Changes in Definitions and Operationalizations in Advertising Research—Justified or Not?

Abstract

Many advertising constructs vary considerably in how they are defined and operationalized—variations which result from changes in definitions and operationalizations over time and between studies. These changes undermine the comparability of empirical insights, challenge the integration of insights into an existing body of knowledge, and thereby impede knowledge development in advertising research. This negative effect is why it is crucial to know whether and when changes are needed and justified. In this article, we address definition and operationalization change from a theoretical perspective, identify factors driving changes in definitions and operationalizations, and outline when changes are justified. We then offer recommendations for how to avoid unjustified changes.

Recent research has demonstrated that many important advertising constructs, such as ad attitude or brand attitude, vary considerably in whether and how they are defined and operationalized in the literature (e.g., Bergkvist and Langner 2017, 2019; Bruner 1998). This variation results from an evolutionary process where constructs evolve through different stages (Bergkvist and Eisend 2021). The dynamic model of construct evolution suggests that consensus is the ideal end state for constructs, which requires a clear construct definition and agreement on its operationalization and most valid measures. Very few constructs seem to have reached a state of consensus or dominance (but see, e.g., need for cognition), and research practice reveals that most constructs are not (yet) in an ideal state but are characterized by changing definitions and variable and customized operationalizations (Bergkvist and Eisend 2021). While the dynamic model acknowledges the evolving nature of definitions and operationalizations and changes as part of long-term consensus building, not all changes are required and necessary to approach a construct’s ideal state; some changes seem not to contribute to construct improvement and thus are not justified.

Although there is a growing body of research on various aspects of advertising and marketing constructs (e.g., Bergkvist and Langner 2017, 2019; Bergkvist and Eisend 2021; Katsikeas et al. 2016; Ptok, Jindal, and Reinartz 2018), there appears to be no research on when changes in construct definitions and variations in operationalizations are justified. This dearth of research is surprising given the negative consequences that unjustified changes have on marketing and advertising knowledge. Advertising meta-analyses show, for instance, that changing the definition and operationalization of advertising creativity from a multidimensional concept (e.g., as original and relevant) to a single dimension of originality reduces the overall correlation-based effect of creativity on brand attitude by more than 50% (from 0.33 to 0.16 ; see Rosengren et al. 2020). Similar results occur for ethnic identity effects in advertising after operationalization changes are introduced: The effects are twice as strong when ethnic identity is measured by a single-item compared to a multi-item scale (0.33 versus 0.15; Sierra, Hyman, and Heiser 2012).

In this article, we address definition and operationalization change from a theoretical perspective. Drawing on recent literature on definitions and research practice, we identify factors driving changes in definitions and operationalizations and outline when changes are justified, and we show examples of justified definition and operationalization changes from current advertising research. We offer recommendations for dealing with changes in definitions and operationalizations in advertising research.

The Importance and Role of Definitions and Operationalizations

Definitions serve several essential functions in research: They distinguish different constructs from one another (e.g., attitude toward the ad versus attitude toward advertising in general); they provide the basis for categorizing objects (e.g., liked versus disliked ads); they identify antecedents, outcomes, and correlates in a nomological network (e.g., the dual-mediation hypothesis as an attitude toward the ad model); and they guide the development of research manipulations and measurement operationalizations (e.g., the items to assess attitudes or the manipulation of involvement) (Podsakoff, MacKenzie, and Podsakoff 2016). A lack of consensus for definitions reduces the likelihood of consensus about measurement operationalizations, which is critical for knowledge accumulation, as divergent operationalizations may render research results incommensurable (Landy et al. 2020). Operationalizing the same construct with different measures in different studies makes it impossible to know whether factors in the study context or differences in the operationalizations caused differences in research results (Katsikeas et al. 2016), even if the different measures have high levels of convergent validity (Carlson and Herdman 2012; Rossiter 2016).

