Estimating Uptake for Reduced-nicotine Cigarettes Using Behavioral Economics

Abstract

Objectives: Lowering the nicotine content in combustible cigarettes may be a viable strategy for reducing dependence and toxin exposure. Understanding how marketing and education may affect initial uptake is an important avenue of inquiry prior to any policy change. There has yet to be an investigation of how framing reductions in nicotine may affect intentions to purchase and consume these cigarettes using the behavioral economic framework. Methods: Participants from Amazon Mechanical Turk completed several tasks, including the Cigarette Purchase Task and Experimental Tobacco Marketplace, under conditions in which a new, reduced-nicotine cigarette alternative is the only cigarette available. Results: Cigarette purchasing was largely unaffected by stated nicotine concentration, but lower concentrations suggested the potential of small estimated compensatory purchasing. Exposure to a narrative detailing how others have perceived the negative subjective effects of lower nicotine cigarettes (eg, less satisfaction) significantly reduced the perceived value of cigarettes. Conclusions: These results suggest information about nicotine content alone is unlikely to reduce initial uptake without accompanying narratives about the effects of this reduced-nicotine content.

Full text

264
Cigarette smoking is the leading prevent-
able cause of death in the United States
(US)1 and incurs more than $300 billion in
healthcare costs annually.2 Understanding the vari-
ables that maintain cigarette use, especially the role
of nicotine,3 is an important avenue of inquiry that
informs tobacco/nicotine control policies aimed at
reducing smoking.4 Research shows that substantial
reductions in nicotine content in tobacco cigarettes
can result in lower exposure to toxins and reduce
dependence.5,6 Towards this end, the Family Smok-
ing Prevention and Tobacco Control Act, passed
in 2009, expanded the purview of the US Food
and Drug Administration (FDA) to allow broader
policy implementation. e FDA has expressed
interest in investigating the potential policy eects
of reducing nicotine content in cigarettes and also
has released an Advanced Notice of Proposed Rule-
making related to nicotine content reductions.7
Emerging research suggests that some smok-
ers misunderstand the role of nicotine in terms of
health risks8 and addictiveness of reduced-nicotine
cigarettes.9,10 For example, some smokers inac-
curately attribute smoking-related diseases such
as asthma and lung cancer to nicotine.11–14 us,
smokers who hold these incorrect risk perceptions
Brent A. Kaplan, Postdoctoral Associate, Fralin Biomedical Research Institute at VTC, Roanoke, VA. Derek A. Pope, Postdoctoral Associate, Fralin Bio-
medical Research Institute at VTC, Roanoke, VA. William B. DeHart, Postdoctoral Associate, Fralin Biomedical Research Institute at VTC, Roanoke,
VA. Jerey S. Stein, Research Assistant Professor, Fralin Biomedical Research Institute at VTC. Warren K. Bickel, Professor, Fralin Biomedical Research
Institute at VTC, Roanoke, VA. Mikhail N. Koarnus, Research Assistant Professor, Fralin Biomedical Research Institute at VTC, Roanoke, VA.
Correspondence Dr Koarnus; mickyk@vtc.vt.edu
Estimating Uptake for Reduced-nicotine
Cigarettes Using Behavioral Economics
Brent A. Kaplan, PhD
Derek A. Pope, PhD
William B. DeHart, PhD
Jerey S. Stein, PhD
Warren K. Bickel, PhD
Mikhail N. Koarnus, PhD
Objectives: Lowering the nicotine content in combustible cigarettes may be a viable strategy for
reducing dependence and toxin exposure. Understanding how marketing and education may
aect initial uptake is an important avenue of inquiry prior to any policy change. There has yet
to be an investigation of how framing reductions in nicotine may aect intentions to purchase
and consume these cigarettes using the behavioral economic framework. Methods: Participants
from Amazon Mechanical Turk completed several tasks, including the Cigarette Purchase Task
and Experimental Tobacco Marketplace, under conditions in which a new, reduced-nicotine
cigarette alternative is the only cigarette available. Results: Cigarette purchasing was largely
unaected by stated nicotine concentration, but lower concentrations suggested the potential
of small estimated compensatory purchasing. Exposure to a narrative detailing how others have
perceived the negative subjective eects of lower nicotine cigarettes (eg, less satisfaction) sig-
nicantly reduced the perceived value of cigarettes. Conclusions: These results suggest infor-
mation about nicotine content alone is unlikely to reduce initial uptake without accompanying
narratives about the eects of this reduced-nicotine content.
Key words: behavioral economics; nicotine reduction; cigarette purchase task; experimental tobacco marketplace; demand;
cigarettes; humans
Tob Regul Sci.™ 2019;5(3):264-279
DOI: https://doi.org/10.18001/TRS.5.3.5

Kaplan et al
Tob Regul Sci.™ 2019;5(3):264-279 265 DOI: https://doi.org/10.18001/TRS.5.3.5
might be less likely to reduce smoking if they be-
lieve reduced-nicotine cigarettes are less harmful.9
Likewise, smokers may be less likely to switch to
safer nicotine alternatives (eg, nicotine replace-
ment therapies) if they believe reduced-nicotine
cigarettes can be used as a smoking cessation prod-
uct.10 Indeed, research suggests greater perceived
nicotine content is associated with greater per-
ceived risk and harm.8 To date, few studies have
assessed relationships between risk perceptions and
subsequent smoking behavior of reduced-nicotine
cigarettes.8,15 In one study,15 researchers found that
after viewing an unaltered, company-created smok-
ing advertisement for a reduced-nicotine cigarette
before smoking the cigarette, participants per-
ceived the reduced-nicotine cigarette as safer than
conventional cigarettes; however, neither partici-
pants’ beliefs nor subjective ratings of reduced-nic-
otine cigarettes directly aected smoking behavior.
Rather, an interaction between subjective ratings
and beliefs was associated with subsequent smok-
ing, with lower subjective ratings and greater false
beliefs associated with greater smoking. Pacek et
al8 also found that smokers’ perceptions of nico-
tine content, but not actual nicotine content, were
positively associated with perceptions of harm. To
date, this research has focused largely on risk per-
ceptions under conditions where reduced-nicotine
cigarettes are conveyed (or perceived) as very low,
low, moderate, and high. Although the primary
goal of a reduced-nicotine policy would be to re-
duce actual smoking behavior, considerations and
prospective methods for how the general public
may react to such a policy are important.16
Apart from assessing perceptions of reduced-nic-
otine cigarettes, which is an important avenue of in-
quiry for a sweeping public policy initiative, several
studies have examined the abuse liability of these
cigarettes using methods from behavioral econom-
ics.17–21 e Cigarette Purchase Task (CPT), one
rapid assay to model cigarette demand,22,23 allows
for a quick determination of cigarette value and
price sensitivity by asking respondents to estimate
the number of cigarettes they would purchase and
consume at a range of escalating monetary prices.
Whereas risk perceptions may be a useful indicator
of subsequent smoking behavior, the CPT also may
be used to prospectively estimate purchasing and
use based on product descriptions.24
Research investigating the eects of reduced-nico-
tine cigarettes on behavioral economic demand has
done so only after participants experience the ciga-
rettes and research suggests that substantial levels
of cigarette smoking continues for at least 6 weeks
following a switch to reduced-nicotine cigarettes.5
Any policy change that would limit the amount of
nicotine in cigarettes would likely be announced
prior to cigarette smokers sampling the reduced-
nicotine cigarettes15 and messaging that hastens
reductions in smoking after this policy change
may dramatically reduce overall cigarette exposure.
erefore, the purpose of the current study was to
address how current cigarette smokers’ intentions
to purchase reduced-nicotine cigarettes might be
aected by various ways of describing the nicotine
content in these cigarettes compared to the amount
of nicotine in their usual-brand cigarettes. Across 3
experiments, our main research question was how
the framing of nicotine concentration in a new type
of cigarette aected 2 key aspects of behavioral eco-
nomic demand: intensity (purchasing under unre-
stricted cost) and elasticity (purchasing sensitivity
to price; ie, cigarette valuation). ese 2 demand
measures provide insight into how smokers might
perceive and respond (via their purchasing inten-
tions) to a nicotine reduction policy. Although
some research suggests individuals misunderstand
the role of nicotine in cigarettes related to health
risks, based on previous in-lab research examining
reduced-nicotine cigarettes and behavioral eco-
nomic measures, we hypothesized that reductions
in nicotine would be associated with reductions in
demand intensity and elasticity. We also investigat-
ed how participant demographic variables related
to these behavioral economic measures.
Experiment 1 evaluated the eects of a stated
concentration, framed as a nicotine percentage.
Experiment 2 attempted to replicate the results of
Experiment 1 when framing nicotine percentage
as a reduction from participant’s usual-brand ciga-
rette. Finally, Experiment 3 examined the eects of
an undesirable narrative description of the subjec-
tive eects of reduced-nicotine cigarettes. In all 3
experiments, we also examined the eects of nico-
tine concentration on alternative product purchas-
ing in the Experimental Tobacco Marketplace – an
online, simulated virtual marketplace; however, we
observed few direct eects of nicotine framing on
other product purchasing. For openness and trans-

