1

2016, American Marketing Association

Journal of Marketing Research
PrePrint, Unedited
All rights reserved. Cannot be reprinted without the express
permission of the American Marketing Association.

Be Careless with That!

Availability of Product Upgrades Increases Cavalier Behavior toward Possessions

SILVIA BELLEZZA

JOSHUA M. ACKERMAN

FRANCESCA GINO

Silvia Bellezza, Assistant Professor of Marketing at Columbia Business School, Uris Hall 508,

3022 Broadway, New York, NY 10027, +1 212 854 2177, sbellezza@gsb.columbia.edu.

Joshua M. Ackerman, Assistant Professor of Psychology at University of Michigan, 1012 East

Hall, 530 Church Street, Ann Arbor, MI 48109-1043, +1 734 647 9028, joshack@umich.edu.

Francesca Gino, Tandom Family Professor of Business Administration at Harvard Business

School, Baker Library 447, Soldiers Field Road, Boston, MA 02163, +1 617 495 0875,

fgino@hbs.edu.
 2

Acknowledgments

The authors are thankful to the IMEI Detective Company for sharing their dataset on lost Apple

iPhones. The authors are grateful for helpful comments and suggestions received from Eva

Ascarza, Kate Barasz, Chris Foster, John T. Gourville, Eric J. Johnson, Ran Kivetz, Donald R.

Lehmann, Oded Netzer, and Neeru Paharia, participants in seminars and lab groups at Columbia

Business School, Harvard Business School, MIT Sloan School of Management, and the

Association for Consumer Research conference.

 3

ABSTRACT

Consumers are often faced with the opportunity to purchase a new, enhanced product, such as a
new phone, even though the product they currently own is still fully functional. The authors
propose that consumers act more recklessly with their current products when in the presence of
appealing, though not yet attained, product upgrades (not just mere replacements). Carelessness
and neglect toward currently owned products stem from a desire to justify the attainment of
upgrades without appearing wasteful. A series of studies with actual owners of a wide range of
different goods (e.g., durable, consumable, functional, and hedonic products) and evidence from
a real-word dataset of lost Apple iPhones demonstrate how the availability of product upgrades
increases cavalier behavior toward possessions. Moreover, the authors demonstrate that product
neglect in the presence of attractive upgrades can occur without deliberate intentions. Finally,
theoretical and managerial implications of these findings are discussed.

Keywords: carelessness, product upgrade, ownership, justification

 4

In a strange series of events, a consumer microwaves his cell phone instead of a burrito, a

lab worker drops his mobile into a vat of toxic sludge, and a commuter throws her phone into the

backseat of a departing taxi. Oops! As becomes clear at the end of this Virgin Mobile TV

commercial, named Happy Accidents, these phone owners are all intent on accidentally

destroying or losing their devices, thus necessitating an upgrade purchase. Is this simply a

humorous advertisement, or can knowledge about the availability of product upgrades actually

lead consumers to mistreat the products they own? We investigate this question by examining the

potential for consumers to exhibit cavalier behavior toward current possessions when product

upgrades are present, a novel topic in the consumer behavior literature.

Building on previous research investigating product upgrade decisions (Bayus 1991;

Cripps and Meyer 1994; Heath and Fennema 1996; Jacoby, Berning, and Dietvorst 1977; Okada

2001, 2006), we propose that individuals become careless and negligent with their products

when in the presence of appealing, though not yet attained, product upgrades. We label this

phenomenon the upgrade effect and suggest that such careless tendencies are intended to

promote the acquisition of upgrade products by helping consumers justify the new purchase. Our

research is of particular relevance to modern production systems, which have significantly

increased the pace of product innovation and the introduction of newer products over the past

decades (Urban and Hauser 1993). In todays advanced economies, consumers are often faced

with the opportunity to purchase an enhanced product (e.g., a new laptop or phone), even though

the device they currently own is still fully functional. Because consumers keep track of the costs

and benefits of their purchases over time (Gourville and Soman 1998; Heath and Fennema 1996;

Prelec and Loewenstein 1998; Thaler 1980) and display a strong aversion to waste (Arkes 1996;

Bolton and Alba 2012), accidentally damaging a product or running out of it quickly allows
 5

them to write-off the residual value of the product and upgrade without recording a loss or

appearing wasteful. Accordingly, we propose that motivated carelessness is driven by the need to

justify the attainment of upgrade products, consistent with the notion that consumers have a

strong desire to justify their decisions and base their choices on reasons (Hsee 1995; Okada

2005; Shafir, Simonson, and Tversky 1993).

We begin with a discussion of the literature on product upgrades and justification

processes. Integrating these two streams of research, we propose a new conceptualization for

understanding how consumers treat their own belongings in the presence of preferred product

upgrades. Next, we report findings from eight studies examining the consequences of product

upgrade availability. Finally, we discuss theoretical and practical implications of our work.

THEORETICAL FOUNDATIONS

Product Upgrades and Carelessness

In most instances, the purchase of a product is not a first time purchase experience in the

product category (Cripps and Meyer 1994; Fernandez 2001). This is almost always the case for

consumer goods that have become standard items in developed economies (e.g., phones,

sunglasses, shampoo). Upgradesnew and enhanced productsare therefore quite common

consumer purchases. When upgrading, a consumer who owns a product that is still functional

purchases a new and improved version of what currently owned (Okada 2006, p. 92). We suggest

that for product upgrades to induce carelessness it is indeed important that the upgrade product is

an enhanced version of the current one (and not just a mere product replacement) and that

consumers are interested in upgrading. In line with previous research on upgrades and new
 6

products (Bertini, Ofek, and Ariely 2009; Jacoby, Berning, and Dietvorst 1977; Mukherjee and

Hoyer 2001; Okada 2001, 2006), in our studies we examine a wide assortment of goods, such as

durable (e.g., phones, mugs), consumable (e.g., shampoo, laundry softener), functional (e.g.,

eyeglasses, toothpaste), and hedonic (e.g., sunglasses, perfume) products. In the context of

durable goods, such as phones for example, an upgrade may be a newer and improved device

(e.g., a new Apple iPhone model with better camera); in the context of non-durable products,

such as toothpaste or laundry detergent for example, an upgrade may instead be a newer and

improved product formula (e.g., a new Colgate toothpaste with better whitening agent). In

general, when consumers buy new products, they open mental accounts that allow them to

keep track of the costs and benefits of these purchases over time (Gourville and Soman 1998;

Heath and Fennema 1996; Prelec and Loewenstein 1998; Thaler 1980). During their ownership

of a good, consumers mentally amortize the initial purchase price and, when considering an

upgrade, determine whether they got their moneys worth on the last purchase (Heath and

Fennema 1996). If the old product is still functional and not fully depreciated in the mental

account, the decision to upgrade to a new product is painful, as the consumer must write off the

remaining mental book value as a loss (Gourville and Soman 1998; Okada 2001). Thus, the

pain associated with retiring a good before its mental account is fully amortized often hinders the

upgrade purchase. Moreover, even when a product is at the end of its presumed life cycle,

consumers still display a strong aversion to waste and unused utility when deciding whether to

discard the product for good (Arkes 1996; Bolton and Alba 2012).

But what if a product became no longer functional, was damaged, or was used up ante

tempore? In this case, consumers could purchase an upgrade before the expected time without

appearing wasteful or experiencing guilt. Though consumers could simply wait for products they
 7

own to degrade or deplete over time, knowledge about the presence of a desired upgrade is likely

to make waiting less appealing. Acting carelessly, neglecting repair, or even consuming at a

faster rate may all be alternative methods of facilitating the upgrade process. Here, we define

carelessness toward owned products as a failure to give sufficient attention to avoiding damage

and/or preserving the current state of the product. With specific regards to durable and non-

durable products, indicators of carelessness are product neglect (e.g., not looking for the product

in case of loss, neglecting repair), risky behaviors (e.g., endangering the product), and faster

consumption rates (e.g., pouring more product than usual). To confirm our conceptualization and

our intended operationalization of carelessness throughout the studies, we conducted a short pilot

study (online appendix) demonstrating that carelessness for durable and non-durable products is

indeed associated with these outcomes.

We propose that the upgrade effect can encourage carelessness and product neglect even

without deliberate, careless intentions. This assertion is in line with previous research

documenting the non-conscious nature of certain consumer behaviors and effects (Folkes,

Martin, and Gupta 1993; Khan and Dhar 2006; Shiv, Carmon, and Ariely 2005). For example,

Khan and Dhar (2006) show that consumers may be unaware of how their prior virtuous conduct

influences their subsequent choices and that the processes underlying this licensing effect may

be largely non-conscious. Similarly, Shiv, Carmon, and Ariely (2005) show that consumers are

unaware about the extent to which marketing actions, such as pricing, alter the actual efficacy of

products to which they are applied. With respect to the upgrade effect, if consumers act

neglectful and careless as a means of subverting the mental conflict associated with upgrade

purchases, they may in fact primarily do so without conscious deliberateness. Intentionally

damaging or mistreating owned products are seemingly irrational behaviors. Purposefully

 8

engaging in these behaviors may prevent a person from construing a damaged product as a loss-

free write-off and instead encourage perceptions that one is wasteful. Indeed, consumers often

willfully ignore information and strategically manage their cognitions to maintain a positive self-

image when they are motivated to do so (Ehrich and Irwin 2005; Paharia, Vohs, and Deshpand

2013). To empirically validate our claim about the non-deliberateness of the upgrade effect, we

examine the extent to which people are conscious of these potential tendencies in a variety of

ways (Studies 1B, 3A, 3B, and follow-up study).

To resolve conflicts and justify their choices, consumers often seek and construct reasons

for doing so (Shafir, Simonson, and Tversky 1993). Similarly, we suggest, the upgrade effect is

driven by the psychological process of need for justification.

The Need for Justification

In general, replacement purchases can occur as a result of forced or unforced

situations (Bayus 1988; Grewal, Mehta, and Kardes 2004). A product failure, for example,

generates a forced purchase situation; by contrast, unforced situations include replacement of

working units due to changes in style or personal preferences. Decisions made in unforced

situations feature strong individual freedom of choice (I want a new product) and can appear

less justifiable and more personally inconsistent than forced situations involving little freedom

(I need a new product) (Baumeister and Tice 1984; Linder, Cooper, and Jones 1967).

In this research, we propose that consumers who are interested in upgrading are more

likely to act carelessly with their current belongings in order to pass from an unforced (less

justifiable) to a forced (more justifiable) replacement situation. The idea that consumers may

desire a justification for upgrading is rooted in the general finding that people seek to base their
 9

decisions on reasons (Hsee 1995; Kunda 1990; Shafir, Simonson, and Tversky 1993). Indeed,

certain consumers rely more on reasons than other consumers, and certain decision contexts

necessitate more attention to reasons than other contexts. On an individual level, people differ

considerably in their actual tendencies to base their decisions on reason versus feelings (Hsee et

al. 2015; Pham 1998). Specifically, the lay rationalism scale (Hsee et al. 2015) measures the

degree to which individuals use reason rather than feelings to guide decisions (e.g., When

making decisions, I like to analyze financial costs and benefits and resist the influence of my

feelings) and is a proxy for individual-level differences in need for justification. As such, we

expect people high in lay rationalism to rely on product damage as a justification to upgrade and

to neglect their products in the absence of a legitimate justification more than people low in lay

rationalism (those who base decisions relatively more on feelings).

On a situation level, the consumer behavior literature demonstrates that spending on

hedonic products and luxury goods requires more justification than spending on utilitarian

products and necessities (Dhar and Wertenbroch 2000; Keinan, Kivetz, and Netzer 2016; Kivetz

and Simonson 2002; Okada 2005; Strahilevitz and Myers 1998; Xu and Schwarz 2009). For

example, Okada (2005) shows that consumers are more likely to choose a dinner certificate

(hedonic consumption) rather than a grocery certificate (utilitarian consumption) if the situation

allows them to justify the choice. Similarly, Keinan, Kivetz, and Netzer (2016) find that

consumers inflate the perceived value of minor functional features or other utilitarian aspects of

luxury products to justify indulgent and seemingly superfluous purchases. This perspective

would predict that motivated carelessness is more likely to occur for hedonic than for utilitarian

products. In contrast, it is possible that consumers feel a strong need to justify attainment of any

upgrade when that newer product is unnecessary. This justification could be accomplished
 10

through deterioration of the currently owned product. Consistent with this idea, evidence in the

domain of product disposition indicates that consumers sometimes use the malfunctioning of

small and repairable aspects of products as an excuse to purchase entirely new products; for

example, scratches on the lens of a watch may provide sufficient cause for the purchase of a new

watch (Jacoby, Berning, and Dietvorst 1977). Because upgrading when the current product still

works (or is not fully used up) is unnecessary, regardless of the specific utilitarian versus hedonic

nature of the product, we expect to observe the upgrade effect for all types of products

indiscriminately. Indeed, in a wide variety of experiments examining both utilitarian and hedonic

products (i.e., 3A, 3B, 4), we do find directional trends suggesting that the upgrade effect is

relatively stronger for hedonic products, but this product type distinction is not significant in the

current data. Thus, carelessness in the face of upgrades operates for all products indiscriminately,

consistent with the notion that upgrading prematurely is an unforced consumption situation,

regardless of the product type.

