No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
The Significance of Statistical Significance Tests in Marketing Research
Abstract
Classical statistical significance testing is the primary method by which marketing researchers empirically test hypotheses and draw inferences about theories. The authors discuss the interpretation and value of classical statistical significance tests and suggest that classical inferential statistics may be misinterpreted and overvalued by marketing researchers in judging research results. Replication, Bayesian hypothesis testing, meta-analysis, and strong inference are examined as approaches for augmenting conventional statistical analyses.
... 3 2 Problems General to Null Hypothesis Significance Testing 2.1 Preface As noted, the NHST paradigm upon which the status quo and the Benjamin et al. (2018) proposal rest is the default statistical paradigm for research, publication, and discovery in the biomedical and social sciences (see, for example, Morrison and Henkel (1970); Gigerenzer (1987); Sawyer and Peter (1983); McCloskey and Ziliak (1996); Gill (1999); Anderson et al. (2000); Gigerenzer (2004); Hubbard (2004)). Despite this, it has been roundly criticized both inside and outside of statistics over the decades (see, for example, Rozenboom (1960); Bakan (1966); Meehl (1978); Serlin and Lapsley (1993); Cohen (1994); McCloskey and Ziliak (1996); Schmidt (1996);Hunter (1997); Gill (1999); Gigerenzer (2004); Gigerenzer et al. (2004);Briggs (2016); Gal (2016, 2017)). ...
We discuss problems the null hypothesis significance testing (NHST) paradigm poses for replication and more broadly in the biomedical and social sciences as well as how these problems remain unresolved by proposals involving modified p-value thresholds, confidence intervals, and Bayes factors. We then discuss our own proposal, which is to abandon statistical significance. We recommend dropping the NHST paradigm--and the p-value thresholds intrinsic to it--as the default statistical paradigm for research, publication, and discovery in the biomedical and social sciences. Specifically, we propose that the p-value be demoted from its threshold screening role and instead, treated continuously, be considered along with currently subordinate factors (e.g., related prior evidence, plausibility of mechanism, study design and data quality, real world costs and benefits, novelty of finding, and other factors that vary by research domain) as just one among many pieces of evidence. We have no desire to "ban" p-values or other purely statistical measures. Rather, we believe that such measures should not be thresholded and that, thresholded or not, they should not take priority over the currently subordinate factors. We also argue that it seldom makes sense to calibrate evidence as a function of p-values or other purely statistical measures. We offer recommendations for how our proposal can be implemented in the scientific publication process as well as in statistical decision making more broadly.
... Path coefficients, standard errors, t-values and p-values for the structural model were presented using a bootstrapping approach with 5,000 replicate samples as suggested by Hair et al. (2019) and Ramayah et al. (2018). Based on Hahn and Ang's (2017) critique that p-values are not a good criterion for testing the significance of hypotheses, several alternative methods were suggested, including the use of replication studies (Sawyer & Peter, 1983;Singh, Ang, & Leong, 2003), effect size estimates, and confidence intervals (Aguinis et al. 2010;Ely, 1999;Hubbard and Meyer, 2013;Lin, Lucas, and Shmueli, 2013). Therefore, several decision rules were used in the study, including the p-value criterion, effect size, and confidence intervals. ...
The COVID-19 pandemic had an immediate and significant impact on consumer behavior, radically changing attitudes and behaviors, as reflected in sales figures. While some researchers believed the pandemic would lead to lasting changes, others expected a return to pre-pandemic behavioral patterns. Despite this ambiguity, there has been little research on the factors influencing consumer behavior, especially about the demand for everyday goods. This study, rooted in the theory of planned behavior (TPB) and the protection motivation theory (PMT), investigated the determinants of consumer behavior in the purchase of daily necessities following the COVID-19 outbreak. In light of the possibility that the pandemic could have caused a new pattern of consumer behavior during the endemic phase, a convenience sampling of 1438 respondents was conducted. Using Partial Least Square Structural Equation Modelling (PLS-SEM), the present study found that choice, information, and perceived economic stability (PES) positively affected consumer behavior, while COVID-19 and risk had negative effects. Interestingly, stress had no impact on consumer behavior. These findings provide valuable insights for companies desiring to understand and anticipate customer behavior during and beyond the pandemic. The present study not only identifies the factors influencing consumer behavior but also enables salespeople to develop strategies to proactively address changes in consumer behaviors.
... In this context, it should be mentioned that replication research is not the only way to improve confidence in scientific claims. Scholars have argued in favor of research practices such as relying on substantive rather than statistical significance (Sawyer and Peter 1983), requiring preregistration of empirical studies (nosek et al. 2018), and meta-analysis (Schmidt 1992) as a means to increase confidence in research results. However, their advocates acknowledge that these research practices serve as a supplement to replication studies rather than a replacement. ...
Drawing on recent research and debates in social sciences, this paper situates replication in an advertising research context. We clarify the role of replication in the field and outline the challenges inherent in replication studies in advertising research. We further elaborate on how researchers should engage in replication research to increase the truth value of advertising research while overcoming the obstacles to replication research. Finally, we discuss how advertising scholars, reviewers, and editors can facilitate replication research to reduce the share of false-positive results and accumulate knowledge in the discipline. We see replication as critical in advertising research, given the high variability of experimental factors and the applied nature of the field. Therefore, a better understanding of replications and the challenges of advertising research should inspire scholars to engage in more replication attempts and reviewers and editors to consider it for publication.
... Accordingly, two competing hypotheses are being developed and evaluated. Scholars have noted that examining competing hypotheses enhances objectivity (Armstrong et al., 2001) and is particularly useful when prior knowledge allows different plausible justifications (Chaker et al., 2019;Sawyer and Peter, 1983). ...
Purpose
Through activism, brands participate in the sociopolitical controversies that shape society today. Based on social identity theory, this study aims to examine the moderating effects of consumer–brand identification (CBI) and political ideology in explaining consumer responses to brand activism. Furthermore, the role of perceived marginalization that can arise in the case of consumer–brand disagreement is explored.
