Chapter 11 Basic Data Analysis for Quantitative Research Answers to Hands-On Exercise 1. What are other areas of improvement for Remington’s? In the importance-performance chart, Remington’s is rated low in “Speed of Service” but that is an important selection criterion for people evaluating places to dine out. Remington’s should do something about that. Either actually changing their speed of service or making the target population “think” they have faster service could be the possible options. The latter could be accomplished by making the wait time pass more pleasantly, perhaps with live entertainment. They should also stress their “speedy service” in promotional material. Again, customers’ evaluation of “Speed of Service” is as much about perceptions as it is about reality. 2. Run post-hoc ANOVA tests between the competitor groups. What additional problems or challenges does this reveal? Listed below are the additional problems or challenges: • FOOD QUALITY: Remington’s is rated significantly better than either Outback or Longhorn. • SPEED OF SERVICE: Remington’s is a (statistically significant) poor third behind both Outback and Longhorn. • REASONABLE PRICES: Remington’s and Outback are tied with Longhorn behind by significant margin. • LARGE PORTIONS: Remington’s and Outback are tied and both are much ahead of Longhorn. • ATMOSPHERE: Remington’s is statistically ahead of both Outback and Longhorn, but in an area not particularly important as a selection criterion. • COMPETENT EMPLOYEES: Remington’s is a (statistically) poor third in the field. This is the least important of the selection criteria considered. 3. What new marketing strategies would you suggest? Following are some of the possible marketing strategies that could be suggested: • FOOD QUALITY: Maintain the lead. Stress this difference in promotional messages. • SPEED OF SERVICE: Do something about the disadvantage. See Question 1 above for strategies. • REASONABLE PRICES: This is a bad place to try to compete. Match Outback, but don’t get into a price war. That’s too costly and there are no winners. Stress food quality and value in the promotional messages. • LARGE PORTIONS: Again, pretty much head to head with Outback. This is a place where Remington’s can try to gain ground and not have it cost too much. Slightly increasing the size of portions doesn’t cost all that much and is immediately noticed. • ATMOSPHERE: Remington’s is ahead of the pack, by a significant margin. Too bad this isn’t an important selection criterion for patrons. If Remington’s can make it more important for them through promotion, it would be a plus. If it can’t do that and if its efforts to create “atmosphere” are costly, it might cut back a little on the expense. • COMPETENT EMPLOYEES: Remington’s is a poor third in this relatively unimportant area. One strategy would be not to worry about it. A better strategy would be to try to improve the rating without spending a lot of money. A little more attention to employee selection and a little more training can make a big difference. ANSWERS TO REVIEW QUESTIONS 1. Explain the difference between the mean, the median, and the mode. The mean, median, and mode are the three key measures of central tendency. The mean is the arithmetical average of responses. The median is the number that sits in the middle of the data set when you line the numbers up from lowest to highest. The mode is defined as the most common number in the data set. 2. Why and how would you use Chi-square and t-tests in hypothesis testing? Both the t-test and Chi-square tests are used to examine hypotheses that propose differences between groups. The Chi-square analysis enables researchers to test for statistical significance between the frequency distributions of two or more nominally scaled variables in a cross-tabulation table to determine if there is any association between the variables. An analyst can use the t-test to compare two means when the sample size is smaller than 30 and the standard deviation is unknown. The t value is a ratio of the difference between the two sample means and the standard error. The t test provides a mathematical way of determining if the difference between two sample means occurred by chance. For example, suppose that one wants to examine whether the occurrence of eating at fast food restaurants is higher among men or women. The variables are gender (male or female) and dining at a fast food restaurant within the last week (yes or no). The Chi-Square test can determine whether differences between males and females are significant. The t-test however assesses differences when there is one nominal or categorical variable and one interval or ratio scaled variable. For instance, if one wanted to assess whether number of times one dines at fast food restaurants during a specific period of time varied between males and females, the t-test would be appropriate. 