focus groups to frame a later quantitative study. That work resulted in a list of research questions, dealing with topics like program awareness and valued program features. The vignette closes with this question, “is quantitative research necessary to answer these questions?” I. Value of Descriptive and Causal Survey Research Designs (PPT slide 5-3) Some research problems require primary data that can be gathered only by obtaining information from a large number of respondents considered to be representative of the target population. Quantitative methods of collecting primary data generally involve much larger samples, including survey designs used in descriptive and causal research. II. Descriptive Research Designs and Surveys (PPT slide 5-4 to 5-6) Selection of a descriptive research design is based on three factors: The nature of the initial problem or opportunity The research questions The research objectives When the research problem/opportunity is either to describe characteristics of existing market situations or to evaluate current marketing mix strategies, then a descriptive research design is the appropriate choice. If research questions include issues such as who, what, where, when, and how for target populations or marketing strategies, then a descriptive research design also is most appropriate. Finally, if the task is to identify relationships between variables or determine whether differences exist between groups, then descriptive research designs are generally best. Two general approaches are used to collect data for descriptive research: Asking questions Observation Descriptive designs frequently use data collection methods that involve asking respondents structured questions about what they think, feel, and do. Thus, descriptive research designs often result in the use of survey research methods to collect quantitative data from large groups of people through the question/answer process. But with the emergence of scanner data and tracking of Internet behavior, observation is being used more often in descriptive designs. The term “descriptive” is sometimes used to describe qualitative research, but the meaning is different than when the word is used to describe quantitative research. Qualitative research is descriptive in the sense that it often results in vivid and detailed textual descriptions of consumers, consumption contexts, and culture. Quantitative studies are descriptive in the sense that they use numbers and statistics to summarize demographics, attitudes, and behaviors. Survey research methods are a mainstay of quantitative marketing research and are most often associated with descriptive and causal research designs. The main goal of quantitative survey research methods is to provide facts and estimates from a large, representative sample of respondents. The advantages and disadvantages of quantitative survey research designs are summarized in Exhibit 5.1 (PPT slide 5-5). III. Types of Errors in Surveys (PPT slide 5-7 and 5-8) Errors can reduce the accuracy and quality of data collected by researchers. Survey research errors can be classified as being either sampling or nonsampling errors. A. Sampling Errors (PPT slide 5-7) Any survey research design that involves collecting data from a sample will have some error. Sampling error is the difference between the findings based on the sample and the true values for a population. It can be reduced by increasing sample size and using the appropriate sampling method. B. Nonsampling Errors (PPT slide 5-8) Errors that occur in survey research design not related to sampling are called nonsampling errors. Most types of nonsampling errors are from four major sources: Respondent error Measurement/questionnaire design errors Incorrect problem definition Project administration errors Nonsampling errors have several characteristics:. They tend to create “systematic variation” or bias in the data. They are controllable. They cannot be directly measured. They can create other nonsampling errors. Respondent errors occur when either respondents cannot be reached, are unwilling to participate, or intentionally or unintentionally respond to questions in ways that do not reflect their true answers. Respondent errors can be divided into: Nonresponse error—a systematic bias that occurs when the final sample differs from the planned sample. It occurs when a sufficient number of the preselected prospective respondents in the sample refuse to participate or cannot be reached. Response error—respondents have impaired memory or do not respond accurately. It is also termed as faulty recall. Memory is subject to selective perception (noticing and remembering what we want to) and time compression (remembering events as being more recent than they actually were). Respondents sometimes use averaging to overcome memory retrieval problems. IV. Types of Survey Methods (PPT slide 5-9) Improvements in information technology and telecommunications have created new survey approaches. Nevertheless, survey methods can be classified as person-administered, self-administered, or telephone-administered. Exhibit 5.2 provides an overview of the major types of survey methods (PPT slide 5-9). A. Person-Administered Surveys (PPT slide 5-9 and 5-10) Person-administered survey methods have a trained interviewer who asks questions and records the subject’s answers. Exhibit 5.3 highlights some of the advantages