How Concrete Language Shapes Customer

Satisfaction

GRANT PACKARD
JONAH BERGER

Downloaded from [Link] by guest on 29 September 2021

Consumers are often frustrated by customer service. But could a simple shift in
language help improve customer satisfaction? We suggest that linguistic concrete-
ness—the tangibility, specificity, or imaginability of words employees use when
speaking to customers—can shape consumer attitudes and behaviors. Five stud-
ies, including text analysis of over 1,000 real consumer–employee interactions in
two different field contexts, demonstrate that customers are more satisfied, willing
to purchase, and purchase more when employees speak to them concretely. This
occurs because customers infer that employees who use more concrete language
are listening (i.e., attending to and understanding their needs). These findings
deepen understanding of how language shapes consumer behavior, reveal a psy-
chological mechanism by which concreteness impacts person perception, and
provide a straightforward way that managers could help enhance customer
satisfaction.

Keywords: language, concreteness, listening, customer service, social perception

F rustration with customer service is one of life’s most

common complaints (PwC 2018), and employee inat-
tention is a central cause (Price, Arnould, and Tierney
This article examines whether the words employees use
can help. Specifically, we suggest that speaking concretely
can improve customer satisfaction. Consider a consumer
1995; Van Dijk, Kamaruzdin, and Anjum 2019). shopping for a shirt in a store. The salesperson might refer
Consumers often feel like service people don’t care and are to that shirt very concretely (e.g., shirt), very abstractly
just “going through the motions” rather than actually lis- (e.g., that), or somewhere in between (e.g., item, top, or
tening and responding to their needs (Hyken 2018; Parker clothing). While these may seem like trivial variations in
2019). language, we suggest that these small changes can signal
that the employee is listening (i.e., attending to and under-
standing the customer’s needs), which boosts consumer
Grant Packard (gpackard@[Link]) is an associate professor attitudes and behaviors.
of marketing at the Schulich School of Business, York University. Jonah Five studies, combining textual analysis of over a thou-
Berger (jberger@[Link]) is an associate professor of market-
ing at The Wharton School of the University of Pennsylvania. Please ad-
sand real customer service interactions in the field with
dress correspondence to Grant Packard. The authors thank Reihane carefully controlled experiments, explore this possibility.
Boghrati, Yang Li, and Russell Richie for insights and Alexa Brewster, They demonstrate that linguistic concreteness improves
Kenza Chamariq, Skylar Deutsch, Claudia Iglesias, Aneesh Matta, Sally attitudes toward the employee and company—as well as
Shin, and Nukhet Taylor for research assistance. This research is sup-
ported in part by funding from the Social Sciences and Humanities actual consumer spending—and illustrate the psychologi-
Research Council of Canada and by a seed grant from the Lazaridis cal mechanism underlying these effects.
Institute for the Management of Technology Enterprises awarded to the Our findings make four main contributions. First, we
first author. Supplementary materials are included in the web appendix ac-
companying the online version of this article. deepen understanding of how language shapes consumer
behavior. While a growing body of research has begun to
Editor: J. Jeffrey Inman
examine the importance of linguistic devices in consumer
Associate editor: Andrew T. Stephen contexts such as advertisements, online reviews, and social
media (Pogacar, Shrum, and Lowrey 2018), less work has
Advance Access publication July 18, 2020
studied how language shapes consumer beliefs and
C The Author(s) 2020. Published by Oxford University Press on behalf of Journal of Consumer Research, Inc. This is an Open Access article distributed under the terms of the Creative
V
Commons Attribution Non-Commercial License ([Link] which permits non-commercial re-use, distribution, and reproduction in any medium,
provided the original work is properly cited. For commercial re-use, please contact [Link]@[Link]  Vol. 47  2021
DOI: 10.1093/jcr/ucaa038

787
788 JOURNAL OF CONSUMER RESEARCH

behavior in everyday conversations with sales or service Zeithaml, Berry, and Parasuraman 1996), and future pur-
people (Ordenes et al. 2014; Packard, Moore, and chases (Sriram, Chintagunta, and Manchanda 2015).
McFerran 2018). We demonstrate the important role of lin- Not surprisingly, then, thousands of academic articles
guistic concreteness and the underlying process that drives seek to understand how to improve the quality of consumer
its impact in this context. interactions with sales and service people (cf. reviews by
Second, we demonstrate that concrete language can Ladhari 2008; Parasuraman and Zeithaml 2002). Initial re-
shape social perceptions about a speaker’s attention and search examined service quality as a function of the gap
understanding. Prior research shows that a speaker’s biases between customer expectations and firm performance
toward a target (i.e., person or object) impact how con- (Lewis and Booms 1983) and tried to identify broad
cretely the speaker talks about that target, and the infer- dimensions, or features, of importance (e.g., tangibles or
ences others draw about the speaker’s attitudes access; Parasuraman, Zeithaml, and Berry 1985;

Downloaded from [Link] by guest on 29 September 2021

(Schellekens, Verlegh, and Smidts 2010; Semin 2008). Parasuraman et al. 1988). Subsequent work helped estab-
Rather than focusing on biases and attitudes, however, we lish linkages between service quality and various outcomes
demonstrate that concreteness can have a different psycho- (e.g., customer acquisition, satisfaction, and profits;
logical function. Concrete language can generate infer- Dawkins and Reichheld 1990; Smith et al. 1999; Zahorik
ences about whether someone else is listening: whether and Rust 1992).
observers believe someone is attending to, and understand- One goal has been to identify specific behaviors or tac-
ing, their personal needs. tics that boost customer attitudes and intentions (cf. review,
Third, from a substantive perspective, these results have Snyder et al. 2016). Responding quickly, apologizing, or
clear implications for improving the customer experience offering compensation or discounts are all beneficial
(Meyer and Schwager 2007). Small linguistic shifts can (Davidow 2003; Rust and Chung 2006; Zeithaml et al.
have a substantial impact. The field data suggest that in- 1996). Employees can adapt their selling pitch to each cus-
creasing linguistic concreteness by one standard deviation tomer instead of making a “canned” presentation (Weitz,
improves customer satisfaction by 9% and actual spending Sujan and Sujan 1986), use flattery (Chan and Sengupta
by at least 13%.1 Furthermore, compared to traditional 2010), use the right pronouns (“I” rather than “we”;
managerial tools (e.g., sales training or customer compen- Packard et al. 2018), or even just smile more (Wang et al.
sation), encouraging employees to use more concrete 2017), although such overt tactics often backfire
words when referencing the customer’s interests (e.g., that (Campbell and Kirmani 2000).
shirt rather than just that) should be more straightforward Beyond specific behaviors or tactics, might something
and less costly. like the words employees use improve customer attitudes
Fourth, from a methodological perspective, we respond and intentions?
to the call for “nonstandard” methods in consumer research
(Berger et al. 2020; Inman et al. 2018). We move beyond LISTENING
the constrained lexical typology of concreteness in the lin-
guistic category model (LCM) (Semin and Fiedler 1988) We suggest this possibility based on research on listen-
and, using approaches from computational linguistics, con- ing. Listening, conceptualized as attending to and under-
sider tens of thousands of words, blending this with standing what someone has said (Bodie 2012; Steil,
experiments. Barker, and Watson 1983), has been linked to a range of
Next, we review research on customer service, listening, positive outcomes, including salesperson performance, re-
and linguistic concreteness and use these to help build our lationship quality, and customer satisfaction (Castleberry,
predictions. Shepherd, and Ridnour 1999; de Ruyter and Wetzels 2000;
Drollinger, Comer, and Warrington 2006). Early communi-
CUSTOMER SERVICE cations research on listening focused on educational set-
tings, where good listeners were defined as students who
Whether emailing customer support, calling customer were “more attentive during classroom activities” (Nichols
service, or speaking face to face with retail or sales repre- 1948, 160). Subsequent work argued that for listening to
sentatives, interacting with employees is a central feature occur, information not only needs to be attended to but
of customer experiences with brands (Meyer and Schwager also understood, leading to comprehension assessments
2007). Great service is the most direct route to customer through “listening tests” (Bodie 2012, 111; Watson and
satisfaction (Marinova, Singh, and Singh 2018; Barker 1984). Although listening is relatively under-
Parasuraman, Berry, and Zeithaml 1988), which shapes be- researched in psychology, “paying full attention to” and
havioral intentions (Smith, Bolton, and Wagner 1999; “understanding” the speaker are commonly described as
the conditions necessary for listening to have occurred
1 We urge caution in assuming generalizability beyond these firms. (Huang et al. 2017, 432).
PACKARD AND BERGER 789

While marketing research sometimes uses slightly dif- detailed and context specific (Miller et al. 2007; Pan et al.
ferent language, it focuses on the same themes of attending 2018).
and understanding (de Ruyter and Wetzels 2000, 277), for Consider the kind of everyday things that different ser-
example, note that a key aspect of listening is vice providers might say to consumers. A waiter might ask
“attentiveness,” or the extent to which someone is receiv- a customer whether they would like anything, a drink, or
ing a message. Similarly, Castleberry et al. (1999, 31) de- coffee. A call center worker could respond to a caller’s in-
fine the first step in salesperson listening as “the reception quiry by saying the package will be arriving there, at their
of [customer-transmitted] stimuli and attending to the [cus- place, or at their door. An employee at a cellphone retailer
tomer’s] message.” Ramsey and Sohi (1997, 128) further could refer to a new iPhone as a product, device, or phone.
note that agents must first attend to the customer’s mes- In each case, the options increase in specificity and imagi-
sage. Beyond simply attending, though, listening also nability. Concreteness also varies in other parts of speech.

Downloaded from [Link] by guest on 29 September 2021

entails attempting to “understand the message by assigning Describing a vehicle as sporty or red should make it more
meaning to the verbal and nonverbal messages that are vivid. Rather than telling a customer they will go to the
transmitted” (de Ruyter and Wetzels 2000, 277). Ramsey back of the store to look for a larger size, an employee who
and Sohi (1997, 128) suggest agents must not only attend says they will walk to the back is using more imaginable
to what is being said but also “understand what is being and specific language to describe this action.
said.” Castleberry et al. (1999, 31) describe this part of lis- Each of these examples increase in linguistic concrete-
tening as “decoding” the message such that the employee ness (Medical Research Council (MRC) psycholinguistic
can “accurately understand and evaluate the buyer’s database; Brysbaert et al. 2014; Coltheart 1981; Paetzold
message.” and Specia 2016; Spreen and Schulz 1966)2 because they
Importantly, for employee listening to impact the cus- describe things or actions in a more vivid, perceptible, tan-
tomer, the customer must perceive that the employee is lis- gible, detailed, or specific manner. These examples further
tening. Imagine talking to someone, but that person reveal that variations in concreteness manifest across dif-
provides no response. They might have attended to and un- ferent parts of speech, including those describing objects
derstood what was said, but if they don’t respond, either and actions. Thus, rather than a typology in which only a
verbally or otherwise, there is no way to know whether few types of words have discrete membership in a particu-
they heard and processed what was said (Bodie 2012; Steil lar concreteness level (i.e., the LCM; Semin and Fiedler
et al. 1983). For listening to have an impact then, employ- 1988), linguistic concreteness can refer to variation in most
ees must communicate that they are attending and under- words someone might choose to convey similar meaning.
standing through their behavioral responses (Castleberry A large body of work examines when and why people
et al. 1999; de Ruyter and Wetzels 2000; Parasuraman, speak more or less concretely (cf. reviews by Fiedler 2008;
Berry, and Zeithaml 1991; Ramsey and Sohi 1997). Could Semin 2008). Overall, this literature demonstrates that,
a simple language feature help fulfill this role? when communicating positive things about the self or liked
others, people use more abstract language because it sug-
CONCRETE LANGUAGE gests that these positive attributes are generalizable, stable
traits (e.g., “Lisa is kind” rather than “Lisa helped me”;
We suggest that one way employees can signal listening Maas, Ceccarelli, and Rudin 1996). Linguistic concrete-
is through linguistic concreteness. Concreteness describes ness can also shape inferences other people make about the
how much a word refers to an actual, tangible, or “real” en- speaker’s attitudes or traits. Audiences infer that someone
tity, describing objects and behaviors in a manner that who uses more abstract language to describe another per-
seems more specific, familiar, and perceptible to the eye or son positively is biased toward them because of their will-
mind (i.e., imaginable or vivid; Brysbaert, Warriner, and ingness to use more abstract, sweeping generalizations
Kuperman 2014; Semin and Fiedler 1988). While abstract (Douglas and Sutton 2006; Wigboldus, Semin, and Spears
language captures intangible qualities, ideas, or concepts 2000). Other work suggests that concrete medical mes-
(e.g., love or anywhere), concrete language (e.g., kiss or sages might make the source seem more credible (Miller
home) arises from or appeals to immediate experience et al. 2007), although evidence for this is mixed (Toma and
(Fujita et al. 2006; Hansen and Wanke 2010; Langacker D’Angelo 2015).
1987; Semin and Fiedler 1988). These conceptions of lin- While relatively little work has examined concrete lan-
guistic concreteness have been extended to applications of guage in consumer contexts, a handful of papers (Aerts and
Trope and Liberman’s (2010) construal level theory, where
concrete representations of objects or processes are simi- 2 MRC concreteness scores (range ¼ 100–700) for comparative utter-
larly said to include more detailed features of the immedi- ances in this paragraph: anything ¼ 290, drink ¼ 548, coffee ¼ 602;
there ¼ 300, place ¼ 460, door ¼ 594; product ¼ 457, device ¼ 545,
ate situation, and to managerial communications, which phone ¼ 585; go ¼ 309, walk ¼ 497. We discuss this measure of lin-
describe lexical concreteness as language that is more guistic concreteness further in study 1.
790 JOURNAL OF CONSUMER RESEARCH