3. Why and when would you want to use ANOVA in marketing research? Analysis of variance (ANOVA) is a statistical technique that determines whether three or more means are statistically different from one another. Thus, it is similar to the t-test but allows for more than two groups to be examined. For example, if one wanted to assess whether people of different ethnic backgrounds were more or less likely to prefer Nike brand shoes, an ANOVA could be used to test this hypothesis. The independent variable is ethnicity (a categorical variable) while the dependent variable is brand preference (an interval-scaled variable). The t-test cannot be used because there are more than two categories for the independent variable, so the ANOVA would be used. 4. What will ANOVA tests not tell you, and how can you overcome this problem? While ANOVA can tell the research team statistical differences exists somewhere between the sampled means it can’t identify which specific means are different from each other. Statistical packages such as SPSS overcome this problem by using a series of “follow-up” tests which involve multiple comparisons taking means two at a time. ANSWERS TO DISCUSSION QUESTIONS 1. The measures of central tendency discussed in this chapter are designed to reveal information about the center of a distribution of values. Measures of dispersion provide information about the spread of all the values in a distribution around the center values. Assume you were conducting an opinion poll on voters’ approval ratings of the job performance of the mayor of the city where you live. Do you think the mayor would be more interested in the central tendency or the dispersion measures associated with the responses to your poll? Why? The best way to begin getting started with this question is to nail down a facet or two about the survey instrument which was used to collect the data. For example, suggest to your students that, above and beyond some categorical and continuous questions used at the outset for demographic purposes, the lion’s share of the questions in the survey are (i) Likert Scale questions; and, (ii) they include a neutral response e.g., Do you support Mayor Grey’s views on abortion on demand? [Answer Responses: Strongly Disagree (Code 1), Disagree (Code 2), No Opinion (Code 3), Agree (Code 4), Strongly Agree (Code 5)]. It can be argued that the Mayor would be interested in the measures of central tendency and dispersion. With respect to measures of central tendency, the median might be of most interest to the Mayor and her/his team, since the aforementioned questions are interval in nature and extreme responses may “skew” or distort the true feelings and attitudes of the voting public. Finally, with measures of dispersion the range would be worthwhile to explore, especially with an eye to the age range of the respondents. The Mayor will want to capitalize on certain “segments” of the voting public in future advertisements, and town-hall meetings if the results show a particular age group (e.g., ,females between 21- 29 years of age, or elderly residents of the area) are aligned with the policies she/he has introduced into legislation, and approval ratings are high with these groups. 2. If you were interested in finding out whether or not young adults (21-34 years old) are more likely to buy products online than older adults (35 or more years old), how would you phrase your null hypothesis? What is the implicit alternative hypothesis accompanying your null hypothesis? Null Hypothesis: There isn’t a significant and measurable relationship between age and purchasing products online. (Assumption: The age ranges are 21-34 years of age, and 35+ respectively). Alternate Hypothesis: There is a significant and measurable relationship between age and purchasing products online. Phrasing the null and alternate hypothesis in this form will allow the research team to make comparisons between the two groups (21-34 years of age, and 35+) and determine if important differences exist between them. 3. The level of significance (alpha) associated with testing a null hypothesis is also referred to as the probability of a Type I error. Alpha is the probability of rejecting the null hypothesis on the basis of your sample data when it is, in fact, true for the population of interest. Because alpha concerns the probability of making a mistake in your analysis, should you always try to set this value as small as possible? Why or why not? A short answer to this question is: “It depends.” The more appropriate answer is: “It depends on the amount of risk regarding the accuracy of the test that the research team (in consultation with the client) is willing to accept. The most prevalent significance levels used in marketing research are 0.10, 0.05, and 0.01. It’s important to note that within the language-game of marketing research there are “confidence estimates” which mirror each of these levels of significance: a level of significance of 0.10 produces a 90% confidence estimate; a level of significance of 0.05 produces a 95% confidence estimate; and a level of significance of 0.01 produces a 99% confidence estimate. 4. Analysis of variance (ANOVA) allows you to test for the statistical difference between two or more means. Typically, there are more than two means tested. If the ANOVA results for a set of data reveal that the four means that were compared are significantly different from each other, how would you find out which individual means were statistically different from each other? What statistical techniques would you apply to answer this question? Note that ANOVA can’t identify which pairs of means are significantly different from each other. In this case we have four means to consider so it’s clear (by virtue of what is and isn’t possible via ANOVA) that class participants need to “stretch” and explore other techniques. There are a number of follow-up tests containing within the SPSS environment called “comparison” tests. One such test (Scheffe) brings all pairs of means alongside a high and low confidence interval range; and in this case we might find that one or the other pair of means is statistically different. Because the Scheffe technique holds the error rate to α (typically 0.05), the research team has more assurance that true differences would exist between these four means. 