Overview of the Present Research

We test our hypotheses and theoretical framework, depicted in Figure 1, in a series of

studies that employ different types of upgrading manipulations (e.g., the physical presence of a

better product, writing tasks). Throughout our studies, we also investigate different indicators of

carelessness toward owned products, such as product neglect (Studies 1A, 1B, and 4), risky

behaviors (Study 2), and faster consumption rates for consumable goods (Studies 3A and 3B).

Similar to prior research on product replacement and upgrade decisions (Chandler and Schwarz

2010; Jacoby, Berning, and Dietvorst 1977; Okada 2001, 2006), our studies are conducted with

owners of widely diffused consumer goods (e.g., phones, mugs, shampoo, glasses) recruited
 11

from a variety of populations (e.g., community participants, students, online respondents). We

complement the laboratory experiments with evidence from a real-word dataset of lost Apple

iPhones. When possible, we control for length of ownership, price paid, and remaining book

value of the current products to rule out any effects of age, cost, and depreciation on possible

carelessness. Importantly, demonstrating the upgrade effect controlling for these variables allows

us to rule out alternative accounts, such as a rational perspective on ownership (i.e., carelessness

is logical given an objective loss of market value of the product) or hedonic adaptation. 1 In the

general discussion, we also explain how our studies address a potential alternative explanation

involving shifts in consumers reference points.

The first three studies examine carelessness with mobile phones in the face of upgrade

options. Specifically, Study 1A examines an international dataset of about 3,000 lost Apple

iPhones and demonstrates that consumers are less likely to look for their lost phone when a new

model is available in the market. Study 1B conceptually replicates this finding in the lab with a

proxy for product neglect. Then, a follow-up study (online appendix) uses an ad featuring other

people engaging in seemingly carelessness practices with their phones to test whether people

acknowledge the existence of an upgrade effect in others. Next, Study 2 tests a behavioral

outcome of the upgrade effect in the lab. We endow participants with a mug to examine

ownership in a controlled manner; we find that the presence of an appealing upgrade opportunity

increases behavioral risk-taking with an owned product. Studies 3A and 3B extend the

investigation to the domain of perishable and consumable goods and demonstrate that people

consume their products at a faster rate when they are thinking of upgrading (but not just merely

1
Over extended periods of time, individuals exhibit an attenuation of positive affective reactions resulting from
ongoing ownership and usage of a particular good (Frederick and Loewenstein 1999; Wang, Novemsky, and Dhar
2009), which might potentially drive carelessness.
 12

substituting the current product). Moreover, these studies further examine the awareness of the

upgrade effect and show that product neglect may be non-deliberate. Finally, Study 4 delves into

the mechanisms underlying the upgrade effect. This study demonstrates the mediating role of

need for justification by manipulating the extent to which upgrading is framed as justifiable or

not and by examining individual-level differences in need for justification, as measured through

lay rationalism (Hsee et al. 2015).

----------------------------------
Insert Figure 1 about here
----------------------------------

RESEARCH DESIGN AND FINDINGS

Study 1: Carelessness with Mobile Phones in the Face of Upgrade Options

We begin our investigation of product neglect with field data on reported phone losses

(Study 1A). Next, in Study 1B, we conceptually replicate the finding in a lab setting, controlling

for a series of other potentially relevant factors, including acquisition methods and book value of

the phone. In addition, in Study 1B and in the follow-up study (online appendix), we begin

investigating the extent to which people are consciously aware of the effect of upgrades on their

behaviors, an inquiry we further continue in Studies 3A and 3B.

Reported Losses of Apple iPhones (Study 1A). We test our hypotheses using reported

phone losses as an ecologically valid measure of consumers carelessness in the marketplace.

Every Apple iPhone in the world has a unique code, known as the IMEI number. In case of loss,

consumers can go online on the IMEI Detective website (http://www.imeidetective.com/) to

report the loss and check if their phone has been found by others. In this context, not reporting a

loss on the IMEI Detective website is a proxy for product neglect. We propose that fewer people
 13

will go online and look for their phones using the IMEI number when a newer model is about to

be released or is already available for purchase (i.e., when the need to justify upgrading is high).

For example, we expect to observe fewer reported losses of the iPhone5S when the iPhone6 is

about to be released or is already on the market. In contrast, we expect that consumers will be

more eager to report the loss online using the IMEI number and to look for the lost device when

a newer model is not available yet. Specifically, we predict that the number of reported phone

losses will be negatively influenced by the availability of a newer model in the market, even

controlling for the total number of iPhones sold.

Datasets (Study 1A). We analyzed an international dataset of lost Apple iPhones provided

by the company IMEI Detective. The dataset included a total of 2,840 lost phones in 119

countries. Each loss is recorded on a specific date ranging from September 2010 to January

2015. The models included in the dataset were the Apple iPhone 4, 4S, 5, 5C, and 5S. Phones

were grouped by week and model. The focal outcome was the number of phones reported in a

given week on the IMEI Detective website. For each model, we calculated the number of days to

the release of the next model. For the official release dates of each model, we used information

available on the Apple iPhones official Wikipedia page (http://en.wikipedia.org/wiki/IPhone).

For example, the iPhone6 was officially launched on September 19, 2014. Thus, for a lost

iPhone5S reported on IMEI Detective on August 19, 2014 (i.e., a month before the release of the

iPhone6), the number of days to the release of the next model would be 31. Similarly, for a lost

iPhone5S reported on IMEI Detective on October 19, 2014 (i.e., a month after the release of the

iPhone6), the number of days to the release of the next model in this case would be -30.

Because the actual trend of reported phone losses on IMEI may also be dictated by the

actual number of phones sold during a specific period of time (e.g., we observe a decrease in the
 14

number of iPhone4 reported on IMEI Detective because fewer iPhone4 are being sold), we

control for total sales of Apple iPhones in the analysis. To this end, we purchased the

information on global monthly iPhone sales per model from Counterpoint Research, a market

research company specializing in technology. Counterpoint surveys distributors, mobile carriers

and retailers, component suppliers, and major manufacturers worldwide. According to

Counterpoint, this combination of sources allows triangulating iPhone sales with accuracy.

Results (Study 1A). We predicted the number of phones reported in a given week

(dependent variable) as a function of the release date of the next model. We used an OLS

regression with the number of days to the release of the next model and its square as the

independent variables. As a control variable, we also included the number of Apple iPhones sold

in any given month per model. Table 1 shows the regression results. As hypothesized, there were

significant linear ( = 1.15, t(567) = 9.14, p < .001) and quadratic effects ( = -.94, t(567) = -

10.44, p < .001) for the number of days to the release of the next model on number of phones

reported lost in a given week. These coefficients show an initial increase of reported losses

followed by a decrease before the release of the next iPhone. Importantly, these trends control

for the number of Apple iPhones sold, also a significant predictor ( = .45, t(567) = 6.08, p <

.001). For illustrative purposes, Figure 2 (top) shows the trends for the iPhone5 and iPhone5S.

As a robustness check, we also conducted similar regressions for each iPhone model

separately and found similar patterns of results. All the iPhone models show the same inverted

U-shape trend, indicating that the number of reported lost phones systematically decreased close

to the release date of the next model. In addition, we also performed a discontinuity analysis

considering the release of the next model on the market as the cutoff point. As seen in Figure 2

(bottom), we detect a statistically significant discontinuity in the number of reported losses

 15

before and after the release date of the next model (p = .001). The detailed results of the

discontinuity analysis are reported in the online appendix.

----------------------------------
Insert Table 1 and Figure 2 about here
----------------------------------

Study 1B. The objective of this study is to replicate the findings of Study 1A in the lab

with a proxy for product neglect, controlling for a series of other potentially relevant factors,

including acquisition methods and book value. We expect that the availability of the new desired

phone on the market and the interest in upgrading will lead to product neglect even when

controlling for other variables. Importantly, accounting for the real depreciation of phones over

time (which also captures the loss of value due to the launch of new models in the market) allows

us to show that product neglect in the presence of upgrades occurs over and beyond a purely

rational account (i.e., consumers are careless with their phones just because of the diminished

objective value of their device once a newer model is available in the market). To construct a

proxy for product neglect, we consider several types of carelessness indicators stemming from

the literatures on upgrades, product disposition, and self-defeating choices. Specifically, we look

at repair neglect, that is a measure of propensity to repair versus dispose of a product, directly

inspired by existing measures of disposal intentions and probability of replacement (Okada 2001;

Park and Mowen 2007; Shih and Jensen 2011; Trudel, Argo, and Meng 2016). Relatedly, we

also measure willingness to spend to repair (reverse coded), motivated by existing measures of

willingness to pay for upgrades (Okada 2006; Zhu, Chen, and Dasgupta 2008). Finally, we

consider intention to participate in a risky lottery with the product, inspired by existing

measures of self-defeating choices and irrational behaviors (Baumeister 2003). Together, these
 16

items reflect neglect of ones currently owned product (e.g., an aversion to repair or running the

risk to lose the current product).

To gain insight into the extent to which participants have conscious access about the

factors influencing their product behaviors, we ask participants to elaborate on the potential

factors contributing to how they treat their phones, following established methods for testing

consciousness of effects (Folkes, Martin, and Gupta 1993; Khan and Dhar 2006; Shiv, Carmon,

and Ariely 2005). If people consciously connect upgrading with carelessness, they should be able

to state this explicitly. In contrast, if they do not recognize whether or how their preferences are

affected by upgrades, they should not be able to accurately articulate the role of upgrades on how

they treat their belongings.

Method (Study 1B). We recruited 602 mobile phone owners (45% female; Mage = 33,

American) for a paid online survey through Amazon Mechanical Turk. First, participants

responded to a series of questions on their current phones: 1. What is the brand of your current

phone? (Open-ended answer); 2. How much did you pay for it, approximately? (Please

indicate the number of $); 3. Did you buy your current phone or did someone else buy it for

you? (I bought it personally; It was an upgrade from the service provider; Someone else bought

it for me (e.g., it is a present); I got it from my company); 4. Do you have a warranty covering

damage or loss on your current phone? (Yes, No); 5. How long have you had this specific

phone for approximately? (1 month, 3 months, 6 months, 9 months, 1 year, 1 year and a half, 2

years, 3 years). Based on mobile industry estimates of the depreciation of phones over time

(Banks 2014; Hsiao 2013; Priceconomics 2012), which notably account for the loss of value due

to the introduction of newer models, the answers for the length of ownership were converted to

the respective remaining value of the phone as a percentage of the original value (i.e., 1 month =
 17

96%, 3 months = 87.5%, 6 months = 75%, 9 months = 62.5%, 1 year = 50%, 1 year and a half =

40%, 2 years = 24%, 3 years = 25%).

Next, participants answered three questions assessing their level of carelessness for their

current phones. To reduce potential demand effects, the three questions were phrased in opposite

directions, with higher numbers indicating higher care for one item (willingness to pay for

repair) and, vice versa, higher numbers indicating higher carelessness for the other two items.

Specifically, participants answered the following questions: 1. If your current phone was

damaged, would you be more likely to repair it or to replace it? (1 = More likely to repair it, 7 =

More likely to replace it); 2. Imagine your current phone was damaged and you were not able to

use it, what is the most amount of money that you would pay to repair it rather than buying a

new one? (Please indicate the number of $); 3. Imagine there is lottery with a 25% probability

of winning a new phone of your choice. To participate you would have to give up your current

phone. Would you be interested in participating to this lottery? (1 = Not interested at all, 7 =

Extremely interested). Next, we asked participants to answer an open-ended question: Which

factors influence how you treat your phone? Finally, participants responded to a series of

questions about their future phones (we purposely asked these questions after collecting the

open-ended comments on potential factors influencing behaviors not to prime respondents with

thoughts of upgrades): 1. How much are you looking forward to replacing your current phone?

(1 = Not at all, 7 = Very much); 2. Will the new phone be included in some kind of upgrading

plan with your provider? (The new phone will be included in an upgrading plan (you will not

pay extra for it); The new phone will only partly be included in an upgrading plan (you will have

to pay part of it); The new phone will not be included in an upgrading plan (you will have to pay
 18

for all of it); Other (please specify)); 3. Is this model available on the market already? (Yes,

No, About to be released).

Preliminary Analyses (Study 1B). The average length of ownership for current phones

was 15 months and the average price paid was $195. In terms of brands, the most represented

brands in the sample were Apple (34%), Samsung (31%), and LG (11%). The majority of

participants purchased the device themselves (63%); about a fifth of participants received it as an

upgrade from the service provider (21%), and smaller percentages received it from others (15%)

or from their company (1%). The majority of respondents did not have a warranty on their phone

(74%). An analysis of answers regarding the purchase of the next phone revealed that about 43%

of participants indicated they would pay partially for their new phones, 31% would not pay for it

(the phone would be included in an upgrading plan), 23% would pay entirely for it, and 3%

mentioned other acquisition methods. About 58% of participants reported that the model they

wanted to buy next was already available on the market. The rest of participants reported that the

next model was either not available (39%) or upcoming (3%).