Verlegh 2018; Schellekens et al. 2010, 2013) have exam- Trope, and Liberman 2013). Compared to saying “I’ll go
ined linguistic concreteness in word of mouth. Similar to search for that,” saying “I’ll go search for that t-shirt”
work on person description and perception, word of mouth more clearly indicates that the employee is attending to
using abstract language (e.g., “A Katasca car is high qual- what the customer wants in the immediate situation. This
ity” rather than “A Katasca car does not rust”) increases should imply that the employee is listening to the custom-
persuasion because it suggests that the product is good in er’s personal needs.
general (i.e., the car is high quality), while more concrete This increase in perceived employee listening, in turn,
language suggests that the positive aspects do not general- should benefit the firm. It should boost customer satisfac-
ize beyond a specific situation (i.e., the car does not rust). tion (de Ruyter and Wetzels 2000) and product attitudes
Consequently, for positive word of mouth, abstract lan- (Ramsey and Sohi 1997) and increase future purchase vol-
guage boosts purchase intent. ume (Cronin, Brady, and Hult 2000; Zeithaml et al. 1996).

Downloaded from [Link] by guest on 29 September 2021

In sum, we predict that using more concrete language
THE CURRENT RESEARCH should increase customer satisfaction, intentions, and be-
havior toward the firm. This should be driven by the belief
Beyond speaker bias and persuasion, we propose that that the employee is listening (i.e., attending to and under-
concrete language can also serve a different function, sig- standing the customer’s specific needs).
naling how carefully someone is listening. Specifically, Six studies test these predictions. The first two test our
concrete language can suggest that employees are paying theorizing in the field. Study 1 analyzes customer service
attention to, and understanding, customer needs. calls to examine whether customer satisfaction is higher
It is probably easier for employees to use more abstract when employees speak more concretely. Study 2 uses writ-
language. Call center and retail employees deal with doz- ten interactions and a different firm, testing whether con-
ens of customers a day. In the case of an online clothing re- sumers purchase more following an email interaction if
tailer, for example, an employee may go from talking to employees respond with more concrete language. Both
someone who received the wrong size of shoes, to helping studies control for a broad range of features (e.g., aspects
someone find the right t-shirt color. Rather than tailoring
of the customer, employee, issue being discussed, other
language to each, employees may fall in the habit of using
language features, and linguistic mimicry) to rule out alter-
the same stock phrases, e.g., “Sorry about this issue” or
native explanations.
“I’ll go look for that,” whether the this or that in question
To demonstrate linguistic concreteness’s causal impact
is about shipping, a t-shirt, or some other topic. While
and test the underlying process, the next three studies use
more concrete language (i.e., “Sorry that we sent the wrong
experiments. Study 3 manipulates employee concreteness,
size” or “I’ll go look for that t-shirt in grey”) would be
demonstrating that it increases satisfaction and willingness
more focused on the specific situation, using more general,
to purchase because it signals perceived listening (i.e., at-
abstract, generic responses allows employees to save time
tention and understanding). Study 4 further tests listening’s
and effort.
However, while abstract language might be easier for mediating role and uses subtle linguistic manipulations to
employees, consumers may react more positively to con- examine whether slightly increasing concreteness has a
crete language because it suggests employee attention to corresponding impact on satisfaction and purchase intent.
and understanding of their specific needs. As discussed, Study 5 tests the psychological process through modera-
concrete language is more specific and vivid (Hansen and tion. If perceived listening is driving the effects of linguis-
Wanke 2010; Langacker 1987; Semin and Fiedler 1988). tic concreteness, as we suggest, then its effects should be
But it can be difficult for service employees to speak con- mitigated when the language used suggests that the agent
cretely without listening. Just as fake reviews may include is not actually paying attention. Study 5 also examines
fewer concrete details because writers did not experience downstream consequences of perceived listening (e.g., per-
the specific things they are writing about (DePaulo et al. ceived empathy, closeness), further affirming its centrality
1982; Ott et al. 2011), employees can’t respond concretely in driving concreteness’s effect.
if they didn’t actually listen to the customer’s specific, idi-
osyncratic needs. Using concrete language should suggest STUDY 1: SPEAKING CONCRETELY IN
that the agent went to the effort to pay attention to and un- THE FIELD
derstand the customer’s situation—they listened.
More broadly, our prediction is consistent with the Study 1 investigates whether customers are more satis-
maxim of relation (Grice 1975) and construal (Trope and fied when employees speak concretely. Using customer
Liberman 2010): concrete language suggests that the service phone calls to a major online apparel retailer, we
speaker is talking in a more situationally relevant manner examine whether employees who use more concrete lan-
with their attention focused on the here and now (Maglio, guage are seen as more helpful.
PACKARD AND BERGER 791

Data and Method similar scales (e.g., 100–700 for MRC concreteness vs. 18–
100 for customer age vs. 1–4 for customer satisfaction), so
A large American online apparel retailer provided 200
all continuous measures were standardized (z-scores).
audio recordings of customer service calls. As a dependent
Unstandardized results do not differ in sign or significance
measure, they included their key measure of customer sat-
(all key ps < .05 or better).
isfaction, how helpful customers felt the employee was
(1 ¼ not at all helpful; 4 ¼ very helpful), which was cap-
tured moments after the call was over. Perceived employee
Results
helpfulness is a key aspect of performance-based measures As predicted, customers were more satisfied when the
of customer satisfaction (Cronin and Taylor 1992; employee used more concrete language (b ¼ 0.17, SE ¼
Parasuraman et al. 1991). For each of the four levels of this 0.07, t ¼ 2.36, p ¼ .019; table 1, model 1). A one standard
measure, the company provided a random sample of 50 deviation increase in concreteness (5.6% increase) is asso-

Downloaded from [Link] by guest on 29 September 2021

calls. The company also provided control variables we de- ciated with an 8.9% increase in customer satisfaction.
scribe later.
Control Variables. A more extensive model shows this
A professional transcription service converted the audio
relationship persists controlling for a variety of alternative
recordings to text. To control for conversational dynamics
explanations, including aspects of the call, customer, and
such as linguistic mimicry (Moore and McFerran 2017),
employee.
each conversational turn was treated as a separate record
First, we control for aspects of the call. Conversational
(e.g., employee turn 1: “How can I help you?,” customer
topics can shape how people speak (Blankenship and Craig
turn 2: “I need to return something,” and employee turn 3: 2011), so one could argue that, rather than being driven by
“I can help”). At least part of the recording was inaudible concreteness, the results are driven by the contents of the
for 15 calls, leaving 185 for analysis. Conversations lasted call. If employees speak more abstractly when addressing
6.19 minutes on average (SD ¼ 3.97) and contained 66.75 more complicated issues the customer feels more nega-
turns (SD ¼ 44.49). tively about, for example, it could be the issue, rather than
We used natural language processing to measure linguis- concreteness itself, that is driving the effects.
tic concreteness. Words were scored using perceived con- We control for call content in four ways. Most simply,
creteness ratings from Paetzold and Specia (2016), which we include dummies for the six different call Reasons
use a bootstrapped extension of the MRC psycholinguistic noted by the company (e.g., account and shipping). To pro-
database to score over 85,000 modern English words on a vide a more fine-grained control for content, we use cus-
continuous scale of 100–700. This scoring is based on ex- tomer language to uncover the latent mixture of issues
tensive work in which participants rated the perceived con- using Topic Modeling (latent Dirichlet allocation, Blei
creteness of different words (Coltheart 1981; Paivio, 2012; see web appendix for details). We also had two
Yuille, and Madigan 1968; Spreen and Schulz 1966). judges evaluate the call’s Severity (1 ¼ not at all severe,
Words referring to more tangible, specific, imaginable 5 ¼ very severe; Intraclass Correlation Coefficient (ICC)
objects, materials, people (e.g., nouns, pronouns, adjec- (2,1) ¼ 0.720) based on how the customer spoke for the
tives), processes (e.g., verbs, adverbs), or relationships first part of each call. Finally, longer calls might make it
(e.g., prepositions, conjunctions) were perceived as more more difficult for employees to remain attentive (de Ruyter
concrete while words referring to abstract concepts or less and Wetzels 2000), which might lead them to speak more
tangible objects, processes, or relationships were perceived abstractly, so we control for call length in seconds.
as less concrete. Results are the same using an alternative Second, we control for aspects of the customer. To start,
concreteness dictionary of which Paetzold and Specia’s we include all available demographics (i.e., customer age,
(2016) is a direct extension (Brysbaert et al. 2014; see web gender, and geographic region). In addition, loyal or high
appendix for details). value customers may tend to be differentially satisfied or
We computed a concreteness score for each conversa- interact with employees differently, which could shape em-
tional turn (averaging across all words in that turn) and for ployee language or customer satisfaction (Niederhoffer and
each conversational participant (averaging across all words Pennebaker 2002). To control for these possibilities, we in-
over all their turns). Results were the same whether or not clude the customer’s lifetime expenditure. Furthermore, the
stop words commonly excluded from linguistics analyses customer’s preexisting attitude toward other aspects of the
(e.g., but, and) were included. We report results excluding company could shape both their satisfaction and how they
stop words. interact with the employee. To account for this possibility,
Finally, ordinary least squares regression examined the we control for the customer’s attitude toward their pre-call
relationship between employee linguistic concreteness and shopping experience and the website.3
customer satisfaction (N ¼ 185). Research (Marquardt
1980; Rawlings, Pantula, and Dickey 1998) recommends 3 Given these were captured at the end of the call, one could be con-
reporting standardized results when variables do not share cerned that they were impacted by employee concreteness, but this
792 JOURNAL OF CONSUMER RESEARCH

Alternatively, one could wonder whether the results are TABLE 1:

driven by how the customer speaks. To account for the pos- STUDY 1 RESULTS
sibility that satisfied customers may tend to speak more
concretely, leading employees to speak more concretely in DV: employee helpfulness
response (e.g., linguistic mimicry, Moore and McFerran
(1) Base (2) Full
2017), we control for customer concreteness. To control
for other major language features that have been exten- IV: employee concreteness 0.17*(0.07) 0.13*(0.06)
Controls
sively linked to psychological phenomena, we include the Call
psychological process dictionaries from Linguistic Inquiry Reason Included
and Word Count (LIWC; cognition, emotion, sociality, per- Topic model Included
ception, motivation, time, and formality; Pennebaker et al. Severity Included
Length Included