Estimating Uptake for Reduced-nicotine Cigarettes Using Behavioral Economics
266
parency, we include methods, analyses, and results
related to these procedures in the Supplemental
Information (https://osf.io/ebqr8/) but do not dis-
cuss the results here.
EXPERIMENT 1 METHODS
Participants
Participants recruited from Amazon Mechanical
Turk (mTurk) had to: (1) reside in the US; (2) have
a task approval rate of ≥90%; (3) have completed
≥50 approved tasks; and (4) report current smok-
ing on a brief qualication test. Overall, 496 work-
ers participated in the experiment, which required
approximately 24 minutes. Participants were paid
$3.00 for completing the experiment (mean real-
ized hourly wage of $7.54).
Procedures
All tasks were administered via Qualtrics Re-
search Suite (www.qualtrics.com). Participants rst
completed an abbreviated Timeline Followback25
(for use in the Experimental Tobacco Marketplace),
followed by a baseline CPT22 for their usual-brand
cigarettes. Participants reported the number of
cigarettes they would purchase and consume at 16
ascending prices. Participants were presented with
general instructions and constraints (eg, imagine
you have the same income/savings as you do now)
used in previous CPT research26 (see Supplemental
Information for details; https://osf.io/ebqr8/).
After completing the baseline CPT, participants
were randomly assigned to one of 6 groups dier-
ing with respect to cigarette nicotine concentration
associated with a new type of cigarette, referred to
as a percentage of nicotine compared to their usual-
brand cigarette. Nicotine concentrations included
100% (current market control), 60%, 30%, 15%,
8%, and 2%. ese specic percentages (except
60%) were chosen because they approximately
match the nicotine contents of investigational
cigarettes used in previous and ongoing reduced-
nicotine research studies (RTI SPECTRUM Ciga-
rettes, 22nd Century). Participants read a vignette
that described a new type of cigarette on the mar-
ket available from the participants’ usual brand
manufacturer, hereafter termed the variable-nico-
tine cigarette (see Supplemental Information for
the full vignette; https://osf.io/ebqr8/). Below the
instructions, participants were required to type the
percentage amount of nicotine and answer a mul-
tiple-choice attending question to proceed through
the remainder of the task. Participants then com-
pleted another CPT for the variable-nicotine ciga-
rette. e instructions and price sequence were
identical except for one assumption that stated:
“e available cigarettes are the new cigarettes with
XX% the amount of nicotine compared to the old
cigarettes,” where “XX%” was one of the nicotine
percentages listed above associated with random
group assignment. e experiment ended with
the Experimental Tobacco Marketplace, followed
by the Fagerström Test of Cigarette Dependence27
(FTCD) and general demographics.
Data Analysis
All data analyses were conducted in R Statistical
Software Version 3.3.2.28 Participant characteris-
tics (sex, education, employment, age, number of
cigarettes smoked per day, FTCD) were compared
across groups using either chi-square test of inde-
pendence or one-way ANOVA. Responses on both
CPTs were examined for systematic responding
per 3 criteria that are typically indicative of inat-
tention or misunderstanding of the task: trend (ie,
invariant or ascending demand curves), as well as
bounce and reversal from zero criteria (variable or
inconsistent purchasing).29 Individual datasets fail-
ing at least one of the criteria (N = 22; 4.4% of
full sample) were removed from the demand analy-
ses. Additionally, 5 participants reported smoking
>100 cigarettes in one day on the Timeline Follow-
back and were removed from all analyses. For the
demand tasks, we applied an exponentiated func-
tion30 based on the exponential demand31 equation
using the beezdemand package in R:32
Equation 1:
where Q represents cigarettes purchased, Q0 (ie,
intensity) is the estimated number of cigarettes
purchased at free price, k is a weighting parameter
signifying the range of consumption in logarithmic
units, α is the rate of change in elasticity across the
entire curve (ie, elasticity), and C is the price per
cigarette. For all experiments, we used a value of
2.54 for k (calculated as a shared parameter across
all datasets33).
    

Kaplan et al
Tob Regul Sci.™ 2019;5(3):264-279 267 DOI: https://doi.org/10.18001/TRS.5.3.5
We logarithmically transformed elasticity prior
to regression analyses (no changes were made to
intensity), then agged outliers for intensity and
elasticity if they exceeded 3.29 SDs,34 and excluded
them for the relevant analyses. Using multiple re-
gression, we examined the eects of concentration
amount on intensity and elasticity. Based on statis-
tically signicant intercorrelations of various mea-
sures, we included several demographic variables in
the multiple regression to: (1) examine the relations
between these variables and demand measures, and
(2) isolate the potential eects of concentration
amount on demand measures. Partial eta squared
(η"
#)
is reported and was obtained using the sjstats
package.36 Post hoc comparisons of marginal means
between groups were accomplished using the em-
means package,37 with Holm-Bonferroni35 adjust-
ments and weighted cell means.
EXPERIMENT 1 RESULTS
Demographics
e second column of Table 1 displays overall
participant demographics for Experiment 1 (Ta-
ble S1 in the Supplemental Information [https://
osf.io/ebqr8/] displays demographics among the
concentration groups). We did not observe any
statistically signicant dierences in demographic
variables across the 6 groups. Spearman rank-order
correlations between income, age, cigarettes per
day, FTCD, and demand measures are reported in
Table S2.
Eects of Concentration on Cigarette Demand.
Equation 1 provided an excellent t to the data
(Mdn R2 = .97, IQR = .96, .98) resulting in a medi-
an elasticity (α) of 0.0101 (IQR = 0.0056, 0.0191)
and median intensity (Q0) of 20.30 (IQR = 10.73,
25.99). Twenty participants (4% of the full sam-
ple) displayed intensity values exceeding 3.29 SDs
and were excluded along with 2 participants who
reported “other” for their sex (when examining the
eect of a categorical variable such as sex, a small
group size [N = 2] may otherwise obfuscate mean-
ingful main eects), and one participant who did
not report income.
Table 2 depicts the F-statistic and corresponding
(eect size) associated with each predictor variable
used in the multiple linear regression models across
all 3 experiments. We observed no dierences across
groups in derived baseline CPT intensity (see top
third of Table 2) while controlling for sex, income,
age, cigarettes per day, and FTCD score. A statisti-
cally signicant eect of concentration was found
for derived variable-nicotine CPT intensity when
controlling for baseline intensity and the aforemen-
tioned demographic variables. Post hoc comparisons
of marginal means of variable-nicotine CPT intensi-
ty revealed participants in the 100% framing group
estimated purchasing fewer cigarettes if they were
free compared to the other concentration groups
(see Table S3 for all post hoc comparisons). No dif-
ferences in estimated purchasing were found be-
tween any of the other concentration groups – that
is, participants reported purchasing more cigarettes,
but increases in purchasing were not systematically
related to nicotine concentration.
Several participant demographic variables were
signicantly related to CPT intensity. With all else
in the model being equal, males reported great-
er baseline CPT intensity (b = 4.48, SE = 1.48)
compared to females, and older age was associated
with lower baseline CPT intensity (b = -0.15, SE =
0.007). Additionally, cigarettes smoked per day (b
= 0.91, SE = 0.12) and FTCD score (b = 0.92, SE
= 0.38) signicantly positively predicted baseline
CPT intensity. Sex (men reporting 2.25 more ciga-
rettes) and cigarettes per day (b = 0.25, SE = 0.09)
were signicantly associated with variable-nicotine
CPT intensity.
Eight participants had elasticity values exceed-
ing 3.29 SDs and were excluded from the elasticity
analysis. No dierences in either baseline or vari-
able-nicotine CPT elasticity were observed (Table
2) when controlling for demographic variables,
suggesting reductions in nicotine concentration did
not signicantly aect cigarette price sensitivity, the
primary measure of cigarette valuation. In terms of
participant demographics, both cigarettes smoked
per day and FTCD score signicantly predicted
baseline CPT elasticity in the expected direction;
that is, greater number of cigarettes smoked per day
(b = -0.034, SE = 0.008) and higher FTCD scores
(b = -0.113, SE = 0.027) predicted lower elasticity
values and thus, higher cigarette valuation.
EXPERIMENT 1 DISCUSSION
e primary aim of Experiment 1 was to deter-
mine whether the stated concentration of nicotine