Results (Study 1B). First, we ran three regressions on each of the carelessness measures

separately with availability of the new phone on the market as the independent variable and

controlling for initial price paid and remaining book value of the phone. As fully reported in the

online appendix, we find the predicted significant results for availability of the next phone on

each dependent variable. Second, we computed a product neglect composite (with higher

numbers indicating higher product neglect) by first standardizing and then averaging the three

carelessness questions together (willingness to pay to repair is reverse coded in the composite).

We then analyzed responses using a linear regression with product neglect as the dependent

variable and the following independent variables: a dummy variable for availability of the new
 19

phone on the market (coded as 1 for availability of the new phone on the market and 0

otherwise), price paid, remaining value of the phone, interest in upgrading, and a series of

dummy variables for warranty, acquisition methods, and future upgrade methods.

Table 2 (top) shows the regression results. As expected, the analysis revealed the

following significant effects: a positive effect of availability of the new phone on the market ( =

.10, t(580) = 2.82, p = .005), a positive effect of willingness to upgrade ( = .19, t(580) = 5.23, p

< .001), a negative effect of paid price ( = -.39, t(580) = -10.60, p < .001), a negative effect of

book value ( = -.16, t(580) = -4.24, p < .001), and a positive effect for having to fully pay for

the next phone ( = .10, t(580) = 2.37, p = .018). 2 In sum, these results suggest that the

availability of the new desired phone on the market and the interest in upgrading lead to product

neglect even when controlling for a series of other relevant factors, such as price paid,

acquisition method, and depreciation. 3 Importantly, the critical finding on availability of the new

phone in the market, which indicates that the presence of an upgrade product increases current

product neglect, is robust to the presence or absence of the other variables (Table 2, bottom).

----------------------------------
Insert Table 2 about here
----------------------------------

To examine consumer awareness about the upgrade effect, we coded the content of the

open-ended question about the presumed factors influencing behaviors toward phones. As

represented in Figure 3, the content analysis revealed the following most recurring categories:

2
Though we were not predicting a significant effect for having to fully pay for the next phone, this result is
consistent with the notion that people who have to pay for the new device are more careless since the cost to upgrade
(and thus the need to justify the purchase) is higher for them than for those who dont have to pay to upgrade.
3
We acknowledge that people may use subjective book values that differ from objective book values. To partially
account for this heterogeneity in depreciation rates, we included two observed heterogeneity variables (income and
age) and their interactions with both availability of the next phone and interest in upgrading in the same regression
and found similarly significant results for the effect of availability of the next phone and interest in upgrading.
 20

37% of comments mentioned price paid or opportunity cost (e.g., How expensive it is,

Money. I would not want to spend anything on a new phone or to fix this one), 21% of

comments mentioned the importance of the phone and wanting the device to last (e.g., I want it

to last so I treat it pretty carefully, Its very important to me), 20% of comments mentioned

specific personality or environmental factors (e.g., Im generally careful with all my

belongings, If Im in a rush and whether Im at home or not), 9% of comments mentioned

features of the phone itself (e.g., Keeping it safe because the screen is fragile, Its durability

and build quality its got Gorilla Glass so its pretty sturdy), 7% of comments mentioned

length of ownership (How long I have had the phone), 3% of comments mentioned nothing

(e.g., None that I am aware of), 1% of comments mentioned upgrading to a new phone (If I

have an upgrade available, How close I am to getting a new one), and lastly, 2% of comments

listed other miscellaneous factors. In sum, only 1% of participants directly mentioned upgrading

as one of the presumed factors influencing how they treat their phones, suggesting that people

pay little attention to this factor. However, it is possible that participants may be reluctant to

openly admit the role of upgrades. We conducted a follow-up study to address this limitation.

----------------------------------
Insert Figure 3 about here
----------------------------------

Follow-up Study. To complement the findings of study 1B on the non-deliberateness of

the effect and try to address potential social desirability concerns, we conducted a follow-up

study (online appendix) using a real ad featuring consumers purposely mistreating their phones

to upgrade. This follow-up study demonstrates that, even when explicitly presented with other

consumers breaking products on purpose and using tricks to upgrade, the vast majority of people

do not acknowledge the effect of upgrades.

 21

Discussion. In conclusion, using real reported losses of Apple iPhones as a proxy for

consumers careless behavior, Study 1A shows that people are less likely to report the loss of an

owned phone when an upgrade model is available in the market, even when controlling for the

total number of phones sold. Study 1B replicates this finding, controlling for series of other

relevant factors, including the real depreciation of the product. Despite the fact that the

availability of the new desired phone on the market and the interest in upgrading lead to

carelessness, the content analysis of presumed factors influencing behaviors in Study 1B and the

follow-up study reveal that participants have little insight onto the effect of upgrades on how

they treat their own products.

Although these data are correlational, the detected pattern of results is very consistent

with the notion that the availability of product upgrades in the market induces product neglect.

All the following studies demonstrate the effect of upgrades through an experimental approach,

manipulating the presence of upgrades, and further delve into the deliberateness of the effect and

the underlying mechanism of need for justification.

Study 2: Behaviorally Endangering an Owned Product in the Presence of Preferred Options

In Study 2, we design a controlled ownership test in the laboratory by gifting all

participants with a free mug. We manipulate the physical availability of a preferred upgrade,

displaying better mugs (upgrade condition) to some participants but not to others (no-upgrade

condition). In this study, we test a behavioral measure of carelessnessthe extent to which

participants are willing to endanger an owned product by engaging in risky behavior with it. To

this end, participants played an incentive-compatible game of Jenga, with their mug placed atop

a tower of blocks. We predict that participants would choose to continue to play the game until
 22

the tower collapses (consequently dropping the mug) more often in the presence of preferred

product upgrades.

Method. We recruited 92 community and student respondents (42% female, Mage = 32)

who participated in a paid laboratory study at the Massachusetts Institute of Technology. After

receiving a basic ceramic mug (retail value = $1.00; picture in the online appendix) as a gift for

them to keep, all participants were brought, one by one, into a room with a table and wooden

Jenga blocks set up in a standard tower (54 blocks with 18 levels of three blocks each; picture in

the online appendix). The experimenter placed the mug on top of the tower. Participants were

randomly assigned to one of two conditions: no-upgrade versus upgrade. In the no-upgrade

conditions, participants only saw their basic ceramic mug. In the upgrade condition, participants

were shown three additional, nicer-looking mugs (average retail value = $9.67; picture in the

online appendix), which they inspected. We varied the appearance of upgrade mugs to account

for potential differences in taste among participants. To better simulate a real market situation

with upgrade options available for sale, we told participants, You may have the opportunity to

purchase one of these other mugs (whichever is your favorite) at a special price at the end of the

study.

All participants then played the game Jenga with the owned mug atop the block tower. In

the game, players remove individual blocks from a preassembled tower of relatively small,

wooden blocks while attempting to not collapse the tower. To encourage participants to keep

playing, we offered them extra payment for each block removed ($.05 each). Participants were

instructed to inform the research assistants when they wanted to stop the game. To emphasize the

risk of each decision to remove a block, participants were told they were not allowed to catch the

mug if it fell and that they would not receive a new mug if theirs broke. Thus, the primary
 23

dependent variable of interest was whether participants continued the game until the mug

dropped. Once the game was over, the research assistant recorded whether or not the mug fell

and the number of removed blocks. Then all participants were ushered into a computer room for

a follow-up questionnaire. In this second part of the study, all participants rated the basic mug by

answering: How much do you like the mug? (1 = Do not like at all, 9 = Like very much). In

addition, participants in the upgrade condition were asked to pick their preferred upgrade mug

(from the three upgrade mugs) and to rate it using the same liking question.

Results. Analysis of the liking rating revealed that the experimental manipulation of the

preferred upgrade product was successful. While liking of the basic mug did not differ across

conditions (Mupgrade = 4.76 vs. Mno-upgrade = 4.85, t(90) = .21, NS), participants in the upgrade

condition liked the upgrade mugs significantly more than the basic mug (Mupgrade mug = 6.04 vs.

Mbasic mug = 4.76, t(45) = 3.25, p = .002).

The analysis of the main dependent variablewhether or not participants dropped the

mugrevealed that participants in the upgrade condition dropped the mug during the game in

61% of the cases (28 dropping instances out of 46), in contrast to only 37% of cases (17

dropping instances out of 46) in the no-upgrade condition (2(1) = 5.26, p = .022). Thus,

exposure to mug upgrades led individuals to become more behaviorally careless with their

owned mugs. Out of all the dropped mugs, two were physically damaged, both in the upgrade

condition. Interestingly, there was no difference between conditions in the number of blocks

removed (Mupgrade = 12.83, SD = 6.68, vs. Mno-upgrade = 12.89, SD = 5.84). Because dropping was

more frequent in the upgrade condition, this result suggests that participants in the upgrade

condition took larger risks by removing more critical blocks during the game than did control

participants. Importantly, controlling for the number of blocks removed did not eliminate the
 24

difference in dropping behavior between conditions (B = 1.05, Wald = 5.37, p = .021). In

addition, although the option to purchase was simply part of the cover story and no upgrade

mugs were actually sold, at the end of the study, three participants in the upgrade condition

spontaneously expressed interest in purchasing the nicer-looking mugs.

Discussion. Study 2 demonstrates that individuals are more willing to endanger an owned

product in the physical presence of preferred upgrades. We find that participants are more likely

to drop an owned mug while playing a Jenga game when they have previously been exposed to a

set of better-looking mugs. A potential limitation of this study is the absence of a mere

replacement conditionthat is, a condition with other basic (rather than upgrade) mugs

present on the table. In the next study, we try to address this limitation by considering mere

product replacements. In addition, Studies 3A and 3B examine faster consumption rate as an

indicator of carelessness for non-durable goods. Importantly, these studies consider both hedonic

and functional products and further examine the awareness of the effect of upgrades on

carelessness tendencies.

Study 3: Faster Rate of Consumption for Non-Durable Goods in the Face of Upgrade Options

Prior research on the effects of supply on usage (Folkes, Martin, and Gupta 1993)

demonstrates that consumers become more parsimonious and systematically curb their

consumption rates of non-durable products as the amount of product available diminishes (e.g.,

consumers pour less shampoo if the bottle is half empty rather than full). Building on this

research and using the same experimental paradigm, we test the impact of product upgrades on

non-durable and consumable goods. Unlike the typical increase in conservation, we hypothesize
 25

that consumers wanting to upgrade to a new non-durable might consume their current product at

a faster pace than usual to hasten purchase of the newer product.

First, Study 3A replicates the effect of supply on usage (Folkes, Martin, and Gupta 1993)

and demonstrates the upgrade effect for four non-durable products with varying degree of

functionality. Study 3B again shows the upgrade effect for these same products in a purely

between-subjects design and also examines the effect of mere replacements. In addition, we seek

to further understand whether carelessness can occur without explicit awareness. To this end, we

ask participants to compare their intended level of consumption with typical usage. If

participants are unaware of the effect of upgrades on their behavior, they should report no

differences between their current usage and their typical usage.

Pretest. Using the same measures employed in previous research on hedonic versus

functional products (Dhar and Wertenbroch 2000; Okada 2005), we selected the following four

non-durable products, varying respectively from highly utilitarian to highly hedonic: toothpaste,

shampoo, laundry softener, and perfume (full pretest reported in the online appendix).

Method (Study 3A). We recruited 101 respondents (72% female, Mage = 23), who

participated in a paid laboratory study at Columbia Business School. Participants were randomly

assigned to one of three between-subjects conditions (full vs. quarter-full vs. quarter-full-

upgrade) for each of four non-durable products (shampoo, laundry softener, toothpaste, and

perfume). Thus, product is a within-subjects factor in this design. All participants read the same

study introduction: In this study we will ask you to consider a variety of commonly used

products, such as shampoo, laundry detergents, etc. Please read the following scenarios slowly

and carefully. Try to imagine what you would be thinking and feeling if the situation was

happening for real. Subsequently, participants in the full condition saw a picture of a full bottle
 26

and read, Imagine you own the bottle of ____ depicted below. Your bottle of ____ is full.

Participants in the quarter-full condition saw a picture of a quarter-full bottle and read, Imagine

you own the bottle of ____ depicted below. Your bottle of ____ contains about one fourth of the

product. Finally, for participants in the quarter-full-upgrade condition the stimuli were the same

as in the quarter-full condition, but the description further mentioned, You have already

decided that you want to upgrade to a different ____ (see online appendix for a visual

representation of the stimuli employed).

Following Folkes, Ingrid, and Gupta (1993)s procedures, we then measured participants

intended product usage (the dependent measure) through product-specific visual aids. For

example, in the case of shampoo, participants were asked, Imagine you were washing your hair

and the circles below represented the different amount of shampoo you may use for this

shower. How much shampoo would you pour on the palm of your hand? (From 1 to 7). In the

case of laundry detergent, participants were asked, Imagine you were doing laundry, how much

softener would you pour for this load in the measuring cup represented below? (From 1 to 14).