Downloaded from [Link] by guest on 29 September 2021

2015). Customer
Third, we control for aspects of the employee. One could Gender Included
argue that, in addition to speaking more concretely, experi- Age Included
Region Included
enced employees also do other things that increase satisfac- Lifetime expenditures Included
tion. To assess this possibility, we included employee Preexisting attitudes Included
tenure in days and the volume of calls they have handled Customer language
Concreteness Included
(lifetime calls). Furthermore, to account for the possibility LIWC psych. dictionaries Included
that unobservable individual differences drove the results, Employee
we included a random effect for each employee (N ¼ 130). Tenure Included
Lifetime calls Included
Even after accounting for these controls, customers Random effect Included
remained more satisfied when employees used more con- Employee language
crete language (b ¼ 0.12, SE ¼ 0.06, t ¼ 1.95, p ¼ .026; LIWC psych. dictionaries Included
see table 1, model 2 and web appendixtable W2, model 5). Intercept 0.00 (0.07) 0.05*(0.64)
R-squared 0.03
Reduced models that independently account for call, cus- Marginal R-squared 0.47
tomer, and employee features appear in web appendix table Conditional R-squared 0.74
W2 (models 2–4). Concreteness has a significant positive N 185 185
effect under all specifications. Correlations, detailed results Note: *p < .05; standard errors appear in parentheses. Statistical results
for controls are presented in web appendix table W2.
for the 52 control variables, and investigation of conceptu-
ally plausible moderation by selected controls appear in the
web appendix.
Results also persist using penalized regression First, we lagged employee and customer concreteness in
approaches (Lasso and Ridge) that include automatic vari- each turn on their conversation partner’s prior turn(s).
Automatic selection via Akaike Information Criterion sup-
able selection and regularization to help account for collin-
ports a model in which employee concreteness is shaped
earity (see web appendix for details and results), and in an
by customer concreteness in the prior turn (lag 1 b ¼ 0.04,
ordinal logit specification treating customer satisfaction as
SE ¼ 0.01, t ¼ 2.88, p ¼ .004). Customer concreteness is
a four-level discrete outcome with (b ¼ 0.70, SE ¼ 0.33,
only weakly impacted by employee concreteness in the
t ¼ 3.23, p ¼ .001) or without controls (b ¼ 0.62, SE ¼
prior turn (lag 1 b ¼ 0.03, SE ¼ 0.02, t ¼ 1.83, p ¼ .069).
0.26, t ¼ 2.35, p ¼ .020).
Lags greater than one turn were nonsignificant (ps > .15).
Controlling for Conversational Dynamics. We con- Second, to allow for dynamics in concreteness mimicry,
trolled for customer concreteness at the call level, but one we enter these two lag terms as controls in a version of the
might still wonder whether the results are driven by more main regression model that operates at the turn rather than
dynamic linguistic mimicry at turn level. To account for call level. After controlling for the significant lag effect (b
this, we used vector auto-regression to assess the direction- ¼ –0.001, SE ¼ 0.0004, t ¼ –6.88, p < .001), however,
ality and duration of any turn-level mimicry effects customers remained more satisfied when employees used
(N ¼ 12,349 turns). This approach allows us to control for more concrete language (b ¼ 0.002, SE ¼ 0.001, t ¼ 4.45,
the possibility that agent language is driven by what the p < .001) in the prior (lagged) turn. These results suggest
customer said recently or vice versa, which could have a that the positive effect of agent concreteness was robust to
perpetuating or even multiplicative effect on the agent’s mimicry in the dynamic back-and-forth of conversation.
use of concrete language (see Aerts, Smits, and Verlegh
2017). Discussion
was not the case (shopping experience b ¼ 0.13, SE ¼ 0.13, t ¼ 1.07, Study 1 provides preliminary support for our theorizing.
p ¼ .286; website b ¼ 0.003, SE ¼ 0.12, t ¼ 0.03, p ¼ .976). Analysis of almost 200 customer service calls demonstrates
PACKARD AND BERGER 793

that customers were more satisfied when employees used Second, we extracted linguistic concreteness and other
more concrete language. The results were robust to a vari- language variables from the email text using the same NLP
ety of more extensive model specifications. Customers procedures described in study 1.
were more satisfied when employees spoke more con- Third, we applied the regression approaches from study
cretely after accounting for more than 50 features of the 1 to test the relationship between employee concreteness
customer, employee, and the call itself. Finally, the results and downstream customer spending. Because we also ob-
were meaningful. A one standard deviation increase in con- serve the customer’s purchases before the interaction, all
creteness was associated with an almost 9% increase in models include this to control for heterogeneity in baseline
satisfaction. expenditures.5 Not surprisingly, customers who spent more
Note that the automated measure used (MRC) tracks before the interaction also spend more after (b ¼ 0.49, SE
people’s perceptions of concreteness well. First, as dis- ¼ 0.03, t ¼ 15.23, p < .001), but pre-interaction spending

Downloaded from [Link] by guest on 29 September 2021

cussed, this measure was built using actual participant rat- was not linked to employee concreteness (b ¼ –12.57, SE
¼ 33.12, t ¼ –0.38, p ¼ .70). As explained for study 1,
ings of perceived concreteness (Brysbaert et al. 2014;
continuous measures were standardized (z-scores).
Coltheart 1981). Second, when three judges (blind to hy-
Unstandardized results are the same.
potheses) rated a random sample of 100 different agent
responses based on how concrete they are (ICC (2,3) ¼
0.86; F ¼ 7.30, p < .001, 95% Confidence Interval (CI) ¼
Results
0.81, 0.90), judges’ ratings were highly correlated with the As predicted, customers spent more after emails in
automated linguistic measure (r ¼ 0.71, p < .001). which the employee used more concrete language (b ¼
0.08, SE ¼ 0.03, t ¼ 2.86, p ¼ .004). A one standard devia-
tion increase in concreteness corresponded to a more than
STUDY 2: WRITING CONCRETELY IN $10 (30%) increase in average customer spending over the
THE FIELD 90 days following the email (from $32.73 to$42.80) for
this particular firm.
Study 2 has three main goals. First, while the results of Given the size of this result, we also consider a more
study 1 are supportive and cast doubt on many alternative conservative model including a dummy controlling for
explanations, one could still wonder whether there is some “entry” (Blundell, Duncan, and Meghir 1998; Manchanda,
other unobserved factor driving the result. Perhaps there is Packard, and Pattabhiramaiah 2015). This approach
something in the vocal features used by the employee, or accounts for potentially inflated estimates due to customers
something else in the call dynamics that is driving the ob- who were not active shoppers (i.e., zero purchases) during
served relationship. the 90 days before the call and therefore could be “new
To rule out these alternatives, rather than phone calls entrants” to the market or firm rather than customers who
with multiple back-and-forths, study 2 focuses on simpler actually increased their purchases from zero to a nonzero
email interactions that have only two turns. The customer value. Results are the same (b ¼ 0.07, SE ¼ 0.03, t ¼ 2.75,
writes the company and receives a reply. If the benefits of p ¼ .006), albeit smaller ($4 or 13%).
concreteness persist even in this simplified context, it casts
Robustness. This relationship persists after including
doubt on the possibility that vocal aspects or mimicry can controls similar to those used in study 1 (see table 2 and
explain the results. web appendix for full list). Customers still spent more after
Second, one could wonder whether the results are some- interacting with employees that spoke more concretely (b
how driven by idiosyncratic features of the company exam- ¼ 0.13, SE ¼ 0.05, t ¼ 2.54, p ¼ .011; table 2, model 2;
ined. Consequently, we test the generalizability of the results see web appendix for correlations [table W4] and detailed
by examining a different retailer in a different category. results for the controls [table W5]). Penalized regression
Third, we test whether the results extend to purchase approaches also support concreteness’s role (web appendix
behavior. table W6).

Data and Method 4 This time period was provided as the firm indicated most customers
purchased at least quarterly. The statistical significance of results did
First, we worked with a large Canadian multi-channel not change when we included dummies for the minority of customers
who had zero expenditure in the pre, post, or both (zero as a special
retailer of consumer durables to acquire a random sample case). This firm did not link their end-of-call survey results to individ-
of 941 customer service interactions. Each includes a cus- ual customers, so a measure of customer satisfaction (e.g., helpfulness)
tomer email and a reply by a service employee (i.e., two is unavailable.
turns). The firm also provided how much the customer 5 The relationship between employee concreteness and customer
expenditures also replicates excluding baseline purchase volume be-
spent in the 90 days before and after the interactions (in fore the customer service interaction (b ¼ 0.08, SE ¼ 0.03, t ¼ 2.39, p
dollars)4 and some customer controls. ¼ .017).
794 JOURNAL OF CONSUMER RESEARCH

TABLE 2: customers feel like the employee is listening, as we sug-

STUDY 2 RESULTS. gest, then perceptions that the agent is attending to and un-
derstanding customer needs should mediate the effect.
DV: customer expenditures (90 days post)
(1) Base (2) Full
Method
IV: employee concreteness 0.08**(0.03) 0.13* (0.05)
Participants (N ¼ 206, Amazon Mechanical Turk) imag-
Controls ined a customer service interaction. They were told that
Interaction they were shopping online, had ordered some shoes, and
Reason Included
Complaint Included
decided they wanted to add a pair of pants to the order, so
Length in words Included they called customer service for help.
Customer

Downloaded from [Link] by guest on 29 September 2021

The only difference between the two between-subjects
Expendit. (90 days pre) Included Included
Gender Included
conditions was the concreteness of the employee’s re-
Region Included sponse. In the low [high] concreteness condition, partici-
Customer language pants were told the employee said “Unfortunately, I can’t
Concreteness Included
LIWC dictionaries Included
just add something [the pants] from here. I can cancel the
Judged severity Included order [the shoes], and you can submit a new order” (MRC
LDA topic probabilities Included score ¼ 321.59 [350.53]).6 A between-subjects pretest
Employee language
LIWC dictionaries Included
(N ¼ 68) asking “How concrete was the salesperson’s
Intercept 0.01 (0.03) –0.13 (0.16) reply?” (1 ¼ not at all concrete, 7 ¼ very much concrete)7
R-squared 0.21 0.26 confirmed that the language in the high concreteness con-
N 941 941
dition (M ¼ 5.26, SD ¼ 1.60) was perceived as more con-
Note: **p < .01, *p < .05; standard errors in parentheses. Statistical crete than the low concreteness condition (M ¼ 4.09, SD ¼
results for controls are presented in web appendixtable W4.
1.85; F(1, 66) ¼ 7.87, p ¼ .007, g2 ¼ 0.107).
Next, we measured the key dependent variables. On 7-
Discussion point scales, participants completed a two-item measure of
satisfaction with the employee (“How satisfied are you
Study 2 provides further evidence for the relationship
with the employee’s response?,” “How satisfied are you
between concreteness and customer satisfaction.
with your experience with this employee so far?”; r ¼
Examining a different company, and a different mode of
0.90) and willingness to purchase (“How likely are you to
communication (i.e., email rather than phone), customers
purchase something from this store?,” “How likely would
spent more after interacting with employees who used
you be to use this store again?”; r ¼ 0.83) both adapted
more concrete language. This relationship persists control-
from Maxham and Netemeyer (2002). Given the well-
ling for a variety of interaction features, customer demo-
established link between customer satisfaction and behav-
graphics, language features, heterogeneity in consumer’s
ioral intentions toward a firm (Cronin et al. 2000; Zeithaml
baseline purchase behaviors, and model specifications.
et al. 1996), we collapse these into a single four-item mea-
sure (a ¼ 0.94). Results replicate when satisfaction and
STUDY 3: MANIPULATING willingness to purchase are analyzed separately (see web
CONCRETENESS appendix).
We then measured the hypothesized process, whether
The first two studies suggest that customers are more
people believed that the agent was listening (i.e., attending
satisfied and purchase more when service employees use
to and understanding their needs). Participants responded
more concrete language. Furthermore, the fact that these
to two items adapted from Parasuraman et al. (1991) (“The
results hold across different outcomes, communication mo-
service person gave me personal attention,” “The employee
dalities, product categories, and firms underscores their
generalizability. 6 Our experimental stimuli were designed to have similar levels of
That said, one could still wonder whether the relation- concreteness to the field data from study 1 (M ¼ 329.62, SD ¼ 18.74)
ship is truly causal. While we attempted to rule out alterna- and study 2 (M ¼ 350.46, SD ¼ 15.40).
tive explanations, an even stronger test would be to 7 Participants were also given a definition that read, “By concrete, we
mean it used words that describe something you can identify, imagine,
manipulate linguistic concreteness and see whether it or describe more accurately with your senses (e.g., touch, see, hear).
increases customer satisfaction. Study 3 does this using a For example, the word “lizard” is more concrete than “creature,” and
scenario based on the language observed in study 1. “walk” is more concrete than “go.”" On suggestion of a reviewer, we
also tested a definition that read, “By concrete, we mean it used words
Study 3 also tests the hypothesized mechanism. If using that describe something in a more precise, specific, or clear manner.”
concrete language increases satisfaction by making Results were the same under either approach.
PACKARD AND BERGER 795

understood my specific needs”; 1 ¼ not at all true, 7 ¼ very not vary by condition (F(1, 204) ¼ 1.10, p ¼ .296, g2 ¼
much true; r ¼ 0.69).8 0.005).
Finally, we collected 10 additional items from prior re- Second, rather than being driven by perceived listening,
search (see more detail below) to address alternative perhaps concrete language is just more fluent. To test this,
explanations based on mimicry, fluency, trueness, typical- we adapted Lee and Aaker’s (2004)two-item measure of
ity, or memorability. processing fluency (“How easy was it to process the sales-
person’s reply?,” “How easy was it to understand the sales-
Results person’s reply?”; r ¼ 0.87). Casting doubt on this
Customer Satisfaction. As predicted, using more con- alternative, however, fluency also did not vary by condi-
crete language increased customer satisfaction and willing- tion (F(1, 204) ¼ 1.17, p ¼ .281, g2 ¼ 0.006).
ness to purchase (Mhigh ¼ 3.73, SDhigh ¼ 1.59 vs. Mlow ¼ Third, perhaps concrete language seemed more factual