and disadvantages of using person-administered surveys (PPT slide 5-10). An in-home interview is a face-to-face structured question-and-answer exchange conducted in the respondent’s home. Interviews are also occasionally conducted in office environments. This method has several advantages: The interviewer can explain confusing or complex questions and use visual aids. Respondents can try new products or watch potential ad campaigns and evaluate them. Respondents are in a comfortable, familiar environment thus increasing the likelihood of respondents’ willingness to answer the survey’s questions. Disadvantages of in-home interview include: In-home interviewing may be completed through door-to-door canvassing of geographic areas. Interviewers who are not well supervised may skip homes they find threatening or even fabricate interviews. The interviews are expensive and time-consuming. Mall-intercept interviews are face-to-face personal interviews that take place in a shopping mall where the mall shoppers are stopped and asked to complete a survey (PPT slide 5-9). Mall-intercept interviews share the advantages of in-home and in-office interviews, but the environment is not as familiar to the respondent. But mall-intercepts are less expensive and more convenient for the researcher. The disadvantages of mall-intercept interviews are similar to those of in-home or in-office interviews except that interviewer travel time is reduced. Moreover, mall patrons are not likely to be representative of the target population, even if they are screened. Typically, mall-intercept interviews must use some type of nonprobability sampling, which adversely affects the ability to generalize survey results. B. Telephone-Administered Surveys (PPT slide 5-9) Telephone interviews involve question-and-answer exchanges that are conducted via telephone technology (PPT slide 5-9). Compared to face-to-face interviews, telephone interviews are less expensive, faster, and more suitable for gathering data from large numbers of respondents. Telephone survey methods have a number of advantages over face-to-face survey methods: Interviewers can be closely supervised if they work out of a central location. Although there is the added cost of the telephone call, they are still less expensive than face-to-face interviews. Telephone interviews facilitate interviews with respondents across a wide geographic area, and data can be collected relatively quickly. Telephone surveys enable interviewers to call back respondents who did not answer the telephone or who found it inconvenient to grant interviews when first called. Random digit dialing can be used to select a random sample. The telephone method also has several drawbacks: Pictures or other nonaudio stimuli cannot be presented over the telephone. Some questions become more complex when administered over the telephone. Telephone surveys tend to be shorter than personal interviews because some respondents hang up when a telephone interview becomes too lengthy. Telephone surveys also are limited, at least in practice, by national borders; the telephone is seldom used in international research. Many people are unwilling to participate in telephone surveys so refusal rates are high and have increased substantially in recent year. Many people are annoyed by telephone research because it interrupts their privacy, their dinner, or their relaxation time. The increased use of telemarketing and the illegal and unethical act of “sugging,” or selling under the guise of research, has contributed to a poor perception of telephone interviewing among the public. Computer-assisted telephone interviews (CATI) are integrated telephone and computer system in which the interviewer reads the questions from a computer screen and enters respondents’ answers directly into the computer program.. The program automatically skips questions that are not relevant to a particular respondent. CATI systems overcome most of the problems associated with manual systems of callbacks, complex quota samples, skip logic, rotations, and randomization. Although the major advantage of CATI is lower cost per interview, there are other advantages as well: Computer technology can send inbound calls to a particular interviewer who “owns” the interview and who can complete the interview at a later time. CATI eliminates the need for separate editing and data entry tasks associated with manual systems. Results can be tabulated in real time at any point in the study. Use of CATI systems continues to grow because decision makers have embraced the cost savings, quality control, and time-saving aspects of these systems. Wireless phone survey is a method of conducting a marketing survey in which the data are collected using wireless phones. Wireless phone surveys are growing in use due to the: high percentage of wireless phone usage. availability of wireless phone applications (apps). rapid decline in landline phone penetration. Many research firms are expanding usage of wireless phone surveys as a result of two advantages over Internet and landline phone surveys: Immediacy Portability Wireless phone surveys provide immediacy in the sense that consumers can fill out surveys close to the moments of shopping, decision making, and consuming. For example, wireless phone surveys have been used to: capture impulse purchases as they are made and consumed. collect data about side