9 Frontier Constructs

The construct map of marketing is not closed. As markets, technologies, and the texture of consumer life change, new constructs enter the nomological network—some because new phenomena demand them (consumers now react to algorithms, share rather than own, and trade data for personalization), some because old intuitions finally acquire measurement. This chapter surveys six constructs at the current frontier: customer experience, psychological ownership, brand authenticity, algorithm aversion (and the broader family of consumer responses to AI), privacy concern, and access-based consumption. Each has accumulated enough lawful relations to meet the Cronbach–Meehl admission test that opens this part, and each illustrates a distinct frontier—experiential, possessive, identity-authenticity, technological, informational, and ownership-model.

A frontier construct carries a special hazard worth stating before we begin. The newest constructs are, almost by definition, the least validated: their scales are young, their discriminant validity against neighboring constructs is often unestablished, and the literature has had little time to expose them to the jingle–jangle fallacies (calling two different things by one name; calling one thing by two names). The measurement discipline of Chapter 3 therefore weighs more heavily at the frontier, not less. The excitement of a new construct is not evidence that it is distinct from the constructs it joins, and the burden of proof for a frontier construct is the full convergent/discriminant/nomological battery, applied with extra skepticism.

9.1 Customer Experience

The most integrative frontier construct is customer experience (CX). Lemon and Verhoef (2016) define it as the customer’s multidimensional response—cognitive, emotional, sensorial, social, and behavioral—to a firm’s offerings across the entire customer journey, from pre-purchase through purchase to post-purchase, accumulated over many touchpoints the firm only partly controls (brand-owned, partner-owned, customer-owned, and social/external). CX subsumes and reorganizes several constructs in this part: satisfaction (Chapter 4) becomes one episodic, post-purchase slice of experience; perceived quality and value (Chapter 6) are cognitive components; engagement (Chapter 8) is a behavioral one. The construct’s frontier status comes from its measurement difficulty rather than novelty of idea: experience is dynamic (it evolves along the journey), distributed (across touchpoints, many beyond the firm’s control), and multidimensional, so a single cross-sectional survey scale is structurally inadequate to it. Measuring CX well requires journey-level, multi-touchpoint, multi-moment data—exactly the kind of behavioral and unstructured data the methodology and multimodal parts of this book develop—and the gap between the expansive construct and the thin measures usually used for it is the construct-to-measure gap of Figure 3.1 at its widest.

9.2 Psychological Ownership

Psychological ownership is the feeling that a target—a product, a brand, an idea—is “MINE,” in the absence of (or in addition to) legal ownership. The construct matters because that feeling of ownership, not the legal fact, drives valuation: it is the psychological engine behind the endowment effect, the tendency to value a good more once it feels one’s own. Peck and Shu (2009) show that mere touch raises perceived ownership and, through it, valuation and willingness to pay—physical contact, or even vivid imagined contact, induces the feeling of mine-ness. Fuchs, Prandelli, and Schreier (2010) demonstrate that empowerment strategies (letting customers participate in product selection or design) raise psychological ownership of the resulting product, which in turn raises demand—a “I-designed-it-myself” effect that converts participation into preference. Psychological ownership has clear antecedents (control over the target, intimate knowledge of it, investment of the self into it) and clear consequences (higher valuation, stewardship, willingness to pay), making it a well-formed construct with a genuine nomological network.

9.2.1 Ownership and valuation: an endowment simulation

The core nomological claim—psychological ownership raises valuation—is directly simulable as an endowment/touch manipulation. The chunk assigns consumers to an ownership-inducing condition (e.g., touch or self-design) versus a control, and shows the induced ownership shifts willingness to pay upward, the Peck and Shu (2009) and Fuchs, Prandelli, and Schreier (2010) result in miniature.