Estimating Uptake for Reduced-nicotine Cigarettes Using Behavioral Economics
268
in a novel, variable-nicotine cigarette would be re-
lated systematically to demand for cigarettes. To
our knowledge, this is the rst investigation look-
ing at hypothetical outcomes as they relate to dif-
ferent cigarette nicotine concentrations.
Given the literature examining cigarette demand
for reduced-nicotine cigarettes, we had originally
hypothesized that demand would be related to
concentration amount. On the contrary, we did
not nd parametric dierences in either demand
intensity or elasticity as a function of concentra-
tion amount suggesting the stated percentage of
nicotine does not appreciably aect cigarette valu-
ation. Interestingly, we found participants exposed
Table 1
Experiments 1-3 Overall Demographics
Experiment 1
(N = 491)
Experiment 2
(N = 212)
Experiment 3
(N = 178)
Variable (Mean [SD])
Age (years) 36.63 (10.77) 34.57 (9.98) 34.73 (9.12)
Cigarettes Smoked/Day 14.60 (8.49) 14.49 (8.95) 14.66 (6.92)
FTCDa4.34 (2.46) 4.22 (2.63) 4.32 (2.32)
Variable (N [%])
Sex
Women 287 (58.5) 121 (57.1) 79 (44.4)
Men 202 (41.1) 91 (42.9) 99 (55.6)
Other 2 (0.4) 0 (0) 0 (0)
Education*
Less than High School 3 (0.6) 2 (0.9) 0 (0)
High School/GED 67 (13.6) 30 (14.2) 27 (15.2)
Some College 165 (33.6) 67 (31.6) 49 (27.5)
2-Year College Degree (Associates) 84 (17.1) 28 (13.2) 32 (18.0)
4-Year College Degree (BA, BS) 128 (26.1) 70 (33.0) 58 (32.6)
Master’s Degree 34 (6.9) 12 (5.7) 9 (5.1)
Professional Degree (MD, JD, DDS, DVM, PsyD) 4 (0.8) 1 (0.5) 3 (1.7)
Doctorate (PhD, DSc, EdD, DFA) 6 (1.2) 2 (0.9) 0 (0)
Employment
Employed 395 (80.4) 168 (79.2) 163 (91.6)
Unemployed 85 (17.3) 42 (19.8) 0 (0)
Retired 11 (2.2) 2 (0.9) 15 (8.4)
User Type
Cigarette Only 302 (61.5) 147 (69.3) 115 (64.6)
Cigarette & ENDSb Only 82 (16.7) 41 (19.3) 42 (23.6)
Cigarette & NRTc Only 40 (8.1) 7 (3.3) 6 (3.4)
Cigarettes & > 1 Product 67 (13.6) 17 (8.0) 15 (8.4)
Note.
*Only signicant difference detected for Experiment 2, p = .020.
a: FTCD: Fagerström Test of Cigarette Dependence
b: ENDS: Electronic Nicotine Delivery System
c: NRT: Nicotine Replacement Therapy

Kaplan et al
Tob Regul Sci.™ 2019;5(3):264-279 269 DOI: https://doi.org/10.18001/TRS.5.3.5
to the 100% frame estimated they would purchase
fewer cigarettes when cigarettes were free compared
to participants in any of the other concentration
groups but found no evidence suggesting concen-
tration amount inuenced elasticity of demand.
e increased intensity observed in the reduced-
nicotine groups compared to the 100% group may
suggest a perceived need to compensate in order
to obtain the feelings associated with participant’s
usual-brand cigarette, which implicitly contains
100% nicotine.
Because we were unable to detect systematic rela-
tions between demand and concentration amount,
we conducted a second experiment to investigate
Table 2
Multiple Regression Predicting Cigarette Purchase Task Intensity and Elasticity
Baseline CPTa
Intensity
Variable Nicotine
CPT Intensity
Baseline CPT
Elasticity
Variable Nicotine
CPT Elasticity
Regression Term F F F F
Experiment 1
Intercept 11.27* (.03) 9.27* (.02) 276.23* (.38) 0.66 (.00)
Concentration 1.19 (.01) 7.65* (.08) 0.73 (.01) 0.77 (.01)
BL Intensity/Elasticity 659.04* (.60) 998.77* (.69)
Sex 9.12* (.02) 4.09* (.01) 0.09 (.00) 1.69 (.00)
Income 0.76 (.00) 3.60 (.01) 0.06 (.00) 0.82 (.00)
Age 4.82* (.01) 1.87 (.00) 3.40 (.01) 0.44 (.00)
Cigarettes/Day 61.84* (.12) 7.17* (.02) 17.97* (.04) 0.87 (.00)
FTCDb5.69* (.01) 0.15 (.00) 17.56* (.04) 2.63 (.01)
Experiment 2
Intercept 9.51* (.05) 0.45 (.00) 121.69* (.39) 0.48 (.00)
Amount 0.02 (.00) 7.60* (.04) 4.38* (.02) 1.50 (.01)
Frame 0.13 (.00) 0.44 (.00) 0.11 (.00) 0.27 (.00)
BL Intensity/Elasticity 12.19* (.06) 178.02* (.50)
Sex 0.17 (.00) 0.02 (.00) 0.08 (.00) 0.97 (.01)
Income 0.49 (.00) 0.01 (.00) 2.69 (.01) 1.04 (.01)
Age 2.75 (.02) 1.28 (.01) 0.68 (.00) 0.75 (.00)
Cigarettes/Day 21.29* (.11) 17.43* (.09) 15.48* (.08) 0.26 (.00)
FTCD 2.23 (.01) 5.46* (.03) 0.19 (.00) 0.18 (.00)
Experiment 3
Intercept 11.17* (.06) 0.14 (.00) 101.69* (.38) 1.39 (.01)
Concentration 1.18 (.01) 3.68* (.04) 0.50 (.01) 14.85* (.15)
BL Intensity/Elasticity 654.54* (.80) 258.64* (.61)
Sex 0.52 (.00) 0.04 (.00) 1.08 (.01) 0.60 (.00)
Income 3.31 (.02) 0.37 (.00) 5.26* (.03) 0.26 (.00)
Age 2.66 (.02) 0.27 (.00) 1.44 (.01) 0.06 (.00)
Cigarettes/Day 6.66* (.04) 0.00 (.00) 0.96 (.01) 2.12 (.01)
FTCD 0.35 (.00) 1.63 (.01) 6.90* (.04) 0.40 (.00)
Note.
* p < .05
a: CPT = Cigarette Purchase Task
b: FTCD = Fagerström Test of Cigarette Dependence
(
)
(
)
(
)
(
)