Design/methodology/approach
The hypothesized effects were tested in three experiments. Study 1 (n = 262) and Study 2 (n = 322) used a moderation analysis, which was supplemented by a mixed design analysis with repeated measures in Study 1. In Study 3 (n = 383), the mediating effect of perceived marginalization by the brand was tested using a moderated mediation model.
Findings
The results show that strong CBI as well as a conservative ideology buffer the negative effects of consumer–brand disagreement on brand attitude and word-of-mouth intentions. In the case of agreement with a brand’s stance, no direct or interactive effects of brand activism on consumer responses occur. Perceived marginalization by a brand mediates the effects of brand activism.
Originality/value
This study extends the “love is blind” versus “love becomes hate” debate to the realm of brand activism and finds evidence for the former effect. It also contributes to the research on political consumption by highlighting the role of political ideology as an important boundary condition for brand activism. Perceived marginalization is identified as a relevant risk for activist brands.
... The current study used a 5% significance level for stricter tests for statistical significance on the examined parameter relationships. This significance level is commonly used in many marketing-related studies (Hair Jr et al., 2017;Sawyer & Peter, 1983) and other business disciplines of science (Kim & Choi, 2021). Results reveal that perceived economic benefits (ECO) (ß=0.359, ...
The study responds to the paucity of understanding of tourists' attitudes toward peer-to-peer (P2P) accommodation services in less-developed economies. By gaining insights into the extant literature, the study proposes and tests a research model that accents the impact of perceived economic benefits, social incentives, and trust on consumers' attitudes towards P2P accommodation. Notably, it also explores the potential attitudinal effect of the new materialist lifestyle. Data were collected through a web-based survey involving 112 consumers aware of P2P accommodation with prior domestic tourism experience in the Philippines. Partial least squares structural equation modelling technique was employed for data analysis. The findings reveal that only economic benefits and trust perceptions can predict consumers' attitudes towards P2P accommodation services. Considering this, accommodation providers should continue to emphasize the economic value of their services to expand the customer base. Additionally, platform and accommodation providers can also enhance trust-building initiatives to reduce consumers' risk perceptions potentially aggravated by the ongoing pandemic.
... Thus, when a firm offers a future discount the customer may feel more satisfied with the result and more likely to return to the business. Consequently, we view these competing predictions as two sets of opposing hypotheses (e.g., Sawyer and Peter 1983;Tax, Brown and Chandrashekaran 1998) which we will test with our data as follows: ...
... Due to the difficulty of the problem, we were aiming at an accessible level of description and presented the bare backbone of the method. We avoided application domain specific formulations in order to make the knowledge transfer easier to different application areas in data science including biomedical science, economics, management, politics, marketing, medicine, psychology or social science [50][51][52][53][54][55]. ...
Statistical hypothesis testing is among the most misunderstood quantitative analysis methods from data science. Despite its seeming simplicity, it has complex interdependencies between its procedural components. In this paper, we discuss the underlying logic behind statistical hypothesis testing, the formal meaning of its components and their connections. Our presentation is applicable to all statistical hypothesis tests as generic backbone and, hence, useful across all application domains in data science and artificial intelligence.
... A comparison between the 21 cases in which interviews were completed and the other nine cases did not reveal any apparent differences that would affect the interpretation of the results. Though 21 cases may be a small sample by some standards, Sawyer and Peter (1983) persuasively argue that there should not be a bias against appropriately selected small samples, especially if the obtained results are significant (as they are here). ...
Marketing strategists assert that environmental management is a critical marketing task. However, little attention has been given to environmental forces that actively oppose an organization's marketing policies. Protest groups often use boycotts in an attempt to coerce an organization to modify allegedly harmful marketing policies. The author proposes and tests a theory of the effectiveness of such boycotts.
... Therefore, these theoretical discrepancies about alternative service relationships are the motivation behind the study, with a view towards developing more than one competing model. Developing more than one competing model is encouraged by many researchers (Burger and Petty, 1981;Sawyer and Peter, 1983) for assisting in testing alternative relationships and for capturing a broader picture. The use of competing models is suggested for supporting findings through the use of alternative hypotheses (McKenzie, 1998). ...
Purpose
– Considering the significance of customer satisfaction and commitment in the competitive banking industry of developing countries, the purpose of this paper is to investigate the alternative relationships of satisfaction with various types of commitment, benefits, quality and trust. It also attempts to identify the interrelationships among three types of trust, namely, competence, contractual and goodwill, as well as affective, calculative and normative commitment.
Design/methodology/approach
– Structural equation modelling (AMOS-based) is used to analyse data collected from a survey of 212 bank customers in Bangladesh.
Findings
– Findings show that satisfaction has a significant impact on affective, calculative and normative commitment, as well as on competence, contractual and goodwill trust. While core and relational service qualities, and confidence and social benefits influence the level of satisfaction, tangible service quality and special treatment benefit seem to have no significant effect on satisfaction. Finally, all three types of trust are found to have a strong influence on affective commitment; however, contractual trust and competence trust are found to have no impact on calculative and normative commitment.
Practical implications
– The paper provides insights for bank management in order to understand the relative importance of each type of trust, commitment, quality and benefit, and their relationships with satisfaction.
Originality/value
– It is the first to take the initiative to test a large range of relationships in a single study to better understand customer satisfaction in the banking industry in a developing country context.
With this article we hope to achieve two goals. The first is to encourage consumer behavioral researchers to consider Bayesian methods for analyzing experimental and survey data. As such, we provide what we hope will be a persuasive set of arguments for trying Bayes. The second goal is to survey the different uses to which the Bayesian posterior distribution can be put. We organize this survey in terms of loss functions and propose that such loss functions can be chosen so as to simply describe a consumer behavioral phenomenon, to highlight a managerial implication, or to emphasize a theoretical contribution.