Code

set.seed(2009)
n <- 2000

owned <- rbinom(n, 1, 0.5)                      # 1 = touch/self-design condition
base_value <- rnorm(n, mean = 20, sd = 5)       # intrinsic valuation of the good

# Psychological ownership is induced by the condition, with individual variation,
# and ownership raises willingness to pay (the endowment channel).
psych_own <- 0.9 * owned + rnorm(n, sd = 0.4)   # latent felt ownership
wtp <- base_value + 6.0 * psych_own + rnorm(n, sd = 3)

cat(sprintf("Mean WTP, control            : $%.2f\n", mean(wtp[owned == 0])))
#> Mean WTP, control            : $19.98
cat(sprintf("Mean WTP, ownership-induced  : $%.2f\n", mean(wtp[owned == 1])))
#> Mean WTP, ownership-induced  : $25.34
# Mediation: does felt ownership carry the condition's effect on WTP?
tot <- coef(lm(wtp ~ owned))["owned"]
dir <- coef(lm(wtp ~ owned + psych_own))["owned"]
cat(sprintf("Total effect of condition on WTP   : $%.2f\n", tot))
#> Total effect of condition on WTP   : $5.35
cat(sprintf("Direct effect controlling ownership: $%.2f  (mostly mediated)\n", dir))
#> Direct effect controlling ownership: $-0.50  (mostly mediated)

The ownership condition lifts willingness to pay, and the lift is largely mediated by felt ownership—the direct effect shrinks once psychological ownership is controlled—confirming the construct’s role as the mechanism rather than the manipulation being a mere framing trick. This is the endowment effect re-described as a measurable consumer construct with a touchable antecedent.

9.3 Brand Authenticity

Brand authenticity is the consumer’s perception that a brand is genuine, true to itself, and not merely a commercial construction. Morhart et al. (2015) give the construct an integrative framework and a measurement scale with four dimensions—continuity (the brand’s timelessness and historical consistency), credibility (its ability and willingness to deliver what it promises), integrity (its moral purity and care), and symbolism (its serving as a resource for consumers’ identity). The construct is a frontier one because the “authenticity economy” has made perceived genuineness a competitive battleground, and because it is, like all the constructs in this book, perceived: authenticity is not a property a brand objectively has but a judgment consumers render, contaminated by marketing communications that try to manufacture the very impression of un-manufactured-ness. That tension—authenticity claimed through inauthentic means—makes its measurement and its nomological links to trust (Chapter 5), brand love (Chapter 7), and self-connection an active and subtle research area.

9.4 Consumer Responses to AI: Algorithm Aversion and Beyond

A fast-growing construct family concerns how consumers respond to algorithms and artificial intelligence as decision agents. Dietvorst, Simmons, and Massey (2015) document algorithm aversion: people erroneously avoid algorithms after seeing them err, abandoning a statistically superior algorithm for an inferior human judge because they penalize algorithmic mistakes more harshly than identical human mistakes. Longoni, Bonezzi, and Morewedge (2019) sharpen the mechanism in the medical domain—resistance to medical AI—and trace it to uniqueness neglect: consumers resist AI care because they believe an algorithm cannot account for their personal uniqueness the way a human provider can, even when the AI is more accurate. Longoni and Cian (2022) add the word-of-machine effect: consumers prefer AI recommendations for utilitarian attributes (where accuracy dominates) but human recommendations for hedonic/experiential attributes (where they doubt the machine can appreciate subjective experience). These are distinct constructs—an aversion, a resistance, a domain-contingent preference—but they form a coherent family describing the consumer’s trust in and acceptance of non-human agents, a family that did not exist fifteen years ago and that is now central to marketing as AI mediates more of the customer journey (Chapter 65).

9.4.1 Why a superior algorithm gets abandoned

The Dietvorst, Simmons, and Massey (2015) finding hinges on an asymmetry: an identical-sized error damages trust more when an algorithm makes it than when a human does. The chunk simulates consumers choosing between a more-accurate algorithm and a less-accurate human after observing each err, and shows that the heavier trust penalty on algorithmic error drives adoption of the inferior human—algorithm aversion as an emergent consequence of asymmetric error-attribution.