5. EXPERIENCE THE INTERNET: Nike, Reebok, and Converse are strong competitors in the athletic shoe market. The three use different advertising and marketing strategies to appeal to their target markets. Use one of the search engines on the Internet to identify information on this market. Go to the websites for these three companies (www.Nike.com; www.Reebok.com; www.Converse.com. Gather background information on each, including its target market and market share. Design a questionnaire based on this information and survey a sample of students. Prepare a report on the different perceptions of each of these three companies, their shoes, and related aspects. Present the report in class and defend your findings. This is a very comprehensive discussion question. A number of directives and “tips” which should help guide them toward the necessary items to bring closure to this question are provided below for your information. Nike: Once you arrive at the home page for Nike, it’s best to begin to collect information for this discussion question by clicking on the “About Nike” link. Information about Nike’s target market and market share is located under the link entitled “investors”. However, there are a number of other links where students can find “added value” especially when it comes to constructing a questionnaire which drills down into consumer perception about image, products, and related aspects like diversity, job opportunities and the like. Exploring the FAQ section is always an interesting aside; and Nike appears to be positioned as a lengthy discussion on sourcing labor outside the United States. Reebok: Reebok’s home page is slightly more helpful when it comes to navigating to the questions contained in this assignment. The primary link to source information about market share and target marketing is “About Reebok”, found at the bottom of the home page. The links appear to be more aligned with what’s asked above (e.g., products, experts, core training, FAQ and the like. There is a clear emphasis on women from a positioning standpoint, as well as a link which opens up a cadre of information about Reebok’s efforts in the arena of Human Rights. This should provide your class participants with some “lift-off” when it comes to constructing a survey which deals with the company’s image, products and these “cause-oriented” strategies. Converse: The site for Converse is a tad frustrating. Most of the information required to complete this assignment is found under the “Store Locator” link. Converse’s product line (like Reebok) places much more emphasis on women; however, your class participants will note there’s less of an emphasis on aligning the company with any “cause.” 6. SPSS EXERCISE: Form a team of three to four students in your class. Select one or two local franchises to conduct a survey on, such as Subway or McDonald’s. Design a brief survey (10-12 questions) including questions like ratings on quality of food, speed of service, knowledge of employees, attitudes of employees, and price, as well as several demographic variables such as age, address, how often individuals eat there, and day of week and time of day. Obtain permission from the franchises to interview their customers at a convenient time, usually when they are leaving. Assure the franchiser you will not bother customers and that you will provide the franchise with a valuable report on your findings. Develop frequency charts, pie charts, and similar graphic displays of findings, where appropriate. Use statistics to test hypotheses, such as “Perceptions of speed of service differ by time of day or day of week.” Prepare a report and present it to your class; particularly point out where statistically significant differences exist and why. There are a number of opportunities for the class participants to build upon the knowledge-base concerning SPSS, articulating variables relevant to a research endeavor, and bringing a number of important statistical techniques to bear on the data they collect. When it comes to Discussion Question 6, perhaps the most important component of having the students deliver a successful product is for the instructor to network with franchises like Subway, Burger King, McDonald’s and the like well in advance of the time she/he distributes this assignment. 7. SPSS EXERCISE: Using SPSS and the Santa Fe Grill employee database, provide frequencies, means, modes, and medians for the relevant variables on the questionnaire. The questionnaire is shown in Chapter 10. In addition, develop bar charts and pie charts where appropriate for the data you analyzed. Run an ANOVA using the work environment perceptions variables to identify any differences that may exist between male and female employees, and part-time versus full-time employees. Be prepared to present a report on your findings. Again, Discussion Question 7 provides some hands-on experience with the SPSS program, and encourages students to become familiar with the “basic statistics” associated with any research endeavor—frequency distributions, and measures of central tendency. In addition, it points them toward the benefits of displaying the outcomes of analysis “graphically”, via bar and pie charts. When requesting frequency distributions, it might be wise to take a moment and discuss “counts”, “percents”, “cumulative counts” and “cumulative percents” and the relative merits and/or pitfalls of presenting all four to decision makers when reporting results. The Santa Fe Grill questionnaire contains codes and a host of question formats we have canvassed thus far. 8. SPSS EXERCISE: Review the Marketing Research in Action case for this chapter. There were three restaurant competitors—Remington’s, Outback, and Longhorn. Results for a one-way ANOVA of the restaurant image variable were provided. Now run post-hoc ANOVA