Downloaded from [Link] by guest on 29 September 2021

3.16; SDlow ¼ 1.51; F(1, 204) ¼ 7.05, p ¼ .009, g2 ¼ or true (Hansen and Wanke 2010), which increased satis-
0.033).9 faction. We asked participants the extent to which the
salesperson’s reply was “factually true” and “consistent
Testing the Process. Furthermore, using more concrete with reality” (r ¼ 0.84). This also did not vary by condition
language increased the perception that employees were lis- (F(1, 204) ¼ 2.66, p ¼ .104, g2 ¼ 0.013).
tening (i.e., attending to and understanding the customer’s Fourth, it may be that employee language in the more
needs; Mhigh ¼ 4.64, SDhigh ¼ 1.66 vs. Mlow ¼ 4.06; SDlow concrete condition seemed more familiar or typical and
¼ 1.83; F(1, 204) ¼ 5.74, p ¼ .017, g2 ¼ 0.027). this drove the effects. To test this possibility, we used a
Finally, as predicted, mediation analysis (PROCESS three-item measure of linguistic typicality (a ¼ 0.88;
model 4; Hayes 2018) confirmed that perceived listening Kronrod, Grinstein, and Wathieu 2012). Typicality, how-
mediated the effect of concrete language on satisfaction ever, did not mediate the effect of concreteness on satisfac-
and purchase (indirect effect ¼ 0.16, 95% CI ¼ 0.03, tion and intent (indirect effect ¼ 0.07, 95% CI ¼ –0.01,
0.29). Using more concrete language made employees 0.16).
seem more attentive to and understanding of the custom- Fifth, maybe the language in the more concrete condi-
er’s needs (b ¼ 0.29, SE ¼ 0.12, t ¼ 2.40, p ¼ .017), boost- tion was more memorable (Paivio 1995; Sadoski 2001),
ing satisfaction and willingness to purchase (b ¼ 0.54, SE
which drove the effect. To test this possibility, we used a
¼ 0.05, t ¼ 10.79, p < .001).
two-item measure of memorability of the employee’s lan-
Alternative Explanations. Ancillary analyses cast guage (memorable, easily remembered; r ¼ 0.75).
doubt on several alternative explanations. First, rather than Memorability, however, also did not mediate the effect of
being driven by employee language, one could wonder concreteness (indirect effect ¼ 0.02, 95% CI ¼ –0.01,
whether the results are driven by the employee mimicking 0.07).
the customer’s words (Moore and McFerran 2017). To test
this possibility, participants were asked “To what extent Discussion
was the salesperson mimicking what you said to them in
their reply? By mimicking, we mean imitating or copying Study 3 provides direct causal evidence that concrete
what you said” (1 ¼ not at all mimicking, 7 ¼ very much language increases customer satisfaction. Participants were
mimicking). Casting doubt on this alternative, mimicry did more satisfied, and more willing to purchase from the firm
in the future, when the employee used more concrete
8 We do not use service research scales purported to measure listen- language.
ing (e.g., de Ruyter and Wetzels 2000; Ramsey and Sohi 1997) be- Furthermore, study 3 provides initial evidence for the
cause they capture what listeners do rather than what listening is process underlying this effect. Linguistic concreteness
(Bodie et al. 2011). That is, these listening scales largely measure pre-
viously hypothesized behavioral consequences of listening (e.g., “The increases customer satisfaction because it makes customers
agent used short, affirmative words and sounds” or “The agent para- think the employee is listening.
phrased what had been said,” de Ruyter and Wetzels 2000; “Seemed Finally, study 3 casts doubt on a number of alternative
bored” or “Didn’t interrupt me,” Ramsey and Sohi 1997) rather than
listening’s underlying psychological facets (attention and understand- explanations including mimicry, fluency, factual truth, fa-
ing). We do not use the full Parasuraman et al. (1991) scale as other miliarity or typicality, and memorability.
items capture factors unrelated to listening (e.g., “has up to date equi-
pment,” “has operating hours convenient to all customers”).
9 As the firm in study 1 used employee helpfulness as their measure STUDY 4: MULTIPLE LEVELS OF
of satisfaction with the employee, we also asked participants “How CONCRETENESS
helpful was the employee’s response” on a 7-point scale. As expected,
this measure was highly correlated with the satisfaction measure (r ¼ Study 3 provides causal evidence of linguistic concrete-
0.87) and replicates the main result when treated as a dependent mea-
sure for the effect of employee concreteness (Mhigh ¼ 3.87 vs. Mlow ¼ ness’s effect, but one could argue that some other factor is
3.15; F(1, 204) ¼ 7.55, p ¼ .007, g2 ¼ 0.036). driving the results. While the mediation results casts doubt
796 JOURNAL OF CONSUMER RESEARCH

TABLE 3: Results
STUDY 4 EXPERIMENT CONDITIONS. Customer Satisfaction. Consistent with the first three
studies, using more concrete language increased customer
Concreteness means
satisfaction and willingness to purchase (b ¼ 0.10, SE ¼
Condition Employee reply Manip. check MRC 0.02, t ¼ 4.78, p < .001). As shown in figure 1, compared
to when employees used less concrete language (e.g., “I’ll
1 “I’ll go look for that” 4.77 295
2 “I’ll go search for that” 5.00 298 go look for that”), using more concrete language (e.g., “I’ll
3 “I’ll go search for that top” 5.15 328 go search for that t-shirt in grey”) increased customer satis-
4 “I’ll go search for that t-shirt” 5.77 353 faction and willingness to purchase.
5 “I’ll go look for that t-shirt in grey” 6.01 366
6 “I’ll go search for that t-shirt in grey” 6.09 368 Testing the Process. Furthermore, using more concrete

Downloaded from [Link] by guest on 29 September 2021

language increased perceptions that the employee was lis-
on this possibility, to further demonstrate its causal impact, tening (b ¼ 0.11, SE ¼ 0.02, t ¼ 5.05, p < .001; figure 1).
study 4 uses multiple levels of concreteness. As predicted, mediation analysis (PROCESS model 4;
Linguistic concreteness can vary in different ways. Hayes 2018) confirmed that perceived listening mediated
concreteness’s effect on customer satisfaction and willing-
Employees might speak more concretely by using more
ness to purchase (indirect effect ¼ 0.07, 95% CI ¼ 0.05,
concrete words to describe something of interest (e.g., “t-
0.10). Using more concrete language made employees
shirt” instead of “top,” or “top” instead of “that”), adding a
seem more attentive to and understanding of the custom-
word to make that thing more vivid (e.g., “grey t-shirt”), or
er’s personal needs (b ¼ 0.11, SE ¼ 0.02, t ¼ 5.08, p <
describing a process in a more tangible way (e.g., “search”
.001), which increased satisfaction and willingness to pur-
instead of “look”). We examine a series of such subtle
chase (b ¼ 0.66, SE ¼ 0.03, t ¼ 23.67, p < .001).
manipulations, each slightly increasing in concreteness,
and measure the subsequent impact on customer satisfac-
Discussion
tion and willingness to purchase.
Study 4 provides further evidence for the effects of lin-
Method guistic concreteness and the process underlying these
effects. First, consistent with the prior three studies, speak-
Participants (N ¼ 481; MTurk) imagined shopping at a ing more concretely increased customer satisfaction and
clothing store and were randomly assigned to one of the willingness to purchase. Showing these effects over a se-
six between-subject conditions (N ¼ 78–85 per condition). ries of subtle linguistic variations underscores concrete-
They liked a t-shirt but could not find the color they ness’s causal impact.
wanted, so they asked an employee for help. Second, as shown in study 3, these effects were driven
The only difference between conditions was the con- by perceived listening. More concrete language made cus-
creteness of the employee’s reply (table 3). The changes tomer service employees seem more attentive to the cus-
are as simple as going from look to search (conditions 1 tomer and understanding of their needs, which increased
vs. 2) from that to that top (2 vs. 3), top to t-shirt (3 vs. 4), satisfaction and willingness to purchase.
or making the t-shirt more concrete by saying its color (4
vs. 5). STUDY 5: TESTING THE PROCESS
Then, participants completed the satisfaction and will-
ingness to purchase items from study 3 (a ¼ 0.88).
THROUGH MODERATION
To test the underlying process, they completed the two- Study 5 has three main goals. First, to offer further evi-
item measure of agent listening from study 3 (r ¼ 0.72). dence of the hypothesized process, we manipulate it (i.e.,
Finally, as a manipulation check, all participants com- “mediation-via-moderation”; Bullock, Green, and Ha
pleted the perceived concreteness measure from the study 2010; Spencer, Zanna, and Fong 2005). If linguistic con-
3 pretest. As expected, effects coding for the linear se- creteness increases customer satisfaction by making it
quence of six conditions for omnibus analysis (i.e., –3, –2, seem like employees are listening, as we suggest, then its
–1, 1, 2, 3) confirmed that participants viewed the language effect should only occur when that concreteness suggests
as increasing in concreteness over the six conditions (see someone is actually listening. If, however, a response is
table 3 for means, b ¼ 0.24, SE ¼ 0.03, t ¼ 7.93, p < not sufficiently related or relevant to the customer’s in-
.001). Furthermore, using MRC scores, as in studies 1 and quiry, being more concrete should not help (maxim of rela-
2, shows the same pattern (b ¼ 13.63, SE ¼ 0.13, tion or relevance; Grice 1975).
t ¼ 103.66, p < .001; table 3). We used OLS regression To test this possibility, in addition to manipulating con-
and effects coding for the linear sequence of six conditions creteness, we orthogonally manipulate the relation (or rele-
for omnibus analysis. vance) of the response to the customer. Specifically, in
PACKARD AND BERGER 797

FIGURE 1

SUBTLE VARIATIONS IN CONCRETENESS DRIVE CUSTOMER SATISFACTION AND WILLINGNESS TO PURCHASE

Downloaded from [Link] by guest on 29 September 2021

addition to a baseline (high relevance response) condition, be consequences of perceived listening, rather than of con-
we add a condition in which the agent mixes up the things creteness itself. When someone attends to what a speaker
the customer asks for (low relevance response). If our theo- is saying, or tries to understand them, it should make them
rizing about listening is correct, more concrete language seem more caring (de Ruyter and Wetzels 2000; Essen and
should increase satisfaction and purchase intent when the Sjoden 1991) and empathic (Castleberry et al. 1999; Spiro
agent’s response is relevant, but not when it is less and Weitz 1990) and may even suggest social closeness
relevant. (Floyd and Morman 1998). So while using concrete lan-
Second, to further assess robustness, we consider several guage might make service representatives seem more car-
additional measures of listening, which we discuss later. ing, this should occur because it seemed like they were
Third, we test several alternative explanations. Under listening. If the employee didn’t seem to be listening, it is
conditions of high value/risk or uncertainty, concreteness unlikely that concrete language would make it seem like
sometimes increases perceptions of trustworthiness and ex- they care.
pertise (Larrimore et al. 2011; Miller et al. 2007; Pan et al. Study 5 tests whether these various factors, rather than
2018; cf. Toma and D’Angelo 2015 for null effects). While listening, explain the effect.
these perceptions seem less likely to matter in more mun-
dane, everyday service encounters (e.g., ordering a pair of Method
pants), one could argue that they might be driving the Participants (N ¼ 415; MTurk) were asked to imagine
effect. they were shopping online and had ordered black dress
Alternatively, rather than through listening, one might shoes for work but wanted to add a pair of grey pants to the
wonder whether concreteness influences customer satisfac- order and called customer service for help. The study used
tion through impacting perceived caring, empathy, or so- a 2 (concreteness: low, high)  2 (relevance: baseline
cial closeness. Importantly, however, these effects should [high], low) between-subjects design.
798 JOURNAL OF CONSUMER RESEARCH