Estimating Uptake for Reduced-nicotine Cigarettes Using Behavioral Economics
270
the eects of framing nicotine concentration. Spe-
cically, we kept all aspects from Experiment 1 con-
stant, but isolated the 100% and 2% concentration
amounts (as we found no systematic dierences in
demand parameters across the intermediate con-
centrations) and reframed concentration as a re-
duction in the amount of nicotine in the cigarettes
(0% reduction, 98% reduction). We also sought to
replicate Experiment 1’s ndings by using 2 of the
original concentration percentages (100%, 2%).
EXPERIMENT 2 METHODS
Participants
We recruited participants using mTurk as de-
scribed in Experiment 1 and workers who partici-
pated in Experiment 1 were not able to participate
in Experiment 2. Altogether, 214 workers com-
pleted the experiment, which required an average
of approximately 23 minutes to complete. Partici-
pants were paid $3.00 for completing the survey
(mean realized hourly wage of $7.76).
Procedures
Tasks were identical to those in Experiment 1
with the following exception. Two groups received
a modied framing of the concentration, framed as
a reduction in the amount of nicotine, and 2 groups
received the same original framing as in Experiment
1. e 4 groups included: 100% (current market
control), 2%, 0% reduction, and 98% reduction.
Data Analysis
Data analyses were conducted similarly as in Ex-
periment 1, except we compared demand parame-
ters using 2-way analysis of covariance (ie, multiple
regression) for variables of concentration amount
(ie, 100%, 2%) and frame (ie, no frame, reduction
frame). In this experiment, we excluded 2 partici-
pants from all analyses for reporting smoking >100
cigarettes in one day on the Timeline Followback.
Additionally, 12 and 7 participants displayed in-
tensity and elasticity values greater than 3.29 SDs
from the respective means and were excluded from
their respective analyses.
EXPERIMENT 2 RESULTS
Demographics
e third column of Table 1 displays overall par-
ticipant demographics for the current experiment,
which were largely similar to Experiment 1. e
only statistically signicant dierence across the
4 groups was in education (Table S5). Spearman
rank-order correlations among the variables are dis-
played in Table S6.
Eects of Concentration and Framing on
Cigarette Demand
Nine participants failed systematic criteria for ei-
ther version of the CPT, which reected a relatively
small 4.25% of the full sample, and were excluded
from subsequent analyses. Equation 1 provided an
excellent t to the data (Mdn R2 = 0.98, IQR =
0.96, 0.98) resulting in a median elasticity of 0.01
(IQR = 0.0057, 0.0221) and median intensity of
21.07 (IQR = 11.61, 27.24). We observed no sta-
tistically signicant dierences in baseline CPT in-
tensity while controlling for demographic variables
as a function of concentration amount or frame (see
middle of Table 2). Cigarettes per day was the only
statistically signicant predictor of baseline CPT
intensity. Examination of variable-nicotine CPT
revealed amount, but not frame signicantly pre-
dicted intensity. Exposure to the 2% concentration
amount resulted in higher intensity compared to the
100% amount (b = 4.57, SE = 1.66). Additionally,
number of cigarettes smoked per day and FTCD
positively and signicantly predicted variable-nico-
tine CPT intensity (b = 0.60, SE = 0.14; b = 1.05,
SE = 0.45, respectively). ese results are consistent
with those found in Experiment 1 suggesting in-
creased cigarette purchasing was inuenced by the
low, stated concentration and was not aected by
framing nicotine amount as a reduction.
We observed no statistically signicant dierenc-
es in baseline CPT elasticity as a function of frame,
but did with concentration amount when control-
ling for demographic variables. Cigarettes per day
negatively and signicantly predicted baseline CPT
elasticity (b = -0.053, SE = -0.014); thus, greater
cigarettes per day were associated with greater ciga-
rette valuation. Consistent with our ndings from
Experiment 1, we observed no statistically signi-
cant dierences in variable-nicotine CPT elasticity
as a function of concentration amount or frame
when controlling for baseline elasticity and demo-
graphic variables. Baseline CPT elasticity signi-
cantly predicted variable-nicotine CPT elasticity.

Kaplan et al
Tob Regul Sci.™ 2019;5(3):264-279 271 DOI: https://doi.org/10.18001/TRS.5.3.5
EXPERIMENT 2 DISCUSSION
Results from Experiment 2 suggested that con-
centration amount, but not framing signicantly
altered demand intensity; however, neither manip-
ulation inuenced elasticity. Specically, exposure
to the 2% concentration amount resulted in higher
intensity compared to the 100% amount (approxi-
mately 4.6 cigarettes higher; b = 4.57, SE = 1.66).
Notably, these results are consistent with the eects
observed in Experiment 1.
erefore, we conducted a nal follow-up experi-
ment with 2 aims. First, we attempted to replicate
our ndings from Experiments 1 and 2 with re-
spect to dierential changes in demand intensity
based on a specied concentration amount (100%
and 2%). Because results from Experiment 2 sug-
gested a simple reduction framing was not eective
in altering cigarette elasticity, we leveraged ideas
from narrative theory36 and sought to test whether
providing an “undesirable” narrative description
associated with the 2% variable-nicotine cigarettes
would alter cigarette elasticity. Briey, narrative
theory suggests that stories or anecdotes related to
someone else’s experiences may be eective in inu-
encing decision-making, especially when compared
to information alone. For example, these narratives
have been shown to inuence real-world decisions
related to health outcomes (eg, scheduling vaccina-
tions,37 driving while under the inuence of alco-
hol38). Relevant to the current study, however, is
that research has shown narratives are eective for
promoting substitution of electronic cigarettes (a
harm-reduction method24) and reducing cigarette
smoking.39,40 Ne et al39 found media ads featur-
ing negative consequences of smoking cigarettes re-
sulted in increased quit attempts and quit successes
since the inception of the US Centers for Disease
Control and Prevention’s “Tips from Former
Smokers” ad campaign. us, evaluating whether
a narrative based on actual feedback from smokers
who have experienced reduced-nicotine cigarettes
will reduce intentions to smoke would help inform
marketing and education eorts.
EXPERIMENT 3 METHODS
Participants
Participants were recruited from mTurk consis-
tent with Experiments 1 and 2. Overall, 188 work-
ers participated and task duration took an average
of approximately 22 minutes. Participants were
paid $3.00 for completing the experiment, which
resulted in a mean realized hourly wage of $8.04.
Procedures
All tasks used in Experiments 1 and 2 were used
in Experiment 3. We again isolated the 100%
(current market control) and 2% concentration
amounts (using the same vignette as Experiments
1 and 2) but included one group that received an
undesirable narrative about the cigarettes (2% nar-
rative group). Using information from previous
research21 in which participants provided ratings
about reduced-nicotine cigarettes, this 2% narra-
tive group received the following vignette instruc-
tions describing the new cigarette on the market
(additions bolded):
For the following questions, we would like to
imagine that there is a new cigarette on the market.
ese new cigarettes look and smell the same as cig-
arettes out on the market, including those of your
preferred brand. Imagine that your preferred brand
of cigarettes now carries these new cigarettes. e
dierence between these new cigarettes and your
usual cigarettes is that these cigarettes have only
2% the amount of nicotine in them, an amount
of nicotine that is too small to have any posi-
tive eects. Other people who have used these
new cigarettes rate them as less satisfying, less
rewarding, and less eective at reducing cravings
compared to the cigarettes they usually smoke.
Furthermore, we included 2 additional tasks we
believed might be sensitive to decisions related to
cigarette purchasing intentions. e rst task, a
hypothetical cross-price purchase task, was similar
to the CPT, but with both cigarettes concurrently
available for purchase. e new, variable-nicotine
cigarette was set at a xed price ($0.25/cigarette),
whereas the price of the participant’s usual-brand
cigarette increased across trials in the same price
progression used in the CPT. With both cigarette
options presented concurrently, this task allows
us to quantify the degree of substitutability of
reduced-nicotine cigarettes for conventional ciga-
rettes, which is a measure of interchangeability
of purchasing intentions (ie, how much purchas-
ing switches to another product if their preferred
product is unavailable or too expensive). Instruc-
tions (ie, assumptions) in this task were identical