effects in real time from patients who are participating in pharmacological testing. survey wireless customers. A company named Kinesis Survey Research offers researchers the option of attaching a miniature bar code reader to a mobile phone that will discretely collect and store bar codes of purchased items. Wireless phone panels may be especially appropriate for surveying teens, early adopters, and impulse buyers. Researchers primarily survey in either text-based or web-based formats. In short messaging (text messaging) format, the respondent can access the survey and display it as text messages on a wireless phone screen. The texting format is used for simple polling and very short surveys. In Europe, mobile phone penetration rates are very high and text message usage is higher than in the United States, so it is often preferred to wireless web surveys. In the United States, wireless web surveys are used more often than text message surveys. As compared to text messaging, wireless web surveys have the following benefits: The web facilitates a continuous session, with no time delay between questions and receipt of responses. Wireless web surveys tend to be cheaper for both the recipient and administrator of the survey. Wireless web surveys also permit some functionality associated with CATI and Internet surveys to be used, including conditional branching and display of images. When CATI abilities are added to wireless web surveys, the result is called CAMI, or computer aided mobile interviewing. Marketing researchers usually do not call wireless phone users to solicit participation as often as they do over landlines for the following reasons: FCC regulations prevent the use of autodialing. Thus, when potential respondents are called, each one must be dialed by hand. Wireless phone respondents may incur a cost when taking a survey. Wireless phone respondents could be anywhere when they are called, meaning they are likely to be distracted by other activities, and may disconnect in the middle of the call. Safety is a potential issue since the respondents could be driving when they get the call from researchers. A second advantage of wireless surveys is their ability to reach mobile phone only households, which are increasing. Utilizing wireless phone surveys along with other methods enables research firms to reach consumers they otherwise could not include in their surveys. Several challenges facing the use of wireless phone surveys are: Because of limited screen space, wireless phone surveys are not suitable for research that involves long and/or complex questions and responses. Even though mobile phones have some capacity to handle graphics, that capacity is somewhat limited. Wireless panels currently provide relatively small sample sizes. C. Self-Administered Surveys (PPT slide 5-9 and 5-11) A self-administered survey is a data collection technique in which the respondent reads survey questions and records his or her own responses without the presence of a trained interviewer (PPT slides 5-9). The advantages and disadvantages of self-administered surveys are shown in Exhibit 5.4 (PPT slide 5-11). Advantages: Low cost per survey Respondent control No interviewer-respondent bias Anonymity in responses Disadvantages: Minimize flexibility High nonresponse rates Potential response errors Slow data acquisition Lack of monitoring capability There are four types of self-administered surveys: Mail surveys Mail panels surveys Drop-off surveys Online surveys Mail surveys typically are sent to respondents using the postal service. This type of survey is inexpensive to implement as there are no interviewer-related costs such as compensation, training, travel, or search costs. Another advantage is that mail surveys can reach even hard-to- interview people. One major drawback is lower response rates than with face-to-face or telephone interviews, which create nonresponse bias. Another problem is that of misunderstood or skipped questions. Mail surveys are slow since there can be a significant time lag between when the survey is mailed and when it is returned. Mail panel surveys are questionnaires sent to a group of individuals who have agreed to participate in advance. The panel can be tested prior to the survey so the researcher knows the panel is representative. This prior agreement usually results in high response rates. In addition, mail panel surveys can be used for longitudinal research. That is, the same people can be questioned several times over an extended period. This enables the researcher to observe changes in panel members’ responses over time. The major drawback to mail panels is that members are often not representative of the target population at large. Drop-off survey is a self-administered questionnaire that a representative of the researcher hand-delivers to selected respondents; the completed surveys are returned by mail or picked up by the representative. The advantages of drop-off surveys include the availability of a person who can answer general questions, screen potential respondents, and create interest in completing the questionnaire. The disadvantage of drop-offs is they are more expensive than mail surveys. Online surveys are the most frequently used survey methods today in marketing research which collect data using the Internet (Exhibit 5-5). An important advantage for online surveys is they are less expensive per