The baseline condition was similar to study 3. In the low expertise) can explain the effects. We also captured meas-
(high) concreteness versions, the employee said ures of perceived effort and helpfulness.11 See the web ap-
“Unfortunately, I can’t just add the clothing [grey pants] pendix for items and reliabilities.
from here. I can cancel the footwear [black dress shoes],
and you can submit a new order” (MRC score ¼ 345.81 Results
[363.68]).
In the less relevant condition, we simply reversed the po- Customer Satisfaction. A main effect of relevance
sition of the pants and the shoes in the agent’s response. (F(1, 411) ¼ 13.72, p < .001, gp2 ¼ 0.032) was qualified
Specifically, the low (high) concreteness employee said by the predicted concreteness by relevance interaction
“Unfortunately, I can’t just add the footwear [black dress (F(1, 411) ¼ 5.95, p ¼ .015, gp2 ¼ 0.014). As in our prior
shoes] from here. I can cancel the clothing [grey pants], studies, in the baseline condition, concreteness enhanced

Downloaded from [Link] by guest on 29 September 2021

and you can submit a new order.” Thus the agent mistakes customer satisfaction and willingness to purchase (Mhigh ¼
the product the customer needed to add to the order (the 4.10, SEhigh ¼ 1.63 vs. Mlow ¼ 3.62, SElow ¼ 1.59; F(1,
pants) for the one they had already ordered (the shoes), 411) ¼ 4.42, p ¼ .036, gp2 ¼ 0.022). When the employee’s
suggesting they failed to attend to and/or did not under- response was less relevant to the issue at hand, however,
stand the customer’s request. While the language is equally the beneficial impact of linguistic concreteness disap-
concrete (or not, depending on concreteness condition), the peared (Mhigh ¼ 3.10, SEhigh ¼ 1.58 vs. Mlow ¼ 3.41,
fact that the agent’s response is not completely relevant SElow ¼ 1.83; F(1, 411) ¼ 1.82, p ¼ .179, gp2 ¼ 0.008)
should moderate the extent to which the agent seems to be
Listening. Similar effects were found on listening. A
listening.
We checked these manipulations in a 2 (concreteness: main effect of relevance (F(1, 411) ¼ 43.77, p < .001, gp2
low, high)  2 (relevance: baseline [high], low) between- ¼ 0.097) was qualified by the predicted concreteness by
subjects pretest (N ¼ 165). Both the concreteness and rele- relevance interaction (F(1, 411) ¼ 9.74, p ¼ .002, gp2 ¼
vance manipulations were supported (see web appendix). 0.023). As in prior studies, in the baseline condition, con-
Next, we measured the dependent variable, using the creteness increased perceived listening (Mhigh ¼ 4.92,
same satisfaction and intention items from studies 3 and 4 SEhigh ¼ 1.45 vs. Mlow ¼ 4.43, SElow ¼ 1.44; F(1, 411) ¼
(a ¼ 0.95). 4.17, p ¼ .042, gp2 ¼ 0.028). When the response was less
To test the mediating role of listening, in addition to the relevant, however, concreteness no longer had a beneficial
measures from studies 3 and 4 (i.e., agent attention and un- effect (Mhigh ¼ 3.29, SEhigh ¼ 1.80 vs. Mlow ¼ 3.85, SElow
derstanding, r ¼ 0.70), we captured measures to further as- ¼ 2.02; F(1, 411) ¼ 5.63, p ¼ .018, gp2 ¼ 0.021).12
sess robustness. First, we used two items from de Ruyter Furthermore, as predicted, the effect of concreteness on
and Wetzels (2000) said to capture listening’s attention and satisfaction was driven by listening. We performed moder-
understanding components (“The agent made an attentive ated mediation analysis (PROCESS model 8; Hayes 2018),
impression,” “The agent attempted to understand what I incorporating relevance as a moderator of concreteness’s
was saying”; r ¼ 0.72). Second, we collected an item effects on listening (a path) in the main process model. As
intended to capture the signaling consequence of listening, in prior studies, the effect of concreteness was driven by
rather than listening itself, from de Ruyter and Wetzels listening in the baseline conditions (conditional indirect ef-
(i.e., responsiveness; “The agent recognized what I fect ¼ 0.17, SE ¼ 0.07, 95% CI ¼ 0.03, 0.31).
needed”).10 Third, we included an explicit measure of em- Concreteness increased perceptions that the employee had
ployee listening (“The agent was listening to me”). These listened (b ¼ 0.24, SE ¼ 0.12, t ¼ 2.04, p ¼ .042), which
six listening items all formed a single component (all load- led to enhanced customer satisfaction and willingness to
ings >.82) in exploratory factor analysis (principal compo- purchase again in the future (b ¼ 0.71, SE ¼ 0.03,
nents analysis) so we combine them into a single index (a t ¼ 21.16, p < .001). When the agent’s response was less
¼ 0.95). Results are the same if we use our original two-
relevant; however, concreteness had a negative effect on
item measure (r ¼ 0.72), or the mean of four items measur-
perceived listening (b ¼ –0.28, SE ¼ 0.12, t ¼ 2.37, p ¼
ing attention and understanding (a ¼ 0.92).
.018), leading to a negative sign for the mediation effect
Finally, we collected 14 additional items from prior re-
search to test whether a variety of alternative explanations
11 We thank a reviewer for this suggestion.
(e.g., caring, empathy, social closeness, trustworthiness, or 12 While the negative effect of linguistic concreteness in the low rel-
evance response condition is interesting, it should be interpreted with
10 As noted earlier, we only use select items from de Ruyter and caution. It is possible that concreteness hurts perceived listening
Wetzels’ (2000) larger scale because other items in it and similar lis- when the response is less relevant, but this may also be an artifact of
tening scales (Castleberry et al. 1999) measure specific behaviors the particular situation used. In this case, concrete language may
that capture what listeners “do” (e.g., maintain eye contact, do not in- have drawn attention to the fact that the employee was not actually
terrupt, use full sentences, ask questions) rather than what listening paying attention to or understanding the request, but this may not al-
“is” (attention and understanding; Bodie et al. 2011). ways be the case.
PACKARD AND BERGER 799

(conditional indirect effect ¼ –0.20, SE ¼ 0.10, 95% CI ¼ Discussion

–0.39, –0.01).
Study 5 underscores the effect of linguistic concreteness
Alternative Explanations. We also tested whether a and the process driving this effect. First, consistent with
number of alternative processes (i.e., perceptions of agent the prior four studies, using more concrete language
caring, two dimensions of empathy, closeness, helpfulness, boosted customer satisfaction and willingness to purchase.
effort, trustworthiness, and expertise) could explain the Second, the results further reinforce the important role
results. Scale items and reliabilities are in the web of perceived listening in these effects. As hypothesized,
appendix. concrete language only increased satisfaction and purchase
To test the explanatory power of these alternatives, we when the agent’s response was relevant to the customer’s
examined whether any of them mediated the effect of con- needs. When the agent used concrete language in a way
creteness on satisfaction and purchase intent in the baseline that suggested they were not actually listening, concrete-

Downloaded from [Link] by guest on 29 September 2021

condition (PROCESS model 4; Hayes 2018). We removed ness no longer had the same benefit. Keeping the actual
listening from the model, and due to high collinearity words used constant across the baseline (relevant) and low
among the items (see correlation matrix in the web appen- relevance response conditions helps cast doubt on alterna-
dix), ran each potential mediator independently. tive explanations.
None of the alternatives mediated the effect of concrete- Third, ancillary analyses cast doubt on the notion that per-
ness on customer satisfaction (caring indirect effect ¼ ceived caring, perspective taking, empathic concern, social
0.03, SE ¼ 0.03, 95% CI ¼ –0.03, 0.10; perspective-taking closeness, helpfulness, effort, trustworthiness, or expertise
indirect effect ¼ 0.03, SE ¼ 0.03, 95% CI ¼ –0.02, 0.11; can explain the effects. We do not propose that some of these
empathic concern indirect effect ¼ –0.01, SE ¼ 0.03, 95% processes play no role, but, as suggested by the literature and
CI ¼ –0.07, 0.05; social closeness indirect effect ¼ –0.05, our results, rather than alternative explanations, some of these
SE ¼ 0.04, 95% CI ¼ –0.13, 0.02; helpful indirect effect ¼ (i.e., caring, perspective taking, empathic concern, and social
0.03, SE ¼ 0.06, 95% CI ¼ –0.08, 0.14; effort indirect ef- closeness) are downstream consequences of listening. Rather
fect ¼ 0.04, SE ¼ 0.04, 95% CI ¼ –0.03, 0.12; trust indi- than explaining the effect of concreteness on satisfaction,
rect effect ¼ 0.02, SE ¼ 0.02, 95% CI ¼ –0.01, 0.06; they help explain why perceived listening increases satisfac-
expertise indirect effect ¼ 0.00, SE ¼ 0.01, 95% CI ¼ – tion and purchase intentions.
0.03, 0.02). This casts doubt on the possibility that caring, Note that an additional study (study 6, reported in the
perspective taking, empathy, closeness, helpfulness, effort, web appendix) provides further evidence of mediation by
trustworthiness, or expertise are driving the effects. moderation. Providing alternate cues that a customer ser-
While these results cast doubt on these constructs as al- vice representative is (rather than is not) listening weakens
ternative processes, we do not mean to suggest that they the benefit of concrete language because the customer al-
play no role. Indeed, as suggested earlier, although they ready makes that inference.
cannot explain the effect of concreteness on customer satis-
faction and intentions by themselves, many of these may GENERAL DISCUSSION
be downstream consequences of listening. When we in-
clude each of these constructs in serial mediation models Consumers and marketers care a lot about the customer
on the b path between listening and the customer outcomes service experience. Consistent with its importance, decades
(PROCESS model 6; Hayes 2018), consistent with prior of academic research have examined how broad strategies
literature, caring, perspective-taking, empathy, social or specific employee actions like apologies, discounts, and
closeness, helpfulness, and effort all function as perceptual flattery shape consumer attitudes and behaviors. But while
consequences of listening, helping to explain listening’s we know about some broad strategies and tactics that may
positive impact on customer satisfaction and willingness to be effective, less is known about how a simple, more fun-
purchase again in the future (caring indirect effect ¼ 0.08, damental feature of social interactions—the words employ-
SE ¼ 0.04, 95% CI ¼ 0.01, 0.16; perspective-taking indi- ees use when speaking to customers—might enhance the
rect effect ¼ 0.08, SE ¼ 0.03, 95% CI ¼ 0.02, 0.15; em- customer experience.
pathic concern indirect effect ¼ 0.05, SE ¼ 0.02, 95% CI The present research helps to address this gap. First, a
¼ 0.01, 0.09; social closeness indirect effect ¼ 0.06, SE ¼ combination of field data and controlled experiments dem-
0.03, 95% CI ¼ 0.01, 0.11; helpful indirect effect ¼ 0.15, onstrates that speaking concretely boosts customer atti-
SE ¼ 0.06, 95% CI ¼ 0.04, 0.28; effort indirect effect ¼ tudes toward the agent, purchase intentions, and actual
0.09, SE ¼ 0.04, 95% CI ¼ 0.02, 0.18). Trust and expertise downstream purchases. Text analysis of over 1,000 cus-
were not supported as downstream consequences of listen- tomer service interactions finds that customers are more
ing (trust indirect effect ¼ 0.03, SE ¼ 0.02, 95% CI ¼ – satisfied (study 1) and purchase more (study 2) after inter-
0.01, 0.08; expertise indirect effect ¼ 0.02, SE ¼ 0.02, acting with service employees who use more concrete lan-
95% CI ¼ –0.02, 0.07). guage. Manipulating concreteness experimentally (studies
800 JOURNAL OF CONSUMER RESEARCH

TABLE 4:

EXAMPLES OF SPEAKING LESS VERSUS MORE CONCRETELY IN CUSTOMER SERVICE.

Less concrete (worse) More concrete (better) Why?

What(adjectives, Those pants are a great choice. Those blue jeans are a great Saying blue jeans should seem
pronouns, nouns) choice. more detailed and context-spe-
cific than pants.
You will receive your refund shortly. You will receive your money back Receiving money should seem
shortly. more tangible and “real” than a
refund.
Would you like anything else? Would you like coffee or tea? Coffee and tea should seem
more specific and tangible than

Downloaded from [Link] by guest on 29 September 2021

anything.
How(adverbs, verbs) Our pie is really good. Our pie is mouthwateringly good. Mouthwateringly should seem
more vivid and imaginable than
really.
We can get that room for you. We can grab that room for you. Grabbing something should
seem more tangible and “real”
than getting it.
I can try to solve that engine issue. I can try to fix that engine issue. Fixing something should seem
more perceptible and imagin-
able than solving it.