Code

set.seed(2015)
n <- 3000

# The algorithm is genuinely more accurate (smaller errors) than the human.
err_algo  <- abs(rnorm(n, sd = 0.6))
err_human <- abs(rnorm(n, sd = 1.0))

# But algorithmic error is penalized ~2x as heavily (the aversion asymmetry).
penalty_algo  <- 2.0
penalty_human <- 1.0
trust_algo  <- 10 - penalty_algo  * err_algo
trust_human <- 10 - penalty_human * err_human

choose_algo <- trust_algo > trust_human
cat(sprintf("Algorithm is more accurate (mean error %.2f vs %.2f).\n",
            mean(err_algo), mean(err_human)))
#> Algorithm is more accurate (mean error 0.47 vs 0.78).
cat(sprintf("Yet only %.1f%% of consumers choose the (superior) algorithm,\n",
            100 * mean(choose_algo)))
#> Yet only 44.3% of consumers choose the (superior) algorithm,
cat("because algorithmic errors are penalized twice as heavily as human ones.\n")
#> because algorithmic errors are penalized twice as heavily as human ones.
# Remove the asymmetry: penalize both equally -> adoption flips.
fair <- (10 - err_algo) > (10 - err_human)
cat(sprintf("Under symmetric error penalties, %.1f%% choose the algorithm.\n",
            100 * mean(fair)))
#> Under symmetric error penalties, 65.8% choose the algorithm.

Under the asymmetric penalty the superior algorithm is rejected by a majority; remove the asymmetry and adoption flips to favor it. Algorithm aversion is thus not (only) about accuracy beliefs but about a differential trust response to machine error—a construct with direct design implications: deploying AI requires managing the error-attribution asymmetry, not merely proving the AI is better.

9.5 Privacy Concern

As marketing runs on personal data, privacy concern—the consumer’s apprehension about the collection and use of their personal information—has become a first-order construct. Kelly D. Martin and Murphy (2017) provide an integrative framework, organizing privacy as a construct with antecedents (data practices, control, transparency), states (concern, perceived risk, perceived intrusion), and consequences (trust, disclosure, patronage). The defining tension is the personalization–privacy paradox: the same data that lets a firm personalize—raising perceived value (Chapter 6) and relationship quality (Chapter 5)—simultaneously raises privacy concern, so personalization both attracts and repels, and the net effect is contingent on trust and control. The construct ties directly to the firm-performance evidence on data practices and to the deeper treatment of privacy regulation and its marketing consequences in Chapter 24. Its measurement inherits a sharp version of the social-desirability contamination of Chapter 3: stated privacy concern routinely overstates the behavior it predicts (the “privacy paradox” in which people who report high concern nonetheless disclose freely), a convergent-validity gap between the attitudinal measure and revealed behavior that any serious use of the construct must confront.

9.6 Access-Based Consumption

Finally, the rise of the sharing economy has produced a construct that reframes ownership itself. Bardhi and Eckhardt (2012) conceptualize access-based consumption—consumption in which no transfer of ownership occurs (car sharing, streaming, rental, subscription)—and show it is not simply diluted ownership but a distinct mode with its own psychology: consumers relate to accessed objects with less identification and less stewardship, treating them instrumentally rather than as extensions of the self. This inverts the psychological-ownership construct of Section 9.2: where ownership induces mine-ness and care, access deliberately withholds them, and the firm managing an access model must build value without the identification that ownership supplies. The construct matters because the ownership-to-access shift undercuts the self–brand connection and psychological-ownership mechanisms that much of this part relies on, demanding new accounts of how value, loyalty, and brand meaning form when the consumer never owns anything.

9.7 The Frontier-Construct Discipline

These six constructs share a methodological situation that is the chapter’s closing lesson. Frontier constructs are proposed faster than they are validated, and the field’s incentive to name a “new” construct outruns its incentive to prove the construct is distinct from its neighbors. Customer experience risks absorbing satisfaction, quality, value, and engagement without demonstrating it adds variance beyond their sum; brand authenticity overlaps trust and self-connection; the AI-response constructs proliferate (aversion, resistance, word-of-machine, and more) faster than their discriminant validity is established. The remedy is not to distrust new constructs but to hold them to the same standard as old ones—the convergent, discriminant, and nomological evidence of Chapter 3—with extra vigilance for the jingle–jangle fallacies that the rush to name invites. Increasingly, frontier construct measurement is also computational: experience, authenticity, and privacy concern are scored from text, behavior, and multimodal traces rather than surveys, which relocates but does not repeal the validity problem, exactly as Section 3.11 argues. The frontier is where the constructs of tomorrow are formed, and it is precisely there that the measurement theory of this part does its most important work.