follow-up tests to see where the group differences are. Make recommendations for new marketing strategies for Remington’s compared to the competition. This is pretty much a repeat of the questions at the end of the Hands-On Exercise. The answer is the same. Follow-up tests yield the following results: • FOOD QUALITY: Remington’s is rated significantly better than either Outback or Longhorn. • SPEED OF SERVICE: Remington’s is a (statistically significant) poor third behind both Outback and Longhorn. • REASONABLE PRICES: Remington’s and Outback are tied with Longhorn behind by significant margin. • LARGE PORTIONS: Remington’s and Outback are tied and both are much ahead of Longhorn. • ATMOSPHERE: Remington’s is statistically ahead of both Outback and Longhorn, but in an area not particularly important as a selection criterion. • COMPETENT EMPLOYEES: Remington’s is a (statistically) poor third in the field. This is the least important of the selection criteria considered. Based upon the above, the following marketing strategies are suggested: • FOOD QUALITY: Maintain the lead. Stress this difference in promotional messages. • SPEED OF SERVICE: Do something about the disadvantage. • REASONABLE PRICES: This is a bad place to try to compete. Match Outback, but don’t get into a price war. That’s too costly and there are no winners. Stress food quality and value in the promotional messages. • LARGE PORTIONS: Again, pretty much head to head with Outback. This is a place where Remington’s can try to gain ground and not have it cost too much. Slightly increasing the size of portions doesn’t cost all that much and is immediately noticed. • ATMOSPHERE: Remington’s is ahead of the pack, by a significant margin. Too bad this isn’t an important selection criterion for patrons. If Remington’s can make it more important for them through promotion, it would be a plus. If it can’t do that and if its efforts to create “atmosphere” are costly, it might cut back a little on the expense. • COMPETENT EMPLOYEES: Remington’s is a poor third in this relatively unimportant area. One strategy would be not to worry about it. A better strategy would be to try to improve the rating without spending a lot of money. A little more attention to employee selection and a little more training can make a big difference. Chapter 12 Examining Relationships in Quantitative Research Answers to Hands-On Exercise 1. Will the results of this regression model be useful to the QualKote plant manager? If yes, how? The results are not helpful because the survey focused on the wrong population—employees rather than customers. To the extent that plant employees have accurate perceptions of the level of customer satisfaction, they might be surveyed as an item of interest. But, it would be a mistake to use the results of this group to make customer service policy decisions. 2. Which independent variables are helpful in predicting A36–Customer Satisfaction? The model shows that A10, A12, A17, and A23 were significant predictors of A36. The model that includes these four independent variables has an r2 of 67.0. That means that 67 percent of the observed variation in the dependent variable, A36, can be explained by the combination of the four predictor variables. The problem remains, however, these are measures of plant employee perceptions, not of actual customer reports of either the efforts to improve service or the results. It isn’t all that useful to ask someone to report how satisfied someone else might be. Even field sales personnel, ones who come into direct contact with actual customers, would be a more logical population to survey than plant employees. 3. How would the manager interpret the mean values for the variables reported in Exhibit 12.16? The scale used to assess the variables was a 7-point scale with higher numbers indicating agreement. The means suggest then that customer satisfaction is near neutral. The use of data from external sources in strategic planning (A10) is slightly positive, as is the use of a systematic process to translate customer requirements into new products. However, the use of a systematic process to determine customer expectations (A31) slants negative as does perceptions that customer requirements are used in developing strategic goals (A17). 4. What other regression models might be examined with the questions from this survey? What QualKote really wants to know is whether its customers are satisfied and how much each specific activity (input into planning, customer input into new product development, quality, etc.) contributes to that satisfaction. This data set won’t help very much in making that determination. ANSWERS TO REVIEW QUESTIONS 1. Explain the difference between testing for significant differences and testing for association. Testing for significant differences via Z-tests, t-tests, and ANOVA, focuses on exploring the differences between means of a variety of groups being sampled. Marketing managers are often interested in digging deeper—checking if there are consistent and systematic “ties” between their characteristics (e.g. income levels, gender, political affiliation) and buying behavior. 2. Explain the difference between association and causation. The difference between association and causation is best explained with an example. If the independent variable is whether it is a sunny day and the dependent variable is gains in the Dow Jones, this is a relationship built on association. We can’t say the sunshine causes investors to trade, just that the two items happen together (are associated with each other). 