Looking back at 50 years of Journal of Consumer Research methods and interviewing some of the field’s most respected methodologists, this article seeks to craft a core set of best practices for scholars in consumer research. From perennial issues like conceptual validity to emerging issues like data integrity and replicability, the advice offered by our experts can help scholars improve the way they approach their research questions, provide empirical evidence that instills confidence, use new tools to make research more inclusive or descriptive of the “real world,” and seek to become thought leaders.
Null hypothesis significance testing (NHST) is the default approach to statistical analysis and reporting in marketing and the biomedical and social sciences more broadly. Despite its default role, NHST has long been criticized by both statisticians and applied researchers including those within marketing. Therefore, the authors propose a major transition in statistical analysis and reporting. Specifically, they propose moving beyond binary: abandoning NHST as the default approach to statistical analysis and reporting. To facilitate this, they briefly review some of the principal problems associated with NHST. They next discuss some principles that they believe should underlie statistical analysis and reporting. They then use these principles to motivate some guidelines for statistical analysis and reporting. They next provide some examples that illustrate statistical analysis and reporting that adheres to their principles and guidelines. They conclude with a brief discussion.
ABSTRACT
Purpose: The null hypothesis significance test (NHST) is a commonly applied statistical method for detecting effects in science, despite it being repeatedly criticized. Detractors argue that by focusing exclusively on NHST results, scientists fail to consider descriptive results, potentially leading to misinformed policy makers. They also point out that the influence of sample size on statistical power is often overlooked. This paper investigates whether this critique holds true in sport science research by analyzing the conclusions in publications about the relative age effect (RAE)-an effect manifested in biased birth date patterns. Method: In an extensive content analysis, 7,247 samples listed in 647 sources were recorded and analyzed using binary logistic regression. Results: Findings show discrepancies between NHST results and birth patterns. Authors in RAE research rely more heavily on NHST results than on birth patterns to draw their conclusions regarding the presence of a RAE. In addition, findings indicate that NHST results are influenced by sample size, birth pattern, and the interaction of both. This interaction leads to a RAE more often being suspected in large samples than small samples, even though birth patterns are more evenly distributed in large samples. Conclusion: As large samples are more likely to represent recreational sport and small samples are more likely to represent 1 2/16 elite sport, the strong orientation towards NHST results for conclusions can lead to misinformation about the location of substantial RAEs. Applications in Sport: Similar reliance on NHST results and potential misinformation are also to be expected in other topics in sport research, where characteristics like elite status tend to accumulate in certain sample sizes. Decision-makers in sport should contextualize research findings. Researchers should use NHST appropriately and carefully and combine it with other statistical measures.
Statistical hypothesis testing is among the most misunderstood quantitative analysis methods in data science, despite its seeming simplicity. Having originated from statistics, hypothesis testing has complex interdependencies between its procedural components, which makes it hard to thoroughly comprehend. In this chapter, we discuss the underlying logic behind statistical hypothesis testing and the formal meaning of its components and their connections. Furthermore, we discuss some examples of hypothesis tests.
O objetivo deste ensaio é discutir as potencialidades e, sobretudo, os limites do emprego de métodos experimentais e quase-experimentais na avaliação de programas públicos. Inicia-se com uma breve apresentação do desenho experimental clássico, seus requisitos e conceitos relacionados, como validades interna, externa e contrafatual. Em seguida, são abordadas as modalidades de desenhos quase-experimentais de avaliação, modelos que flexibilizam os requisitos do experimento clássico. Em duas seções de natureza crítica, apontam-se as limitações éticas, políticas e operacionais dos desenhos experimentais e quase-experimentais na avaliação de programas e, depois, as motivações político- -institucionais da resiliência dessa abordagem diante de conhecidos e recorrentes problemas de robustez.
In light of the recent statements on the p-value criterion made by the American Statistical Association (ASA), it is now clear that that the current paradigm of statistical significance and its decision rule should be modified, in many fields of science including the business disciplines. We need a new or modified paradigm for “thoughtful, open and modest” research. This paper explains why we should adopt the ASA recommendations, and proposes a range of possible alternatives that may be included in a new paradigm for statistical research. As an application, the effect of seasonal affective disorder on stock return is re-evaluated.
Für die Marktforschung ist die Verwendung von Stichproben ganz typisch. Beispielsweise soll auf Basis einer Befragung von wenigen hundert Konsumenten auf Präferenzen, Verhaltensweisen etc. aller Konsumenten geschlossen werden. Dazu bedarf es spezieller statistischer Schlussweisen, auf die im Kap. 8 eingegangen wird. Dabei geht es einerseits um Schätzverfahren, mit deren Hilfe man die Genauigkeit und Sicherheit von Ergebnissen berechnen kann. Mithilfe statistischer Tests trifft man Entscheidungen, insbesondere darüber ob Zusammenhänge zwischen Variablen oder Unterschiede zwischen Gruppen durch Zufälligkeiten bei der Stichprobenziehung zu erklären oder systematisch begründet sind.
Anhand von Hypothesen lassen sich Theorien empirisch prüfen. Nachstehend werden das Wesen von Hypothesen und die Vorgehensweise bei der Überprüfung von Hypothesen, den empirischen Tests, dargestellt. Dabei wird insbesondere auch auf die Probleme der Signifikanztests von Null-Hypothesen und der in der Wissenschaft immer wichtiger werdende Beziehung zwischen Signifikanztests und Effektstärken (oder etwas allgemeiner: zwischen statistischer und substantieller Signifikanz) eingegangen sowie auf die Problematik der Post-hoc-Hypothesentests. Mit der Modellierung, die empirisch beispielsweise mit Regressionsanalysen arbeitet, aber auch mit Strukturgleichungsmodellen, können mehrere Hypothesen gleichzeitig getestet werden.