9.8 Key Takeaways

The construct map is open; frontier constructs earn admission by the same Cronbach–Meehl test as established ones, and—being youngest—carry the highest discriminant-validity and jingle–jangle risk, so the discipline of Chapter 3 weighs more, not less.
Customer experience (Lemon and Verhoef 2016) is a multidimensional response across the whole journey and touchpoints; it subsumes satisfaction/quality/value/engagement and is hard to measure because it is dynamic, distributed, and multidimensional.
Psychological ownership—felt mine-ness without legal title—drives valuation via the endowment effect, induced by mere touch (Peck and Shu 2009) and empowerment/self-design (Fuchs, Prandelli, and Schreier 2010) (Section 9.2.1).
Brand authenticity (Morhart et al. 2015) is a perceived four-dimensional judgment (continuity, credibility, integrity, symbolism), contested because it is often manufactured through inauthentic means.
Consumer responses to AI form a construct family—algorithm aversion (Dietvorst, Simmons, and Massey 2015), resistance to medical AI via uniqueness neglect (Longoni, Bonezzi, and Morewedge 2019), and the word-of-machine effect (Longoni and Cian 2022)—unified by asymmetric trust in machine error/judgment (Section 9.4.1).
Privacy concern (Kelly D. Martin and Murphy 2017) centers the personalization–privacy paradox and a stated-vs-revealed convergent-validity gap; it links to Chapter 24. Access-based consumption (Bardhi and Eckhardt 2012) inverts psychological ownership, withholding identification and stewardship and forcing new accounts of value and loyalty.
Frontier measurement is increasingly computational, relocating but not repealing the validity problem (Section 3.11).

9.9 Further Reading

Customer experience is Lemon and Verhoef (2016). Psychological ownership runs through Peck and Shu (2009) and Fuchs, Prandelli, and Schreier (2010); brand authenticity through Morhart et al. (2015). The consumer-AI family is Dietvorst, Simmons, and Massey (2015), Longoni, Bonezzi, and Morewedge (2019), and Longoni and Cian (2022), developed further in Chapter 65. Privacy concern is Kelly D. Martin and Murphy (2017) with the firm-performance evidence in Kelly D. Martin, Borah, and Palmatier (2017) and the regulatory treatment in Chapter 24; access-based consumption is Bardhi and Eckhardt (2012). The overarching methodological frame—that frontier and computational constructs are governed by, not exempt from, classical measurement theory—is laid out in Chapter 3 and Section 3.11.

Bardhi, Fleura, and Giana M. Eckhardt. 2012. “Access-Based Consumption: The Case of Car Sharing.” Journal of Consumer Research 39 (4): 881–98. https://doi.org/10.1086/666376.

Dietvorst, Berkeley J., Joseph P. Simmons, and Cade Massey. 2015. “Algorithm Aversion: People Erroneously Avoid Algorithms After Seeing Them Err.” Journal of Experimental Psychology: General 144 (1): 114–26. https://doi.org/10.1037/xge0000033.

Fuchs, Christoph, Emanuela Prandelli, and Martin Schreier. 2010. “The Psychological Effects of Empowerment Strategies on Consumers’ Product Demand.” Journal of Marketing 74 (1): 65–79. https://doi.org/10.1509/jmkg.74.1.65.

Lemon, Katherine N., and Peter C. Verhoef. 2016. “Understanding Customer Experience Throughout the Customer Journey.” Journal of Marketing 80 (6): 69–96. https://doi.org/10.1509/jm.15.0420.