Estimating Uptake for Reduced-nicotine Cigarettes Using Behavioral Economics
272
to those of the CPT, and at each price combina-
tion participants were asked: “How many of each
of the following would you purchase and consume
at the indicated prices?” e only datasets excluded
in this analysis were for decreasing or inconsistent
responding for the new cigarette alternative. In-
consistent responding occurred anytime purchas-
ing decreased and then subsequently increased on
more than one instance.
e second task was a concurrent choice task17
in which participants indicated their preference for
purchasing either the old, usual-brand cigarette at
increasing prices ($0.13, 0.25, 0.50, 1.00, 2.00,
4.00, 8.00/cigarette) or the new, variable-nicotine
cigarette (100%, 2%) at a xed price ($0.25/ciga-
rette). Similar to the hypothetical cross-price pur-
chase task, at each price combination, participants
were asked: “Which would you prefer to pur-
chase?” However, rather than reporting a quantity
measure, participants indicated their relative pref-
erence for each of the alternatives at each price,
allowing us to model the likelihood of switching
cigarettes at each price.
Data Analysis
Data analysis was conducted similarly as the
previous experiments. Five participants displayed
unsystematic demand trends for either versions of
the CPT (2.81% of the full sample). For the con-
current choice task, we used a generalized logis-
tic model with a logit link function and binomial
distribution to predict the probability of choosing
the new, variable-nicotine cigarette at each price.
For the hypothetical cross-price purchase task, we
used Equation 1 to examine usual-brand cigarette
demand (ie, the xed-price alternative) and we
used an exponentiated version of the exponential
cross-price equation33,41 to t substitution curves:
Equation 2:
Where Q represents purchasing of the new,
variable-nicotine cigarette, QAlone is the estimated
number of new, variable-nicotine cigarettes pur-
chased when the price of the variable-price alter-
native (usual-brand cigarette) approaches innity,
I is the interaction coecient, β is purchasing
sensitivity of the new, variable-nicotine cigarette
to price of the variable-price alternative, and C is
the price per usual-brand cigarette. An extra sum-
of-squares F-test was conducted to compare QAlone
derived from this model. Finally, in this experi-
ment, we excluded 3 participants from all analyses
Figure 1
Derived Intensity (Q0; top panel) and
Elasticity (α; bottom panel) for
Variable-Nicotine Cigarettes after
Accounting for Baseline Cigarette Purchase
Task Intensity and Elasticity, Respectively
Note.
Although intensity increased nominally with concentra-
tion group, elasticity was statistically signicantly higher
after exposure to the Narrative. Symbols and error bars
indicate mean and standard error of the mean. Note the
logarithmic y-axis (bottom panel).
Q =  Q$%&'( ∗ 10,∗(-./

Kaplan et al
Tob Regul Sci.™ 2019;5(3):264-279 273 DOI: https://doi.org/10.18001/TRS.5.3.5
for reporting smoking >100 cigarettes in a day on
the Timeline Followback.
EXPERIMENT 3 RESULTS
Demographics
e nal column of Table 1 displays overall par-
ticipant demographics for Experiment 3, which
were similar to those of the previous experiments.
ere were no statistically signicant dierences
among the 3 groups.
Eects of Concentration on Cigarette Demand
Equation 1 provided an excellent t to the data
(Mdn R2 = .97, IQR = .95, .98) resulting in a medi-
an elasticity of 0.0078 (IQR = 0.0044, 0.0127) and
median intensity of 21.02 (IQR = 15.55, 30.74).
No participants displayed intensity values exceed-
ing 3.29 SDs. Baseline CPT intensity was not sig-
nicantly dierent across the 3 groups (see bottom
of Table 2). When predicting variable-nicotine
CPT intensity (ie, following group assignment),
we found a statistically signicant eect of concen-
tration, as well as baseline CPT intensity. Post hoc
comparisons indicated variable-nicotine CPT in-
tensity was signicantly higher under the 2% nar-
rative condition compared to the 100% condition
(see top panel of Figure 1; t[164] = 2.78, p = .018).
Additionally, although we observed intensity was
higher for the 2% group compared to the 100%
group and for the 2% narrative compared to the
2% group, these comparisons were not signicant
(ps = .315). Only cigarettes per day signicantly
predicted baseline CPT intensity, such that more
cigarettes smoked per day predicted higher baseline
CPT intensity (b = .99, SE = 0.39). ese results
are largely consistent with the ndings from the
previous 2 experiments, suggesting concentration
amount inuenced initial purchasing intentions.
Four participants had elasticity values exceeding
3.29 SDs on either CPT and were excluded from
the following analysis. We observed no statistically
signicant dierences in baseline CPT elasticity
between groups when controlling for demographic
variables. Concentration group and baseline elas-
ticity both signicantly predicted variable-nicotine
CPT elasticity. Post hoc tests indicated exposure to
the narrative resulted in statistically signicantly
higher elasticity values (see bottom panel of Figure
1) compared to both the 100% (t[160] = 4.66, p
< .001) and 2% groups (t[160] = 4.96, p < .001),
but these latter 2 groups were not dierent from
each other (t[160] = 0.65, p = .518). Income (b =
-7.6x10-6, SE = -3.0x10-6) and FTCD score (b =
-0.10, SE = -0.047) were both negatively and sig-
nicantly associated with baseline elasticity.
Taken together, these results suggest the poten-
tial for compensatory purchasing as evidenced by
increasing intensity values. at is, we observed
the same directional trend in intensity as we did
in Experiments 1 and 2, but with the 2% narrative
resulting in the most compensation. In contrast to
the previous experiments, in this experiment we
observed an increase in elasticity, which is indica-
tive of the narrative decreasing the perceived value
of the new, variable-nicotine cigarettes. In other
words, participants exposed to the narrative dem-
onstrated greater sensitivity to price and cigarette
purchasing decreased at a relatively faster rate as
compared to the other 2 groups.
Usual-brand Cigarette Demand and Variable-
nicotine Cigarette Substitution in the
Hypothetical Cross-price Purchase Task
Concentration amount was positively related to
the degree to which variable-nicotine cigarettes
substituted for usual-brand cigarettes and usual-
brand demand intensity (F[2,1464] = 3.11, p =
.045, = .004) but did not inuence demand
elasticity (F[2,1464] = 0.58, p = .563, = .001).
As Figure 2 shows, the number of variable-nicotine
cigarettes purchased at the lowest usual-brand ciga-
rette price (the y-intercept) diered across groups.
Participants in the 2% narrative group purchased
fewer variable-nicotine cigarettes (M = 6.59, SEM
= 1.67) compared to the 100% (M = 10.88, SEM
= 1.95) and 2% (M = 13.20, SEM = 1.72) groups.
Consistent with our previous ndings, the 2% con-
centration amount resulted in slightly more ciga-
rettes being purchased at unrestricted cost (ie, free)
compared to the 100% concentration amount, and
the narrative resulted in fewer cigarettes purchased
compared to the 2 other groups.
In addition, tted QAlone (ie, the terminal intensity
of substitution; far right side of Figure 2) was sig-
nicantly dierent across the 3 groups (F[2,1461]
= 6.21, p = .002, = .008). Post hoc comparisons
indicated QAlone for the 2% narrative was signi-
η"
#
η"
#
η"
#

