MKT605
Why is qualitative research needed?
If the subject is new and there is no research yet.
Why do researchers use quasi experimental designs?
Used when researchers can't exercise a high degree of control. They are also quicker and less expensive.
The author talked about dividing the sample into two parts on p. 573. What is the analysis sample? What is the validation sample? Why do researchers need both samples?
- Analysis sample: part of the total sample that is used for estimation of the discriminant function. - Validation sample: that part of the total sample used to check the results of the estimation sample. -->Researchers need both to see if their estimate is valid: double check.
What are the differences between content analyses and trace analyses?
- Content analysis: the objective, systematic, and quantitative description of the manifest content of a communication. It includes observation as well as analysis (for example looking at mass media). This is an appropriate method when the phenomenon to be observed is communication, rather than behavior or physical objects. - Trace analysis: an approach in which data collection is based on physical traces, or evidence, of past behavior (for example going thought someone's garbage).
We discussed multivariate dependence techniques in Weeks 4 and 5. How do dependence techniques differ from interdependence techniques?
- Dependence techniques: multivariate techniques appropriate when one or more of the variables can be identified as dependent variables and the remaining as independent variables. - Interdependence techniques: multivariate statistical techniques that attempt to group data based on underlying similarity, and thus allow for interpretation of the data structures. No distinction is made as to which variables are dependent and which are independent.
What are the null and alternative hypotheses in regression?
- H0 = absence of an effect from an independent variable to the dependent variable (no difference) à sign. val. ≥ 0.05 - H1: = presence of an effect from an independent variable to the dependent variable (difference) à sign. val. <0.05
Are there null and alternative hypotheses in a factor analysis? Why or why not?
- H0= absence of a multivariate association à sign. value ≥ 0.05 o If we support the null hypotheses, we do not start the factor analysis. - H1= presence of a multivariate association 00> sign. value < 0.05 o If we support the alternative hypotheses, we do start the factor analysis.
What are the differences between mail interview and mail panel?
- Mail interviews: a group of individuals that are willing to participate in mail survey. You usually have a database with willing individuals and the way of contacting them and sending them the email is through (e)mail. - Mail panel: consists of a large, nationally representative sample of households that have agreed to participate in periodic mail questionnaires, product tests and telephone surveys. The households are compensated with various incentives.
What are the null and alternative hypotheses in t-Test? What are the null and alternative hypotheses in one-way analysis of variance (i.e., ANOVA)?
- T-Test: o H0 = absence of an effect (no difference between two groups) à sign. val. ≥ 0.05 o H1: = presence of an effect (difference between two) à sign. val. <0.05 - One-way analysis of variance (ANOVA): o H0 = absence of an effect (no difference among three or more groups) à sign. val. ≥ 0.05 o H1: = presence of an effect (difference among three or more groups) à sign. val. <0.05
We discussed the classification of univariate techniques and the classification of multivariate techniques last week. Are there any univariate techniques in Figure 16.1? Why or why not? Are there any multivariate techniques in Figure 16.1? Why or why not?
- Univariate techniques in figure 16.1 (max 2 variables): t-test & one-way analysis of variance - Multivariate techniques in figure 16.1 (min 3 variables): regression & N-way analysis of variance
How many factors are identified if the eigenvalues are 3.171, 1.256, 1.056, .438, .071, and .009?
3
What is a constant sum scale? Are constant sum scales interval or ratio scales?
A constant sum scale is a comparative scaling technique in which respondents are required to allocate a constant sum of units. Since this allows researchers to rank-order the objects and compare differences, this is a ratio scale.
What is a continuous rating scale? Are continuous rating scales interval or ratio scales? Explain your answer.
A continuous rating scale is a measurement scale that has the respondents rate the objects by placing a mark at the appropriate position on a line that runs from one extreme to the criterion variable to the other. This is an interval scale.
The author discussed dummy variables in regression on p. 557. What are dummy variables? Are dummy variables interval variables? Why nor why not?
A dummy variable is a respecification procedure using variables that take on only two values, usually 0 or 1. A dummy variable is a numerical variable used in regression analysis to represent subgroups of the sample in your study. -->So no, dummy variables are not interval variables but nominal variables.