Longoni, Chiara, Andrea Bonezzi, and Carey K Morewedge. 2019. “Resistance to Medical Artificial Intelligence.” Journal of Consumer Research 46 (4): 629–50. https://doi.org/10.1093/jcr/ucz013.

Longoni, Chiara, and Luca Cian. 2022. “Artificial Intelligence in Utilitarian Vs. Hedonic Contexts: The Word-of-Machine Effect.” Journal of Marketing 86 (1): 91–108. https://doi.org/10.1177/0022242920957347.

Martin, Kelly D, Abhishek Borah, and Robert W Palmatier. 2017. “Data Privacy: Effects on Customer and Firm Performance.” Journal of Marketing 81 (1): 36–58.

Martin, Kelly D., and Patrick E. Murphy. 2017. “The Role of Data Privacy in Marketing.” Journal of the Academy of Marketing Science 45 (2): 135–55. https://doi.org/10.1007/s11747-016-0495-4.

Morhart, Felicitas, Lucia Malär, Amélie Guèvremont, Florent Girardin, and Bianca Grohmann. 2015. “Brand Authenticity: An Integrative Framework and Measurement Scale.” Journal of Consumer Psychology 25 (2): 200–218. https://doi.org/10.1016/j.jcps.2014.11.006.

Peck, Joann, and Suzanne B. Shu. 2009. “The Effect of Mere Touch on Perceived Ownership.” Journal of Consumer Research 36 (3): 434–47. https://doi.org/10.1086/598614.