Estimating Uptake for Reduced-nicotine Cigarettes Using Behavioral Economics
274
Table 3
Generalized Logistic Regression
Choosing Variable-Nicotine Cigarette
Regression Term Odds Ratio 95% Condence Interval Standard Error p
Intercept 3.67 2.69, 5.16 0.61 <.001
Price 18.99 10.70, 35.93 5.85 <.001
Group: 100% 2.46 1.41, 4.46 0.72 <.01
Group: 2% Narrative 0.24 0.16, 0.36 0.05 <.001
Price X Group: 100% 5.68 2.00, 17.21 3.11 .001
Price X Group: 2% Narrative 0.68 0.31, 1.46 0.27 .329
Observations 1267
AIC 1036.89
Null Deviance 1664.62
Residual Deviance 1024.89
χ2
deviance p < .001
Family (link) Binomial (Logit)
Note.
2% coded as reference group
Figure 2
Usual-brand Cigarette Demand (square symbols) and Variable-nicotine Cigarette
Substitution (circle symbols), Fixed at $0.25 per Cigarette, as a Function of Increasing
Price of Usual-brand Cigarettes
Note.
Variable-nicotine cigarettes in all three groups served as partial substitutes for usual-brand cigarettes with cigarettes
under the Narrative condition demonstrating least substitution. Symbols and error bars indicate mean and standard
error of the mean. Note the logarithmic x-axis.

Kaplan et al
Tob Regul Sci.™ 2019;5(3):264-279 275 DOI: https://doi.org/10.18001/TRS.5.3.5
cantly lower compared to QAlone for both the 100%
(F[1,939] = 10.15, p = .002, = .021) and 2%
groups (F[1,969] = 6.27, p = .013, = .012). No
dierences were found between the 100% and 2%
groups (F[1,1014] = 1.88, p = .171, = .004).
ese ndings provide further support for the ef-
cacy and domain specicity of the narrative to
inuence estimated purchasing of variable-nicotine
cigarettes, but not usual-brand cigarettes.
Eects of Concentration on Concurrent Choice
Task
Results from the concurrent choice task indicat-
ed a signicant price by group interaction, χ2(2) =
20.48, p < .001 (Table 3 and Figure 3). As the price
of the usual-brand cigarette increased, the odds
of switching to the variable-nicotine cigarette in-
creased more quickly for participants in the 100%
group compared to those in the 2% (OR = 5.68,
p = .001) and 2% narrative groups (OR = 8.36, p
< .001). Although the rate at which participants
switched to the variable-nicotine cigarettes was not
dierent between those in the 2% group compared
to the 2% narrative group (OR = 1.47, p = .329),
participants in the 2% group were more likely to
purchase the variable-nicotine cigarette overall
(OR = 4.10, p < .001).
EXPERIMENT 3 DISCUSSION
e purpose of this experiment was to determine
the eects of concentration amount and an unde-
sirable narrative associated with the 2% concentra-
Figure 3
Estimated Probability of Choosing Variable-nicotine Cigarettes (set at a xed $0.25
per cigarette) Over Usual-brand Cigarettes (increasing in price) Based on Group
Note.
Participants in the 100% group switched to the variable-nicotine cigarette more rapidly as price increased
compared to participants in the 2% and 2% Narrative groups. Participants in the 2% group were more
likely to purchase variable-nicotine cigarettes regardless of price compared to the 2% Narrative group.
Shaded curves represent 95% condence intervals.
η"
#
η"
#
η"
#

Estimating Uptake for Reduced-nicotine Cigarettes Using Behavioral Economics
276
tion amount on demand indices and substitution.
Although we did not observe a concentration
amount (100% vs 2%) eect on either demand
intensity or elasticity, we found that intensity was
nominally higher under the 2% amount com-
pared to the 100% amount, as well as a consistent
and pronounced eect of the narrative in alter-
ing the value of variable-nicotine cigarettes across
a number of tasks. at is, demand elasticity for
these cigarettes was higher (greater price sensitiv-
ity) in the variable-nicotine CPT, and the price of
usual-brand cigarettes had to be suciently high
for participants to switch to the variable-nicotine
cigarettes.
GENERAL DISCUSSION
Emerging evidence suggests a tobacco regulatory
policy limiting the amount of nicotine in cigarettes
may result in a socially signicant reduction in ciga-
rette use and dependence.6 Whereas much of this
research has used experiential contexts, the current
set of experiments explored initial intentions of cig-
arette uptake by measuring how cigarette smokers
estimate their cigarette purchasing under dierent
scenarios. We approximated a realistic scenario in
which the participant’s usual-brand cigarette manu-
facturer replaced their old cigarettes with a new cig-
arette containing some variable amount of nicotine
(variable-nicotine cigarette) and that the only dif-
ference was the amount of nicotine in the cigarette.
When the stated concentration was 100% of their
usual-brand cigarette, we observed a reduction in
the estimated number of these new cigarettes par-
ticipants would purchase when they were free.
However, relative to the 100% group, our cur-
rent market control, participants in all other
groups tended to show an increase in the estimated
number of cigarettes purchased if cigarettes were
free. is nding was unexpected and may indicate
some minimal amount of compensatory smok-
ing behavior based solely on perceptions of what
it means to have a reduction in nicotine content,
but not necessarily commensurate with the degree
of nicotine reduction. Consistent with previous re-
search,8,15 the observed increases in purchasing may
reect participants’ misconceptions of the role of
nicotine such that any reductions in nicotine com-
pared to their usual brand are associated with de-
creased health risks. If this were the case, however,
we would expect that decreases in concentration
amount would be systematically associated with in-
creases in purchasing, similar to the results found by
Pacek et al.8 However, this was not the case as par-
ticipants in the very low concentration groups did
not purchase relatively more cigarettes compared
to the more intermediate groups. In addition, we
note that neither the vignette describing the new
type of cigarette nor the narrative included any in-
formation about changes in health eects; rather
the narrative described dierences in the subjective
feelings associated with the new cigarette. Whether
dierences in risk perceptions between the dierent
nicotine concentrations mediated estimated uptake
is unknown, but this knowledge could be of value
when designing marketing or education campaigns
associated with a nicotine reduction policy.
Related to concentration amount, another ma-
jor nding was the lack of inuence the stated per-
centage had on altering cigarette elasticity, one of
the main measures of cigarette valuation. Across
all the experiments conducted, we did not observe
any changes in demand elasticity for any group as
a function of nicotine framing alone. Only when
we described the new cigarette scenario associated
with an undesirable narrative did cigarette demand
elasticity increase, which is reective of relatively
rapid declines in purchasing as price increases. is
eect was captured across a variety of tasks, which
may speak to the power of the narrative in inu-
encing decision making. Moreover, even though
the narrative itself was relatively short (only one
sentence) and only the 2% concentration was
shown when participants were completing the de-
mand and choice tasks, the narrative maintained its
eectiveness. ese ndings suggest information
alone about any changes in nicotine content will
not reduce either smoking intentions or cigarette
valuation, and may actually lead to smokers pur-
chasing more cigarettes. Rather, an eective pol-
icy would consider not only providing narratives
about the cigarettes’ undesirable subjective eects,
but also would include targeted information about
the cigarettes’ health risks. Such a campaign could
dampen both initial smoking intentions as well as
alter initial cigarette valuation prior to experiencing
the cigarettes. is combinatorial approach would
be consistent with the ndings of Mercincavage et
al15 where subsequent smoking of reduced-nicotine
cigarettes was predicted by a combination of sub-