Purpose
This paper aims to examine how conversing about experiences and objects affects consumer happiness. In contrast to previous research focusing on conversation frequency, this paper explores how each conversation instance influences happiness.
Design/methodology/approach
Four experiments use three different methodologies, namely, actual talking behavior (Study 1), recalled and mental framing interventions and measurement of the focal variables (Studies 2 and 3) and manipulation of purchase conversationality (Study 4).
Findings
Consumers derive equivalent levels of happiness from each material or experiential conversation they have. When the object is highly conversational (when it generates as much conversation as experiences do), it advances as much happiness as experiences.
Research limitations/implications
The findings inform precisely how the purchase conversationality model unfolds; clarify previous claims made in the literature; establish the direction of causal effect; and reveal a novel boundary condition of happiness superiority of experiences.
Practical implications
The findings inform marketing managers how to optimally allocate their world-of-mouth (WOM) resources to advance consumer happiness. Additionally, this work shows a mental framing strategy able to increase WOM for objects – i.e. a tool for the manager.
Originality/value
This is the first investigation to disentangle the frequency of conversation from each conversation’s ability to advance happiness. It is also the first to engage participants in an actual conversation and measure changes in their happiness, and therefore, conclusively establish the direction of the effect. Additionally, by manipulating purchase conversationality, this work demonstrates a new boundary condition associated with conversationality.
In buyer-supplier relationships, salespeople engage in behaviors that buyers may (or may not) view as deceptive. Despite the salesperson’s underlying miscreant or innocent motives, buyers have a difficult time attributing the intentionality of salesperson behaviors after the fact. While extant research has explicated various governance mechanisms to mitigate the occurrence of opportunistic behaviors a priori, scholarship is not as well-versed in understanding (a) the relational factors that influence the buyer’s interpretation of debatably opportunistic salesperson behaviors and (b) the buyer’s retributive response within the ongoing relationship. To bring clarity to these issues, the authors examine relationship factors that influence the likelihood of perceived salesperson opportunism following equivocal salesperson acts. Utilizing data from industrial buyers in the U.S. healthcare industry, this study shows that buyer specific investments are related positively to attributions of salesperson guile, whereas contractual agreements are related negatively to attributions of guile. Relationship solidarity moderates these effects. Further, we find that attributions of salesperson guile lead to perceived salesperson opportunism, which in turn results in buyers lowering their expectations of relationship continuity and increasing their retributive responses. We corroborate these findings with cross-sectional survey data from a sample of industrial buyers. Collectively, these findings hold implications with regard to the role of the buyer’s attribution of salesperson guile for a specific behavior as a determinant of perceived opportunism in general, while also outlining conditions under which buyers are inclined to engage in retributive opportunism.
This thesis investigates marketing's role in new product development (NPD) in commercial, investment and merchant banks. It examines how marketing inputs contribute to new product development success. NPD success can be measured at two levels of analysis - at the program and at the project level. Our study is concentrated at the program level at which sustained product development success is examined, rather than one-off project success. Successful product developers are identified as those banks with a better record of being first to market with new products. This measure of product development success is important in the financial risk management market in which commercial, investment and merchant banks compete fiercely. Based on peer evaluation seventeen banks were identified as innovative, that is to say; active new productdevelopers in the financial risk management market from a universe of almost 130 U.K. and foreign bankswith established operations in London. From these seventeen eight participated in our research study. Data was collected in two stages. First, personal interviews were conducted with the heads of the treasury divisions or the heads of derivatives desks to collect background information for control purposes. Second, detailed questionnaires were administered to two further members of each bank who were involved with the development of financial risk management products. The quesionnaires consisted of statements for which respondents were invited to indicate agreement or disagreement on 5-point Likert type scales. Our findings show that it is not the trapping but the quality of marketing inputs that contribute to program success. Quality of marketing inputs comprises the quality of approach adopted and the quality of execution. The most important finding is that successful product developers adopt higher quality marketing than do less successful product developers. Successful product developers place great emphasis on getting both their approach and their execution right.(DX180997)
For banks operating in the fiercely competitive derivatives market, the difference between the leaders and the also-rans often boils down to their respective approaches to product innovation. To achieve market leadership in this highly volatile field, a bank must develop and continually refine the processes and the expertise necessary for identifying new areas of business and attracting and retaining customers. Although the development of derivatives is a complex, expensive process, such development efforts provide the means for satisfying existing clients as well as attracting new customers.F. Axel Johne and M. Panos Pavlidis examine the managerial practices of banks that are acknowledged leaders in bringing new derivatives to market ahead of their competitors. In particular, they examine how those first-mover banks apply their marketing expertise and they review the advantages those banks enjoy as a result of their success in product innovation.The banks studied fall into one of two groups: highly active innovators and less active innovators. The study reveals several significant differences between the two groups with respect to managing marketing inputs for product innovation purposes. First, compared to less active banks, the highly active innovators take a more sophisticated, market-based approach to identifying innovation opportunities. Rather than looking for innovations that offer a close fit with existing products and competencies, they analyze the benefits sought by target clients and initiate innovation efforts based on those analyses.Highly active innovators recognize the important role that internal marketing plays in encouraging functional specialists to work together for the purpose of identifying follow-on development opportunities. Internal marketing also helps to ensure that all parties understand and can support the planned innovation.The highly active innovators in this study do not take a formulaic approach to the development of new derivatives; instead, they rely on marketing expertise to identify and capitalize on business opportunities. Rather than concentrate solely on improving the core technical features of a product, the highly active innovators also recognize the importance of product augmentation innovation (to ensure the appropriate support for various market segments), process innovation (so they can reduce prices, when necessary), and market innovation (to ensure that they pursue the optimal mix of markets)
Based on hypotheses, theories can be examined empirically (see Sect. 5.2). This chapter illustrates the nature of hypotheses and the procedure for testing them empirically. The chapter specifically addresses the increasingly important relationship between significance tests and effect sizes in research, as well as the problem of post hoc hypothesis tests. Modeling can test several hypotheses simultaneously, by working empirically with regression analyses or with structural equation models.