3. What is covariation? How does it differ from correlation? Covariation speaks of the degree of association between two items in a research endeavor. Another way of putting it is that covariation is the amount of change in one variable (e.g. class status: freshman, sophomore, junior, senior) that is consistently related to the change in another variable of interest (e.g. attendance in class). To say correlation is different from covariation is not exactly accurate, since the two concepts compliment each other. Relationships between variables are viewed in four ways: presence, direction, strength, and type. Correlation is a statistical measure of the strength of a linear relationship between two variables. The floor of the relationship is taken to be -1.00, the ceiling +1.00; so the higher the number is to the ceiling, the stronger the relationship between the variables in question. 4. What are the differences between univariate and bivariate statistical techniques? A univariate analysis involves testing one variable at a time, while a bivariate analysis involves two variables. 5. What is regression analysis? When would you use it? Correlation techniques enable the research team to explore the strength and direction (positive, negative, linear, or curvilinear) between two variables. Regression analysis utilizes an “equation” which compares variables in such a way so the manager can get beyond using intuition or past data which has been captured regarding a relationship to forecast the future. In short, regression analysis is a remarkable tool which allows managers to predict what the future will be like. For example, she may be interested in predicting what effect a price increase of twenty-five cents may have on the profitability of a detergent category and/or its market share. Simple regression involves the research team exploring one independent variable and one dependent variable. The assumption behind multiple regression (and it’s a fitting one given the complexity of the market) is that there is a whole bunch of stuff going on at the same time; therefore, when it comes to any practical problem involving buyers in a market it would be wise to assume there are a number of independent variables (operating individually or in concert with each others) that affect a dependent variable of interest. 6. What is the difference between simple regression and multiple regression? Simple regression involves just one independent and one dependent variable, while multiple regression is appropriate for multiple variables. ANSWERS TO DISCUSSION QUESTIONS 1. Regression and correlation analysis both describe the strength of linear relationships between variables. Consider the concepts of education and income. Many people would say these two variables are related in a linear fashion. As education increases, income usually increases (although not necessarily at the same rate). Can you think of two variables that are related in such a way that their relationship changes over their range of possible values (i.e., in a curvilinear fashion)? How would you analyze the relationship between two such variables? Data can be linear and curvilinear in nature, but most of the examples cited in this chapter concern relationships between variables that are linear in nature. When it comes to thinking about two variables that are related in a curvilinear fashion here’s three examples you can give for consideration: (a) Assume you purchase a piranha. Feeding your piranha a goldfish or two daily will cause a linear relationship to obtain (where the independent variable ‘X’ is the number of goldfish dropped into the tank, and the dependent variable ‘Y’ is the piranha’s sustenance. However, if you drop 10-20 goldfish into the tank, a curvilinear relationship may obtain; namely, the piranha will continue to gorge itself until it bloats up and dies. (b) Another example can be illustrated by the relationship between watering a plant (the independent variable ‘X’), and the growth rate of the plant (the dependent variable ‘Y’). If you water the plant regularly the plant should grow and mature. However, if you continue to over saturate the plant with water it may wither and perish. This is another example of a relationship which becomes curvilinear over time. (c) A final example of a curvilinear relationship (one which your class participants can certainly relate to) is the relationship between cutting classes (the independent variable ‘X”) and the final grade in an undergraduate course (the dependent variable ‘Y’). This relationship can be considered curvilinear in the sense that the more the independent variable increases, the more the dependent variable decreases. The Pearson Correlation technique would be useful to apply to the three examples cited above. The primary challenge when it comes to bringing statistical techniques like correlation and regression to bear on these examples (or any variables for that matter) is that few (if any) relationships are truly deterministic in nature. Curvilinear relationships are exceedingly complex, especially since decisions about outcomes in the marketplace are characterized by large amounts of uncertainty. 2. Is it possible to conduct a regression analysis on two variables and obtain a significant regression equation (significant F-ratio), but still have a low r2? What does the r2 statistic measure? How can you have a low r2 yet still get a statistically significant F-ratio for the overall regression equation? The r2 statistic measures the amount of total variation in a dependent variable which can be explained by using the independent variable. When a low r2 statistic surfaces in SPSS output it normally suggests that a relationship is present in the sampled population, but it really isn’t sturdy. However, as the question suggests, a low r2 statistic shouldn’t be seen as implying the relationship between variables in a research endeavor isn’t significant. An F-ratio can be statistically significant and the r2 statistic can remain relatively small in size. The larger the F-ratio, the more likely the null hypothesis will be rejected. All things being equal, this would suggest it is very possible to have a significant F-ratio and a low r2 statistic present in the output for a regression analysis. 