Two recent studies relying on crowdsourcing of researchers have demonstrated the importance of definitions and operationalization for research results. Schweinsberg et al. (2021) enlisted 19 researchers to define and operationalize the constructs in two hypotheses concerning online scientific discussions, without restrictions on how the constructs could be defined or operationalized, and to test the hypotheses on the same data set. Results across 29 analyses showed considerable dispersion with the hypotheses rejected, supported, or inconclusive results, and the authors concluded that “subjective researcher choices make a critical contribution to the results obtained” (Schweinsberg et al. 2021, p. 243). In a similar study, Landy et al. (2020) crowdsourced 15 research teams who independently created stimuli for five research questions relating to moral judgment, negotiation, and implicit cognition. Experiments by the research teams and subsequent replication studies obtained directionally inconsistent results for three of the five hypotheses, and analyses found considerable heterogeneity in the effect sizes. The authors conclude that “idiosyncratic choices in stimulus design have a very large effect on observed results, over and above the overall support (or lack thereof) for the hypothesis in question” (Landy et al. 2020, p. 469). The conclusions from these studies align with the meta-analytic findings in the introduction that show, for instance, altering the definition and operationalization of advertising creativity changes the effects for the same hypothesis (i.e., advertising creativity increases brand attitudes) by more than 50% (for further discussion of the importance of definitions, see Harford 2021).

Thus, there are compelling theoretical and empirical grounds to state that consensus regarding definitions and operationalizations is fundamental for comparing, integrating, and interpreting research findings and consequently for knowledge accumulation in marketing and advertising research. However, studies of research practice show that changes in definitions and operationalizations occur frequently, and it is often unclear whether these changes are justified (Bergkvist and Langner 2017; Bergkvist and Eisend 2021).

Change in Construct Definitions

Defining a construct involves specifying its necessary and/or sufficient characteristics (Podsakoff, MacKenzie, and Podsakoff 2016). Specifically, for a measurement construct, this involves specifying the attribute (the “judgment” or rating), the object of measurement, and the rater (Rossiter 2011). It also involves specifying the domain of the construct (Nunnally and Bernstein 1994). Drawing on Nunnally and Bernstein (1994) and Rossiter (2011), we define the domain as (1) the application area (e.g., social media advertising) and (2) the theoretical context (e.g., attitude toward the influencer influences brand attitude). Together, the elements of the definition (attribute, object, rater) and the two domain elements include those entities that directly bear on the construct’s operationalization, making it possible to outline factors that motivate changes in construct definitions.

There are two categories of reasons for changing construct definitions. First, if there are limitations in existing definitions, they should be changed. Limitations in definitions have multiple negative consequences, such as invalid measurement operationalizations, construct proliferation, and lacking discriminant validity (Podsakoff, MacKenzie, and Podsakoff 2016; Teas and Palan 1997), and it is imperative that advertising scholars, as a collective, work toward construct improvement and consensus over time. Second, changes in the domain of the construct (application area, theoretical context) have consequences for the definition, and definitions should be changed accordingly (Nunnally and Bernstein 1994).

Limitations in the Existing Definition that Motivate Change

Drawing on the literature on definitions in marketing and related fields, we identified seven ways in which definitions can have limitations that motivate change, and we found examples of each limitation in top advertising journals (International Journal of Advertising, Journal of Advertising, and Journal of Advertising Research; refer to Table 1). These limitations fall into three categories. First, some definitions lack specification of the construct’s characteristics. Definitions in this category do not completely specify the construct’s necessary characteristics, include ambiguous or vague terms, or are tautological (Podsakoff, MacKenzie, and Podsakoff 2016; Teas and Palan 1997; Wacker 2004). These definitions do not clarify the meaning of the construct, which can lead to poor measurement operationalizations. When improving definitions in this category, scholars should ensure that definitions include all of the construct’s necessary characteristics and use precise terms that specify the construct’s meaning. Second, some definitions have unclear boundaries toward other constructs (Podsakoff, MacKenzie, and Podsakoff 2016; Teas and Palan 1997; Wacker 2004). These definitions may not include all construct instances or overlap with other constructs, in the latter case leading to construct proliferation (Podsakoff, MacKenzie, and Podsakoff 2016) or construct confusion (Bergkvist and Langner 2019). Scholars should consolidate overlapping constructs and clarify the boundaries of constructs. Third, some definitions are overly specific (Summers 2001; Wacker 2004). These definitions are not parsimonious or include relationships that are true by definition. Scholars should remove redundant terms and avoid relationships between the construct and its antecedents or consequences in the definition.

Table 1. Limitations in definitions with commented examples, descriptions of the problems, and recommendations.