Serious concerns have been raised that false positive findings are widespread in empirical research in business disciplines. This is largely because researchers almost exclusively adopt the ‘p‐value less than 0.05’ criterion for statistical significance; and they are often not fully aware of large‐sample biases which can potentially mislead their research outcomes. This paper proposes that a statistical toolbox (rather than a single hammer) be used in empirical research, which offers researchers a range of statistical instruments, including a range of alternatives to the p‐value criterion such as the Bayesian methods, optimal significance level, sample size selection, equivalence testing and exploratory data analyses. It is found that the positive results obtained under the p‐value criterion cannot stand, when the toolbox is applied to three notable studies in finance.
This study focuses on organizational efforts to constrain ex post transaction costs in interorganizational exchange. The theoretical model frames opportunism as a determinant of transaction costs and implicates cooperation and formalization as control structures that alleviate opportunism. The model also examines whether the proposed theoretical relationships are enduring. Franchisee-franchisor relationships in the Norwegian distribution system of a multinational oil refiner provide the context for analysis. A test of the model using multisample data across two time periods indicates that opportunistic behavior consistently increases transaction costs. Furthermore, cooperative interaction curbs bargaining costs, and formalization reduces opportunism. The authors discuss implications for interorganizational theory and franchising management.
Several distribution channel studies have examined the effects of various sources of a channel member's power on such phenomena as power, conflict, and satisfaction. However, as causal relationships among these power sources have not been considered, the author investigates some aspects of this issue. Specifically, the effects of reward and coercion on the expert, referent, and legitimate power sources in a marketing channel are identified. Results indicate that (1) a supplier's application of reward and punishment does affect the strength of its other three power sources and (2) these relationships exert a major influence on whatever impact reward and coercion may have on other channel phenomena such as supplier power and dealer satisfaction. The implication of this finding of a compound effect of the use of certain power sources is that channel entities should be even more circumspect about such action than would be advised on the basis of prior research in this area.
The authors attempt to assess what has been learned from econometric models about the effect of advertising on sales. Short-term and long-term advertising response as well as model fit are analyzed for 128 econometric models involving the impact of advertising on sales. The approach, a form of meta-analysis called “replication analysis,” treats the studies as imperfect experimental replications and uses ANOVA to identify sources of systematic variation. For short-term advertising elasticities, systematic variability is found related to model specification, estimation, measurement, product type, and setting of study. For advertising carryover and model goodness of fit, the “quasi-experimental design” is so imperfect that a high degree of sharing of explained variance among explanatory factors makes it difficult to identify the impact of a particular factor. Because the studies mostly address mature products in the U.S., suggestions are made for research needs crucial to better understanding of how advertising affects sales.
Several authors have observed that average repeat rates are lower after a promotion purchase. One interpretation of this result is that a media-coupon or cents-off promotion undermines the consumer's repeat purchase rate. The authors describe an alternative explanation. The explanation is that the promotion temporarily attracts a disproportionate number of households with low purchase probabilities. When the repeat rates of these households are averaged with the repeat rates of those that would have bought the brand even without a promotion, the average rate after a promotion purchase is lower. This effect is demonstrated by means of a numerical example, a closed-form equation for repeat rates, a Monté Carlo purchase simulation, and a logit choice model. An exploratory empirical analysis supports these arguments. Implications for marketing management, research methodology, and future research are discussed.
The authors discuss several issues in the timing, construction, and analysis of manipulation and confounding checks in marketing experiments. A review of 34 experiments involving latent independent variables reported in the Journal of Marketing Research over the past decade suggests that most researchers are familiar with the concept of manipulation checks but few systematically evaluate potential sources of confounding in experimental manipulations. Three alternative approaches for assessing the construct validity of experimental manipulations also are discussed.
In two experiments, the authors manipulated the audience's level of involvement in processing an advertisement and whether the advertisement was open-ended (i.e., did not include an explicit conclusion) or closed-ended. Results of both experiments show the open-ended advertisement to be superior to the closed-ended one for an involved audience in terms of brand attitude, purchase intention, and choice, whereas only a small and statistically insignificant difference between the two types of advertisements is found for the uninvolved audience. Moreover, the second experiment shows that the choice advantage of the open-ended ad for the involved audience persisted when measured one week later.
Hunt's 1990 interpretation of scientific realism is compared with the relativistic/constructionist view previously introduced in this journal. Though some similarities are found, the two views differ most strikingly in terms of the nature of reality, the nature of truth, and the value of the concept of incommensurability. Several implications are offered for marketing scientists about what difference it makes which view they accept. However, the author concludes that philosophically oriented marketing theorists need to further demonstrate the value of their work for practicing marketing scientists.
Marketing universals are defined as consumer behaviors within a segment and toward a particular product category that are invariant across cultures. Using several definitions of culture and three different criteria for universality, the authors evaluate whether the use of brand, price, retailer reputation, and physical product appearance as signals of quality are marketing universals for consumer electronics products. Using a sample representing 38 nationalities, they find that there are few differences in the use of quality signals across cultures for a high priority segment of consumers. They draw conclusions for the adaptation versus standardization debate and argue that certain behaviors are likely to be universal, whereas others are not. Understanding such differences is essential to designing international marketing strategies.
The Lakatosian sophisticated methodological falsification framework is introduced, compared with other philosophies of science, and applied to reconstruct marketing science. From this the location and intensity of extant theoretical conflicts are pinpointed and explained, marketing's scientific status evaluated, and implications for research methodology in the discipline discussed.