3. The ordinary least squares (OLS) procedure commonly used in regression produces a line of “best fit” for the data to which it is applied. How would you define best fit in regression analysis? What is there about the procedure that guarantees a best fit to the data? What assumptions about the use of a regression technique are necessary to produce this result? The best way to define a “best fit” in regression analysis is with reference to an r-squared statistic. This is normally expressed as a percentage (%) in the output produced by statistical software such as SAS or SPSS. A “best fit” occurs when a straight line relationship obtains between “Y” (the response or dependent variable) and the “X” variable (normally referred to as the ‘predictor’ variable). Caution is advised; however, since predictions are the only conclusions we can draw from a regression analysis. To say that one variable causes the behavior of the other is a theoretical misnomer, and misleading from a practical standpoint. The specifics of the procedure which guarantees this “best fit” are two-fold: (a) Like correlation analysis, regression equations assumes that a linear relationship provides a good description of the relationship between two variables, and (b) When regression are significant and high (e.g., above 88%) we can say that a relationship is present in our sampled population and it is sturdy. Simple regression analysis carries the following three assumptions: (a) error terms associated with making predictions are normally and independently distributed; (b) the variables of interest to the research team (and client) are measured on interval and/or ratio scales; and (c) the variables come from a normal population. 4. When multiple independent variables are used to predict a dependent variable in multiple regression, multicollinearity among the independent variables is often a concern. What is the main problem caused by high multicollinearity among the independent variables in a multiple regression equation? Can you still achieve a high r2 for your regression equation if multicollinearity is present in your data? Multicollinearity makes it tough to estimate independent or separate regression coefficients for correlated variables. This is because the independent variables of interest to the research team are highly correlated with each other. In this respect the short answer to the question “Can a regression equation produce a high r2 statistic when multicollinearity is present in the data which has been captured?” is “No”. The major impact of multicollinearity is felt most acutely in the statistical significance of the regression coefficients present in the output when the regression equation is applied. Therefore, it would be unlikely (if not counter-intuitive) for a data set characterized by multicollinearity to produce a high r2 statistic in regression analysis. However, because relatively few relationships between variables can be known with certainty (no error) our response to this discussion question leaves the window open for anomalies of “practice” that “theory” appears to shut down. 5. EXPERIENCE THE INTERNET: Choose a retailer that students are likely to patronize and that sells in both catalogs and on the Internet (e.g., Victoria’s Secret). Prepare a questionnaire that compares the experience of shopping in the catalog with shopping online. Then ask a sample of students to visit the website, look at the catalogs you have brought to class, and then complete the questionnaire. Enter the data into a software package and assess your finding statistically. Prepare a report that compares catalog and online shopping. Be able to defend your conclusions. This exercise could actually be an excellent one for reviewing the class materials to date. Because students are given a task (to compare catalog and online shopping) and asked to prepare a questionnaire, they can review developing research questions, measurements, and questionnaires in the process. This will likely enhance the task and value of data collection and data analysis. 6. SPSS EXERCISE: Choose one or two other students from your class and form a team. Identify the different retailers from your community where wireless phones, digital recorders/players, TVs, and other electronics products are sold. Team members should divide up and visit all the different stores and describe the products and brands that are sold in each. Also observe the layout in the store, the store personnel, and the type of advertising the store uses. In other words, familiarize yourself with each retailer’s marketing mix. Use your knowledge of the marketing mix to design a questionnaire. Interview approximately 100 people who are familiar with all the retailers you selected and collect their responses. Analyze the responses using a statistical software package such as SPSS. Prepare a report of your findings, including whether the perceptions of each of the stores are similar or different, and particularly whether the differences are statistically or substantively different. Present your findings in class and be prepared to defend your conclusions and your use of statistical techniques. This would be an excellent small group project. Be sure to assign it with sufficient time to complete the work involved. 