Finally, we measured how the selected amount compared to typical usage. For example,

in the case of shampoo, respondents were asked, Is the amount of shampoo you decided to use

more or less than what you typically pour when washing your hair? (1= Less than usual; 4 =

Same as usual; 7 = More than usual). 4 Consistent with our hypothesis that people do not engage

in carelessness deliberately, we do not expect to detect significant differences among conditions

on this measure.

Results (Study 3A). Because the four usage questions were on different scales, we first

standardized (z-scores) the answers for the analysis. We analyzed results using a mixed model

4
The sensitivity and understanding of this scale were pre-validated (test reported in the online appendix).
 27

ANOVA with condition (full vs. quarter-full vs. quarter-full-upgrade) as the between-subjects

factor and product-specific usage as the repeated measure. As shown in Figure 4 (top left), there

was a significant effect of condition (F(2, 97) = 7.11, p < .001). As a further check, we

conducted the analysis including gender and age as covariates and found similarly significant

results for the effect of condition. Pairwise comparisons revealed that participants willingness to

use the product was significantly lower in the quarter-full condition (Mquarter-full = -.30) than in

the full (Mfull = .16, p = .003, Bonferroni adjusted), thus replicating Folkes, Ingrid, and Guptas

(1993) finding on parsimony. Importantly, participants willingness to use the product was

significantly higher in the quarter-full upgrade condition (Mupgrade = .10) than in the quarter-full

condition (Mquarter-full = -.30, p = .007, Bonferroni adjusted), thus demonstrating the predicted

upgrade effect for non-durable goods. Finally, there was no significant difference between the

full and quarter-full-upgrade conditions. Although the interaction between product type and

condition was not significant (F(2, 97) = .95, NS), the rate of acceleration (i.e., the difference

between the usage levels in the quarter-full vs. quarter-full-upgrade conditions) was

directionally higher for the two hedonic products (Mhedonic = .54) than for the two utilitarian ones

(Mutilitarian = .23).

Next, we performed the same mixed model ANOVA on current usage as compared to

usual usage as the second dependent variable. Again, this measure tests whether participants are

aware about reporting consumption levels different than usual. As expected, the effect of

condition was non-significant (F(2, 97) = 1.25, NS) and, consistently, none of the pairwise

comparisons revealed a significant difference between conditions (Figure 4, top right).

Method (Study 3B). We recruited 302 respondents (66% female, Mage = 24), who

participated in a paid laboratory study at Columbia Business School and Harvard Business
 28

School. To attain more than 50 people per condition for each product, we complemented the lab

sample with 368 additional respondents (55% female; Mage = 38, American) recruited online

through Amazon Mechanical Turk, leading to a total sample size of 670 individuals (about 56

per condition). 5

Participants were randomly assigned to one of 12 conditions in a 3 (quarter-full-

replacement vs. quarter-full vs. quarter-full-upgrade) x 4 (shampoo vs. laundry softener vs.

toothpaste vs. perfume) between-subjects design. Instructions, stimuli, and questions were

identical to Study 3A, though this time we tested a quarter-full-replacement condition (instead of

the full condition). Specifically, the end of the description for participants in the quarter-full-

replacement condition read, You have already decided that you want to use the same ____ in

the future.

Results (Study 3B). After standardizing the four usage questions, we conducted an

ANOVA with intended product usage as the dependent measure and condition (quarter-full-

replacement, quarter-full, quarter-full-upgrade) and product type as factors. As predicted, the

only significant effect was for condition (F(2, 658) = 5.84, p = .003). As a further check, we also

conducted the same analysis including gender and age as covariates and found similarly

significant results for the effect of condition. Replicating the previously detected upgrade effect

for non-durable goods, planned contrasts revealed that participants willingness to use the

product was significantly higher in the quarter-full-upgrade condition (Mupgrade = .19) than in the

quarter-full condition (Mquarter-full = -.17; t(667) = 3.96, p < .001), as also seen in Figure 4

(bottom left). Usage ratings in the quarter-full-replacement condition fell in between the

quarter-full and quarter-full-upgrade conditions and were significantly lower than in the

5
We also conducted all the analyses on the two samples (lab and online) separately and found the same results.
 29

quarter-full-upgrade condition (Mupgrade = .19 vs. Mreplacement = -.04, t(667) = 2.40, p = .016) and

non-significantly higher than ratings in the quarter-full condition (Mreplacement = -.04 vs. Mquarter-full

= -.17, t(667) = 1.39, NS).

As in Study 3A, although the interaction between condition and product type was not

significant (F(6, 658) = 1.28, NS), the rate of acceleration (i.e., the difference between the usage

levels in the quarter-full vs. quarter-full-upgrade conditions) was again directionally higher for

the two hedonic products (Mhedonic = .44) than for the two utilitarian ones (Mutilitarian = .23).

Next, we performed the same ANOVA on current usage as compared to usual usage as

the second dependent variable. As expected, the effect of condition was non-significant (F(2,

657) = 1.16, NS) and, consistently, none of the planned contrasts revealed a significant

difference between conditions (Figure 4, bottom right).

----------------------------------
Insert Figure 4 about here
----------------------------------

Discussion. Studies 3A and 3B document carelessness in the domain of non-durable

goods by examining the consumption rate at which consumers are willing to use their products.

Results from both studies show that consumers accelerate the rate of consumption and become

less parsimonious with their products when the thought of an upgrade (and not just a mere

replacement) is salient. Importantly, these studies further demonstrate that carelessness can occur

without deliberateness. Despite the results supporting an upgrade effect, participants did not

acknowledge using more product than usual.

The next study delves into the underlying mechanisms of the upgrade effect by

investigating the mediating process of justification and examining a moderating effect of

individual differences in need for justification.

 30

Study 4: The Need to Justify Upgrades

The main objective of this study is to test whether product neglect in the presence of

upgrade cues stems from a need to justify the attainment of new, enhanced products. First, a pilot

study demonstrates the premise that product damage (even if just aesthetic damage) does indeed

operate as justification to upgrade (especially for lay rationalistic people) for both hedonic and

utilitarian products. Second, as recommended for evaluating psychological processes that are

easier to manipulate than to measure (Spencer, Zanna, and Fong 2005), such as need for

justification (see Okada 2005 for a comparable approach), we provide evidence in favor of our

proposed justification process through a moderation-of-process design by manipulating the

extent to which upgrading is framed as justifiable (i.e., low need for justification) or not

justifiable (i.e., high need for justification). Marketing research has highlighted how intended

future plans for current products after upgrading (e.g., donating or recycling) impact consumers

mental accounts and potentially mitigate their feelings of guilt (Dahl, Honea, and Manchanda

2003; Jacoby, Berning, and Dietvorst 1977; Okada 2001; Roster and Richins 2009). For

example, Okada (2001) finds that giving away a reusable product as a gift attenuates the mental

cost of the upgrade purchase. Accordingly, we predict that product neglect in the presence of

upgrades will be substantially weakened when a reason to attain the new product without feeling

guilty (i.e., giving the current one to someone else) is exogenously provided. Moreover, we

expect that people who chronically experience a high need to justify and rationalize their

decisions (i.e., individuals with high levels of lay rationalism) will be particularly sensitive to the

justification manipulation. Again, we consider two types of products, eyeglasses (functional) and
 31

sunglasses (hedonic) based on a pretest (online appendix). Finally, we manipulate upgrading

cues through a writing task about either a currently owned product or a potential upgrade.

Pilot Study (Justifying Product Upgrades in the Face of Damage). In an online study

with 958 American participants, we randomly assigned participants to one of 12 conditions in a 3

(intact vs. minor-damage vs. major-damage physical product state) x 2 (functional vs. hedonic

product) x 2 (self vs. other framing) between-subjects design. Due to space limits, we briefly

summarize the findings below and visually display the main results in Figure 5. The detailed

analyses are fully reported in the online appendix.

Regardless of the product type and of the framing, participants were more willing to

upgrade and felt more justified to upgrade when the product was damaged, even just lightly so,

as compared to when the product was intact (Figure 5, top). In a multi-categorical independent

variables mediation analysis, ratings of justifiability mediated willingness to upgrade (Figure 5,

middle). As predicted, we also found two significant interactions with lay rationalism.

Specifically, high lay rationalistic participants saw upgrading in the case of damage as more

justifiable and appropriate than low lay rationalistic participants (Figure 5, bottom). The overall

model of mediated moderation with willingness to upgrade as dependent variable, justifiability

as mediator, and lay rationalism as moderator was also significant, suggesting that the mediating

process of justifiability was particularly strong for lay rationalistic people who chronically

experience a high need to rationalize and justify their decisions.

----------------------------------
Insert Figure 5 about here
----------------------------------

Method. We recruited 325 owners of eyeglasses or sunglasses for a paid online survey

(63% female; Mage = 39, American) through Amazon Mechanical Turk. Owners of both
 32

eyeglasses and sunglasses (144 respondents) answered the survey for both products (and were

randomly re-assigned to one condition for the second product), thus leading to a total of 469

observations.

First, all participants were asked to visualize their pair of glasses: Think about a pair of

eyeglasses [sunglasses] that you own and use often. Once you have a good picture of this in your

mind, go to the next page. Next, we randomly assigned participants to one of three between-

subject conditions (no-upgrade vs. upgrade vs. justification-to-upgrade). In the no-upgrade

condition, participants read: Take a minute to write a couple of sentences in the space below

about the characteristics of the eyeglasses [sunglasses] you thought about. In the upgrade

condition, participants read an additional sentence mentioning an upgrade product: Imagine that

you recently saw a new and enhanced pair of eyeglasses [sunglasses] at the store. You would like

to buy this new pair. Take a minute to write a couple of sentences in the space below about the

characteristics of this new pair of eyeglasses [sunglasses]. Finally, in the justification-to-

upgrade condition, participants read the same statements as in the upgrade condition and, after

the writing task, they read an additional sentence mentioning an exogenous justification: Once

you will buy the new pair of eyeglasses [sunglasses], you are planning on giving your

current pair to someone else, such as a relative, a friend, or a person you care about.

After completing the writing task, participants answered the same questions on

carelessness (repair neglect), length of ownership, and price of the currently owned product from

Study 1B. The answers for the length of ownership were converted to the respective numbers of

months (i.e., 3 months = 3; 6 months = 6; 9 months = 9; ; 3 years = 36). Finally,

participants completed the six-item (Cronbachs = .77) lay rationalism scale (Hsee et al. 2015).
 33

Results. The average number of words written by participants was comparable across

conditions: 22 in the no-upgrade condition, 23 in the upgrade condition, and 22 in the

justification-to-upgrade condition. We first examine the effect of the upgrade manipulation

controlling for price, length of ownership, and product type, and then explore moderation with

lay rationalism. An ANCOVA on product neglect as a dependent measure, with condition and

product type as factors, their interaction, and price and length of ownership as covariates,

revealed significant effects only for condition (F(2, 460) = 5.97, p = .003), product type (F(1,

460) = 10.06, p = .002), 6 and price (F(1, 460) = 18.03, p < .001). 7 As hypothesized, planned

contrasts revealed that product neglect was higher in the upgrade condition (Mupgrade = 5.89) than

in the no-upgrade (Mno-upgrade = 5.49, F(1, 308) = 3.95, p = .048) and justification-to-upgrade

(Mjustification = 5.17, F(1, 305) = 12.9, p < .001) conditions. There was no significant difference

between the no-upgrade and justification-to-upgrade conditions (Mno-upgrade = 5.49 vs. Mjustification

= 5.17, F(1, 308) = 2.19, NS).

Moderation. Because again there was no interaction between conditions and product

type, we collapsed the results for the two products (i.e., eyeglasses and sunglasses). To explore

the moderating role of lay rationalism, responses were analyzed using an ANCOVA with product

neglect as the dependent variable, a dummy variable for no-upgrade condition and a dummy

variable for justification-to-upgrade condition as factors, the lay rationalism scale as covariate,

and two-way interaction terms between each of the dummies and the lay rationalism scale. The

analysis revealed significant effects only for the no-upgrade dummy (F(1, 463) = 3.91, p = .048)

and, most importantly, for the interaction between justification-to-upgrade condition and lay

6
On average, participants were more careless with the hedonic product than the functional one (Msunglasses = 5.93 vs.
Meyeglasses = 5.04, however, there was no significant interaction with the upgrade manipulation (F(2, 460) = .04, NS).
7
We also conducted the same analysis considering only respondents who completed the survey for one product only
and found again a significant effect of condition (p = .001).
 34

rationalism (F(1, 463) = 4.19, p = .041). To further explore the significant interaction between

justification-to-upgrade condition and lay rationalism, we applied the Johnson-Neyman

procedure (Spiller et al. 2013). We find a significant effect of condition on product neglect at and

above 4.56 of the lay rationalism scale (B = -.48, SE = .25, t(307) = 1.97, p = .05), but not for

any level below this point (figure in online appendix). Thus, for highly lay rationalistic people,

the upgrade cue produced greater product neglect than when the upgrade cue was combined with

an exogenous justification to upgrade (i.e., passing the current product to someone else).