The concept of filter questions is explained on p. 308. What are filter questions? Are they needed for the study illustrated in Figure 10.2? Why or why not?
A filter question is an initial question in a questionnaire that screens potential respondents to ensure they meet the requirements of the sample. They are needed for that study.
What is the objective to do frequency distribution? What statistics are used to measure location? What statistics are used to measure variability?
A frequency distribution is a mathematical distribution whose objective is to obtain a count of the number of responses associated with different values of one variable and to express these counts in percentage terms. Objective is to know how many people belong to a certain category. Measures of location: mode, median, mean. Measures of variability: range, interquartile range, standard deviation, coefficient of variation.
What do consumers need? Provide some examples.
A need is an opportunity to deliver a benefit to a customer. Competition is intense: companies compete for the same customers trying to satisfy their needs. Product needs could be things like functionality, design, performance etc., whereas service needs it could be transparency, or accessibility
What is a nominal scale? Provides some examples. Are nominal scales comparative or noncomparative scales? Explain your answer.
A nominal scale is a scale whose numbers serve only as labels or tags for identifying and classifying objects. This is a trick question because the concept of comparative and noncomparative scales doesn't apply to nominal scales
Read your Survey Analysis questionnaire. How many categories are there in Q7 Retirement Status? How many dummy variables can Q7 be converted to? How to convert Q7 to these dummy variables?
According to the book, four categories means three dummy variables, however, according to SPSS, three categories means three dummy variables.
What is an ordinal scale? Provides some examples. Are ordinal scales comparative or noncomparative scales? Explain your answer.
An ordinal scale is a ranking scale in which numbers are assigned to objects to indicate the relative extent to which some characteristic is possessed. Since this scale implies it ranking one object over another, it is a comparative scale
What is analysis of covariance (i.e., ANCOVA)? Why do researchers run ANCOVA?
Analysis of covariance includes at least one categorical independent variable and at least one interval or metric independent variable. -->ANCOVA = ANOVA + covariates Sometimes ANOVA is not enough because you want to add questions into the analysis, which is why researches would use ANCOVA
There are three types of itemized rating scales. Which one will be the most appropriate if we need to measure brand attitude? Explain your answer.
Attitudes are often measured using the Likert scale. All three can be used depending on what you want to research. Usually, you want to check out previous research so you can see how those researchers did it and they do what they did.
There are five approaches in Approach to the Problem in Fig 2.1. Which one was used in the "Real Research Major League Baseball Majors in Research" case?
Research questions was used: whether they should get rid of non-price promotions.
Are traditional telephone surveys, in-home surveys and personal mall intercept surveys outdated? Why or why not?
Yes, they are outdated. It was popular years ago, because back then, there was no internet and artificial intelligence yet. Now that there is, this technique is outdated.
What is discriminant analysis? What are the null and alternative hypotheses in discriminant analysis?
Discriminant analysis is a technique for analyzing marketing research data when the criterion or dependent variable is categorical, and the predictor or independent variables are interval in nature. - H0 = absence of an effect from an independent variable to the dependent variable (no difference) à sign. val. ≥ 0.05 - H1: = presence of an effect from an independent variable to the dependent variable (difference) à sign. val. <0.05
Why do researchers include dummy variables in regression?
Due to an insufficiency in ANOVA: post hoc analysis is disabled when covariates are added into the analysis, which is why you'd have to use dummy variables in regression.
What are the differences between email surveys and internet surveys?
E-mail surveys: survey is written within the body of the e-mail message. The e-mails are sent out over the Internet. E-mail surveys use pure text (ASCII). Respondents come from a list of e-mail addresses obtained. Internet surveys use HTML, the language of the Web, and are posted on a Web site, respondents are not recruited; they happen to be visiting the Web site where the survey is posted.
The author discussed an elbow criterion when discussing the decision on the number of dimensions. If, based on the elbow criterion, 4 dimensions are identified in an MDS study, what shall we do?
Elbow criterion: a plot of stress versus dimensionality used in MDS. The point at which an elbow or a sharp bend occurs indicates an appropriate dimensionality. -->We live in a three-dimensional world, there is no use in doing a four-dimensional study. So the answer is: do nothing; the study is useless.