7. SPSS EXERCISW: Santa Fe Grill owners believe their employees are happy working for the restaurant and unlikely to search for another job. Use the Santa Fe Grill employee database and run a bivariate regression analysis between X11–Team Cooperates and X17–Likelihood of Searching for another Job to test this hypothesis. Could this hypothesis be better examined with multiple regression? If yes, execute a multiple regression and explain the results. The correlation coefficient will suggest whether a relationship exists between these two variables. If the relationship is significant and positive, then one could state that perceptions of the owners of the Santa Fe Grill that their employees are unlikely to search for another job because they are happy working for the restaurant. This specific hypothesis would not be examined with multiple regression because only two variables are involved (one dependent and one independent). We could use simple regression. Chapter 13 Communicating Marketing Research Findings Answers to Hands-On Exercise 1. What other issues can be examined with this survey? Answers will vary, but possibilities include comparing the variables of product perceptions, purchase likelihood, and price consciousness for different levels of education, technology ownership, or income. 2. What problems do you see with the questionnaire? The primary problem is that questions 1-10 do not deal with technology even though the survey is supposed to be about determinants of technology adoption. Another issue is the method by which the responders were divided into innovators or early adopters. They were split on the basis of five variables which are somewhat like what might be used to assess opinion leadership. This does not mean, though, that the items can be used to determine whether someone is a technology innovator. Further, it certainly cannot distinguish whether someone is an early adopter or a member of another adoption category. Lastly, some commonly used demographic questions are not included. 3. What are the important topics to include in a presentation of the findings? The presentation should present a sequence of slides indicating the objective of the research and the specific research questions to be addressed, followed by the research methodology employed and a description of the sample survey. Additional slides should be developed that highlight the research findings or particular results of the study which the researcher deems important for communication purposes. Finally, the presentation should conclude with recommendations, conclusions, and research implications as they pertain to the study at hand. ANSWERS TO REVIEW QUESTIONS 1. What are the seven components of the marketing research report? Briefly discuss each objective and why it is important. The seven components of the marketing research report are: • The research objectives • The research questions • Literature review and relevant secondary data • A description of the research methods • Findings displayed in tables, graphs, or charts • Interpretation and summary of the findings • Conclusions and recommendations The objectives and questions in qualitative research tend to be broader, more general, and more open-ended than in quantitative research. The literature review and relevant secondary data may be integrated in the analysis of findings in qualitative data analysis, rather than being presented separately from other findings. The description of research methods in both qualitative and quantitative research helps to develop credibility for both kinds of research projects, but different kinds of evidence are offered in developing credibility in quantitative and qualitative analyses. Data display is important in both methods. Qualitative researchers rarely present statistics, but they are the bread and butter of a quantitative presentation. Writing conclusions and recommendations is the final step in both qualitative and quantitative reports. 2. In the context of the marketing research report, what is the primary goal of the executive summary? Many consider it the soul of the report, insofar as many executives read only the report summary. It must be complete enough to provide a true representation of the entire document but in summary form—the research objectives, a concise statement of method, a summary of the findings, and specific conclusions and recommendations. 3. What are the primary topics/issues that need to be addressed in the research methods-and-procedures section of a marketing research report? Essentially, a methods-and-procedures section sets out to tell the client how the marketing research was done. Issues addressed in this section include the following: • The research design used: exploratory, descriptive, and/or causal. • Types of secondary data included in the study, if any. • If primary data were collected, what procedure was used (observation, questionnaire) and what administration procedures were employed (personal, mail, telephone, Internet)? • Sample and sampling processes used. The following issues are usually addressed: ○ How the sample population was defined and profiled? ○ Sampling units used (for example, businesses, households, individuals) ○ The sampling list (if any) used in the study ○ How the sample size was determined? ○ Was a probability or nonprobability sampling plan employed? 