Discussion. In conclusion, this study provides evidence in favor of our proposed

justification process and demonstrates that product neglect toward owned products is

significantly mitigated in the presence of an exogenous justification to upgrade without

appearing wasteful, such as donating the product to others. As hypothesized, this effect is

particularly accentuated for those people who chronically experience a high need to rationalize

their decisions (i.e., individuals with high levels of lay rationalism).

GENERAL DISCUSSION

The present research documents the novel phenomenon of consumer carelessness toward

owned products. The empirical evidence from both field data and laboratory studies demonstrate

that consumers become careless with their current products when in the presence of desired

product upgrades. At a process level, our investigation reveals that some forms of consumer

carelessness are driven by the need to justify attainment of the upgrade product. The desire to

acquire new products can motivate changes in how consumers treat the products they own,
 35

changes that can occur without their knowledge and that may appear undesirable from a rational

perspective on ownership.

Our theoretical framework (Figure 1) and findings contribute to a diverse set of

literatures, including product upgrades (Bayus 1988, 1991; Mukherjee and Hoyer 2001; Okada

2001, 2006), ownership (Carmon and Ariely 2000; Strahilevitz and Loewenstein 1998; Thaler

1980), consumption rates (Ailawadi and Neslin 1998; Chandon and Wansink 2002), and mental

accounting (Gourville and Soman 1998; Heath and Fennema 1996). In addition, this work

advances prior research on cognitive processes that intensify consumers reluctance to dispose of

their possessions (Brough and Isaac 2012; Haws et al. 2012; Lastovicka and Fernandez 2005;

Price, Arnould, and Curasi 2000; Roster and Richins 2009; Trudel, Argo, and Meng 2016) by

focusing on how the connection with owned products can actually be weakened through

exposure to upgrades. Finally, our findings inform previous research on justification processes,

motivated reasoning, and guilt (Dahl, Honea, and Manchanda 2003; Hsee 1995; Hsee et al. 2015;

Kivetz and Simonson 2002; Okada 2005; Shafir, Simonson, and Tversky 1993; Strahilevitz and

Myers 1998) by demonstrating that carelessness toward owned products in the presence of

upgrades is driven by a need to justify attainment of the upgrade.

Directions for Future Research

The present research provides the seed for further investigation on the topic of

carelessness. We consider future research directions pertaining to products characteristics,

ownership processes, and carelessness. Finally, we discuss a potential alternative explanation.

Product Characteristics. With respect to products, a fruitful direction for further research

pertains to the study of product domains that consumers perceive to be particularly meaningful
 36

and symbolic. Indeed, more than a fifth of participants in Study 1B mentioned the importance of

the phone as the main factor influencing how they treat their device. Products are perceived as

symbolic to the extent that they are able to express the identity of the individual using them

(Berger and Heath 2007; Escalas and Bettman 2005). Hence, future research could examine

whether the perceived symbolism of products moderates the upgrade effect, such that consumers

may exhibit carelessness primarily for those products that are less meaningful to the self.

Certainly, consumers are more likely to recycle (rather than dispose of) a product if it is linked to

their identity because placing an identity-linked product in the trash is symbolically similar to

trashing a part of the self, a situation consumers want to avoid (Trudel, Argo, and Meng 2016).

In addition, the current research could be extended to examine the specific kinds of

upgrades that are more likely to generate careless tendencies. For example, how new or enhanced

does the product upgrade need to be to activate carelessness? Prior work has classified upgrades

depending on the degree of similarity and novelty relative to the original product (Bertini, Ofek,

and Ariely 2009; Okada 2006). Future research could directly manipulate such dimensions and

investigate their impact on carelessness.

Ownership Processes. With respect to process-level factors involved in ownership,

questions remain about the psychological power of upgrades for affecting both ownership biases

and various forms of product care. Consumers become attached to products both through

ownership (Kahneman, Knetsch, and Thaler 1990; Thaler 1980) and through the effort they

invest in acquiring or constructing those products (Norton, Mochon, and Ariely 2012). The

presence of preferred upgrades may influence the degree to which factors such as effort elicit

increased valuation of belongings; conversely, the effort invested in current products may

insulate consumers against the lure of upgrades. Furthermore, research on ownership processes
 37

suggests that consumers anthropomorphize and see human schema in products (Aggarwal and

McGill 2007). Future work could test whether the level of product anthropomorphism inhibits

the emergence of product neglect even when in the presence of desired upgrades.

Carelessness. Finally, careless behavior itself is a topic that has received relatively little

attention in the ownership literature. Future work addressing the scope of this behavior would

thus be quite useful. For instance, under what circumstances do consumers engage in

carelessness through passive omission behaviors (e.g., carrying a phone without a protective

case) or through more active commission acts (e.g., causing damage by dropping or handling the

product roughly)? Because consumers have a strong aversion to appearing wasteful (Arkes

1996), we would expect that omission is more common than commission behavior. This

prediction is consistent with research on the omission bias, which suggests that harmful

omissions are perceived as less blameworthy than corresponding commissions (Spranca, Minsk,

and Baron 1991). Omission may help to remove the stigma associated with acting recklessly

while at the same time promoting individual justification motives.

In addition, it would be interesting to explore the signaling dimension of carelessness and

contribute to recent research examining alternative signals of status (Bellezza, Gino, and Keinan

2014; Bellezza, Paharia, and Keinan 2017). Because conspicuous consumption is tightly linked

to the idea of resource waste (Veblen 1899/2007), destroying products or consuming them

lavishly may sometimes operate as a signal of status in the eyes of others.

Alternative Explanation. An explanation for the upgrade effect may involve shifts in

consumers reference points. In the context of product upgrades, a reference point account would

suggest that careless outcomes stem from unmotivated changes to product evaluation as a result

of considering upgrades. That is, instead of evaluating a current product in absolute terms,
 38

consumers might evaluate the current product in comparison to the upgrade product, resulting in

lowered evaluations and careless behavior. While reference points are interesting and relevant

aspects to consider, this account could not fully explain our findings. In Study 2, the presence of

upgrade mugs did not influence liking of the currently owned good. The observed non-

significant difference between conditions in liking of the basic mugs rules out the potential

alternative reference point explanation for subsequent game-playing behavior. In addition, Study

4 also helps to address the alternative reference point account. If thinking of upgrade options

lowered the perceived valuation of the currently owned goods relative to these more appealing

products, this mechanism should theoretically apply to all the scenarios featuring the upgrade,

independent of a justification element. Yet, the significant difference between the two upgrade

conditions in Study 4 is inconsistent with the reference point explanation.

Managerial Implications

This research has clear relevance for actual consumer behavior, as emphasized by our

investigation of real owners of a wide array of goods. How might marketing managers use these

findings to better meet consumer needs and desires? Should marketers promote careless behavior

or attempt to prevent it? The next paragraphs discuss the implications of our findings for framing

promotions that address consumer-product relationships and for marketing of products designed

to fail, a strategy referred to as planned obsolescence (Bulow 1986).

Marketers face the complex dilemma of wanting to encourage product and brand

attachment in order to drive choice while, at the same time, persuading owners not to hang on to

owned products for too long. For example, IKEAs television commercial Death of a Lamp

encourages consumers to disregard their sadness when replacing an old but functioning lamp to
 39

which they feel attached with a newer one: Many of you are feeling bad for this lamp; that is

because youre crazy! It has no feelings, and the new one is much better! Emphasizing similar

shifts in focus toward function over sentiment may be especially useful when done in

comparison to upgrades. Making explicit comparisons about upgrade features or style could

simplify the upgrade justification process, especially in the case of hedonic goods (Okada 2005)

and ambiguous product features (Hsee 1995).

As noted earlier, several brands (e.g., Virgin Mobile, eBay, T-Mobile) recently have

taken the more direct approach of promoting careless behavior with owned products. Although

commercials following this approach tend to be humorous and considered unrealistic (as found

in the follow-up study in the online appendix), their message may sway owners searching for

upgrade justifications to behave similarly. Of course, encouraging owners to donate or sell their

products may produce similar outcomes with less waste and loss of welfare. For example,

H&Ms long live fashion campaign encourages consumers to dispose of their old clothes in

donation bins in their stores so that the company may reduce waste and give old products a new

life. As seen in Study 4, this is an effective strategy for mitigating guilt when buying new

products. To consider another example in the domain of non-durable goods, Vichy (LOral) in

Italy has been running the toss your old foundation campaign, offering a substantial discount

on new Vichy foundations to those consumers trading in their unfinished products.

This research also has implications for managers and designers interested in capitalizing

on product use cycles. The principle of planned obsolescence is one method of purposely

creating a limited lifespan for products (Bulow 1986). This strategy can, however, create

backlash from consumers (Packard 1960). For example, Apples practice of remotely distributing

software updates that owners can directly (and irreversibly) download on their devices has
 40

negatively been labeled a trap to dampen the performance of older products and thus promote

acquisition of the latest devices (Rampell 2013). Since new operating systems require more

resources to function properly, they typically diminish product performance of older devices (as

in the case of installing the latest operating system onto an older iPhone model). However, our

findings suggest that planned obsolescence might actually be beneficial for upgrade-minded

consumers by making it easier for them to damage or detect functional flaws in owned

products.

In conclusion, we hope our work is a first step toward providing insights into the novel

phenomenon of motivated consumer carelessness toward owned belongings. Contrary to the

prevailing notion that consumers highly value and care for their possessions, the current research

demonstrates that consumers exhibit cavalier behavior toward owned products when in the

presence of appealing product upgrades.

 41

REFERENCES
Aggarwal, Pankaj and Ann L. McGill (2007), Is That Car Smiling at Me? Schema Congruity as
a Basis for Evaluating Anthropomorphized Products, Journal of Consumer Research, 34
(4), 46879.
Ailawadi, Kusum L. and Scott A. Neslin (1998), The Effect of Promotion on Consumption:
Buying More and Consuming It Faster, Journal of Marketing Research, 35 (3), 39098.
Arkes, Harl (1996), The Psychology of Waste, Journal of Behavioral Decision Making, 9
(January), 21324.
Banks, Roland (2014), The Best Time to Buy a Flagship Smartphone?, Mobile Industry
Review.
Baumeister, Roy F. (2003), The Psychology of Irrationality: Why People Make Foolish, Self-
defeating choices, The psychology of economic decisions, 1, 316.
and Diane M. Tice (1984), Role of Self-Presentation and Choice in Cognitive
Dissonance Under Forced Compliance: Necessary or Sufficient Causes?, Journal of
Personality and Social Psychology, 46 (1), 513.
Bayus, Barry L. (1988), Accelerating the Durable Replacement Cycle with Marketing Mix
Variables, Journal of Product Innovation Management, 5 (September), 21626.
(1991), The Consumer Durable Replacement Buyer, Journal of Marketing, 55 (1), 42
51.
Bellezza, Silvia, Francesca Gino, and Anat Keinan (2014), The Red Sneakers Effect: Inferring
Status and Competence from Signals of Nonconformity, Journal of Consumer Research,
41 (1), 3554.
, Neeru Paharia, and Keinan (2017), Conspicuous Consumption of Time: When
Busyness and Lack of Leisure Time Become a Status Symbol (forthcoming), Journal of
Consumer Research.
Berger, Jonah and Chip Heath (2007), Where Consumers Diverge from Others: Identity
Signaling and Product Domains, Journal of Consumer Research, 34 (8), 12135.
Bertini, Marco, Elie Ofek, and Dan Ariely (2009), The Impact of Add-On Features on
Consumer Product Evaluations, Journal of Consumer Research, 36 (1), 1728.
Bolton, Lisa E. and Joseph W. Alba (2012), When Less is More: Consumer Aversion to Unused
Utility, Journal of Consumer Psychology, 22 (3), 36983.
Brough, Aaron R. and Mathew S. Isaac (2012), Finding a Home for Products We Love: How
Buyer Usage Intent Affects the Pricing of Used Goods, Journal of Marketing, 76 (July),
7891.
Bulow, Jeremy (1986), An Economic Theory of Planned Obsolescence, Quarterly Journal of
Economics, 4 (101), 72949.
Carmon, Ziv and Dan Ariely (2000), Focusing on the Forgone: How Value Can Appear So
Different to Buyers and Sellers, Journal of Consumer Research, 27 (December), 36070.
Chandler, Jesse and Norbert Schwarz (2010), Use Does Not Wear Ragged the Fabric of
 42