The author discussed the concept of eigenvalues when discussing the determination of factor numbers (p. 611). What is an eigenvalue? How many eigenvalues will be generated if 6 variables are used in a factor analysis?
Eigenvalue tells you how many variables a factor is going to be able to represent à in other words: how many variables you're going to be able to put into one group. In theory, when you have six variables, you are going to have six eigenvalues, but the values that are below 1 are useless.
We discussed scale reliability and validity when going over Figure 9.5. Are reliability and validity related to the determination of model fit? Why or why not?
Yes, they are related: - Reliability is about internal consistency - Validity is about construct convergent.
Which type of projective techniques in Figure 5.2 was used in the "Real Research What Will the Neighbors Say" case? Was it appropriate? Why or why not?
Expressive techniques (third-person technique) were used. It was appropriate because respondents gave less socially desirable answers.
What is generalizability? How to achieve generalizability?
Generalizability is the degree to which a study is based on a sample applies to a universe of generalizations. To achieve generalizability, we need to make sure the sample truly represents the population well; we use sighs what signs to get the right people into our sample. -->Generalizability is achieved through test and retest and study samples from a particular universe
What are the null and alternative hypotheses in cross-tabulation analyses?
H0 = absence of an effect (no association) à sign. val. ≥ 0.05. H1: = presence of an effect (association) à sign. val. <0.05
What are the null and alternative hypotheses in a difference analysis on two independent samples?
H0 = absence of an effect (no difference) à sign. val. ≥ 0.05 H1: = presence of an effect (difference) à sign. val. <0.05
What are the null and alternative hypotheses in tests of association? What are the null and alternative hypotheses in tests of differences?
H0= absence of an effect (no association/no difference) --> sign. val. ≥ 0.05 H1= presence of an effect (association/difference) --> sign. val. <0.05
What are the null and alternative hypotheses in the Kolmogorov-Smirov(K-S) one-sample test?
H0= assumption met (normality) --> sign. val. ≥ 0.05 (distribution normal). H1= assumption not met (no normality) --> sign. val. <0.05 (distribution not normal).
There are two clustering procedures: hierarchical and nonhierarchical. In which method do researchers have to pre-specify the number of clusters?
In the non-hierarchical cluster.
How are branching questions used in the study illustrated in Figure 10.2? Are there any issues? Why or why not?
In the study, questions have different answers like yes, no, and cash, credit or other. Depending on the answer given, the interviewer is guided to a different spot in the questionnaire. Yes, there are issues. For example after saying "no" and "no", that should be the end of the survey. However, the survey continues to ask whether you want a store card.
The author discussed the concept of interactions on p. 509. What is an interaction? Is there just one interaction in a two-way ANOVA? Why or why not?
Interaction effect: the effect of an independent variable on a dependent variable is different for different categories or levels of another independent variable. Yes, there is just one interaction in a two-way ANOVA
Did the airlines benefit from the research? Why or why not?
It benefited American Airlines, because it helped them to effectively position the airline and design a strategy. The airlines could benefit from the research if the managers have the resources to implement the research findings.
Is rotation of factors a required step in factor analysis? Why or why not?
It depends on the situation: - If there is some kind of controversy as to where to assign the variable (for example do you assign Jeep to US brands or to cars?), we do rotation. - If there is no controversy (straightforward assignment), rotation is not necessary.
The Classification of Marketing Research Design was covered in Module 1. Is the Classification of Marketing Research Design different from or similar to the Classification of Marketing Research Data in Fig 5.1? Explain your answer.
It's similar. Exploratory research is the same as qualitative and conclusive research is the same as quantitative.
What depth-interviewing techniques were used in the "Real Research Hidden Issues and Hidden Dimensions in Air Travel" case? Who were interviewed?
Laddering was used; the male middle managers were interviewed.
Was simple random sampling used in the "Real Research Online Retirement Plans Are On" case? Why or why not?
No, because the sample was stratified the income and ago so they didn't use the simple stratified sampling technique. Simple random sampling would mean you treat everyone as one group, which cannot be the case here.