Definition Limitation	Example^a	Comment	Problem	Recommendation
The definition is lacking in the specification of the construct’s characteristics
1. The definition fails to specify all the construct’s necessary characteristics (Podsakoff, MacKenzie, and Podsakoff 2016).	“In the hospitality, tourism, and travel context, risk is defined as what is perceived and experienced by tourists during their process of considering, selecting, and consuming travel services”	This definition refers to risk as anything that is “perceived and experienced” and thus fails to specify the characteristics of risk.	Incomplete definitions can lead to incomplete measurement operationalizations and unclear boundaries toward other constructs (Teas and Palan 1997).	Make sure that the definition specifies all the necessary attributes of the definition.
2. The definition includes ambiguous and vague terms (Teas and Palan 1997; Wacker 2004).	“[D]efine the concept [consumer engagement] as ‘a state of being involved, occupied, fully absorbed, or engrossed in something’”	This definition is vague as it includes four different states or activities, separated by “or”; these states/activities are vague or ambiguous.	Ambiguous and vague definition terms entail poor measurement operationalizations (Wacker 2004).	Scholars can address this limitation by using more precise terms or specifying the meaning of terms.
3. The definition is tautological (i.e., it defines the construct in terms of synonyms; Podsakoff, MacKenzie, and Podsakoff 2016).	“Congruency is defined as entities ‘going well together’”	The phrase “going well together” is synonymous with congruency.	A tautological definition does not clarify the meaning of the construct as it refers to the construct itself (Podsakoff, MacKenzie, and Podsakoff 2016).	Scholars should define constructs using primitive terms that clarify the constructs’ meanings (Podsakoff, MacKenzie, and Podsakoff 2016; Teas and Palan 1997; Wacker 2004).
The definition has unclear boundaries toward other constructs.
4. The construct is not uniquely defined and overlaps with other constructs (Podsakoff, MacKenzie, and Podsakoff 2016; Teas and Palan 1997; Wacker 2004).	“The celebrity/product fit . . . refers to the harmony of the match between the celebrity endorser and the product being endorsed” “[S]ource congruence, that is, the match between the celebrity’s image and the brand’s image”	Definitions of fit and congruence tend to overlap and essentially refer to the same construct. Thus, neither construct is uniquely defined.	Overlapping definitions lead to construct proliferation (Podsakoff, MacKenzie, and Podsakoff 2016) and construct confusion (Bergkvist and Langner 2019).	Consolidate overlapping constructs under one name and treat them as the same construct (Bergkvist and Langner 2019).
5. The definition offers examples of the construct but does not define it (Summers 2001).	“Brand authenticity . . . can be characterized by terms such as genuine, natural, honest, and real; and the extent of authenticity in any situation is subjectively assessed by the perceiver”	This definition relies on examples of what characterizes the construct.	Definitions based on examples do not include all instances of the construct and fail to clarify what is not included in the construct (Summers 2001).	Avoid examples and clarify the meaning and boundaries of the construct.
The definition is overly specific.
6. The definition is not parsimonious and includes too many terms (Wacker 2004).	“Internet literacy, defined as the ability to assess, analyze, evaluate, and create online content”	The terms assess and evaluate are synonyms, and analyze subsumes both of them. Thus, there is redundancy in the definition, and it could have been expressed more succinctly.	Parsimony is a virtue of good theory, and shorter definitions are generally preferred to longer (Wacker 2004).	Remove redundant terms from the definition.
7. The definition includes hypotheses or specifies relationships between the construct and its antecedents or consequences (Summers 2001; Wacker 2004).	“[T]he concept of ad–brand congruity that can be defined as the extent to which advertising information conforms to consumers’ expectations about a brand, based on consumers’ previous beliefs, schemata, and knowledge of the brand”	This definition assumes a causal relationship between previous beliefs, schemata, and brand knowledge, and ad–brand congruity.	Including hypotheses or causal relationships in definitions means that these relationships are true by definition and thus are not testable (Summers 2001).	Outline the relationships between the construct and its antecedents and consequences in the conceptualization but do not include them in the definition.

^a The examples are from recent issues of leading advertising journals (International Journal of Advertising, Journal of Advertising, and Journal of Advertising Research). They are cited without sources as they refer to common problems, and multiple articles have used identical or similar definitions. Consequently, there is no need to single out authors or articles associated with a particular limitation.

Thus, existing definitions can have several limitations that motivate change. Scholars should scrutinize definitions early in their research and carefully consider whether definitions need change. If change is needed, scholars should report how they changed the definition and justify it by outlining the limitations in the existing definition, as suggested in Table 1.