Kaplan et al
Tob Regul Sci.™ 2019;5(3):264-279 277 DOI: https://doi.org/10.18001/TRS.5.3.5
jective taste ratings and degree of false beliefs.
Finally, we note the potential utility of using tasks
grounded in the behavioral economic paradigm,
which among others include self-administration,
simulated purchase tasks, and discrete choice tasks,
for assessing the ecacy of reduced-nicotine ciga-
rettes.6,42 Indeed, these tasks have been used success-
fully in recent studies investigating reduced-nicotine
cigarettes17,19,21 and their results hold promise for
shedding insight into the cigarettes’ abuse liability
and public reactions to policy changes.
Limitations and Future Directions
Two aspects of the current experiments diered
from previous investigations of reduced-nicotine
cigarettes, including the hypothetical nature of the
tasks and our description of the new cigarette sce-
nario. Much of the research on reduced-nicotine
cigarettes has been conducted using experiential
procedures (ie, participants experience the sub-
jective eects of the reduced-nicotine cigarettes)
and participants are typically blinded to the ciga-
rette concentration.5,6,19 Here, we conveyed how
these cigarettes diered by indicating the nicotine
content as a percentage of their usual-brand ciga-
rette and describing a realistic scenario in which
only reduced-nicotine cigarettes are available. We
sought to isolate the potential inuence of cigarette
concentration amount by describing the new, vari-
able-nicotine cigarettes as similar to participants’
usual-brand cigarette. It is plausible that tobacco
companies would try to market reduced-nicotine
cigarettes as being similar in characteristics to con-
ventional-nicotine cigarettes. To our knowledge,
we are not aware of any experiential research on
reduced-nicotine cigarettes that has assessed esti-
mates of purchasing prior to and following the ex-
perience of these cigarettes. In addition, we did not
measure either participants’ knowledge of reduced-
nicotine cigarettes nor participants’ perceptions of
the health risks (eg, Perceived Health Risks scale43)
associated with our hypothetical reduced-nicotine
cigarette so the relations between individual knowl-
edge, risk perceptions, and estimated purchasing is
unknown. Future research may benet from ex-
amining a potential moderating role of knowledge
and/or perceived health risks and prospective pur-
chasing intentions on the CPT, as well as subse-
quent correspondence of uptake.
Taken together, the results of the current study
suggest estimated uptake of variable-nicotine ciga-
rettes is largely unaected by a specied nicotine
concentration amount alone, and if anything results
in small, but consistent, compensatory purchasing.
Importantly, narratives describing variable-nicotine
cigarettes as less satisfying, less rewarding, and less
eective at reducing cravings signicantly reduced
the value of cigarettes indicating a potential mecha-
nism for reducing cigarette purchasing. Our results
suggest a public policy initiative reducing nicotine
content aimed at reducing cigarette smoking might
benet from careful marketing and education, and
our results provide content that may be important
to include in such endeavors.
IMPLICATIONS FOR TOBACCO
REGULATION
Lowering the nicotine content in combustible
cigarettes may be a viable strategy for reducing de-
pendence and toxin exposure, however our results
suggest information about nicotine content alone
is unlikely to reduce the number of cigarettes pur-
chased without accompanying narratives about
the eects of this reduced nicotine content. ere-
fore, policymakers should not market or describe
reduced-nicotine cigarettes in terms of the nicotine
percentage alone. Rather, policymakers should also
market and describe reduced-nicotine cigarettes
with respect to their subjective eects (eg, less
satisfying, less eective at reducing cravings). Re-
searchers should also consider utilizing tasks from
the behavioral economic framework (eg, purchase
tasks, substitution tasks) to prospectively assess
policy change initiatives.
Human Subjects Statement
e treatment of human participants was in ac-
cordance with ethical standards, all study proce-
dures were approved by the Institutional Review
Board at Virginia Tech (IRB #17-311), and all
participants provided informed consent. Addition-
ally, the current study meets the ethical standard
outlines in Helsinki Declaration of 1975 as revised
in 2000.
Conict of Interest Statement
W.K.B. is a principal of HealthSim, LLC and