The author discussed the concept of loading when discussing the rotation of factors (pp. 612-613). What is loading? Can the value of a loading larger than 1 or smaller than -1? Why is loading related to the rotation of factors?
Loading shows the variance explained by the variable on that particular factor. -->It has to be between -1 and 1; it cannot exceed 1 or be below -1 (max 1, min -1).
The author also discussed multivariate analysis of variance (MANOVA) on p. 519? How does MANOVA differ from ANOVA?
MANOVA differs from ANOVA in terms of the number of dependent variables. - If we run ANOVA there is only one dependent variable. - If we run MANOVA we can add more than one dependent variable. -->And running multiple ANOVA's with a different dependent variable each time instead of one MANOVA with multiple dependent variables results in a higher probability of making Type 1 Errors (and thus a lower confidence level).
What is marketing?
Marketing is the activity, set of institutions, and processes for creating, communicating, delivering, and exchanging offerings that have value for customers, clients, partners and society at large.
What is α? What is β?
Probability of alpha error (type I error) and probability beta error (type II error). It's okay to make mistakes, but there is a limit: α: 0.05, β: 0.20. We also use alpha and beta to determine sample size (online calculator).
What is marketing research?
Marketing research links the consumer, customer and public to the marketer through information. Marketing research is the systematic and objective identification, collection, analysis, dissemination, and use of information for the purpose of improving decision making related to the identification and solution of problems and opportunities in marketing. Marketing research supports the marketing strategy.
What statistics are used to measure the shape of a distribution? Is the Kolmogorov-Smirov(K-S) one-sample test described under One Sample on p. 478 relevant to the shape of a distribution? Why or why not?
Measures of shape: skewness, kurtosis. The Kolmogorov-Smirov(K-S) one-sample test is relevant to the shape of distribution, because it compares the cumulative distribution function for a variable is specified distribution
What are missing responses?
Missing responses represent values of a variable that are unknown, because these respondents did not provide unambiguous answers to the question.
What are multivariate techniques? Are there any similarities between multivariate and univariate techniques? Why or why not?
Multivariate techniques: statistical techniques suitable for analyzing data when there are two or more measurements on each element and the variables are analyzed simultaneously. Univariate techniques use one-way ANOVA which is the same as the multivariate technique analysis of variance. Multivariate's cross-tabulation is the same as univariate's chi-square.
Are there perfect solutions to missing responses? Why or why not?
No, all four techniques have their flaws, so no there is no perfect solution to missing responses.
If the alternative hypothesis of a three-group one-way ANOVA is supported, does it mean that the three groups differ from one another? Why or why not?
No, there are different kind of combinations in this post hoc analysis. So, it may not be true that the three groups different from one another, even if that was supposed to be the alternative hypothesis.
Are there dependent, extraneous, and independent variables in a factor analysis? Why or why not?
No, these concepts do not apply to factor analysis. Factor analysis is an interdependence variable.
The concept of sampling frame is explained on p. 341. Is the sampling frame in the "Real Research Online Retirement Plans Are On" case appropriate? Why or why not?
No, this sampling frame is not appropriate because you will not participate during office hours, so that's a problem
Can observation methods replace survey methods in research? Why or why not?
No, we do observation to get the how part and use the survey to get the why part. Together we have a complete understanding of what is happening. We need both methods and cannot replace one with another
Do we always need the 6 steps in the Marketing Research Process? Why or why not? Be specific.
No, we don't always have to do step 4 (fieldwork/data collection). It depends on whether or not we have to talk to people or can use artificial intelligence
There are two types of conclusive research, descriptive and causal, in Fig 3.1. Were both of them used in the "Real Research Women's Golf Apparel Market Is in 'Full Swing'" case? Explain your answer
No. Causal research (with experiments) was not used, but they did do descriptive research.
There are two kinds of descriptive research, cross-sectional and longitudinal, in Fig 3.1. Were both of them used in the "Real Research Women's Golf Apparel Market Is in 'Full Swing'" case? Explain your answer.
No. They only used longitudinal designs based on panels. Cross-sectional means they only interviewed participants once, which was not the case here as participants were followed for years
Convenience sampling was used in all the surveys conducted by the IOC, and the results from all these surveys were positive. Do you feel surprised? Why or why not?