Changes in the Construct’s Domain That Prompt a Definition Change

Changes in the external world or the theoretical context frequently impact a construct’s domain and can prompt a change of the definition, as the object of measurement, rater, or attribute parts of the definition may become inadequate due to the change of domain. We identified three types of external changes that may prompt changes in construct definitions (Table 2). First, societal changes in gender roles, acceptance of sexual identities, and other socially constructed phenomena influence the characteristics defining related constructs (e.g., whom or what to include in the definition). Changes in the defining characteristics, in turn, have implications for the rater and object of measurement parts of the operational definition and the operationalization of the construct (Nunnally and Bernstein 1994; Rossiter 2011). Accordingly, definitions should be changed to reflect changes in society. Second, different types of technological development (e.g., artificial intelligence, augmented reality) could extend or reduce a construct’s application area, which, in turn, could influence the object of measurement or the attribute in the operationalization. Consequently, definitions should be updated to reflect relevant technological development. Third, advances in advertising practice, such as personalization, influencer marketing, and social messaging apps, could extend the application area of many advertising constructs (Thorson and Rodgers 2006; Tsai and Men 2013). Definitions of advertising constructs should be updated to reflect relevant advances in marketing practice.

Table 2. Changes in the construct’s domain, their effects on definitions, examples, and recommendations.

Domain Change	Consequences for the Definition	Examples	Recommendation
1. Societal changes such as female workforce participation, the growth of rainbow families, and acceptance of diverse sexual identities.	Societal changes could affect the characteristics of a construct (e.g., whom or what to include).	Societal changes impact family forms (Gerson and Torres 2015) and household characteristics (Casimir and Tobi 2011). These changes affect definitions of the family and household and related constructs (e.g., whom to include in a family or household). The definition of diversity in advertising (Lee, Williams, and Ferle 2004) addresses societal changes by broadening the definition to new diversity dimensions (e.g., sexual identity).	Analyze the effects of societal changes on the construct’s characteristics and update the definition to capture the domain change by removing, adding, or changing domain-related characteristics of the definition.
2. Technological development such as online media, artificial intelligence, and augmented reality.	Technological changes could extend a construct’s domain by adding new application areas or, alternatively, reducing the domain by rendering previous technologies and behaviors obsolete.	Definitions of online privacy concerns could address technological developments by including privacy threats due to artificial intelligence (for an overview of how technological development has shaped the definition of privacy, see Rapp et al. 2009). New online media formats could add the attribute degree of social interaction to the ad perception construct (MacKenzie and Lutz 1989). Young consumers replace traditional TV with streaming services and online media (Rubenking and Bracken 2021; Scholz 2021), adding new application areas and potentially removing a previous application area.	Analyze the effects of technological development on the construct’s application areas and update the definition by adding new application areas or removing obsolete areas.
3. Advances in marketing practice such as personalization, influencer marketing, and social messaging apps.	Advances in marketing practice could affect the construct’s characteristics and application areas.	Influencer marketing widens the scope of contexts available for brand placements, celebrity endorsements, and other forms of marketing communications (Rundin and Colliander 2021). Social media apps offer an additional touch point for customer interactions. Personalized ads raise privacy concerns concerning data collection and targeting.	Analyze the effects of advances in marketing practice on the construct’s characteristics and application areas and update the definition by changing the terminology to consider advances in practice.

Thus, societal changes, technological development, and changes in advertising practice frequently motivate changes in definitions. Scholars should be observant of relevant changes in the external world and update definitions when needed.

Unjustified Definition Changes

Some construct definition changes are not justifiable, and these are changes undertaken for reasons other than limitations in existing definitions or changes in the domain. There are two primary types of unjustified definition changes. First, definitions tend to drift over time, taking on additional or different meanings than the original definition (Suddaby 2010). This drift is unintended and could be the result of drifting paraphrases (i.e., a definition is restated rather than cited and critical terms are subtly changed; e.g., “brand” is changed to “product”), scholars broadening the meaning of a definition when attempting to clarify its terms (Wacker 2004), or chance factors such as typos or misinterpretations of existing definitions. Second, scholars sometimes change definitions deliberately without justifying doing so. For example, when scholars make ad hoc changes to definitions according to specific research needs (e.g., changing “customer” to “consumer” in the definition of “customer-based brand equity” if the data set includes consumers rather than customers).

Unjustified definition changes adversely affect knowledge accumulation within a field, and scholars should avoid these (Bergkvist and Eisend 2021). If definition changes are justified, these should be explicitly motivated; if they are not, they should be avoided.