# Frontier Constructs {#sec-frontier} The construct map of marketing is not closed. As markets, technologies, and the texture of consumer life change, new constructs enter the nomological network—some because new phenomena demand them (consumers now react to *algorithms*, share rather than own, and trade data for personalization), some because old intuitions finally acquire measurement. This chapter surveys six constructs at the current frontier: **customer experience**, **psychological ownership**, **brand authenticity**, **algorithm aversion** (and the broader family of consumer responses to AI), **privacy concern**, and **access-based consumption**. Each has accumulated enough lawful relations to meet the Cronbach–Meehl admission test that opens this part, and each illustrates a distinct frontier—experiential, possessive, identity-authenticity, technological, informational, and ownership-model. A frontier construct carries a special hazard worth stating before we begin. The newest constructs are, almost by definition, the *least validated*: their scales are young, their discriminant validity against neighboring constructs is often unestablished, and the literature has had little time to expose them to the **jingle–jangle fallacies** (calling two different things by one name; calling one thing by two names). The measurement discipline of @sec-construct-vs-variable therefore weighs *more* heavily at the frontier, not less. The excitement of a new construct is not evidence that it is distinct from the constructs it joins, and the burden of proof for a frontier construct is the full convergent/discriminant/nomological battery, applied with extra skepticism. ## Customer Experience {#sec-fr-experience} The most integrative frontier construct is **customer experience** (CX). @lemon2016 define it as the customer's *multidimensional response*—cognitive, emotional, sensorial, social, and behavioral—to a firm's offerings across the entire **customer journey**, from pre-purchase through purchase to post-purchase, accumulated over many **touchpoints** the firm only partly controls (brand-owned, partner-owned, customer-owned, and social/external). CX subsumes and reorganizes several constructs in this part: satisfaction (@sec-satisfaction) becomes one episodic, post-purchase slice of experience; perceived quality and value (@sec-evaluative) are cognitive components; engagement (@sec-engagement) is a behavioral one. The construct's frontier status comes from its measurement difficulty rather than novelty of idea: experience is *dynamic* (it evolves along the journey), *distributed* (across touchpoints, many beyond the firm's control), and *multidimensional*, so a single cross-sectional survey scale is structurally inadequate to it. Measuring CX well requires journey-level, multi-touchpoint, multi-moment data—exactly the kind of behavioral and unstructured data the methodology and multimodal parts of this book develop—and the gap between the expansive construct and the thin measures usually used for it is the construct-to-measure gap of @fig-cm-gap at its widest. ## Psychological Ownership {#sec-fr-ownership} **Psychological ownership** is the feeling that a target—a product, a brand, an idea—is "MINE," in the absence of (or in addition to) legal ownership. The construct matters because that *feeling* of ownership, not the legal fact, drives valuation: it is the psychological engine behind the **endowment effect**, the tendency to value a good more once it feels one's own. @peck2009 show that *mere touch* raises perceived ownership and, through it, valuation and willingness to pay—physical contact, or even vivid imagined contact, induces the feeling of mine-ness. @fuchs2010 demonstrate that **empowerment** strategies (letting customers participate in product selection or design) raise psychological ownership of the resulting product, which in turn raises demand—a "I-designed-it-myself" effect that converts participation into preference. Psychological ownership has clear antecedents (control over the target, intimate knowledge of it, investment of the self into it) and clear consequences (higher valuation, stewardship, willingness to pay), making it a well-formed construct with a genuine nomological network. ### Ownership and valuation: an endowment simulation {#sec-fr-ownership-sim} The core nomological claim—psychological ownership raises valuation—is directly simulable as an endowment/touch manipulation. The chunk assigns consumers to an ownership-inducing condition (e.g., touch or self-design) versus a control, and shows the induced ownership shifts willingness to pay upward, the @peck2009 and @fuchs2010 result in miniature. ```{r} #| message: false #| warning: false #| label: fr-ownership-endowment set.seed(2009) n <- 2000 owned <- rbinom(n, 1, 0.5) # 1 = touch/self-design condition base_value <- rnorm(n, mean = 20, sd = 5) # intrinsic valuation of the good # Psychological ownership is induced by the condition, with individual variation, # and ownership raises willingness to pay (the endowment channel). psych_own <- 0.9 * owned + rnorm(n, sd = 0.4) # latent felt ownership wtp <- base_value + 6.0 * psych_own + rnorm(n, sd = 3) cat(sprintf("Mean WTP, control : $%.2f\n", mean(wtp[owned == 0]))) cat(sprintf("Mean WTP, ownership-induced : $%.2f\n", mean(wtp[owned == 1]))) # Mediation: does felt ownership carry the condition's effect on WTP? tot <- coef(lm(wtp ~ owned))["owned"] dir <- coef(lm(wtp ~ owned + psych_own))["owned"] cat(sprintf("Total effect of condition on WTP : $%.2f\n", tot)) cat(sprintf("Direct