From a survey of the papers published in leading accounting journals in 2014, we find that accounting researchers conduct significance testing almost exclusively at a conventional level of significance, without considering key factors such as the sample size or power of a test. We present evidence that a vast majority of the accounting studies favour large or massive sample sizes and conduct significance tests with the power extremely close to or equal to one. As a result, statistical inference is severely biased towards Type I error, frequently rejecting the true null hypotheses. Under the ‘p‐value less than 0.05’ criterion for statistical significance, more than 90% of the surveyed papers report statistical significance. However, under alternative criteria, only 40% of the results are statistically significant. We propose that substantial changes be made to the current practice of significance testing for more credible empirical research in accounting.
In science publishing and many areas of research, the status quo is a lexicographic decision rule in which any result is first required to have a p-value that surpasses the 0.05 threshold and only then is consideration--often scant--given to such factors as prior and related evidence, plausibility of mechanism, study design and data quality, real world costs and benefits, novelty of finding, and other factors that vary by research domain. There have been recent proposals to change the p-value threshold, but instead we recommend abandoning the null hypothesis significance testing paradigm entirely, leaving p-values as just one of many pieces of information with no privileged role in scientific publication and decision making. We argue that this radical approach is both practical and sensible.
This meta-analysis investigates the effects of research design on the reliability of rating scale measures used in marketing research. Three sets of independent variables—sampling characteristics, measure characteristics, and measure development processes—are investigated. The overall results suggest that measure characteristics have a major influence on obtained reliability estimates and that sampling characteristics and measure development processes have little impact. These results support measurement theory. The implications for measure development are discussed.
Should a uniform checklist be adopted for methodological and statistical reporting? The current article discusses this notion, with particular attention to the use of old versus new statistics, and a consideration of the arguments brought up by Von Roten. The article argues that an overly exhaustive checklist that is uniformly applied to all submitted papers may be unsuitable for multidisciplinary work, and would further result in undue clutter and potentially distract reviewers from pertinent considerations in their evaluation of research articles.
This study summarizes econometric findings from time-series research on the relationships between cigarette consumption and advertising, price, and income using a meta-analysis methodology. The study also investigates the trends in these relationships over time, controlling for possible confounding effects of design variables in the studies analyzed. The results show that the advertising and income elasticities are positive and significant on average across studies, and the average price elasticity is significantly negative. The elasticities vary across countries and design factors and have become smaller over time. Implications for public policy and future research are discussed.
We describe how statistical predictive models might play an expanded role in educational analytics by giving students automated, real-time information about what their current performance means for eventual success in eLearning environments. We discuss how an online messaging system might tailor information to individual students using predictive analytics. The proposed system would be data-driven and quantitative; e.g., a message might furnish the probability that a student will successfully complete the certificate requirements of a massive open online course. Repeated messages would prod underperforming students and alert instructors to those in need of intervention. Administrators responsible for accreditation or outcomes assessment would have ready documentation of learning outcomes and actions taken to address unsatisfactory student performance. The article's brief introduction to statistical predictive models sets the stage for a description of the messaging system. Resources and methods needed to develop and implement the system are discussed.
Sponsorships capture a significant proportion of marketing budgets. In firm evaluations of the effectiveness of sponsorship engagements, image improvements represent the most important company objective. This study develops and tests a framework for explaining how exposure and activity involvement moderate the effects of event image, event-sponsor fit, and event commercialization on sponsor image. Using empirical data collected at a large sporting event with multiple sponsors, the authors show that increased sponsorship exposure reduces sponsor image if respondents perceive a low fit between the event and sponsor or high levels of event commercialization. Involvement in the sponsored activity improves the sponsor's image, in that the effect of event commercialization is positive for highly involved persons. This paper concludes with some reasons for the findings, implications for the choice and design of sponsorships, and further areas for research.
Purpose
– This article aims to draw on the contingency theory to develop a conceptual model of compatibility between corporate environmental responsibility and business strategy that reflects heterogeneity in this relationship. Four dimensions of compatibility are explored: trade-off, ambidexterity, synergy and symbiosis.
Design/methodology/approach
– The intended contribution is essentially conceptual. A company case study is included to contribute to the development of the four dimensions of compatibility and support the practical relevance of the model. Twelve in-depth interviews with six managers in different functions of the company were conducted. A grounded theory approach was used to identify and express the patterns of compatibility that emerge from the qualitative data and how these patterns are grounded in managers’ meaning-in-use.
Findings
– The contribution of the compatibility framework is essentially made to the literature on environmental strategy management, evolved from an implicit and, at most, two-dimensional (win–win and win–lose) conceptualisation of the relationship between green and business strategy into an explicit and multi-dimensionally grounded identification of processes and strategic challenges of corporate environmental and social responsibility. The resulting model contributes to a better understanding of corporate greening as a strategic and moral concern to individuals acting on behalf of business organisations and a greater understanding of the linkages between green and business strategies and operations.
Originality/value
– By clarifying the construct of corporate environmental sustainability and providing useful directions for theory and practice, this research claims to inform green management decision-making. While the compatibility model is not intended to explain all pathways by which firms may elicit contingencies of relevance to environmental and social responsibility, it is suggested that the model paints a more complete and contextualized picture of environmental management mechanisms in business.
Une analyse empirique a été effectuée sur l'importance des effets (c'est-à-dire la force d'une relation ou l'amplitude d'une différence entre variables) dans les expériences sur le comportement du consommateur publiées entre 1970 et 1982. Au total, 118 expériences et 1 036 résultats expérimentaux ont été analysés. En utilisant ω² comme mesure de l'importance des effets, les résultats de l'enquête révèlent que, en moyenne, environ 11% de la variance de la variable dépendante sont expliqués par un effet statistiquement significatif (p ≤ 0.05) Cependant, ce pourcentage varie selon les caractéristiques méthodologiques de l'expérience.