Not surprised since convenience samples are not representative of any definable population. Of course people who attend the Olympic Games don't think these are too commercial, because they are contributing to this themselves by attending the event.
Are audits a secondary research method or a primary research method? Explain your answer.
Often secondary, but it depends on the type of audit/what you are trying to get from the audit.
What type of focus groups were used in the "Real Research Enhancing the Utility of Sports Utility Vehicles" case? Was it appropriate? Why or why not?
Online focus groups. It was appropriate, because they were targeting a young and active kind of population for the SUV. Using online focus groups is probably the best way to target that audience as they all have access to internet.
Two major methods of factor analysis are principal components analysis and common factor analysis. How do they differ from each other?
Only in the case of principal components analysis is it possible to compute exact factor scores & only the overlapping information is used; in common factor analysis estimates of these scores are obtained & all information we have is used.
What was the management decisions problem in the "Real Research Major League Baseball Majors in Research" case? Was it appropriate? Why or why not?
Research problem was whether or not they could get rid of non-price promotion. It was not appropriate, because it has special meaning to fans.
The Classification of Marketing Research Data was covered in Module 2. Do researchers use the scaling techniques in Figure 8.2 to collect quantitative data only? Can the scaling techniques be used to collect observational data? Explain your answer.
Scaling means collecting numbers; we cannot collect qualitative data (measures of types) through scaling. So yes, researchers use scaling techniques to collect quantitative data (measures of values) only. And yes, scaling techniques can be used to collect observational data, but when we observe we do not ask any questions so we have to use scaling techniques where no questions are asked.
What is stepwise regression? When to use it?
Stepwise regression: regression procedure in which the predictor variables enter or leave the regression equation one at a time. The purpose of stepwise regression is to select, from a large number of predictor variables, a small subset of variables that account for most of the variation in the dependent or criterion variable.
What is marketing strategy?
Target market; competition; proper 'mix' of the product/service, price, promotion and distribution system
Two studies were conducted to understand why some people did not fly in the "Real Research What Will the Neighbors Say" case? Were the answers from the second study more truthful than those from the first study? Why or why not?
The answers from the second study were more truthful because they didn't give socially desirable answers.
How does discriminant analysis differ from ANOVA? How does discriminant analysis differ from regression?
The difference is about the kind of variables used: - For discriminant analysis we have categorical variables as the dependent variable, and for the independent variable, interval variables are used. - Whereas ANOVA uses the categorical variable as the independent variable and the dependent variables are the interval variables. - For regression the dependent variable has to be the interval variable, whereas the dependent variable for discriminant analysis has to be the nominal or categorical variable.
Respondents' inability to remember is discussed on pp. 308-309. Is the inability to remember an issue for the study illustrated in Figure 10.2? Why or why not?
The inability to remember would be an issue for the study. You don't have time to check your bank statements during the study so you need to remember. And it's hard for respondents to remember what they did two months ago, so it's an issue.
What is the marketing concept?
The market concept is a business philosophy that holds that the key to achieving organizational goals consists of the company's being more effective than competitors in creating, delivering and communicating customer value to its chosen target market
The author discussed product moment correlation, partial correlation, and bivariate regression. What is the relationship among them?
The relationship among them is the linear association between the two variables (if we increase one variable, we are going to see an effect on the other variable).
There are two categories of research designs, exploratory and conclusive, in Fig 3.1. Were both of them used in the "Real Research Women's Golf Apparel Market Is in 'Full Swing'" case? Explain your answer.
They used both exploratory and conclusive research (surveys and panels).
There are four methods in Tasks Involved in Fig 2.1. Which one was used in the "Real Research Major League Baseball Majors in Research" case?
They used discussion with decision makers because they wanted to see if they could get rid of non-price promotion.
How many factors are there in a two-way analysis of variance (i.e., two-way ANOVA)?
Two-way ANOVA means we are looking at more than two variables. There are two factors.
What is Type I error? What is Type II error? Among the four cells (A, B, C, & D) in the table below, which one indicates the Type 1 error and which one indicates the Type II error?