Change in Construct Operationalizations

Justified Operationalization Changes

Marketing scholars have long argued that marketing measures should be developed and validated in scale development studies and then used unchanged in subsequent studies (e.g., Churchill 1979). However, scholars cannot expect operationalizations to remain fixed over time, and several reasons may make changes justifiable, or in some cases, justify replacing existing operationalizations. These changes may be motivated by construct definition changes or limitations in the properties of the existing operationalization. We identified four situations that motivate operationalization change (Table 3). First, if the construct definition is justifiably changed (as outlined previously), the operationalization should change accordingly, as operationalizations should follow from definitions (Podsakoff, MacKenzie, and Podsakoff 2016). Leaving the operationalization of a construct unchanged when the definition is changed entails a considerable risk of a mismatch between the two (i.e., the operationalization is not valid). Accordingly, operationalizations should be updated to reflect changes in definitions. Second, if repeated usage of an accepted operationalization suggests that it has limitations, the operationalization should be updated or replaced (Summers 2001). Scholars should strive to use optimal operationalizations of their constructs (Rossiter 2016) and thus improve or replace them when results in previous studies identify weaknesses. Third, if a comparative study shows that a new operationalization is more valid than the previously accepted operationalization, then scholars should use the new operationalization. Developing improved operationalizations is an important methodological contribution (Summers 2001), and scholars should use the most valid operationalization. Fourth, if a more parsimonious operationalization of equal or better validity could replace an existing operationalization, scholars should use the former. Operationalizations with fewer items save respondents’ time and increase answer quality (Bergkvist 2015) and should be preferred over operationalizations with more items.

Table 3. Situations that justify changes in operationalizations with examples and recommendations.

Situation	Justification of Change	Examples	Recommendation
1. The construct definition changes, and the change is justified.	Operationalizations should follow from definitions (Podsakoff, MacKenzie, and Podsakoff 2016), and justified changes in definitions should lead to corresponding operationalization changes.	Older definitions of privacy (Smith, Milberg, and Burke 1996) did not include online concerns. Later work expanded the definition and included measures of online concerns (Malhotra, Kim, and Agarwal 2004).	Add, replace, change, or remove measurement items to reflect the definition change.
2. Repeated usage of an accepted operationalization suggests that it has limitations (e.g., measurement model evaluations in structural equation models from different data sets suggest that certain items do not fit in the factor structure).	Improving existing operationalizations is an important methodological contribution (Summers 2001).	Several studies using the brand personality scale developed by Aaker (1997) found limitations in the scale. Geuens, Weijters, and De Wulf (2009) developed an updated scale in response to these limitations.	Add, replace, change, or remove measurement items to improve the operationalization (i.e., the validity of measurement); if not possible, consider developing a new measure.
3. A comparative study shows that a new operationalization is more valid than the previously accepted operationalization.	Developing improved operationalizations is an important methodological contribution (Summers 2001).	Graf, Mayer, and Landwehr (2018) developed an improved processing fluency measure and demonstrated higher validity than previous measures.	Use new operationalization in future studies; consider operationalization change in reviewing and combining research results of studies with prior and new measurements.
4. A more parsimonious operationalization of equal or better validity could replace an existing operationalization.	Operationalizations with fewer items save respondents’ time and increase answer quality (Bergkvist 2015).	The need for a shorter cognition scale (Cacioppo, Petty, and Kao 1984) replaced the original, lengthier scale (Cacioppo and Petty 1982). Single-item measures of attitude toward the ad and attitude toward the brand are equally valid as multiple-item measures (Bergkvist and Rossiter 2007, 2009).	Use the parsimonious operationalization in future studies; consider operationalization change in reviewing and combining research results of studies with prior and new, more parsimonious measurements.

Consistency in construct operationalizations is a precondition for knowledge accumulation within a field (Bergkvist and Langner 2017; Bergkvist and Eisend 2021). Thus, operationalizations should change only in response to changes in definitions, demonstrated weaknesses in existing operationalizations, or if better or more parsimonious measures exist, and subsequent studies should use the new or improved operationalizations.

Unjustified Operationalization Changes

There are also situations in which construct operationalization changes occur that are not justified (i.e., when there are no grounds for changing the operationalization). Studies of research practice (Bergkvist and Langner 2017; Bergkvist and Eisend 2021) have identified two situations in which unjustified operationalization change take place. First, operationalizations are changed (e.g., measurement items are dropped or added) ad hoc without corresponding definition changes or justifications as explained previously. Validated operationalizations should be left unchanged, as even minor changes may have substantive effects on research results (see examples in Rossiter 2011, p. 46; Weijters, Geuens, and Baumgartner 2013, p. 377). For example, multiple studies refer to Zaichkowsky’s (1985) definition of involvement, but many use only some items in her Personal Involvement Inventory (PII) to measure the construct. Second, scholars drop items from a validated operationalization based on a single study’s measurement model evaluation, which is inherently perilous, as a single study can provide random results that cannot be generalized (Summers 2001). Changes in operationalizations should result from repeated evaluations showing a consistent and generalizable pattern (as noted previously). If scholars detect a problem with a measurement operationalization, they could report results for both the complete and the reduced operationalization (perhaps in a Web appendix) to ensure the availability of results comparable to other studies.