Friendship: Thinking of Objects as Alive Makes People Less Willing to Replace Them,
Journal of Consumer Psychology, 20 (2), 13845.
Chandon, Pierre and Brian Wansink (2002), When Are Stockpiled Products Consumed Faster?
A Convenience-Salience Framework of Postpurchase Consumption Incidence and
Quantity, Journal of Marketing Research, 39 (3), 32135.
Cripps, John D. and Robert J. Meyer (1994), Heuristics and Biases in Timing the Replacement
of Durable Products, Journal of Consumer Research, 21 (2), 30418.
Dahl, Darren W., Heather Honea, and Rajesh V. Manchanda (2003), The Nature of Self-
Reported Guilt in Consumption Contexts, Marketing Letters, 14 (3), 15971.
Dhar, Ravi and Klaus Wertenbroch (2000), Consumer Choice Between Hedonic and Utilitarian
Goods, Journal of Marketing Research, 37 (1), 6071.
Ehrich, Kristine R. and Julie R. Irwin (2005), Willful Ignorance in the Request for Product
Attribute Information, Journal of Marketing Research, 42 (3), 26677.
Escalas, Jennifer E. and James R. Bettman (2005), Self-Construal, Reference Groups, and
Brand Meaning, Journal of Consumer Research, 32, 37889.
Fernandez, Viviana P. (2001), Observable and Unobservable Determinants of Replacement of
Home Appliances, Energy Economics, 23 (3), 30523.
Folkes, Valerie S., Ingrid M. Martin, and Kamal Gupta (1993), When to Say When: Effects of
Supply on Usage, Journal of Consumer Research, 20 (3), 46777.
Frederick, Shane and George Loewenstein (1999), Hedonic adaptation., in Well-being: The
Foundations of Hedonic Psychology, Russell Sage Foundation, ed., New York, NY, 302
29.
Gourville, John T. and Dilip Soman (1998), Payment Depreciation: the Behavioral Effects of
Temporally Separating Payments From Consumption, Journal of Consumer Research, 25
(2), 16074.
Grewal, Rajdeep, Raj Mehta, and Frank R. Kardes (2004), The Timing of Repeat Purchases of
Consumer Durable Goods: The Role of Functional Bases of Consumer Attitudes, Journal
of Marketing Research, 41 (1), 10115.
Haws, Kelly L., Rebecca Walker Naylor, Robin A. Coulter, and William O. Bearden (2012),
Keeping It All Without Being Buried Alive: Understanding Product Retention Tendency,
Journal of Consumer Psychology, 22 (2), 22436.
Heath, C. and M. Fennema (1996), Mental Depreciation and Marginal Decision Making,
Organizational Behavior and Human Decision Processes, 68 (2), 95108.
Hsee, Christopher K. (1995), Elastic Justification: How Tempting but Task-Irrelevant Factors
Influence Decisions, Organizational Behavior and Human Decision Processes, 62 (3),
33037.
, Yang Yang, Xingshan Zheng, and Hanwei Wang (2015), Lay Rationalism: Individual
Differences in Using Reason Versus Feelings to Guide Decisions, Journal of Marketing
Research, LII (February), 13446.
Hsiao, Aron (2013), How Quickly Do Smartphones Lose Their Value?, Terapeak Trends.
 43

Jacoby, Jacob, Carol K. Berning, and Thomas F. Dietvorst (1977), What about Disposition?,
Journal of Marketing, 41 (2), 2228.
Kahneman, Daniel, Jack L. Knetsch, and Richard H. Thaler (1990), Experimental tests of the
endowment effect and the Coase Theorem, Journal of Political Economy, 99, 13251348.
Keinan, Anat, Ran Kivetz, and Oded Netzer (2016), The Functional Alibi, Journal of the
Association for Consumer Research, 1 (4).
Khan, Uzma and Ravi Dhar (2006), Licensing Effect in Consumer Choice, Journal of
Marketing Research, 43 (2), 25966.
Kivetz, Ran and Itamar Simonson (2002), Earning Toward the Right to Indulge: Effort
Determinant of Customer Rewards, Journal of Marketing Research, 39 (2), 15570.
Kunda, Ziva (1990), The Case for Motivated Reasoning, Psychological Bulletin, 108
(November), 48098.
Lastovicka, John L. and Karen V. Fernandez (2005), Three Paths to Disposition: The
Movement of Meaningful Possessions to Strangers, Journal of Consumer Research, 31 (4),
81323.
Linder, Darwyn, Joel E. Cooper, and Edward E. Jones (1967), Decision Freedom as a
Determinant of the Role of Incentive Magnitude in Attitude Change, Journal of
Personality and Social Psychology, 6 (3), 245254.
Mukherjee, Ashesh and Wayne D. Hoyer (2001), The Effect of Novel Attributes on Product
Evaluation, Journal of Consumer Research, 28 (3), 46272.
Norton, Michael I., Daniel Mochon, and Dan Ariely (2012), The IKEA Effect: When Labor
Leads to Love, Journal of Consumer Psychology, (22), 45360.
Okada, Erica Mina (2001), Trade-ins, Mental Accounting, and Product Replacement
Decisions, Journal of Consumer Research, 27 (4), 43346.
(2005), Justification Effects on Consumer Choice of Hedonic and Utilitarian Goods,
Journal of Marketing Research, 42 (1), 4353.
(2006), Upgrades and New Purchases, Journal of Marketing, 70 (4), 92102.
Packard, Vance (1960), The Waste Makers, (D. McKay, ed.), New York.
Paharia, Neeru, Kathleen D. Vohs, and Rohit Deshpand (2013), Sweatshop Labor Is Wrong
Unless the Shoes Are Cute: Cognition Can Both Help and Hurt Moral Motivated
Reasoning., Organizational Behavior and Human Decision Processes, 121 (1), 8188.
Park, Sojin and John C. Mowen (2007), Replacement Purchase Decisions: On the Effects of
Trade-ins, Hedonic versus Utilitarian usage goal, and Tightwadism, Journal of Consumer
Behavior, 6, 12331.
Pham, Michel Tuan (1998), Representativeness, Relevance, and the Use of Feelings in Decision
Making, Journal of Consumer Research, 25 (2), 14459.
Prelec, Drazen and George Loewenstein (1998), The Red and the Black: Mental Accounting of
Savings and Debt, Marketing Science, 17 (1), 428.
Price, Linda L., Eric J. Arnould, and Carolyn Folkman Curasi (2000), Older Consumers
 44

Disposition of Special Possessions, Journal of Consumer Research, 27 (2), 179201.

Priceconomics (2012), Your Phone Loses Value Pretty Fast (Unless Its an iPhone).
Rampell, Catherine (2013), Cracking the Apple Trap, The New York Times.
Roster, Catherine A. and Marsha L. Richins (2009), Ambivalence and Attitudes in Consumer
Replacement Decisions, Journal of Consumer Psychology, 19 (1), 4861.
Shafir, Eldar, Itamar Simonson, and Amos Tversky (1993), Reason-Based Choice, Cognition,
49 (12), 1136.
Shih, Eric and Hope Jensen (2011), To Justify or Not to Justify: The Role of Anticipated Regret
on Consumers Decisions to Upgrade Technological Innovations, Journal of Retailing, 87
(2), 24251.
Shiv, Baba, Ziv Carmon, and Dan Ariely (2005), Placebo Effects of Marketing Actions:
Consumers May Get What They Pay For, Journal of Marketing Research, 42 (4), 38393.
Spencer, Steven J., Mark P. Zanna, and Geoffrey T. Fong (2005), Establishing a Causal Chain:
Why Experiments Are Often More Effective Than Mediational Analyses in Examining
Psychological Processes, Journal of Personality and Social Psychology, 89 (6), 84551.
Spiller, Stephen A., Gavan J. Fitzsimons, John G. Jr Lynch, and Gary H. McClelland (2013),
Spotlights, Floodlights, and the Magic Number Zero: Simple Effects Tests in Moderated
Regression, Journal of Marketing Research, 2 (50), 27788.
Spranca, Mark, Elisa Minsk, and Jonathan Baron (1991), Omission and Commission in
Judgment and Choice, Journal of Experimental Social Psychology, 27 (1), 76105.
Strahilevitz, Michal A. and George Loewenstein (1998), The Effect of Ownership History on
the Valuation of Objects, Journal of Consumer Research, 25 (December), 27689.
Strahilevitz, Michal and John G. Myers (1998), Donations to Charity as Purchase Incentives:
How Well They Work May Depend on What You Are Trying to Sell, Journal of
Consumer Research, 24 (4), 43446.
Thaler, Richard H. (1980), Toward a Positive Theory of Consumer Choice, Journal of
Economic Behavior and Organization, (1), 3690.
Trudel, Remi, Jennifer J. Argo, and Matthew D. Meng (2016), The Recycled Self: Consumers
Disposal Decisions of Identity-Linked Products, Journal of Consumer Research, (2013).
Urban, Glen L. and John R. Hauser (1993), Design and Marketing of New Products, Upper
Saddle River, NJ: Prentice Hall.
Veblen, Thorstein (2007), The Theory of the Leisure Class, New York: Oxford University Press.
Wang, Jing, Nathan Novemsky, and Ravi Dhar (2009), Anticipating Adaptation to Products,
Journal of Consumer Research, 36 (2), 14959.
Xu, Jing and Norbert Schwarz (2009), Do We Really Need a Reason to Indulge?, Journal of
Marketing Research, 46 (1), 2536.
Zhu, Rui (Juliet), Xinlei (Jack) Chen, and Srabana Dasgupta (2008), Can Trade-Ins Hurt You?
Exploring the Effect of a Trade-In on Consumers Willingness to Pay for a New Product,
Journal of Marketing Research, XLV (April), 15970.
 45

TABLE 1: PREDICTORS OF REPORTED LOSSES OF APPLE IPHONES ON THE IMEI

DETECTIVE WEBSITE (STUDY 1A)

Unstandardized Standardized
Predictors Coefficient Std. Error Coefficient t-value p-value

Intercept 3.159 0.385 8.206 0.001

Days to the release of the next model 0.010 0.001 1.148 9.142 0.001
2
Days to the release of the next model -0.001 0.001 -0.939 -10.442 0.001
iPhone sales monthly 0.539 0.089 0.451 6.078 0.001

2
R = 0.162
 46

TABLE 2 (TOP AND BOTTOM): PREDICTORS OF PRODUCT NEGLECT FOR MOBILE

PHONES (STUDY 1B)

Unstandardized Standardized
Predictors Coefficient Std. Error Coefficient t-value p-value Tolerance

Intercept 0.026 0.148 0.175 0.861

Market availability of next model 0.148 0.052 0.103 2.822 0.005 0.845
Interest in upgrading 0.071 0.014 0.187 5.231 0.000 0.888
Price paid -0.001 0.000 -0.393 -10.602 0.000 0.825
Book value (% of original value) -0.005 0.001 -0.157 -4.240 0.000 0.825
Dummy - Warranty on current phone -0.066 0.059 -0.042 -1.130 0.259 0.835
Dummy - This phone was an upgrade -0.098 0.066 -0.056 -1.487 0.138 0.794
Dummy - This phone was a present -0.009 0.069 -0.005 -0.133 0.894 0.898
Dummy - This phone was given by my company 0.264 0.210 -0.043 -1.259 0.208 0.957
Dummy - Next phone will be partially paid by me -0.044 0.067 -0.026 -0.658 0.511 0.707
Dummy - Next phone will be fully paid by me 0.144 0.061 0.101 2.368 0.018 0.624
Dummy - Next phone other acquisition method 0.156 0.142 0.039 1.102 0.271 0.903

R 2 = 0.342
Bold indicates predictor is significant (p-value column < .05)

Predictors Standardized Coefficient

Market availability of next model 0.295 0.253 0.158 0.122

Interest in upgrading 0.230 0.214 0.174
Price paid -0.407 -0.395
Book value (% of original value) -0.157
2
R change 0.087 0.051 0.156 0.021
2
Bold indicates betas and R changes are significant (all p-values < .001)
 47

FIGURE 1: THEORETICAL FRAMEWORK

 48

FIGURE 2: APPLE IPHONE 5 AND 5S REPORTED LOSSES ON THE IMEI DETECTIVE

WEBSITE (TOP) AND DISCONTINUITY ANALYSIS (BOTTOM) (STUDY 1A)

Note. Number of Apple iPhones 5 and 5S reported lost on IMEI Detective over time. The vertical lines represent the
introduction dates of the iPhone5S (September 20, 2013) and the iPhone6 (September 19, 2014).

Note. Release of the next model on the market is the cutoff point.
 49

FIGURE 3: CONTENT ANALYSIS OF FACTORS INFLUIENCING PRODUCTS

TREATMENT (STUDY 1B)

Which factors influence how you treat your phone?

Price paid, opportunity cost 37%

Importance of the phone, wanting it to last 21%

Personality, situation 20%

Phone itself (durability, functionality) 9%

Lengh of ownership 7%

Nothing 3%

Other 2%

Upgrade 1%

0% 10% 20% 30% 40%

 50

FIGURE 4: INTENDED USAGE ACROSS PRODUCTS (STUDIES 3A, TOP, AND 3B,
BOTTOM)
 51

FIGURE 5: PILOT STUDY MAIN EFFECTS OF PHYSICAL PRODUCT STATE (TOP),

MEDIATION (MIDDLE), AND INTERACTION WITH LAY RATIONALISM (BOTTOM)
 52

ONLINE PAPER APPENDIX

Pilot Study on Carelessness (Introduction)

The objective of this pilot study was to confirm that carelessness for durable and non-

durable products is associated with product neglect (e.g., neglecting repair, not looking for the

product in case of loss,), risky behaviors (e.g., endangering the product), and faster consumption

rates (e.g., pouring more product than usual).