This article examines the nature of positioning strategies pursued by companies (domestic and foreign) conducting business in a liberalized developing African economy, Ghana, over a ten-year period. Specifically, this research deals with the aggregate behaviors of firms operating in Ghana's post-SAP (structural adjustment program)-induced environment. The findings reveal that the dominant positioning strategies are the brand name (branding tactics) and value for money (affordability). However, more recently, attractiveness, which implies elegance about the offering, has also emerged as a prominent positioning strategy. Evidence that supports the robustness of companies' adoption of positioning strategies within sub-Saharan African economies is also developed. The results also suggest that firms doing business in sub-Saharan African marketplaces characterized by an open and liberalized business climate have no choice but to become competitive in their positioning and with the aim of targeting the mass market and the lower-middle-class target audiences. © 2011 Wiley Periodicals, Inc.
The statistical testing approaches of Hypothesis Testing (Fisher) and Significance Testing (Neyman-Pearson) are briefly reviewed. Highly related notions of alpha-values, P-values, and magnitude-of-effect are discussed. A statistical testing approach that is a hybrid of Hypothesis Testing and Significance Testing is then advanced. This approach takes into consideration the two decision-error probabilities as well as an alternative hypothesis characterization of interest.
Ambiguity surrounds any importance measure in cases in which predictor variables are correlated. However, a new measure is proposed that has attractive properties, such as providing individual contributions that are both non-negative and sum to R2. The new measure is compared with four other commonly used measures and its advantages over each of them are pointed out.
Statistical power and effect size are not considered sufficiently by marketing researchers. The authors discuss how better attention to these two factors can improve the planning, execution, and reporting of marketing and consumer research. Suggestions are offered about how to increase effect size and improve statistical power.
The effects of price and advertising expenditures on new-product sales are examined across four test-market experiments. All experiments involved two levels of advertising expenditure tested across cities and two or three price levels tested across stores within city. Mean sales results over a 24–week period are reported for each treatment combination. Analysis of variance results are interpreted as supporting a negative price main effect, a positive advertising main effect, a negative price-advertising interaction, and significant but variable time effects. Assessments of these results in terms of design adequacy and their implications for alternative theoretical explanations are offered.
The authors analyze the impact of data collection period on the coefficient of the lagged dependent variable in 69 econometric studies of the sales-advertising relationship. The null hypothesis that the coefficient is independent of data collection period is not rejected. Reasons for this result are offered and discussed.
Weinberg and Weiss observe that the coefficients of the lagged dependent variable in Clarke's survey are similar regardless of data interval. In this note Clarke demonstrates that the similarity of these coefficients substantiates rather than contradicts his conclusion that there is data interval bias in his survey results. A different test for data interval bias than the one in the survey is presented which further substantiates the presence of data interval bias in the survey results.
The author explains how decision theory can be applied to the design and evaluation of advertising experiments and discusses its application to marketing problems.
Research on multi-attribute attitude models has relied on correlational methods. An analysis of variance paradigm applied to individual level data supports the multiplicative assumption of multi-attribute models. A single-attribute form of the Fishbein model is found to be superior to a similar version of the “adequacy-importance” model.
A survey of the econometric literature is undertaken to determine the duration of cumulative advertising effect on sales. The surveyed studies yield conflicting estimates of the duration interval. The data interval is shown to have a powerful influence on the implied duration of advertising effect. The evidence leads to the conclusion that the cumulative effect of advertising on sales lasts for only months rather than years.
The author presents a model designed to explain the performance, job satisfaction, and other behavioral outcomes experienced by salespeople. By building on the model of man proposed by Lewin and modern interactionist and social learning theories of personality, the behavior of salespeople is shown to be a function of the person, the interactions the person has with significant others in his or her role set, and forces in the situation. Hypotheses are tested on data gathered on industrial salespeople.
A comprehensive literature review of techniques used to increase response rates to questionnaires was conducted. Conclusions were based on arithmetic combination of 497 response rates found in 93 journal articles. Response rates were found to be increased by personal and telephone (versus mail) surveys, either prepaid or promised incentives, nonmonetary premiums and rewards, and increasing amounts of monetary reward. Other facilitators that increased responding were preliminary notification, foot-in-the-door techniques, personalization, and followup letters. Several facilitators were found to be ineffective. An estimated magnitude of effect was calculated for each design variation.
The accuracy of marketing research results is affected by sampling and nonsampling errors. While the former can be computed objectively using sampling theory, the latter depend on subjective estimates. Such estimates can be facilitated by examining relevant empirical evidence. This article presents a model for integrating the subjective estimates about nonsampling error into estimates of total survey error and demonstrates its use through an example.
This paper deals with what is known as opposed to what is thought about marketing. The distinction is not subtle. The knowledge of a field is defined in terms of scientific generalizations, whereas theory or thought is necessarily speculative. We argue that marketing generalizations logically precede marketing theories and that marketing generalizations are marketing knowledge.
Analysis of panel data for two consumer packaged goods indicates that media-distributed coupons and cents-off deals induce brand switching and result in less loyalty when retracted than if no deal is offered. In contrast, package coupons stimulate brand loyalty which is maintained when they are retracted. The extent to which economic utility theory and self-perception theory order these findings is evaluated and the implications of the results for managing demand are discussed.
What factors account for the wide variation in performance with in a salesforce, and what is the relative importance of these factors? The authors attempt to answer that question by first developing a framework that identifies the major constructs to be included in a model of territory response, and then testing the model in three different corporate settings. The empirical results provide a basis for useful generalizations about the importance of territory characteristics, span of control, and other factors in explaining the variability in salesperson performance.
The research was supported in part by grants from the Marketing Science Institute and the American Cyanamid Company. The authors thank the executives of the companies that participated in this study for their help in providing data, and also thank Dick R. Wittink and a JMR reviewer for their helpful comments on a draft of this article. The research assistance of Tsuneo Yahagi is gratefully acknowledged.
Statistical power and effect size are not considered sufficiently by marketing researchers. The authors discuss how better attention to these two factors can improve the planning, execution, and reporting of marketing and consumer research. Suggestions are offered about how to increase effect size and improve statistical power.