Unjustified operationalization changes risk causing inconsistency among studies, which imperils knowledge accumulation within the field (Bergkvist and Langner 2020; Bergkvist and Eisend 2021). Scholars should refrain from adapting, changing, or replacing existing measures unless there are solid grounds for doing so and report these grounds.

Discussion

Our article discusses when changes in definitions and operationalizations are justified and when to avoid them. Advertising scholars should change definitions and operationalizations only when justified to ensure that knowledge can accumulate and our field can progress. Scholars should explicitly state why they have made changes, reporting previous definitions and their limitations or the domain changes that motivate change. Scholars should prevent unjustified definition and operationalization changes from spreading and impacting future studies.

Tables 1 through 3 outline the factors that justify changes in definitions and operationalizations and how they should be improved. If a definition is lacking in the specification of the construct’s characteristics, the improved definition should specify the necessary attributes of the definition using precise primitive terms (or specify the meaning of the terms) that clarify the construct’s meaning. If a definition has unclear boundaries toward other constructs, the boundaries toward other constructs should be clarified and overlapping constructs consolidated. Overspecific definitions should be simplified and relationships excluded (Table 1). If the construct’s domain changes, the impact on the definition should be analyzed, and the definition changed accordingly (Table 2). If a definition has been (justifiably) changed, a change in the operationalization should follow. Similarly, demonstrated limitations in existing operationalizations and the development of better or more parsimonious operationalizations justify changes in operationalizations (Table 3).

Scholars should justify the definition change by outlining the limitations in the existing definitions and how the new definition resolves these. Thus, studies changing definitions should analyze existing definitions carefully, outline their limitations, and put forward the updated definition together with explanations of how the new definition addresses the limitations in the existing definition (for a recent example of this process, see Bergkvist and Taylor, forthcoming). The process for operationalization change should follow a similar logic.

Importantly, scholars should always include definitions and corresponding references, as perhaps the most prominent problem is that key constructs are not defined. Bergkvist and Langner (2017) found that about one-third of advertising studies did not define attitude toward the ad, attitude toward the brand, or purchase intention, even if their minimum requirement for a definition was a reference to a previous study that had defined the construct. Similarly, Katsikeas et al. (2016) found that less than 10% of marketing studies defined the marketing performance construct. One straightforward step toward ensuring construct transparency would be to require that manuscripts submitted for publication include a table with definitions and operationalizations, including references and changes, for all constructs in the study (Bergkvist and Langner 2019).

An ideal solution for promoting consensus-based evolution toward optimal definitions and operationalizations would be for advertising scholars to organize an online database with definitions and operationalizations of the field’s constructs. The database would store all updates to construct definitions and operationalizations and explain the reasons for the changes. The categories provided in Tables 1 through 3 can serve as an orientation to assign the correct reasons for a change. A construct database can help other scholars identify the changes made to a construct and whether these changes are justified and support scholars in finding the best available definition and operationalization for a construct. This support would be beneficial for scholars who apply different constructs in their work, including constructs they are less familiar with and they lack knowledge about the status of a construct in the construct life cycle. Setting up and maintaining a construct database would be a demanding task requiring advertising scholars’ concerted effort, and it might not be attainable in the short run. However, the American Academy of Advertising and the European Academy of Advertising could take the first step by forming a joint task force to do a feasibility study and propose a plan for setting up the database.

Working toward better construct definitions and operationalizations is the advertising research community’s responsibility and is in its shared interest. We hope the structure for improvement that we have proposed can contribute to this ongoing process by increasing scholars’ awareness of the issues and suggesting how to address them.

Disclosure Statement

The authors have no conflicts of interest to declare.

Additional information

Notes on contributors

Lars Bergkvist

Lars Bergkvist (PhD, Stockholm School of Economics, Sweden) is a professor of marketing, Norwegian School of Hotel Management, University of Stavanger.

Martin Eisend

Martin Eisend (PhD, Free University Berlin, Germany) is a professor of marketing, European University Viadrina.