effect controlling ownership: $%.2f (mostly mediated)\n", dir)) ``` The ownership condition lifts willingness to pay, and the lift is largely *mediated* by felt ownership—the direct effect shrinks once psychological ownership is controlled—confirming the construct's role as the mechanism rather than the manipulation being a mere framing trick. This is the endowment effect re-described as a measurable consumer construct with a touchable antecedent. ## Brand Authenticity {#sec-fr-authenticity} **Brand authenticity** is the consumer's perception that a brand is *genuine, true to itself, and not merely a commercial construction*. @morhart2015 give the construct an integrative framework and a measurement scale with four dimensions—**continuity** (the brand's timelessness and historical consistency), **credibility** (its ability and willingness to deliver what it promises), **integrity** (its moral purity and care), and **symbolism** (its serving as a resource for consumers' identity). The construct is a frontier one because the "authenticity economy" has made perceived genuineness a competitive battleground, and because it is, like all the constructs in this book, *perceived*: authenticity is not a property a brand objectively has but a judgment consumers render, contaminated by marketing communications that try to manufacture the very impression of un-manufactured-ness. That tension—authenticity claimed through inauthentic means—makes its measurement and its nomological links to trust (@sec-relational), brand love (@sec-brand-self), and self-connection an active and subtle research area. ## Consumer Responses to AI: Algorithm Aversion and Beyond {#sec-fr-ai} A fast-growing construct family concerns how consumers respond to **algorithms and artificial intelligence** as decision agents. @dietvorst2015 document **algorithm aversion**: people erroneously avoid algorithms after seeing them err, abandoning a statistically superior algorithm for an inferior human judge because they penalize algorithmic mistakes more harshly than identical human mistakes. @longoni2019 sharpen the mechanism in the medical domain—**resistance to medical AI**—and trace it to **uniqueness neglect**: consumers resist AI care because they believe an algorithm cannot account for their personal uniqueness the way a human provider can, even when the AI is more accurate. @longoni2022wordofmachine add the **word-of-machine effect**: consumers prefer AI recommendations for *utilitarian* attributes (where accuracy dominates) but human recommendations for *hedonic/experiential* attributes (where they doubt the machine can appreciate subjective experience). These are distinct constructs—an aversion, a resistance, a domain-contingent preference—but they form a coherent family describing the consumer's *trust in and acceptance of* non-human agents, a family that did not exist fifteen years ago and that is now central to marketing as AI mediates more of the customer journey (@sec-ai-ml). ### Why a superior algorithm gets abandoned {#sec-fr-ai-sim} The @dietvorst2015 finding hinges on an *asymmetry*: an identical-sized error damages trust more when an algorithm makes it than when a human does. The chunk simulates consumers choosing between a more-accurate algorithm and a less-accurate human after observing each err, and shows that the heavier trust penalty on algorithmic error drives adoption of the *inferior* human—algorithm aversion as an emergent consequence of asymmetric error-attribution. ```{r} #| message: false #| warning: false #| label: fr-algorithm-aversion set.seed(2015) n <- 3000 # The algorithm is genuinely more accurate (smaller errors) than the human. err_algo <- abs(rnorm(n, sd = 0.6)) err_human <- abs(rnorm(n, sd = 1.0)) # But algorithmic error is penalized ~2x as heavily (the aversion asymmetry). penalty_algo <- 2.0 penalty_human <- 1.0 trust_algo <- 10 - penalty_algo * err_algo trust_human <- 10 - penalty_human * err_human choose_algo <- trust_algo > trust_human cat(sprintf("Algorithm is more accurate (mean error %.2f vs %.2f).\n", mean(err_algo), mean(err_human))) cat(sprintf("Yet only %.1f%% of consumers choose the (superior) algorithm,\n", 100 * mean(choose_algo))) cat("because algorithmic errors are penalized twice as heavily as human ones.\n") # Remove the asymmetry: penalize both equally -> adoption flips. fair <- (10 - err_algo) > (10 - err_human) cat(sprintf("Under symmetric error penalties, %.1f%% choose the algorithm.\n", 100 * mean(fair))) ``` Under the asymmetric penalty the superior algorithm is rejected by a majority; remove the asymmetry and adoption flips to favor it. Algorithm aversion is thus not (only) about accuracy beliefs but about a *differential trust response to machine error*—a construct with direct design implications: deploying AI requires managing the error-attribution asymmetry, not merely proving the AI is better. ## Privacy Concern {#sec-fr-privacy} As marketing runs on personal data, **privacy concern**—the consumer's apprehension about the collection and use of their personal information—has become a first-order construct. @martinmurphy2017 provide an integrative framework, organizing privacy as a construct with antecedents (data practices, control, transparency), states (concern, perceived risk, perceived intrusion), and consequences (trust, disclosure, patronage). The defining tension is the **personalization–privacy paradox**: the same data that lets a firm personalize—raising perceived value (@sec-evaluative) and relationship quality (@sec-relational)—simultaneously