The authors analyze the issue of comparative price advertising from a behavioral perspective. Because public policy recognizes that comparative pricing may lead to consumer misperceptions, the authors review the regulatory setting and pose several research questions that need to be addressed. A complex experiment and replication examining some of these questions is reported.
Ambiguity surrounds any importance measure in cases in which predictor variables are correlated. However, a new measure is proposed that has attractive properties, such as providing individual contributions that are both non-negative and sum to R². The new measure is compared with four other commonly used measures and its advantages over each of them are pointed out.
The comparison of classical statistics and decision theory approaches to evaluation of advertising experiments is extended beyond Blattberg's focus on incorporation of prior judgment. Explicit consideration of risk provides a clearer comparison of these two analytic approaches.
The evaluation of total error in survey findings is largely neglected. Confidence intervals and other common tools deal only with random and occasionally other limited errors. This article presents a new method by which market researchers and clients can interpret their information and judgment about all survey errors.
Bayesian statistics points the way to an articulation of business judgment and statistical research in approaching marketing problems.
We have always known that neither executive nor statistician is sufficient in himself, but now for the first time we have formal tools that show that the two can be brought closer together. This article shows how this may be accomplished.
Also, see the succeeding article, “Bayesian Decision Theory in Pricing Strategy,” by Paul E. Green, pp. 5–14.
To increase the validity and reliability of survey data, one must minimize total error and its components, sampling and nonsampling error. This article examines and empirically compares the components of total survey error for several research designs and data collection methods. The consistent finding is that nonsampling error is the major contributor to total survey error, while random sampling error is minimal. On this basis guidelines for improving quality of survey data are provided.
The effects of price and advertising expenditures on new-product sales are examined across four test-market experiments. All experiments involved two levels of advertising expenditure tested across cities and two or three price levels tested across stores within city. Mean sales results over a 24-week period are reported for each treatment combination. Analysis of variance results are interpreted as supporting a negative price main effect, a positive advertising main effect, a negative price-advertising interaction, and significant but variable time effects. Assessments of these results in terms of design adequacy and their implications for alternative theoretical explanations are offered.
Weinberg and Weiss observe that the coefficients of the lagged dependent variable in Clarke's survey are similar regardless of data interval. In this note Clarke demonstrates that the similarity of these coefficients substantiates rather than contradicts his conclusion that there is data interval bias in his survey results. A different test for data interval bias than the one in the survey is presented which further substantiates the presence of data interval bias in the survey results.
The author presents a model designed to explain the performance, job satisfaction, and other behavioral outcomes experienced by salespeople. By building on the model of man proposed by Lewin and modern interactionist and social learning theories of personality, the behavior of salespeople is shown to be a function of the person, the interactions the person has with significant others in his or her role set, and forces in the situation. Hypotheses are tested on data gathered on industrial salespeople.
The accuracy of marketing research results is affected by sampling and nonsampling errors. While the former can be computed objectively using sampling theory, the latter depend on subjective estimates. Such estimates can be facilitated by examining relevant empirical evidence. This article presents a model for integrating the subjective estimates about nonsampling error into estimates of total survey error and demonstrates its use through an example.
The authors analyze the impact of data collection period on the coefficient of the lagged dependent variable in 69 econometric studies of the sales-advertising relationship. The null hypothesis that the coefficient is independent of data collection period is not rejected. Reasons for this result are offered and discussed.
A survey of the econometric literature is undertaken to determine the duration of cumulative advertising effect on sales. The surveyed studies yield conflicting estimates of the duration interval. The data interval is shown to have a powerful influence on the implied duration of advertising effect. The evidence leads to the conclusion that the cumulative effect of advertising on sales lasts for only months rather than years.
To increase the validity and reliability of survey data, one must minimize total error and its components, sampling and nonsampling error. This article examines and empirically compares the components of total survey error for several research designs and data collection methods. The consistent finding is that nonsampling error is the major contributor to total survey error, while random sampling error is minimal. On this basis guidelines for improving quality of survey data are provided.
The author explains how decision theory can be applied to the design and evaluation of advertising experiments and discusses its application to marketing problems.
Several statistical problems in the design of research are discussed: (1) The use of statistical tests and the search for causation in survey research are examined; for this we suggest separating four classes of variables: explanatory, controlled, confounded, and randomized. (2) The relative advantages of experiments, surveys, and other investigations are shown to derive respectively from better control, representation, and measurement. (3) Finally, three common misuses of statistical tests are examined: "hunting with a shot-gun for significant differences," confusing statistical significance with substantive importance, and overemphasis on the primitive level of merely finding differences.
Research on multi-attribute attitude models has relied on correlational methods. An analysis of variance paradigm applied to individual level data supports the multiplicative assumption of multi-attribute models. A single-attribute form of the Fishbein model is found to be superior to a similar version of the "adequacy-importance" model.
Belief in the possibility of demonstrating ESP on an experimental task and evaluation of ESP were independently manipulated in two laboratory studies investigating the relationship between these variables and ESP task performance. The results of Experiment I indicated that subjects exposed to a positive evaluation of ESP scored higher on the ESP task than those exposed to a negative evaluation of ESP. Furthermore, a significant belief × sex of subject interaction was found. The interaction indicated that males exposed to the disbelief manipulation scored significantly higher on the task than males exposed to the belief manipulation or females exposed to the disbelief manipulation. In general, the ESP task performance results for females followed the pattern predicted from the Ajzen and Fishbein (1972) attitude-behavior model, while those for males did not. Experiment II was a strict methodological replication of Experiment I. In Experiment II there were no significant effects present upon ESP task performance. The implications of these contradictory results for ESP research specifically and significance levels generally are both discussed.
This paper deals with what is known as opposed to what is thought about marketing. The distinction is not subtle. The knowledge of a field is defined in terms of scientific generalizations, whereas theory or thought is necessarily speculative. We argue that marketing generalizations logically precede marketing theories and that marketing generalizations are marketing knowledge.