Type I error: alpha error, it occurs when the sample results lead to the rejection of a null hypothesis that is in fact true. In table, cell C. Type II Error: beta error, it occurs when the sample results lead to the nonrejection of a null hypothesis that is in fact false. In table, cell B.
What are univariate techniques? Are the techniques in Figure 14.6 relevant to the scaling techniques in Figure 8.2 (covered in Module 3)? Why or why not?
Univariate techniques are statistical techniques appropriate for analyzing data when there is a single measurement of each element in the sample or, if there are several measurements on each element, but each variable is analyzed in isolation. The paired t-test (univariate, two samples, related) is connected to the paired comparison, so yes the techniques are relevant.
How to interpret factors?
When we try to interpret factors, we try to assign a name to the factors. Making sense of factors means naming factors.
Can causality be investigated through statistical designs? Why or why not?
Yes, causality can be investigated through statistical design
Can cluster sampling be used in the "Real Research Online Retirement Plans Are On" case? Why or why not?
Yes, cluster sampling can be used, but you have to use it together with stratified sampling.
Can the convenience sampling technique in the "Real Research Olympic Convenience" case be replaced with the other nonprobability sampling techniques? Why or why not?
Yes, convenience sampling can be replaced with other nonprobability sampling techniques. For example with judgment sampling.
Can the other techniques in Figure 5.2 be used in the "Real Research Enhancing the Utility of Sports Utility Vehicles" case? Why or why not?
Yes, depth interviews and projective techniques could also be used, but online focus groups are the best option.
There are two types of errors in measurement: systematic and random. Does reliability address random error only? Does validity address both errors? Explain your answer.
Yes, reliability only affects random errors because systematics has adverse effects because they affect the measurement in a consistent way that doesn't need inconsistency. Validity does not address both errors.
The author discussed similarity judgments when discussing input data. Are similarity judgments based on paired comparisons?
Yes, similarity judgements are based on paired comparisons. We do comparisons to see how similar or different things are. Low difference means high similarity and vice versa so it's really about making a judgement on similarity.
In addition to the scales shown in Figure 8.2, we also discussed another four scales, nominal, ordinal, interval, and ratio, in Module 3. Are the techniques in Figure 14.6 relevant to these four scales as well? Why or why not?
Yes, the techniques are relevant to the four scales of nominal ordinance interval and ratio because univariate techniques are classified based on whether the data our metric which would be interval or ratio skills or non-metric nominal or ordinance scales
The concepts of reliability, validity, and generalizability were covered in Figure 9.5. Are these concepts also relevant to the sampling techniques in Figure 11.2? Why or why not?
Yes, they are relevant. If you want to have high quality information, reliability, validity and generalizability, you want to use probability sampling techniques (due to the use of science in these techniques). And vice versa.
Can pre-experimental designs be flawed? Why or why not?
Yes, they can be flawed. They could be subject to threats of validity because they do not employ randomization procedures, which control for extraneous factors. There is no benchmark so we can't compare the findings to anything.
Can projective techniques be used in focus groups or depth interviews? Why or why not?
Yes, we could use the technique in both of them (projective technique is just a fancy way of asking questions), but depth interviews is probably the best option, considering the social pressure in focus groups.
Can the convenience sampling technique in the "Real Research Olympic Convenience" case be replaced with one of the four probability sampling techniques? Why or why not?
Yes, you can replace it with (scientific) probability sampling. But, in this case, convenience sampling isn't even a problem because everyone likes the Olympics so it doesn't matter what method we use. Then, convenience sampling would be the most cost effective. You could also use the other sampling methods. We will probably get the same results as with convenience sampling. That is an exception however; usually convenience sampling is not good.
Is the Classification of Marketing Research in Fig 1.1 relevant to the Marketing Concept Model? Why or why not?
Yes. Marketers need to have some idea to start/develop marketing strategy. To develop a marketing strategy, marketers must understand trend, also known as problem identification research. We need to translate the need/trend into a product. We need numbers to confirm/indicate the validity of the strategy: problem-solving research. Then, we need to assess the actual performance of the strategy: monitoring marketing performance.