raises privacy concern, so personalization both attracts and repels, and the net effect is contingent on trust and control. The construct ties directly to the firm-performance evidence on data practices and to the deeper treatment of privacy regulation and its marketing consequences in @sec-privacy. Its measurement inherits a sharp version of the social-desirability contamination of @sec-construct-vs-variable: stated privacy concern routinely *overstates* the behavior it predicts (the "privacy paradox" in which people who report high concern nonetheless disclose freely), a convergent-validity gap between the attitudinal measure and revealed behavior that any serious use of the construct must confront. ## Access-Based Consumption {#sec-fr-access} Finally, the rise of the sharing economy has produced a construct that reframes ownership itself. @bardhi2012 conceptualize **access-based consumption**—consumption in which no transfer of ownership occurs (car sharing, streaming, rental, subscription)—and show it is *not* simply diluted ownership but a distinct mode with its own psychology: consumers relate to accessed objects with less identification and less stewardship, treating them instrumentally rather than as extensions of the self. This inverts the psychological-ownership construct of @sec-fr-ownership: where ownership induces mine-ness and care, access deliberately withholds them, and the firm managing an access model must build value without the identification that ownership supplies. The construct matters because the ownership-to-access shift undercuts the self–brand connection and psychological-ownership mechanisms that much of this part relies on, demanding new accounts of how value, loyalty, and brand meaning form when the consumer never owns anything. ## The Frontier-Construct Discipline {#sec-fr-discipline} These six constructs share a methodological situation that is the chapter's closing lesson. Frontier constructs are proposed faster than they are validated, and the field's incentive to name a "new" construct outruns its incentive to prove the construct is distinct from its neighbors. Customer experience risks absorbing satisfaction, quality, value, and engagement without demonstrating it adds variance beyond their sum; brand authenticity overlaps trust and self-connection; the AI-response constructs proliferate (aversion, resistance, word-of-machine, and more) faster than their discriminant validity is established. The remedy is not to distrust new constructs but to hold them to the *same* standard as old ones—the convergent, discriminant, and nomological evidence of @sec-construct-vs-variable—with extra vigilance for the jingle–jangle fallacies that the rush to name invites. Increasingly, frontier construct measurement is also *computational*: experience, authenticity, and privacy concern are scored from text, behavior, and multimodal traces rather than surveys, which relocates but does not repeal the validity problem, exactly as @sec-cvv-frontiers argues. The frontier is where the constructs of tomorrow are formed, and it is precisely there that the measurement theory of this part does its most important work. ## Key Takeaways {#sec-fr-takeaways} - The construct map is open; **frontier constructs** earn admission by the same Cronbach–Meehl test as established ones, and—being youngest—carry the *highest* discriminant-validity and jingle–jangle risk, so the discipline of @sec-construct-vs-variable weighs more, not less. - **Customer experience** [@lemon2016] is a multidimensional response across the whole journey and touchpoints; it subsumes satisfaction/quality/value/engagement and is hard to measure because it is dynamic, distributed, and multidimensional. - **Psychological ownership**—felt mine-ness without legal title—drives valuation via the endowment effect, induced by mere touch [@peck2009] and empowerment/self-design [@fuchs2010] (@sec-fr-ownership-sim). - **Brand authenticity** [@morhart2015] is a perceived four-dimensional judgment (continuity, credibility, integrity, symbolism), contested because it is often manufactured through inauthentic means. - Consumer **responses to AI** form a construct family—algorithm aversion [@dietvorst2015], resistance to medical AI via uniqueness neglect [@longoni2019], and the word-of-machine effect [@longoni2022wordofmachine]—unified by *asymmetric trust in machine error/judgment* (@sec-fr-ai-sim). - **Privacy concern** [@martinmurphy2017] centers the personalization–privacy paradox and a stated-vs-revealed convergent-validity gap; it links to @sec-privacy. **Access-based consumption** [@bardhi2012] inverts psychological ownership, withholding identification and stewardship and forcing new accounts of value and loyalty. - Frontier measurement is increasingly **computational**, relocating but not repealing the validity problem (@sec-cvv-frontiers). ## Further Reading {#sec-fr-further} Customer experience is @lemon2016. Psychological ownership runs through @peck2009 and @fuchs2010; brand authenticity through @morhart2015. The consumer-AI family is @dietvorst2015, @longoni2019, and @longoni2022wordofmachine, developed further in @sec-ai-ml. Privacy concern is @martinmurphy2017 with the firm-performance evidence in @martin2017data and the regulatory treatment in @sec-privacy; access-based consumption is @bardhi2012. The overarching methodological frame—that frontier and computational constructs are governed by, not exempt from, classical measurement theory—is laid out in @sec-construct-vs-variable and @sec-cvv-frontiers.