respondent than other survey methods. Surveys are self-administered, and no coding is necessary. The ability of Internet surveys to collect data from hard-to-reach samples is another important reason for the growth of online surveys. Other advantages of online surveys include the improved functional capabilities of website technologies over pen and pencil surveys. One functional improvement is the ability to randomize the order of questions within a group, so that the effects of question order on responses is removed. Another important improvement over other types of surveys is that missing data can be eliminated. Also, marketing research firms are now learning how to use the improved graphic and animation capabilities of the web. In addition to using online panels created by marketing research firms, companies may survey their own customers using their existing e-mail lists to send out invitations to participate in surveys. While the benefits of online surveys include low cost per completed interview, quick data collection, and the ability to use visual stimuli, Internet samples are rarely representative and nonresponse bias can be high. United States have home access to the Internet, which limits the ability to generalize to the general population. Propensity scoring can be used to adjust the results to look more like those a representative sample would have produced, but the accuracy of this procedure must be evaluated. With propensity scoring the responses of underrepresented sample members are weighted more heavily to adjust for sampling inadequacies. The primary disadvantage with propensity scoring is that respondents who are demographically similar may be otherwise different. V. Selecting the Appropriate Survey Method (PPT slide 5-12 to 5-14) Researchers must consider situational, task, and respondent factors when choosing a survey method. A. Situational Characteristics (PPT slide 5-12) In an ideal situation, researchers could focus solely on the collection of accurate data. However, in reality, researchers must cope with the competing objectives of budget, time, and data quality. In choosing a survey research method, the goal is to produce usable data in as short a time as possible at the lowest cost. It is easy to generate large amounts of data in a short time if quality is ignored. But excellent data quality often can be achieved only through expensive and time-consuming methods. In selecting the survey method, the researcher commonly considers the following situational characteristics in combination (PPT slide 5-12): Budget—includes all the resources available to the researcher. The budget is considered along with data quality and time in selecting a survey method. Completion time frame—long time frames give researchers the luxury of selecting the method that will produce the highest quality data. In many situations, however, the affordable time frame is much shorter than desired, forcing the researcher to choose a method that may not be ideal. Some surveys, such as direct mail or personal interviews, require relatively long time frames. Other methods, such as Internet surveys, telephone surveys, or mall intercepts, can be done more quickly. Quality requirements—data quality is a complex issue that encompasses issues of scale measurement, questionnaire design, sample design, and data analysis. Three key issues that help explain the impact of data quality on the selection of survey methods are: Completeness of data—refers to the depth and breadth of the data Data generalizability—data that are generalizable accurately representing the population being studied and can be accurately projected to the target population Data precision—refers to the degree of accuracy of a response in relation to some other possible answer. B. Task Characteristics (PPT slide 5-13) Researchers ask respondents to engage in tasks that take time and effort. Task characteristics include: Task difficulty Stimuli needed to elicit the response Amount of information needed from the respondent Research topic sensitivity—topic sensitivity is the degree to which a specific survey question leads the respondent to give a socially acceptable response. When asked about a sensitive issue, some respondents will feel they should give a socially acceptable response even if they actually feel or behave otherwise. C. Respondent Characteristics (PPT slide 5-14) Since most marketing research projects target prespecified groups of people, the third major factor in selecting the appropriate survey method is the respondents’ characteristics (PPT slide 5-14). The extent to which members of the target group of respondents share common characteristics influences the survey method selected. Diversity of respondents refers to the degree to which respondents share characteristics. The more diverse the respondents the fewer similarities they share. The less diverse the respondents the more similarities they share. The incidence rate is the percentage of the general population that is the focus of the research. Sometimes researchers are interested in a segment of the general population that is relatively large and the incidence rate is high. Normally, the incidence rate is expressed as a percentage. Thus, an incidence rate of 5 percent means that 5 out of 100 members of the general population have the qualifying characteristics sought in a given study. Respondent participation involves