We recruited 100 respondents (50% female; Mage = 36, American) for a paid online

survey through Amazon Mechanical Turk. All participants were told: Imagine the following

two people and answer the questions that follow. Person A appears to be more careless than

average with her products. Person B appears to be less careless than average with her products.

Participants were then asked five questions regarding persons A and B. To reduce potential

demand effects, questions were phrased in opposite directions, with higher numbers indicating

higher care for the question on willingness to look for the lost phone and, vice versa, higher

numbers indicating higher carelessness for the other questions. Specifically, participants

answered the following questions: 1. In case of loss of a product (e.g., a phone), how likely is

this person to check online whether someone found the product and possibly returned it? (1 =

Not likely at all, 7 = Extremely likely); 2. How likely is this person to endanger her products?

(1 = Not likely at all, 7 = Extremely likely); 3. In case of product damage, would this person be

more likely to repair or to replace her products? (1 = More likely to repair, 7 = More likely to

replace); 4. When using consumable products (e.g., shampoo, perfume), how quickly does she

use them? (1 = Very slow, 7 = Very fast); 5. When using consumable products (e.g., shampoo,
 53

perfume), how much does she pour each time? (1 = Very little amount of product, 7 = Very big

amount of product).

Paired sample t-tests revealed that the more careless person was seen as less likely to

look for lost products (MA = 3.79 vs. MB = 5.4, t(97) = 5.84, p < .001), more likely to endanger

her products (MA = 5.47 vs. MB = 2.97, t(98) = 8.69, p < .001), and more likely to dispose of,

rather than maintain, her products (MA = 4.98 vs. MB = 3.66, t(98) = 4.17, p < .001). In addition,

participants inferred that the more careless person would also consume her non-durable products

in a faster fashion (MA = 5.45 vs. MB = 3.16, t(98) = 9.42, p < .001) and use more of them each

time (MA = 5.47 vs. MB = 3.21, t(97) = 9.81, p < .001).

In sum, these results indicate that people associate carelessness for durable and non-

durable products with product neglect, risky behaviors, and faster consumption rates. We use

these different behaviors throughout our studies as ways to measure carelessness.

Discontinuity Analysis (Study 1A)

To complement the analysis of Study 1A, we performed a discontinuity analysis

considering the release of the next model on the market as the cutoff point. We used an OLS

regression to predict the number of phones reported in a given week (dependent variable) as a

function of the release date of the next model (same continuous variable as in the regression

analysis reported in the paper), the release of the next phone (dummy variable coded as 0 if the

next model has not been released yet and 1 if the next model has already been released), and the

interaction between these two variables. The table below shows the regression results. As

hypothesized, we detect a statistically significant discontinuity in the number of reported losses

 54

before and after the release date of the next model ( = -0.32, t(303) = -3.21, p = .001). As a

further check, we also conducted the same analysis controlling for the number of Apple iPhones

sold worldwide and found again significant effects for all the independent variables.

TABLE: PREDICTORS OF REPORTED LOSSES OF APPLE IPHONES ON THE IMEI

DETECTIVE WEBSITE, DISCONTINUITY ANALYSIS (STUDY 1A)

Unstandardized Standardized
Predictors Coefficient Std. Error Coefficient t-value p-value

Intercept 2.231 0.149 14.971 0.001

Days to the release of the next model (IV 1 ) 0.003 0.001 1.176 3.594 0.001
Release of the next model (IV 2 ) -0.574 0.179 -0.320 -3.209 0.001
Interaction IV 1 *IV 2 -0.004 0.001 -1.135 -4.015 0.001

2
R = 0.079

Regression Results (Study 1B)

Results. We analyzed responses using linear regression with repair neglect as the first

dependent variable and with the following independent variables: a dummy variable for

availability of the new phone on the market (coded as 1 for availability of the new phone on the

market and 0 otherwise), paid price, and remaining book value of the phone as a percentage of

the original value. The analysis revealed a significant and positive effect of availability of the

new phone on the market ( = .09, t(588) = 2.31, p = .021), a significant effect of paid price ( =

-.25, t(588) = -6.57, p < .001), and a significant effect of book value ( = -.25, t(588) = -6.44, p <

.001). Next, we conducted the same analysis on willingness to pay to repair the current phone as

the second dependent variable. The analysis revealed a significant and negative effect of

availability of the new phone on the market ( = -.082, t(585) = -2.36, p = .018), a significant
 55

effect of paid price ( = .54, t(585) = 16.08, p < .001), and a significant effect of book value ( =

.15, t(585) = 4.49, p < .001). Finally, we conducted the same analysis on interest in participating

in a risky lottery as the third dependent variable and found a significant and positive effect of

availability of the new phone on the market ( = .12, t(588) = 2.66, p = .008), a non-significant

effect of paid price ( = -.04, t(588) = -.98, NS), and a non-significant effect of book value ( = -

.03, t(588) = -.58, NS).

Follow-up Study: The Perceived Realism of Ads on Carelessness (Study 1B)

The objective of this follow-up study was to complement the results of Study 1B on the

extent to which people are consciously aware of the effect of upgrades and address potential

social desirability concerns in Study 1B. In this follow-up, we use an ad featuring other people

engaging in seemingly carelessness practices with their phones to test whether people

acknowledge the existence of an upgrade effect in others. Asking about other consumers should

reduce the extent to which participants are willing to admit carelessness in the presence of

upgrades due to social desirability worries.

Method. We recruited 124 respondents (51% female; Mage = 39, Italian) for a paid online

survey through Qualtrics. The advertising and survey questions were in Italian, the native

language of respondents. First, we asked participants to watch a 30-second ad

(http://bit.ly/1SorckM) by Telecom Italia Mobile (TIM), one of the biggest telecommunication

companies in Italy. Similar to the Virgin Mobile ad described at the beginning of the paper, this

ad features a series of characters damaging their phones accidentally. For example, one girl

drops her phone in a river, and one man throws his mobile under a cement roller (see figure
 56

below for two representative screenshots). The narrator says, Sometimes our passion for

technology makes us do crazy things. Like coming up with the most improbable excuses to get a

rid of our old phone and get a new model. Enough with these tricks! With TIM, you can get a

new phone! The ad was airing on national TV when we conducted the study (July 2015). After

showing them the video, we asked participants to answer the following questions: 1. How

realistic is the idea that consumers would damage their products on purpose? (1 = Not realistic

at all, 7 = Extremely realistic) 8; and 2. Why might the people in the advertising behave like

this? (Open-ended).

Results. Participants did not think that the ad was realistic (M = 2.95) and, in absolute

terms, the rating was lower than the scale midpoint (t(123) = 6.85, p < .001). We coded the

content of the open-ended question on the presumed reasons to explain the behaviors depicted in

the ad. The content analysis revealed the following most recurring categories: 30% of comments

mentioned that the video was just an unrealistic advertising idea (e.g., Just to grab attention with

something unrealistic, but fun, It is just an advertising); 25% of comments mentioned

superficiality and exhibitionism (e.g., Futility, To show off, Because people are slave of

consumerism); 15% of comments mentioned not knowing why people would behave this way;

12% of comments mentioned upgrading (e.g., They want to change phone, To get the new

Samsung); 10% of comments listed other miscellaneous reasons (e.g., Anger), and lastly 8%

of comments mentioned wanting a justification to upgrade (e.g., So that they are justified to

upgrade, To avoid guilt). This follow-up study demonstrates that, even when explicitly

presented with other consumers breaking products on purpose and using tricks to upgrade, the

8
The sensitivity and understanding of this scale were validated vis--vis three other test ads with varying degree of
realism (full details reported in this appendix right after this study).
 57

vast majority of people do not acknowledge the effect of upgrade. Only a fifth of participants

mentioned either willingness to upgrade (12%) or need for justification (8%) in their comments

FIGURE: SCREENSHOTS FROM TIM ADVERTISING

Scale Validation Test for Ad Realism Measure

The objective of this test was to validate the sensitivity and understanding of the question

on the realism of the TIM ad employed in the follow-up study previously reported in this

appendix.

Method. We recruited 103 participants (50% female; Mage = 39, American) for a paid

online survey through Amazon Mechanical Turk. We selected three commercials aired during
 58

the Super bowl game 2016 as the study stimuli. We deliberately picked a very unrealistic ad

(Puppy-Monkey-Baby by Mountain Dew, https://www.youtube.com/watch?v=ql7uY36-LwA),

a moderately unrealistic ad (Distant memory by Advil

https://www.youtube.com/watch?v=xryoyvdTVnw), and a realistic ad (Good morning by

McDonalds (https://www.youtube.com/watch?v=SEbrNUXGq0Q). Each participant was

assigned to watching two commercials randomly selected from among the three ads. After

watching each ad, participants were asked to rate the realism of the commercial. Specifically, for

Puppy-Monkey-Baby, participants answered, How realistic is the idea that consumers would

follow a puppy-monkey-baby? In the case of Distant memory, participants rated, How

realistic is it that consumers would feel so energetic? For Good morning, participants

indicated, How realistic is it that would eat and drink as in the ad? As in the main study, all

answers were rated on a 7-point scale (1 = Not realistic at all, 7 = Extremely realistic).

Results. As expected, Puppy-Monkey-Baby was rated as very unrealistic (M = 1.98)

and, in absolute terms, the rating was lower than the scale midpoint (t(52) = 9.88, p < .001).

Distant memory was also seen as unrealistic (M = 3.15) and rated below the midpoint (t(51) =

4.26, p < .001). Finally, Good morning was seen as a realistic (M = 5.66) and rated above the

midpoint (t(49) = 8.69, p < .001). The picture below offers a visual display of the scores

(including TIM from the main study) on the 7-point question scale. In sum, the scale employed

in the follow-up study to rate the extent to which respondents saw the TIM commercial as

plausible is a sensitive and valid measure.

 59

FIGURE: SCALE VALIDATION RESULTS

Material for the Jenga Game (Study 2)

A: Basic Mug (retail value = $1.00)

B: Upgrade Mugs (average retail value = $9.67)

 60

C: Jenga Tower
 61

Pretest: Functional and Hedonic Products (Studies 3A, 3B, 4)

The objective of this pretest was to examine a series of products (durable and non-

durable) varying from utilitarian to hedonic.

Method. We recruited 61 American respondents for a paid online survey through Amazon

Mechanical Turk. Participants were asked to rate a randomized list of six products: eyeglasses,

sunglasses, toothpaste, shampoo, laundry softener, and perfume. To minimize demand effects,

the focal products were interspersed with six other commonly used products (i.e., wallet,

keychain, laundry detergent, toilet paper, body cream, shaving cream). Using the same measures

employed in previous research on hedonic versus functional products (Dhar and Wertenbroch

2000; Okada 2005), participants rated each product on the following two dimensions: 1. How

functional is this product? (1 = Not at all functional, 7 = Extremely functional), and 2. How

hedonic is this product? (1 = Not at all hedonic, 7 = Extremely hedonic).

Results. For each product, we reverse-coded the rating on functionality and averaged it

with the hedonism rating. As expected in the domain of durable products, the sunglasses were

rated as a more hedonic product than eyeglasses (Msunglasses = 3.35 vs. Meyeglasses = 2.16, t(60) =

7.44, p < .001). Moreover, for non-durable products, perfume was the product most highly rated

as hedonic (Mperfume = 5.03), followed by laundry softener (Msoftener = 3.49), shampoo (Mshampoo =

2.78), and finally toothpaste (Mtoothpaste = 2.34; all pairwise comparisons p-values < .001).
 62

FIGURE: SELECTED STIMULI, CONDITIONS, AND QUESTIONS (STUDIES 3A, 3B)

 63

Scale Validation Test for Usage Measure (Studies 3A, 3B)

The objective of this test was to validate the sensitivity and understanding of the question

on intended usage of the product compared to usual employed in Studies 3A and 3B.

Method. We recruited 103 participants (50% female; Mage = 39, American) for a paid

online survey through Amazon Mechanical Turk. Participants read four product scenarios, one

per product (shampoo, toothpaste, perfume, and laundry softener). To test the sensitivity of the

usage-compared-to-usual measure, two scenarios were intentionally framed to elicit higher usage

than usual; conversely, the other two were framed to elicit lower usage than usual. In the case of

shampoo, participants read, Imagine you are leaving for a one-week trip. You are washing your

hair before departure. The circles below represented the different amount of shampoo you may

use. You will not be able to wash your hair during the trip (i.e., for one week). How much

shampoo would you pour on the palm of your hand? (from 1 to 7). Subsequently, participants

were asked, Is the amount of shampoo you decided to use more or less than what you typically

use when washing your hair? (1= Less than usual; 4 = Same as usual; 7 = More than usual). In

the case of laundry softener, participants read, Imagine you were doing laundry and your load

this time is very light; how much softener would you pour in the measuring cup represented

below? (from 1 to 14). Subsequently, participants were asked, Is the amount of softener you

decided to use more or less than what you typically use when laundering your clothes? (same

scale as shampoo). In the case of toothpaste, participants read, Imagine you about to wash your

teeth and you wont be able to wash them again for 24 hours. How much toothpaste would you

put on your toothbrush? (from 1 to 7). Next, participants were asked, Is the amount of

toothpaste you decided to use more or less than what you typically use when washing your
 64

teeth? (same scale as shampoo). Finally, in the case of perfume, participants read, Imagine you

were going to the gym, how many times would you spray the perfume on you? (from none to

seven times or more). Next, participants were asked, Is the amount of perfume you decided to

use more or less than what you typically use? (same scale as shampoo). The visual aids for

these questions were identical to those employed in the Studies 3A and 3B.