3–5) provides more direct causal evidence of its impact. Finally, the fact that the results persist across different
Using concrete language boosts customer satisfaction and communication modalities (i.e., email vs. phone) and con-
willingness to purchase (studies 3–5), even over a series of texts (different firms, customer requests, issues, and lan-
six variations in concreteness within the same short utter- guage) speaks to their generalizability. Whether dealing
ance (study 4). Notably, subtle changes in concreteness with a problem or searching for a product, speaking or
were enough to produce these effects—as little as one writing more concretely increased customer satisfaction,
changed or added word enhanced consumer attitudes and purchase intentions, and actual purchase behavior. We also
intentions. replicated these results in a retail service encounter sce-
Second, using both mediation and moderation, the nario (see study 7 in the web appendix).
results highlight the role of listening in driving these
effects. Speaking more concretely led participants to per- Contributions and Implications
ceive employees as more attentive to and understanding of This research makes several contributions. First, we ex-
their needs, which, in turn, mediated the effect of concrete tend research on language and cognition by revealing that
language on consumer attitudes and intentions (studies 3– concrete language shapes inferences about a speaker’s at-
5). Furthermore, providing conversational evidence that tention and understanding. While prior work has consid-
the employee was not attending to them (study 5) moder- ered why people write or speak more or less concretely
ated the effect. based on the distance of the topic (Fujita et al. 2006;
Third, the studies cast doubt on numerous alternative Snefjella and Kuperman 2015), or how someone’s attitudes
explanations. The effects persisted in the face of over 50 toward a target shape whether they use more or less con-
controls for customer and employee observables including crete language to describe that target (Douglas and Sutton
demographics, tenure or history, language, and other 2006; Schellekens et al. 2013), the present research reveals
interaction-specific features that might otherwise explain that concreteness can signal not only a speaker’s attitudes
these outcomes. Experimental evidence showed that proc- but also their attention (i.e., whether they are listening). A
essing fluency, mimicry, truth perceptions, familiarity, re- speaker’s concreteness generates inferences that they are
call, perceived trustworthiness, and expertise all have attending to and understanding the topics raised by a con-
trouble explaining the relationship between concreteness versational partner. This work contributes to the growing
and customer satisfaction. literature examining the importance of subtle linguistic
Fourth, ancillary results help explain why perceptions devices on information processing, perception, and persua-
that an agent is listening increases customer satisfaction sion (Pogacar et al. 2018).
(study 5). Employees who seem to listen are seen as more Second, from a methodological perspective, using MRC
caring, empathetic, helpful, and hard-working. Good listen- and related linguistic lexicons and text analysis methods is
ers also seem more socially close, offering what is to our a nonstandard (Inman et al. 2018) or boundary-spanning
knowledge a new potential means of communicating psy- approach that can help enrich consumer research (Berger
chological distance. et al. 2020; Moorman et al. 2019). Prior social science
PACKARD AND BERGER 801

research on concrete language often applies the LCM Third, these results have clear implications for improv-
(Semin and Fiedler 1988). The LCM operationalizes con- ing the customer experience. American companies spend
creteness as a four-level typology starting with three verb over a trillion dollars a year recruiting and training employ-
types of increasing abstractness (descriptive action verbs, ees to deliver a great experience (Cespedes and Wallace
e.g., “Starbucks pours coffee,” interpretative action verbs, 2017; Morgan 2017). Most frontline employees are trained
e.g., “Starbucks makes coffee,” and state verbs, e.g., to be knowledgeable about their company’s offering, give
“Starbucks has coffee”) followed by a “most abstract” help, and “show they care” through a variety of exemplar
level involving adjectives (“Starbucks has great coffee”). actions (e.g., apologies, asking questions; Buttle 1996;
Studies in this paradigm present participants with varia- Zeithaml, Berry, and Parasuraman 1996). Many firms even
tions on a single idea (e.g., about Starbucks coffee) across train employees to speak a certain way or use specific
the four levels, and observe participant’s attitudes toward phrases (Bacal 2011; Chen 2011).

Downloaded from [Link] by guest on 29 September 2021

these statements, the object, or the speaker (Schellekens However, the field appears largely unaware of the bene-
et al. 2010, 2013; Semin and Fiedler 1988). fit of concrete language. A review of the customer service
While this approach has shed light on important linguis- literature and correspondence with customer service direc-
tic biases and persuasive effects, it has clear limits. The tors at three firms (an airline, a bricks and mortar retailer,
LCM ignores concreteness variation within each of its lev- and an online retailer) suggested no usage of this idea. This
els. Some descriptive actions (e.g., pouring coffee), for ex- perspective is corroborated by a survey we conducted with
ample, may be more or less concrete than others (e.g., people who have been customer service managers
grinding or roasting coffee), yet the LCM treats all these (N ¼ 209) or employees (N ¼ 100; MTurk). While almost
verbs as the same. What is more, rather than being more all reported having received or given training on how to
abstract, some adjectives (e.g., brown, steaming) may de- speak to customers (97.1% of managers/94.0% of employ-
scribe things more concretely than some verbs (e.g., create, ees), and a majority trained on specific phrases or words to
produce). Consequently, focusing on part of speech catego- use or avoid (68.9%/66.0%), none (0%) reported they had
been trained to speak concretely (unaided recall). When
ries, rather than on the level of words, ignores potentially
subsequently prompted with “speak using concrete words”
important variation. Starbucks could find, for example, that
as one of the 10 key customer service principals, a strong
speaking concretely matters more when talking about in-
majority (82.8%/80.9%) recalled eight items commonly
store processes (e.g., those executed by the barista, such as
described as positive customer service principals (e.g.,
grinding the beans or tamping the grounds) than those that
“ask questions,” “apologize,” “speak clearly”), but a much
happen far from the store (e.g., picking or roasting the
smaller percentage suggested that using concrete words
beans).
was part of their training (28.8%/38.8%). Furthermore, this
There are other issues as well. The LCM excludes words
number was similar to a bogus item included to check for
like nouns, which themselves vary in concreteness (e.g., false recall (“talk quietly”; 32.2%/17.3%) suggesting that a
Starbucks makes coffee vs. beverages vs. drinks). Further, much smaller percentage were actually given such
the same utterance may seem more abstract when it instructions.
describes the target using nouns (Starbucks is a success) The present research suggests that, just by using more
rather than adjectives (Starbucks is successful), suggesting concrete language (e.g., t-shirt rather than top or that),
a potential “fifth level” of concreteness overlooked by the employees can demonstrate that they are listening to the
LCM. This argument is supported by evidence that a word customer’s needs and interests. Rather than saying “That
stem used as a noun may describe people more concretely will be coming right up” or “Your pants are right over
than the same stem as an adjective (e.g., “She’s an athlete” here,” employees should explicitly mention the conversa-
vs. “She’s athletic”; Carnaghi et al. 2008). Thus, Starbucks tion’s subject matter, instead saying “One scoop of Chunky
might find advertisements that say “Starbucks is a success” Monkey coming right up” or “Your grey slacks are right
are more effective than those that say “Starbucks is over here.” Similarly, when taking action on the custom-
successful.” er’s behalf, a waiter might say, “I’ll talk to the chef about
Using the large, machine-readable MRC concreteness your nut allergy” rather than “I’ll tell them about your
lexicon and text analysis methods from computational lin- issue.”
guistics, scholars can now examine linguistic concreteness This relatively subtle approach to signaling attention and
in greater breadth (more parts of speech) and depth (over understanding may be particularly valuable, given consum-
85,000 scored words) than ever before. While text analysis ers often assume employees are trying to manipulate them
is becoming more prevalent in consumer research (Berger (i.e., “schemer schema,” Friestad and Wright 1994).
and Packard 2018; Humphreys and Wang 2018; Packard Generalized suspicion toward employee actions may even
and Berger 2017), the present research offers a small step lead one of the central language tactics taught to firm
toward understanding linguistic concreteness and psycho- employees—the apology—to backfire if not used carefully
logical functions of language more broadly. (Marinova et al. 2018; Wooten 2009). In comparison, using
802 JOURNAL OF CONSUMER RESEARCH

more concrete language seems less likely to backfire. It is We attempted a similar analysis in study 4, the only
unlikely to receive as much conscious attention and, even study that offered sufficient variation in specific parts of
if it is consciously processed, may be less likely to cue con- speech across experimental conditions. Although there was
cerns as it seems unlikely to be seen as intended to flatter, no significant difference between study 4 conditions in
ingratiate, or persuade. verb concreteness,13 there were differences in perceived
Encouraging employees to speak concretely could also concreteness for nouns (that vs. that t-shirt; conditions 2
help improve actual employee listening. First, to be able to vs. 4; F(1, 473) ¼ 11.57, p < .001, gp2 ¼ 0.076) and adjec-
talk about the concrete, specific interests of the customer, tives (t-shirt vs. t-shirt in grey; conditions 4 vs. 6; F(1,
employees would have to attend to the customer’s words 473) ¼ 2.81, p ¼ .094, gp2 ¼ 0.018). Furthermore, both
more carefully. As such, by asking employees to use more these differences were linked to increased satisfaction and
concrete language in referring to the customer’s interests, willingness to purchase (noun contrast conditions 2 vs. 4,

Downloaded from [Link] by guest on 29 September 2021

they should naturally be more attentive and understanding. F(1, 473) ¼ 2.84, p ¼ .093, gp2 ¼ 0.018; adjective contrast
Second, work on behavioral priming under the LCM sug- conditions 4 vs. 6, F(1, 473) ¼ 4.10, p ¼ .043, gp2 ¼
gests that exposure to concrete language can produce a 0.032).
stronger focus on specific, context-dependent features of a We did not design these studies to test the causal impacts
decision or task (Malkoc, Zauberman, and Bettman 2010). of different parts of speech, so these results should be inter-
While it may be difficult for employees to pay attention to preted with caution. They do, however, provide some sug-
individual customers over the long hours of a frontline gestion that nouns and potentially adjectives may more
shift, speaking concretely should help encourage them to strongly drive linguistic concreteness’s effects. Given these
cognitively “tune in” to the specifics of the present situa- words’ respective linguistic roles, it could be that demon-
tion rather than just going through the motions. Additional strating the employee’s attention and understanding of the
examples for managers are offered in table 4. topic of discussion (nouns, adjectives) matters more than
Linguistic concreteness could even be measured to eval- potential behavioral responses; that is, doing things (verbs)
uate employees. As a part of employee assessment and de- on behalf of customers. The concreteness of quantities
velopment, text analysis can be used to evaluate employee (e.g., “many packages” vs. “eleven packages”; Pan et al.
concreteness in relation to other important language fea- 2018) could also be important. Future research should con-
tures (e.g., negativity, Ordenes et al. 2014; pronouns, sider how different parts of speech and other language fea-
Packard et al. 2018). Given most firms record service calls tures shape concreteness’s impact.
and maintain databases of text interactions, transcribing It is also important to remember that these relationships
and analyzing frontline employee language could be done may depend on the context in which they were observed.
at each point in the customer service experience. Verbs could be more important than nouns when customers
are focused on getting employees to take action on their
Limitations and Future Research behalf. If a customer has taken a product back to a store for
a return, whether the employee will take, accept, handle,
This work also suggests some directions for future re- or process the return (verbs) may be more important than
search. One interesting question is whether the effects of focusing on the product (noun). Future research could even
concreteness are moderated by parts of speech (e.g., nouns, examine when concreteness matters more within each part
verbs, and adjectives). Concrete nouns and adjectives may of speech. Take the words we use for people, places, or
lead consumers to infer that the employee is focused on things (i.e., common and proper nouns). While our ap-
objects that are relevant to them personally, while more proach only manipulates or measures common nouns, it
concrete verbs could signal the employee’s attention to im- seems likely that proper nouns would be beneficial over
portant actions (e.g., shipping, providing physical common nouns when they refer to the specific places or
assistance). objects relevant to the customer (e.g., “New York City”
To begin to consider this question, we separately exam- rather than “the city,” “Nike shoes” rather than “athletic
ined the relationship between concreteness and customer shoes”).
satisfaction (study 1) and expenditures (study 2) for nouns, Future research might also consider when and why dif-
verbs, and adjectives using part of speech tagging. In study ferent parts of speech or specific words produce the out-
1, nouns (b ¼ 0.004, SE ¼ 0.002, t ¼ 1.76, p ¼ .080) and comes observed here. Research at the part of speech level
adjectives (b ¼ 0.005, SE ¼ 0.002, t ¼ 1.99, p ¼ .048) but could also inform other social and psychological
not verbs (b ¼ 0.002, SE ¼ 0.003, t ¼ 0.51, p ¼ .611) were
independently linked to customer satisfaction. In study 2, 13 The verb concreteness contrast (conditions 1 and 5 vs. 2 and 6)
nouns (b ¼ 0.37, SE ¼ 0.16, t ¼ 2.31, p ¼ .021) and verbs failed to shift satisfaction and willingness to purchase significantly
(b ¼ 0.66, SE ¼ 0.24, t ¼ 2.68, p ¼ .008) but not adjectives (F(1, 473) ¼ 1.26, p ¼ .26), but this may be due to a weaker shift in
perceived concreteness (F(1, 473) ¼ 0.77, p ¼ .38) between the par-
(b ¼ 0.17, SE ¼ 0.11, t ¼ 1.45, p ¼ .147) were associated ticular verbs we chose to manipulate in the stimuli (look vs. search)
with increased customer spending. (see table 3).
PACKARD AND BERGER 803