Estimating Uptake for Reduced-nicotine Cigarettes Using Behavioral Economics
278
Notius, LLC; a scientic advisory board member
of Sober Grid, Inc. and DxRx, Inc.; and a consul-
tant for ProPhase, LLC and Teva Branded Pharma-
ceutical Products R&D, Inc. B.A.K. and W.K.B.
are principals of BEAM Diagnostics, Inc.
Acknowledgements
Research reported in this publication was sup-
ported by NIDA/NIH grant R01DA042535 and
FDA Center for Tobacco Products (CTP). e
content is solely the responsibility of the authors
and does not necessarily represent the ocial views
of the NIH or the Food and Drug Administration.
All authors have contributed, read, and approved
this version of the manuscript.
Portions of this study were presented at the 2017
International Study Group Investigating Drugs as
Reinforcers Annual Meeting and at the 2018 An-
nual Meeting of the Society for Research on Nico-
tine and Tobacco.
References
1. US Department of Health and Human Services (USD-
HHS).e Health Consequences of Smoking: 50 Years of
Progress. A Report of the Surgeon General. Atlanta, GA: US-
DHHS, US Centers for Disease Control and Prevention,
National Center for Chronic Disease Prevention and
Health Promotion, Oce on Smoking and Health; 2014.
2. Xu X, Bishop EE, Kennedy SM, et al. Annual healthcare
spending attributable to cigarette smoking: an update.
Am J Prev Med. 2015;48(3):326-333.
3. Benowitz NL. Pharmacologic aspects of cigarette
smoking and nicotine addiction. N Engl J Med.
1988;319(20):1318-1330.
4. Benowitz NL, Henningeld JE. Reducing the nicotine
content to make cigarettes less addictive. Tob Control.
2013;22(Suppl 1):i14-i17.
5. Donny EC, Denlinger RL, Tidey JW, et al. Randomized
trial of reduced-nicotine standards for cigarettes. N Engl J
Med. 2015;373(14):1340-1349.
6. Tidey JW, Cassidy RN, Miller ME, Smith TT. Behavioral
economic laboratory research in tobacco regulatory sci-
ence. Tob Regul Sci. 2016;2(4):440-451.
7. US Food and Drug Administration. Tobacco Product
Standard for Nicotine Level of Combusted Cigarettes.
Fed Regist. 2018;83:11818-11843. Available at: https://
www.federalregister.gov/d/2018-05345. Accessed March
6, 2019.
8. Pacek LR, McClernon FJ, Denlinger-Apte RL, et al. Per-
ceived nicotine content of reduced nicotine content ciga-
rettes is a correlate of perceived health risks. Tob Control.
2018;27(4):420-426.
9. O’Brien EK, Nguyen AB, Persoskie A, Homan AC. U.S.
adults’ addiction and harm beliefs about nicotine and low
nicotine cigarettes. Prev Med. 2017;96:94-100.
10. Mercincavage M, Lochbuehler K, Villanti AC, et al. Ex-
amining risk perceptions among daily smokers naïve to
reduced nicotine content cigarettes. Nicotine Tob Res.
2018 Apr 28. doi: 10.1093/ntr/nty082. [Epub ahead of
print]
11. Bansal-Travers M, Cummings KM, Hyland A, et al. Edu-
cating smokers about their cigarettes and nicotine medi-
cations. Health Educ Res. 2010;25(4):678-686.
12. Bobak A, Shiman S, Gitchell JG, et al. Perceived safety
of nicotine and the use of nicotine replacement products
among current smokers in Great Britain: results from two
national surveys. J Smok Cessat. 2010;5(2):115-122.
13. Giovino GA, Sidney S, Gfroerer JC, et al. Epidemiology
of menthol cigarette use. Nicotine Tob Res. 2004;(6 Suppl
1):S67-S81.
14. Mooney ME, Leventhal AM, Hatsukami DK. Attitudes
and knowledge about nicotine and nicotine replacement
therapy. Nicotine Tob Res. 2006;8(3):435-446.
15. Mercincavage M, Saddleson ML, Gup E, et al. Reduced
nicotine content cigarette advertising: how false beliefs
and subjective ratings aect smoking behavior. Drug Al-
cohol Depend. 2017;173:99-106.
16. Pacek LR, Oliver JA, Sweitzer MM, McClernon FJ.
Young adult dual combusted cigarette and e-cigarette us-
ers’ anticipated responses to a nicotine reduction policy
and menthol ban in combusted cigarettes. Drug Alcohol
Depend. 2018;194:40-44.
17. Higgins ST, Heil SH, Sigmon SC, et al. Addiction po-
tential of cigarettes with reduced nicotine content in
populations with psychiatric disorders and other vul-
nerabilities to tobacco addiction. JAMA Psychiatry.
2017;74(10):1056-1064.
18. Higgins ST, Reed DD, Redner R, et al. Simulating de-
mand for cigarettes among pregnant women: a low-risk
method for studying vulnerable populations. J Exp Anal
Behav. 2017;107(1):176-190.
19. Smith TT, Cassidy RN, Tidey JW, et al. Impact of smok-
ing reduced nicotine content cigarettes on sensitivity to
cigarette price: further results from a multi-site clinical
trial. Addiction. 2017;112(2):349-359.
20. Tidey JW, Pacek LR, Koopmeiners JS, et al. Eects of
6-week use of reduced-nicotine content cigarettes in
smokers with and without elevated depressive symptoms.
Nicotine Tob Res. 2017;19(1):59-67.
21. Tucker MR, Laugesen M, Grace RC. Estimating demand
and cross-price elasticity for very low nicotine content
(vlnc) cigarettes using a simulated demand task. Nicotine
Tob Res. 2018;20(7):843-850.
22. Jacobs EA, Bickel WK. Modeling drug consumption in
the clinic using simulation procedures: demand for her-
oin and cigarettes in opioid-dependent outpatients. Exp
Clin Psychopharmacol. 1999;7(4):412-426.
23. Roma PG, Reed DD, DiGennaro Reed FD, Hursh SR.
Progress of and prospects for hypothetical purchase task
questionnaires in consumer behavior analysis and public
policy. Behav Anal. 2017;40(2):329-342.
24. DeHart WB, Kaplan BA, Pope DA, et al. e experimen-
tal tobacco marketplace: narrative inuence on electronic
cigarette substitution. Exp Clin Psychopharmacol. 2018
Nov 5. doi: 10.1037/pha0000233. [Epub ahead of print]
25. Sobell LC, Sobell MB. Timeline followback: a technique
for assessing self-reported alcohol consumption. In Litten

Kaplan et al
Tob Regul Sci.™ 2019;5(3):264-279 279 DOI: https://doi.org/10.18001/TRS.5.3.5
RZ, Allen JP, eds. Measuring Alcohol Consumption: Psycho-
social and Biological Methods. Totowa, NJ: Humana Press;
1992:41-72.
26. MacKillop J, Murphy JG, Ray LA, et al. Further valida-
tion of a cigarette purchase task for assessing the relative
reinforcing ecacy of nicotine in college smokers. Exp
Clin Psychopharmacol. 2008;16(1):57-65.
27. Fagerström K. Determinants of tobacco use and renam-
ing the ftnd to the Fagerström test for cigarette depen-
dence. Nicotine Tob Res. 2012;14(1):75-78.
28. R: a language and environment for statistical computing
[computer program]. Version 3.5.1; 2018.
29. Stein JS, Koarnus MN, Snider SE, et al. Identica-
tion and management of nonsystematic purchase task
data: toward best practice. Exp Clin Psychopharmacol.
2015;23(5):377-386.
30. Koarnus MN, Franck CT, Stein JS, Bickel WK. A modi-
ed exponential behavioral economic demand model to
better describe consumption data. Exp Clin Psychophar-
macol. 2015;23(6):504-512.
31. Hursh SR, Silberberg A. Economic demand and essential
value. Psychol Rev. 2008;115(1):186-198.
32. Kaplan BA, Gilroy SP, Reed DD, et al. e R package
beezdemand: behavioral economic easy demand. Perspect
Behav Sci. 2018 Dec 18. pii: s40614-018-00187-7 [Epub
ahead of print]
33. Koarnus MN, Kaplan BA. Clinical models of deci-
sion making in addiction. Pharmacol Biochem Behav.
2018;164:71-83.
34. Tabachnick BG, Fidell LS. Using Multivariate Statistics.
5th ed. Needham Heights, MA: Allyn & Bacon; 2007.
35. Holm S. A simple sequentially rejective multiple test pro-
cedure. Scand Stat eory Appl. 1979;6(2):65-70.
36. Bickel WK, Stein JS, Moody LN, et al. Toward narrative
theory: interventions for reinforcer pathology in health
behavior. In Stevens JR, ed. Impulsivity: How Time and
Risk Inuence Decision Making. Berlin, Germany: Spring-
er International Publishing; 2017:227-267.
37. Frank LB, Murphy ST, Chatterjee JS, et al. Telling stories,
saving lives: creating narrative health messages. Health
Commun. 2015;30(2):154-163.
38. Moyer-Gusé E, Jain P, Chung AH. Reinforcement or re-
actance? Examining the eect of an explicit persuasive
appeal following an entertainment-education narrative. J
Commun. 2012;62(6):1010-1027.
39. Ne LJ, Patel D, Davis K, et al. Evaluation of the national
tips from former smokers campaign: the 2014 longitudi-
nal cohort. Prev Chronic Dis. 2016;13:E42.
40. CDCTobaccoFree. Tips From Former SmokersTM. Avail-
able at: https://www.cdc.gov/tobacco/campaign/tips/in-
dex.html. Accessed September 12, 2018.
41. Hursh SR, Roma PG. Behavioral economics and empiri-
cal public policy. J Exp Anal Behav. 2013;99(1):98-124.
42. Smith TT, Sved AF, Hatsukami DK, Donny EC. Nico-
tine reduction as an increase in the unit price of cigarettes:
a behavioral economics approach. Prev Med. 2014;68:23-
28.
43. Hatsukami DK, Vogel RI, Severson HH, et al. Perceived
health risks of snus and medicinal nicotine products. Nic-
otine Tob Res. 2016;18(5):794-800.