Results. For the two product scenarios eliciting higher consumption than average (i.e.,

shampoo and toothpaste), participants indicated wanting to use more than usual (Mshampoo = 4.58;

Mtoothpaste = 4.53; both ratings higher than 4, p-values < .001). In contrast, for the two product

scenarios eliciting lower consumption than average (i.e., softener and perfume), participants

indicated wanting to use less than usual (Msoftener = 3.37; Mperfume = 3.02; both ratings lower than

4, p-values < .001). In sum, the scale employed in the paper to rate intended product usage as

compared to average is a sensitive and valid measure.

Pilot Study: Justifying Product Upgrades in the Face of Damage (Study 4)

This pilot study examines the basic premise that individuals perceive damage to owned

products as a validation and justification for purchasing new, upgraded products. We manipulate

the physical state of a pair of glasses within three experimental conditions (intact vs. minor-

damage vs. major-damage physical product state) and ask participants to rate their willingness to

upgrade to a newer pair. Based on the pretest (reported earlier in this this online appendix), we

selected both a utilitarian and hedonic product (eyeglasses vs. sunglasses) and, to avoid social

desirability biases, we framed the scenario in two different ways (self vs. other). We predict that

participants will be more likely to upgrade when the product is either damaged or broken
 65

because they perceive this decision to be justifiable. In contrast, we expect that participants will

disapprove of upgrading when the product is intact. Moreover, we examine individual-level

differences in lay rationalism (Hsee et al. 2015), and we expect these effects to be particularly

accentuated for people who chronically experience a high need to rationalize and justify their

decisions. All the results figures are reported in the main paper.

Method. We decided in advance to recruit 950 participants (about 80 per condition). We

recruited 958 respondents (47% female; Mage = 35, American) for a paid online survey through

Amazon Mechanical Turk. We randomly assigned participants to one of 12 conditions in a 3

(intact vs. minor-damage vs. major-damage physical condition) x 2 (functional vs. hedonic

product type) x 2 (self vs. other framing) between-subjects design. The framing manipulation

was included to account for potential differences between the two scenarios due to social

desirability concerns. In the self-framing conditions, participants were asked to imagine that they

owned a functional product, namely a pair of eyeglasses (or a pair of sunglasses in the hedonic

product conditions) and were randomly assigned to one of the three physical product state

conditions. In the intact condition, participants read, The lenses of your eyeglasses are intact;

participants in the minor-damage condition read, The lenses of your eyeglasses are mildly

scratched; and participants in the major-damage condition read, The lenses of your eyeglasses

are extremely scratched. Subsequently, all participants were told about a decision to upgrade:

You just found out that a new version of your current eyeglasses model has been released. You

decide to upgrade to the new model. In the other-framing conditions, participants were

randomly assigned to six analogous scenarios (i.e., three for eyeglasses and three for sunglasses)

about another person. In the case of eyeglasses, for example, participants read: Imagine a

middle-aged man whose name is John. John owns a pair of eyeglasses. The lenses of Johns
 66

eyeglasses are intact [mildly scratched / extremely scratched]. He just found out that a new

version of his current eyeglasses model has been released. After reading the description,

participants answered two questions assessing their propensity to upgrade and the degree of

approval towards the upgrading decision (Cronbachs = .87): 1. If a pair of glasses you owned

was really in this condition, would you get a new pair? (1 = Definitely no, 4 = Maybe, 7 =

Definitely yes); 2. To what extent do you agree with the decision to upgrade? (1 = Strongly

disagree, 4 = Neither agree nor disagree, 7 = Strongly agree). Next, all participants rated two

items on the justifiability of the decision (Cronbachs = .62): 9 Is upgrading a justifiable

decision? 1 = Not justifiable at all, 4 = Somewhat justifiable, 7 = Very justifiable); 2. Would

you feel the need to explain yourself if you bought the new pair? (1 = Definitely no, 4 = Maybe,

7 = Definitely yes; reverse coded). Finally, participants completed the six-item (Cronbachs =

.80) lay rationalism scale (Hsee et al. 2015).

Results. We conducted a full factorial between-subjects ANOVA on willingness to

upgrade as the dependent variable. The analysis revealed a significant main effect of physical

state of the glasses (F(2, 946) = 282.6, p < .001), a significant main effect of product type (F(1,

946) = 20.0, p < .001), and a significant main effect of framing (F(1, 946) = 10.7, p = .001).

None of the interactions was significant. Planned contrasts showed that participants in the major-

damage condition (Mmaj-dam = 6.07) were more willing to upgrade relative to participants in the

minor-damage condition (Mmin-dam = 5.31, t(955) = 7.16, p < .001) and in the intact condition

(Mintact = 3.62, t(955) = 22.9, p < .001). Propensity to upgrade in the minor-damage condition fell

in between the major-damage and intact conditions, and was significantly higher than the intact

condition (t(955) = 15.72, p = .001). The main effects for product type and framing indicate that,

9
Owing to the low reliability of the justifiability items, we also performed all the following analyses with the two
items separately. We find the same pattern of results when each justifiability item is considered separately.
 67

on average, participants were more willing to upgrade in the case of the functional product than

the hedonic one (Meyeglasses = 5.19 vs. Msunglasses = 4.8, t(956) = 3.52, p < .001), and when the

scenario was framed in terms of another individual rather than in personal terms (Mother = 5.12

vs. Mself = 4.87, t(956) = 2.34, p = .019).

We then conducted the same analyses on justifiability and found the same pattern of

results. A full factorial between-subjects ANOVA revealed a significant main effect of physical

state of the product (F(2, 946) = 179.8, p < .001), a significant main effect of product type (F(1,

946) = 9.5, p = .002), and a significant main effect of framing (F(1, 946) = 12.6, p < .001). None

of the interactions was significant. Planned contrasts showed that participants in the major-

damage condition (Mmaj-dam = 6.04) perceived the decision to upgrade as significantly more

justifiable relative to participants in the minor-damage condition (Mmin-dam = 5.58, t(955) = 4.36,

p < .001) and the intact condition (Mintact = 4.16, t(955) = 17.93, p < .001). Rates of justifiability

in the minor-damage condition fell in between results for the major-damage and intact

conditions and were significantly higher than rates of justifiability in the intact condition (t(955)

= 13.56, p < .001). The main effects for product type and framing indicate that, on average,

participants justified more the decision to upgrade for the functional product than for the hedonic

one (Meyeglasses = 5.39 vs. Msunglasses = 5.13, t(956) = 2.56, p = .010) and that they justified the

decision more when the scenario was framed for others than for themselves (Mother = 5.39 vs.

Mself = 5.12, t(956) = 2.82, p = .005).

In sum, the decision to upgrade was seen as more appropriate and justifiable when the

product was damaged, even just lightly so, than when the product was intact, especially for

necessary products and in the case of others upgrading. Since there were no interactions with the

manipulations of product type and framing, we collapsed these conditions and concentrated the
 68

analysis on the focal independent variable of physical state of the product (i.e., intact vs. minor-

damage vs. major-damage conditions) when testing mediation and moderation.

Mediation. We examined whether justifiability mediated the relationship between the

physical state of the product and willingness to upgrade following the Hayes and Preacher (2012)

method for mediation with multi-categorical independent variables. We created two dummy-

coded variables (one with the major-damage condition coded as 1 and the other conditions coded

as 0; the other with the minor-damage condition coded as 1 and the other conditions coded as 0)

and included them simultaneously as predictors in the regression analysis. The effects of both

major and minor damage on propensity to upgrade were significantly reduced (from Bmaj-dam =

2.45, t(955) = 22.9, p < .001, to Bmaj-dam = 1.23, t(954) = 12.88, p < .001, for the major-damage

condition; and from Bmin-dam = 1.69, t(955) = 15.72, p < .001, to Bmin-dam = .76, t(954) = 8.41, p <

.001 for the minor-damage condition) when justifiability was included in the mediation model

and justifiability was a significant predictor (B = .65, t(954) = 25.45, p < .001). Furthermore, the

bias-corrected confidence interval of the indirect effects through justifiability excluded zero for

both the difference between the intact condition and the major-damage condition (indirect effect

= 1.22; 95% CI = 1.06 to 1.41) and for the difference between the intact condition and the minor-

damage condition (indirect effect = .92; 95% CI = .77 to 1.09). In summary, the results support

the notion that when a product is broken or (importantly) slightly damaged, people find

upgrading justifiable and thus are inclined to upgrading to a newer version.

Moderation. Responses were analyzed using a linear regression with willingness to

upgrade as the first dependent variable and with the following independent variables: a variable

for major-damage (coded as 1 for major-damage condition and -1 otherwise), a variable for

minor-damage (coded as 1 for minor-damage condition and -1 otherwise), the lay rationalism
 69

scale (z-scores), and the two interactions between the scale and each of the damage conditions.

The analysis revealed a significant main effect of major-damage condition (B = 1.22, SE = .05,

t(952) = 23.27, p < .001), a significant main effect of minor-damage condition (B = .83, SE =

.05, t(952) = 15.92, p < .001), a non-significant main effect of lay rationalism (B = .08, SE = .05,

t(952) = 1.49, NS), a significant interaction between major-damage condition and lay rationalism

(B = .30, SE = .05, t(952) = 5.96, p < .001), and a significant interaction between minor-damage

condition and lay rationalism (B = .27, SE = .05, t(952) = 5.08, p < .001).

The same analysis performed on justifiability as the second dependent variable revealed

an analogous pattern of results. The analysis revealed a significant main effect of major-damage

condition (B = .93, SE = .05, t(952) = 18.14, p < .001), a significant main effect of minor-

damage condition (B = .71, SE = .05, t(952) = 13.78, p < .001), a significant main effect of lay

rationalism (B = .21, SE = .05, t(952) = 3.92, p < .001), a significant interaction between major-

damage condition and lay rationalism (B = .27, SE = .05, t(952) = 5.38, p < .001), and a

significant interaction between minor-damage condition and lay rationalism (B = .29, SE = .05,

t(952) = 5.60, p < .001).

Mediated Moderation. As reported above, lay rationalism moderated both the dependent

variable (willingness to upgrade) and the mediator (justifiability). We then conducted a mediated

moderation analysis using model 8 in PROCESS (Hayes 2013). The interaction between lay

rationalism and the dependent variable (willingness to upgrade) was significantly reduced when

the mediator (justifiability) was included in the model (indirect effect = .18; 95% CI: .12, .25).

Justifiability was a stronger mediator at higher levels (indirect effect = .72; 95% CI: .62, .84)

than at lower levels of lay rationalism (indirect effect = .35; 95% CI: .26, .45).
 70

Discussion. In conclusion, the results of this pilot study support the notion that when a

product is broken or (importantly) slightly damaged, people find upgrading justifiable and

therefore upgrade to a newer version regardless of the product type (functional vs. hedonic) and

of the framing (self vs. other). Consistent with our expectations, we also found a significant

interaction with lay rationalism. Specifically, high lay-rationalistic participants saw upgrading in

the case of damage as more justifiable and appropriate than low lay-rationalistic participants. In

contrast, in the case of an intact product, high lay-rationalistic participants saw upgrading as less

justifiable and appropriate than participants scoring low on lay rationalism.

Moderation Results (Study 4)

FIGURE: MODERATION RESULTS, FLOODLIGHT ANALYSIS (STUDY 4)

Note. Gray horizontal line fixed at JohnsonNeyman point (4.56).

 71

REFERENCES

Aiken, Leona S., Stephen G. West, and Raymond R. Reno (1991), Multiple Regression: Testing
and Interpreting Interactions, Newbury Park, CA: SAGE.
Dhar, Ravi and Klaus Wertenbroch (2000), Consumer Choice Between Hedonic and Utilitarian
Goods, Journal of Marketing Research, 37 (1), 6071.
Hayes, Andrew F. (2013), Introduction to Mediation, Moderation, and Conditional Process
Analysis: A Regression-Based Approach, Guilford Press.
and Kristopher J. Preacher (2012), Statistical Mediation Analysis with a
Multicategorical Independent Variable, British Journal of Mathematical and Statistical
Psychology.
Hsee, Christopher K., Yang Yang, Xingshan Zheng, and Hanwei Wang (2015), Lay
Rationalism: Individual Differences in Using Reason Versus Feelings to Guide Decisions,
Journal of Marketing Research, LII (February), 13446.
Okada, Erica Mina (2005), Justification Effects on Consumer Choice of Hedonic and Utilitarian
Goods, Journal of Marketing Research, 42 (1), 4353.