phenomena linked to concrete language under the LCM Aerts, Goele and Peeter Verlegh (2018), “Say Your Name: How
and related paradigms (e.g., cognitive biases and persua- Language Concreteness in Online Consumer Reviews
sion). More broadly, analyzing linguistic devices at the Mediates the Effect of Identification on Review
Helpfulness,” International Conference on Research in
part of speech level may represent a rich opportunity to ex- Advertising (ICORIA), Valencia.
pand understanding on how language is processed and per- Bacal, Robert (2011), Perfect Phrases for Customer Service, sec-
suades (Pogacar et al. 2018). ond edition, New York, NY: McGraw-Hill.
It would also be interesting to explore customer interac- Berger, Jonah (2020), The Catalyst: How to Change Anyone’s
tions initiated by the employee. With the possible excep- Mind, New York, NY: Simon & Schuster.
tion of study 4, in which a store employee approached a Berger, Jonah, Ashlee Humphreys, Stephan Ludwig, Wendy Moe,
Oded Netzer, and David Schweidel (2020), “Uniting the
customer, our field data and experimental interactions were
Tribes: Using Text for Marketing Insight,” Journal of
started by the customer. Future research might consider Marketing Research, 84 (1), 1–25.

Downloaded from [Link] by guest on 29 September 2021

whether these effects hold in situations where the customer Berger, Jonah, and Grant Packard (2018), “Are Atypical Things
might not have salient needs or wants to begin with, such More Popular?,” Psychological Science, 29 (7), 1178–84.
as in door-to-door sales visits, cold-calling by telemark- Blankenship, Kevin L., and Traci Y. Craig (2011), “Language Use
eters, or mall intercepts. In these situations, employee con- and Persuasion: Multiple Roles for Linguistic Styles,” Social
creteness might make a potential need or want more “real” and Personality Psychology Compass, 5 (4), 194–205.
Blei, David M. (2012), “Probabilistic Topic Models,”
and tangible to the consumer, increasing their own atten- Communications of the ACM, 55 (4), 77–84.
tion to the conversation. Blundell, Richard, Alan Duncan, and Costas Meghir (1998),
Finally, future marketing research may fruitfully exam- “Estimating Labor Supply Responses Using Tax Reforms,”
ine other verbal, nonverbal, or behavioral cues that signal Econometrica, 66 (4), 827–61.
listening. Things like eye contact, head nods, and assent Bodie, Graham D., [Link]. Cyr, Michelle Pence, Michael Rold,
language such as “uh-huh” and “ok” can suggest someone and James Honeycutt (2011), “Listening Competence in
Initial Interactions I: Distinguishing between What Listening
is paying attention (Gardiner 1971) and asking open-ended is and What Listeners Do,” International Journal of
questions can signal one is trying to gain deeper under- Listening, 26 (1), 1–28.
standing (Huang et al. 2017). Alternatively, it could be Bodie, Graham D. (2012), “Listening as Positive
beneficial to simply take more pauses while speaking Communication,” in The Positive Side of Interpersonal
(Berger 2020; Murphy 2020). Such breaks can signal Communication. ed. T. Socha and M. Pitts, New York: Peter
someone truly wants to hear what their conversation part- Lang, 109–25.
ner has to say. Bullock, John G., Donald P. Green, and Shang E. Ha (2010), “Yes,
but What’s the Mechanism? (Don’t Expect an Easy
In conclusion, the current research demonstrates that Answer),” Journal of Personality and Social Psychology, 98
subtle variations in how service employees talk to custom- (4), 550–8.
ers have important consequences for consumer attitudes, Buttle, Francis (1996), “SERVQUAL: Review, Critique, Research
intentions, and purchases. In doing so, this work deepens Agenda,” European Journal of Marketing, 30 (1), 8–32.
our understanding of language effects in the marketplace Brysbaert, Marc, Amy Beth Warriner, and Victor Kuperman
and on consumer behavior more broadly. (2014), “Concreteness Ratings for 40 Thousand Generally
Known English Word Lemmas,” Behavior Research
Methods, 46 (3), 904–11.
DATA COLLECTION INFORMATION Campbell, Margaret C., and Amna Kirmani (2000), “Consumers’
Use of Persuasion Knowledge: The Effects of Accessibility
The second author collected the field data used in study and Cognitive Capacity on Perceptions of an Influence
1 in late 2017 and early 2018 in the USA. The first author Employee,” Journal of Consumer Research, 27 (1), 69–83.
collected the field data for study 2 in 2004 in Canada. Carnaghi, Andrea, Anne Maass, Sara Gresta, Mauro Bianchi,
Mara Cadinu, and Luciano Arcuri (2008), “Nomina Sunt
Studies 3 and 4 were collected by the first author in the
Omina: On the Inductive Potential of Nouns and Adjectives
winter and spring of 2019, and study 5 was collected by in Person Perception,” Journal of Personality and Social
the first author in the winter of 2020. The survey reported Psychology, 94 (5), 839–59.
in the general discussion was collected in the summer of Castleberry, Stephen B., C. David Shepherd, and Rick Ridnour
2019. All experiments and the survey used Amazon (1999), “Effective Interpersonal Listening in the Personal
Mechanical Turk participants located in the USA. Analysis Selling Environment: Conceptualization, Measurement, and
for all studies was performed by the first author. Nomological Validity,” Journal of Marketing Theory and
Practice, 7 (1), 30–8.
Cespedes, Frank V. and Christopher Wallace (2017), “Executives and
REFERENCES Salespeople Are Misalaigned—And the Effects Are Costly,”
Harvard Business Review, [Link]
Aerts, Goele, Tim Smits, and Peeter Verlegh (2017), “How salespeople-are-misaligned-and-the-effects-are-costly.
Online Consumer Reviews Are Influenced by the Language Chan, Elaine, and Jaideep Sengupta (2010), “Insincere Flattery
and Valence of Prior Reviews: A Construal Level Actually Works: A Dual Attitudes Perspective,” Journal of
Perspective,” Computers in Human Behavior, 75, 855–64. Marketing Research, 47 (1), 122–33.
804 JOURNAL OF CONSUMER RESEARCH

Chen, Adrien (2011), “Working at the Apple Store: Tales from the Personality and Social Psychology Bulletin, 36 (11),
Inside,” [Link] 1576–88.
tales-from-the-inside-5812622. Hayes, Andrew F. (2018), Introduction to Mediation, Moderation
Coltheart, Max (1981), “The MRC Psycholinguistic Database,” and Conditional Process Analysis: A Regression-Based
The Quarterly Journal of Experimental Psychology Section Approach, second edition. New York, NY: The Guilford
A, 33 (4), 497–505. Press.
Cronin, J. Joseph, and Steven A. Taylor (1992), “Measuring Huang, Karen, Michael Yeomans, Alison Wood Brooks, Julia
Service Quality: A Reexamination and Extension,” Journal Minson, and Francesca Gino (2017), “It Doesn’t Hurt to Ask:
of Marketing, 56 (3), 55–68. Question-Asking Increases Liking,” Journal of Personality
Cronin, J. Joseph, Michael K. Brady, and G. Tomas Hult (2000), and Social Psychology, 113 (3), 430–52.
“Assessing the Effects of Quality, Value and Customer Humphreys, Ashlee, and Rebecca Jen-Hui Wang (2018),
Satisfaction on Consumer Behavioral Intentions in Service “Automated Text Analysis for Consumer Research,” Journal
Environments,” Journal of Retailing, 76 (2), 193–218. of Consumer Research, 44 (6), 1274–306.

Downloaded from [Link] by guest on 29 September 2021

Davidow, Moshe (2003), “Organizational Responses to Customer Hyken, Shep (2018), “What Customers Want and Expect,”
Complaints: What Works and What Doesn’t,” Journal of Forbes, [Link]
Service Research, 5 (3), 225–50. 05/what-customers-want-and-expect/#5a4c6f7a7701.
Dawkins, Peter, and Frederick Reichheld (1990), “Customer Inman, Jeffrey, Margaret C. Campbell, Amna Kirmani, and Linda
Retention as a Competitive Weapon,” Directors and Boards, L. Price (2018), “Our Vision for the Journal of Consumer
14 (4), 42–7. Research: It’s All about the Consumer,” Journal of
DePaulo, Bella M., Robert Rosenthal, Judith Rosenkrantz, and Consumer Research, 44 (5), 955–9.
Carolyn Rieder Green (1982), “Actual and Perceived Cues to Kronrod, Ann, Amir Grinstein, and Luc Wathieu (2012), “Enjoy!
Deception: A Closer Look at Speech,” Basic and Applied Hedonic Consumption and Compliance with Assertive
Social Psychology, 3 (4), 291–312. Messages,” Journal of Consumer Research, 39 (1), 51–61.
de Ruyter, Ko, and Martin G. M. Wetzels (2000), “The Impact of Ladhari, Riadh (2008), “Alternative Measures of Service Quality:
Perceived Listening in Voice-to-Voice Service Encounters,” A Review,” Managing Service Quality: An International
Journal of Service Research, 2 (3), 276–84. Journal, 18 (1), 65–86.
Douglas, Karen M., and Robbie M. Sutton (2006), “When What Langacker, R. W. (1987), “Nouns and Verbs,” Language, 63 (1),
You Say about Others Says Something about You: Language 53–94.
Abstraction and Inferences about Describers’ Attitudes and Larrimore, Laura, Li Jiang, Jeff Larrimore, David Markowitz, and
Goals,” Journal of Experimental Social Psychology, 42 (4),
Scott Gorski (2011), “Peer to Peer Lending: The Relationship
500–8.
between Language Features, Trustworthiness, and Persuasion
Drollinger, Tanya, Lucette B. Comer, and Patricia T. Warrington
Success,” Journal of Applied Communication Research, 39
(2006), “Development and Validation of the Active
(1), 19–37.
Empathetic Listening Scale,” Psychology and Marketing, 23
Lee, Angela Y., and Jennifer L. Aaker (2004), “Bringing the
(2), 161–80.
Essen, Louise, and Per-Olow, Sjoden (1991), “Patient and Staff Frame into Focus: The Influence of Regulatory Fit on
Perceptions of Caring: Review and Replication,” Journal of Processing Fluency and Persuasion,” Journal of Personality
Advanced Nursing, 16 (11), 1363–74. and Social Psychology, 86 (2), 205–18.
Fiedler, Klaus (2008), “The Implicit Meta-Theory That Has Lewis, Robert C. and Bernard H. Booms (1983), “The Marketing
Inspired and Restricted LCM Research: Why Some Studies Aspects of Service Quality,” in Emerging Perspectives on
Were Conducted but Others Not,” Journal of Language and Services Marketing, ed. L. Berry, G. Shostack, and G. Upah,
Social Psychology, 27 (2), 182–96. Chicago: American Marketing Association, 99–107.
Floyd, Kory, and Mark T. Morman (1998), “The Measurement of Maas, Anne, Roberta Ceccarelli, and Samantha Rudin (1996),
Affectionate Communication,” Communication Quarterly, “Linguistic Intergroup Bias: Evidence for in-Group-
46 (2), 144–62. Protective Motivation,” Journal of Personality and Social
Friestad, Marian, and Peter Wright (1994), “The Persuasion Psychology, 71 (3), 512–26.
Knowledge Model: How People Cope with Persuasion Maglio, Sam J., Yaacov Trope, and Nira Liberman (2013), “The
Attempts,” Journal of Consumer Research, 21 (1), 1–31. Common Currency of Psychological Distance,” Current
Fujita, Kentaro, Marlone D. Henderson, Juliana Eng, Yaacov Directions in Psychological Science, 22 (4), 278–82.
Trope, and Nira Liberman (2006), “Spatial Distance and Malkoc, Selin A., Gal Zauberman, and James R. Bettman (2010),
Mental Construal of Social Events,” Psychological Science, “Unstuck from the Concrete: Carryover Effects of Abstract
17 (4), 278–82. Mindsets in Intertemporal Preferences,” Organizational
Gardiner, James C. (1971), “A Synthesis of Experimental Studies Behavior and Human Decision Processes, 113 (2), 112–26.
of Speech Communication Feedback,” Journal of Manchanda, Puneet, Grant Packard, and Adithya Pattabhiramaiah
Communication, 21 (1), 17–35. (2015), “Social Dollars: The Economic Impact of Customer
Grice, Herbert P. (1975), “Logic and Conversation,” in Syntax and Participation in a Firm-Sponsored Online Customer
Semantics, Vol 3, Speech Acts, ed. Peter Cole and Jerry L. Community,” Marketing Science, 34 (3), 367–87.
Morgan, New York: Academic Press, 41–58. Marinova, Detelina, Sunil K. Singh, and Jagdip Singh (2018),
Hansen, Jochim, and Michaela Wanke (2010), “Truth from “Frontline Problem-Solving Effectiveness: A Dynamic
Language and Truth from Fit: The Impact of Linguistic Analysis of Verbal and Nonverbal Cues,” Journal of
Concreteness and Level of Construal on Subjective Truth,” Marketing Research, 55 (2), 178–92.
PACKARD AND BERGER 805