three components: The respondent’s ability to participate—ability to participate refers to the ability of both interviewers and respondents to get together in a question-and-answer interchange. The ability of respondents to share thoughts with interviewers is an important selection consideration. The respondent’s willingness to participate—willingness to participate is the respondent’s inclination or disposition to share his or her thoughts. The respondent’s knowledge—knowledge level is the degree to which the selected respondents feel they have the knowledge or experience to answer questions about the survey topic. Respondents’ knowledge levels play a critical role in whether or not they agree to participate, and directly impacts the quality of data collected. Marketing researchers have developed “best practices” to increase participation levels some of which include offering some type of incentive and personally delivering questionnaires to potential respondents. VI. Causal Research Designs (PPT slide 5-15 and 5-16) Causal research designs differ from exploratory or descriptive designs in several ways: The primary focus of causal research is to obtain data that enables researchers to assess the “cause-effect” relationships between two or more variables. The concept of causality between several independent variables (X) and one dependent variable (Y) in research designs specifies relationships that are investigated in causal research studies and stated as “If X, they Y.” Three fundamental conditions must exist in order to accurately conclude that a cause-effect relationship exists between variables: Researchers must establish that there is temporal order between the independent X and the dependent Y variables such that variable X (or a change in X) must occur prior to observing or measuring variable Y. Researchers must establish that collected data confirm there is some type of meaningful association between variable X and variable Y. Researchers must account for (or control for) all other possible variables other than X that might cause a change in variable Y. Another difference between causal and descriptive research is that causal research requires that researchers collect data using experimental designs. An experiment involves carefully designed data collection procedures in which researchers manipulate a proposed causal independent variable, and observe (measure) the proposed effect on a dependent variable, while controlling for all other variables. Exploratory and survey research designs typically lack the “control” mechanism of causal designs. Typically, researchers use either a controlled laboratory environment where the study is conducted in an artificial setting where the effect of all, or nearly all, uncontrollable variables is minimized. In a field environment, researchers use a natural setting similar to the context of the study in which one or more of the independent variables are manipulated under conditions controlled as carefully as the situation will permit. Finally, while exploratory and descriptive designs almost always involve data collection using surveys, experimental designs collect data using both surveys and observation. In fact, in recent years one of the most often executed experimental designs is online research that observes online activity to determine which marketing mix variables are likely to influence website traffic patterns and ultimately purchases. A third difference is the framing of research questions for causal designs. In exploratory and survey research designs, initial research questions are typically framed broadly and hypotheses focus on the magnitude and/or direction of the association, and not on causality. In contrast, questions examining causal relationships between variables are framed with the focus being on the specific impact (or influence) one variable causes on another variable. A. The Nature of Experimentation (PPT slide 5-17 and 5-18) Exploratory and descriptive research designs are useful for many types of studies. But they do not verify causal links between marketing variables. In contrast, experiments are causal research designs and can explain cause-and-effect relationships between variables/constructs and determine why events occur. Marketing research often involves measurement of variables. A variable is a concept or construct that can vary or have more than one value (PPT slide 5-17). In marketing, variables include demographics such as age, gender and income, attitudes such as brand loyalty and customer satisfaction, outcomes such as sales and profits, and behaviors such as media consumption, website traffic, purchase, and product usage. When conducting an experiment, researchers attempt to identify the relationships between variables of interest. Experimental research is primarily a hypothesis-testing method that examines hypotheses about relationships between independent and dependent variables. Researchers develop hypotheses and then design an experiment to test them. To do so, researchers must identify the independent variables that might bring about changes in one or more dependent variables. Experiments and other causal designs are most appropriate when the researcher wants to find out why certain events occur, and why they happen under certain conditions and not others. Experiments provide stronger evidence of causal relationships than exploratory or descriptive