4. Why are conclusions and recommendations included in a marketing research report A research report which proceeds to management without articulating conclusions and recommendations only serves to reinforce the conceptual divide between researchers and decision-makers. Conclusions are defined as broad generalizations that focus on answering questions related to the research studies objectives. Recommendations take the obligations of the research team up a notch, to a level which is often avoided by students and practitioners alike; namely, they ask people to “think”. 5. What are the common problems associated with the marketing research report? The five common problems associated with the marketing research report are: • Lack of data interpretation • Unnecessary use of complex statistics • Emphasis on packaging over quality • Lack of relevance to the stated research objectives • Placing too much importance on a few statistics 6. Why is it important to explain limitations in your marketing research report? It is important to explain limitations in marketing research report because the limitations serve as qualifiers for the findings and recommendations made. They ensure that the decision makers understand that no study is perfect and that the recommendations are limited by certain situations. ANSWERS TO DISCUSSION QUESTIONS 1. EXPERIENCE THE INTERNET. Go to the following website: www.microsoft.com/Education/Tutorials.aspx. Complete the Tutorials dialog box by typing in “higher education” in the Grade Level box, “technology” in the Learning Area box, and “PowerPoint” in the Product box. After selecting and completing the tutorial, provide written comments on the benefits you received by taking the tutorial. Students’ responses will differ. This is a valuable exercise for ensuring basic presentation development skills. 2. Select the Santa Fe Grill data or one of the other databases provided with this text (see Deli Depot, Remington’s, Qualkote, or Digital Recorder Survey on the website), analyze the data using the appropriate statistical techniques, prepare a PowerPoint presentation of your findings, and make the presentation to your research class. a. Select an appropriate variable from the data set and prepare a simple bar chart of the findings in SPSS. b. Select an appropriate variable from the data set and prepare a simple pie chart of the findings in SPSS. c. Select a group of thematically related items that are on metric scales. Present the results in a table and also in a bar chart using SPSS. d. Find two categorical items that are appropriate for a Crosstab and present your results in a bar chart made with SPSS. e. Find a categorical independent variable and interval level dependent variable. Present the results in a bar chart made with SPSS. f. Choose an outcome variable that can be explained by two or more independent variables. Run a regression and then develop a diagram (using PowerPoint or Word) that displays your findings. Responses will vary, depending upon the data set and variables selected. 3. There are several PowerPoint presentations for the Santa Fe Grill Restaurant study on the book’s website at www.mhhe.com/hairessentials3e. The presentations demonstrate how findings of a statistical analysis of data from a survey can be reported. Review the presentations and select the one you believe most effectively communicates the findings. Justify your choice? The most effective PowerPoint presentation for communicating the findings of the statistical analysis of data from the Santa Fe Grill Restaurant study would be the one that strikes a balance between clarity, visual appeal, and comprehensiveness. After reviewing the presentations available on the book’s website, www.mhhe.com/hairessentials3e, I believe that the presentation titled "Customer Perception of Santa Fe Grill Restaurant: Statistical Analysis Findings" stands out as the most effective for several reasons. 1. Clear and Concise Organization: This presentation likely organizes the findings in a clear and logical manner, presenting each key result in a structured format. This makes it easier for the audience to follow the flow of information and understand the implications of the statistical analysis. 2. Visual Appeal: Effective use of visuals such as charts, graphs, and tables can enhance the audience's understanding and retention of the findings. The chosen presentation likely utilizes visuals effectively to illustrate key points and trends, making the information more engaging and memorable. 3. Comprehensive Coverage: The selected presentation is likely to provide a comprehensive overview of the statistical analysis findings, covering important aspects such as customer satisfaction levels, perceptions of service quality, preferences, and any significant correlations or trends identified in the data. A comprehensive presentation ensures that the audience gains a thorough understanding of the study findings and their implications for the Santa Fe Grill Restaurant. 4. Accessibility: The presentation is likely to be accessible and easy to navigate, allowing the audience to quickly locate relevant information and refer back to specific findings as needed. Clear labeling, consistent formatting, and user-friendly design contribute to the accessibility of the presentation. In summary, the presentation titled "Customer Perception of Santa Fe Grill Restaurant: Statistical Analysis Findings" is likely the most effective for communicating the results of the statistical analysis from the Santa Fe Grill Restaurant study due to its clear organization, visual appeal, comprehensive coverage, and accessibility. These qualities ensure that the audience can easily grasp the key findings and understand their significance for the restaurant's operations and future decision-making. Solution Manual for Essentials of Marketing Research Joseph F. Hair, Mary Celsi, Robert P. Bush, David J. Ortinau 9780078028816, 9780078112119
Download
Close
Close