Marquardt, Donald W. (1980), “You Should Standardize the Straight (Most of the Time): Top Managers’ Use of Concrete
Predictor Variables in Your Regression Models,” Journal of Language and Its Effect on Investor Reactions,” Strategic
the American Statistical Association, 75 (367), 87–91. Management Journal, 39 (8), 2204–25.
Maxham, James G., III, and Richard G. Netemeyer (2002), “A Parasuraman, A., Leonard L. Berry, and Valarie A. Zeithaml
Longitudinal Study of Complaining Customers’ Evaluations (1988), “SERVQUAL: A Multiple-Item Scale for Measuring
of Multiple Service Failures and Recovery Efforts,” Journal Consumer Perceptions of Service Quality,” Journal of
of Marketing, 66 (4), 57–71. Retailing, 64 (1), 12–40.
Meyer, Andre, and Chris Schwager (2007), “Understanding ——— (1991), “Refinement and Reassessment of the SERVQUAL
Customer Experience,” Harvard Business Review, 2, https:// Scale,” Journal of Retailing, 67 (4), 420–51.
[Link]/2007/02/understanding-customer-experience. Parasuraman, A. and Valarie A. Zeithaml (2002), “Understanding
Miller, Claude H., Lindsay T. Lane, Leslie M. Deatrick, Alice M. and Improving Service Quality: A Literature Review and
Young, and Kimberly A. Potts (2007), “Psychological Research Agenda,” in Handbook of Marketing, ed. Barton A.
Reactance and Promotional Health Messages: The Effects of Weitz and Robin Wensley, Thousand Oaks, CA: Sage,

Downloaded from [Link] by guest on 29 September 2021

Controlling Language, Lexical Concreteness, and the 339–55.
Restoration of Freedom,” Human Communication Research, Parasuraman, A., Valarie A. Zeithaml, and Leonard L. Berry
33 (2), 219–40. (1985), “A Conceptual Model of Service Quality and Its
Moore, Sarah G., and Brent McFerran (2017), “She Said, She Implications for Future Research,” Journal of Marketing, 49
Said: Differential Interpersonal Similarities Predict Unique (4), 41–50.
Linguistic Mimicry in Online Word of Mouth,” Journal of Parker, Maggie (2019), “Is Customer Service the New Therapy?,”
the Association for Consumer Research, 2 (2), 229–45. The New York Times, [Link]
Moorman, Christine, Harald J., van Heerde, C. Page, Moreau, and well/mind/[Link].
Robert W., Palmatier (2019), “Challenging the Boundaries of Pennebaker, James W., Ryan L. Boyd, Kayla Jordan, and Kate
Marketing,” Journal of Marketing, 83 (5), 1–4. Blackburn (2015), “The Development and Psychometric
Morgan, Blake (2017), “Customer Service is a $350 Billion Properties of LIWC2015,” Retrieved from [Link]
Industry, and It’s a Mess,” Forbes, [Link] [Link]/bitstream/handle/2152/31333/
sites/blakemorgan/2017/09/25/customer-service-is-a-350b- LIWC2015_ [Link].
industry-and-its-a-mess/#40e31d3c11be. Pogacar, Ruth, L. J. Shrum, and Tina Lowrey (2018), “The Effects
Murphy, Kate (2020), You’re Not Listening: What You’re Missing of Linguistic Devices on Consumer Information Processing
and Why It Matters, New York, NY: MacMillan (Celadon). and Persausion: A Language Complexity x Processing Mode
Niederhoffer, Kate G., and James W. Pennebaker (2002), Framework,” Journal of Consumer Psychology, 28 (4),
“Linguistic Style Matching in Social Interaction,” Journal of 689–711.
Language and Social Psychology, 21 (4), 337–60. Price, Linda L., Eric J. Arnould, and Patrick Tierney (1995),
Nichols, R. (1948), “Factors in Listening Comprehension,” “Going to Extremes: Managing Service Encounters and
Speech Monographs, 15 (2), 154–63. Assessing Provider Performance,” Journal of Marketing, 59
Ordenes, Francisco Villarroel, Babis Theodoulidis, Jamie Burton, (2), 83–97.
Thorsten Gruber, and Mohamed Zaki (2014), “Analyzing PwC (2018), “PwC Future of Customer Experience,” [Link]
Customer Experience Feedback using Text Mining: A [Link]/us/en/advisory-services/publications/consumer-in-
Linguistics-Based Approach,” Journal of Service Research, telligence-series/pwc-consumer-intelligence-series-cus-
17 (3), 278–95. [Link]#page¼8.
Ott, Myole, Yejin Choi, Claire Cardie, and Jeffrey T. Hancock Ramsey, Rosemary P., and Ravipreet S. Sohi (1997), “Listening to
(2011), “Finding Deceptive Opinion Spam by Any Stretch of Your Customers: The Impact of Perceived Salesperson
the Imagination,” in Proceedings of the 49th Annual Meeting Listening Behavior on Relationship Outcomes,” Journal of
of the Association for Computational Linguistics, 309–19. the Academy of Marketing Science, 25 (2), 127–37.
Packard, Grant, and Jonah Berger (2017), “How Language Shapes Rawlings, John O., Sastry G. Pantula, and David A. Dickey
Word of Mouth’s Impact,” Journal of Marketing Research, (1998), Applied Regression Analysis: A Research Tool, sec-
54 (4), 572–88. ond edition. New York, NY: Springer.
Packard, Grant, Sarah G. Moore, and Brent McFerran (2018), Rust, Roland T., and Tuck Siong Chung (2006), “Marketing
“(I’m) Happy to Help (You): the Impact of Personal Models of Service and Relationships,” Marketing Science, 25
Pronouns Use in Customer-Firm Interactions,” Journal of (6), 560–80.
Marketing Research, 55 (4), 541–55. Sadoski, Mark (2001), “Resolving the Effects of Concreteness on
Paetzold, G. H., and L. Specia (2016), “Inferring Psycholinguistic Interest, Comprehension, and Learning Important Ideas from
Properties of Words,” in Proceedings of the North American Text,” Educational Psychology Review, 13 (3), 263–81.
Association for Computational Linguistics-Human Language Schellekens, Gaby A. C., Peeter Verlegh, and Ale Smidts (2010),
Technologies 2016, 435–40. “Language Abstraction in Word of Mouth,” Journal of
Paivio, Allan (1995), “Imagery and Memory,” in The Cognitive Consumer Research, 37 (2), 207–23.
Neurosciences, ed. ed. M. S. Gazzaniga, Cambridge, MA: ——— (2013), “Linguistic Biases and Persuasion in
MIT Press, 977–86. Communication about Objects,” Journal of Language and
Paivio, Allan, John C. Yuille, and Stephen A. Madigan (1968), Social Psychology, 32 (3), 291–310.
“Concreteness, Imagery, and Meaningfulness Values for 925 Semin, Gun R. (2008), “Language Puzzles: A Prospective
Nouns,” Journal of Experimental Psychology, 76 (1, Pt. 2), Retrospective on the Linguistic Category Model,” Journal of
1–25. Language and Social Psychology, 27 (2), 197–209.
Pan, Lingling, Gerry, McNamara, Jennifer J., Lee, Jerayr (John), Semin, Gun R., and Klaus Fiedler (1988), “The Cognitive
Haleblian, and Cynthia E., Devers (2018), “Give It to Us Functions of Linguistic Categories in Describing Persons:
806 JOURNAL OF CONSUMER RESEARCH

Social Cognition and Language,” Journal of Personality and Trope, Yaacov, and Nira Liberman (2010), “Construal-Level
Social Psychology, 54 (4), 558–68. Theory of Psychological Distance,” Psychological Review,
Smith, Amy K., Ruth N. Bolton, and Janet Wagner (1999), “A 117 (2), 440–63.
Model of Customer Satisfaction with Service Encounters Van Dijk, Olwen, Jazli Kamaruzdin, and Ateeba Anjum (2019),
Involving Failure and Recovery,” Journal of Marketing “The Global Fashion Benchmark Study 2017/2018,” https://
Research, 36 (3), 356–72. [Link]/wp-content/uploads/Global
Snefjella, Bryor, and Victor Kuperman (2015), “Concreteness and %20Fashion%20Benchmark%20Report%202017-2018.v5.3.
Psychological Distance in Natural Language Use,” [Link].
Psychological Science, 26 (9), 1449–60. Wang, Ze, Huifang Mao, Yexin Jessica Li, and Fan Liu (2017),
Snyder, Hannah, Lars Witell, Anders Gustafsson, Paul Fombelle, “Smile Big or Not? Effects of Smile Intensity on Perceptions
and Per Kristensson (2016), “Identifying Categories of of Warmth and Competence,” Journal of Consumer
Service Innovation: A Review and Synthesis of the Research, 43 (5), 787–805.
Literature,” Journal of Business Research, 69 (7), 2401–8. Watson, Kittie W., and Larry L. Barker (1984), “Listening

Downloaded from [Link] by guest on 29 September 2021

Spencer, Steven J., Mark P. Zanna, and Geoffrey T. Fong (2005), Behavior: Definition and Measurement,” Annals of the
“Establishing a Causal Chain: Why Experiments Are Often International Communication Association, 8 (1), 178–97.
More Effective Than Mediational Analysis in Examining Weitz, Barton A., Harish Sujan, and Mita Sujan (1986),
Psychological Processes,” Journal of Personality and Social
“Knowledge, Motivation, and Adaptive Behavior: A
Psychology, 89 (6), 845–51.
Framework for Improving Selling Effectiveness,” Journal of
Spiro, Rosann L., and Barton A. Weitz (1990), “Adaptive Selling:
Conceptualization, Measurement, and Nomological Marketing, 50 (4), 174–91.
Validity,” Journal of Marketing Research, 27 (1), 61–9. Wigboldus, Daniel H. J., Gun R. Semin, and Russell Spears
Spreen, Otfried, and Rudolph W. Schulz (1966), “Parameters of (2000), “How Do We Communicate Stereotypes? Linguistic
Abstraction, Meaningfulness, and Pronunciability for 329 Bases and Inferential Consequences,” Journal of Personality
Nouns,” Journal of Verbal Learning and Verbal Behavior, 5 and Social Psychology, 78 (1), 5–18.
(5), 459–68. Wooten, David B. (2009), “Say the Right Thing: Apologies,
Sriram, S., Pradeep K. Chintagunta, and Puneet Manchanda Reputability, and Punishment,” Journal of Consumer
(2015), “Service Quality Variability and Termination Psychology, 19 (2), 225–35.
Behavior,” Management Science, 61 (11), 2739–824. Zahorik, Anthony J. and Roland T. Rust (1992), “Modeling the
Steil, Lyman K., Larry L. Barker, and Kittie W. Watson (1983), Impact of Service Quality on Profitability: A Review,” in
Effective Listening: Key to Your Success. New York: Advances in Services Marketing and Management, ed. Teresa
McGraw-Hill. A. Swartz, David E. Bowen, and Stephen W. Brown,
Toma, Catalina L., and Jonathan D. D’Angelo (2015), “Tell-Tale Greenwich, CT: JAI Press.
Words: Linguistic Cues Used to Infer the Expertise of Online Zeithaml, Valerie A., Leonard L. Berry, and A. Parasuraman
Medical Advice,” Journal of Language and Social (1996), “The Behavioral Consequences of Service Quality,”
Psychology, 34 (1), 25–45. Journal of Marketing, 60 (2), 31–46.