designs because of the control made possible by causal research designs. Experiments enable marketing researchers to control the research situation so that causal relationships among the variables can be examined. In a typical experiment, the independent variable is manipulated (changed) and its effect on another variable (dependent variable) is measured and evaluated. Control variables are variables that the researcher does not allow to vary freely or systematically with independent variables. These should not change as the independent variable is manipulated. Extraneous variables are any variables that experimental researchers do not measure or control but that may affect the dependent variable. Extraneous variables include the respondent’s mood or feelings, the temperature of the room in which the experiment is taking place, or even the general weather conditions at the time of the experiment. Exhibit 5.6 explains these concepts in more detail. B. Validity Concerns with Experimental Research (PPT slide 5-19) In any type of research design, but particularly in causal designs, researchers must understand validity and take steps to ensure they have achieved it. Validity is the extent to which the conclusions drawn from a particular research design, such as an experiment are true (PPT slide 5-19). Internal validity refers to the extent to which the research design accurately identifies causal relationships. In other words, internal validity exists when the researcher can rule out competing explanations for the conclusions about the hypothesized relationship. External validity means the results of the experiment can be generalized to the target population. Random selection of subjects and random assignment to treatment conditions are usually necessary for external validity, but they are not necessarily sufficient to confirm that the findings can be generalized. D. Comparing Laboratory and Field Experiments (PPT slide 5-20) Marketing researchers use two types of experiments (PPT slide 5-20): Laboratory—a laboratory (lab) experiment is conducted in an artificial setting. It enables the researcher to control the setting and therefore achieve high internal validity. But the trade-off is that laboratory experiments lack external validity. Field—field experiment is a causal research design that manipulates the independent variables in order to measure the dependent variable in a natural setting. It is performed in a natural or “real” setting. But high levels of realism mean that independent and extraneous variables are difficult to control. Besides realism and control, the three other issues to consider when deciding whether to use a field experiment are: Time frames—field experiments take longer to complete than laboratory experiments. The planning stage—which can include determining which test market cities to use and which retailers to approach with product experiments, securing advertising time, and coordinating the distribution of the experimental product—adds to the length of time needed to conduct field experiments. Costs—field experiments are more expensive to conduct than laboratory experiments because of the high number of independent variables that must be manipulated. Competitive reactions—because field experiments are conducted in a natural setting, competitors can learn about the new product almost as soon as it is introduced and respond by using heavy promotional activity to invalidate the results of the experiment or by rushing similar products to market. If secrecy is desired, then laboratory experiments are generally more effective. E. Test Marketing (PPT slide 5-21) Test marketing is the use of experiments to obtain information on market performance indicators (PPT slide 5-21). For example, marketing mix variables (product, price, place, and promotion) are manipulated and changes in dependent variables such as sales volume or website traffic are measured. Test marketing, often referred to as a controlled field experiment, has three broad applications in marketing research: Test marketing has long been used to pilot test new product introductions or product modifications. Test marketing is used to explore different options of marketing mix elements. Product weaknesses or strengths, or inconsistencies in the marketing strategies are frequently examined in test marketing. The main objectives of test marketing are to: predict sales. identify possible customer reactions. anticipate adverse consequences of marketing programs. Test marketing measures the sales potential of a product or service and evaluates variables in the marketing mix. MARKETING RESEARCH DASHBOARD RIDERS FITS NEW DATABASE INTO BRAND LAUNCH (PPT slide 5-22 and 5-23) The Marketing Research Dashboard provides a discussion of how Lee Apparel Company used market test data from a field experiment to build a customer database which then helped to launch a new brand of jeans under the name Riders. The approach used a promotion to generate product trial while collecting basic customer contact information. Three months after the initial promotion, customers were contacted with a telephone survey. This provided more detailed information. The information was used to effectively position the new brand. Instructor Manual for Essentials of Marketing Research Joseph F. Hair, Mary Celsi, Robert P. Bush, David J. Ortinau 9780078028816, 9780078112119
Download
Close
Close
