See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/288057267

Development and Initial Validation of the Caffeine Consumption Questionnaire-

Revised

Article  in  Journal of Caffeine Research · December 2015

DOI: 10.1089/jcr.2015.0012

CITATIONS READS
7 2,038

5 authors, including:

Jessica G. Irons Drew T Bassett

James Madison University Auburn University
37 PUBLICATIONS   367 CITATIONS    13 PUBLICATIONS   19 CITATIONS   

SEE PROFILE SEE PROFILE

R. Eric Landrum Adrienne Heinz

Boise State University VA Palo Alto Health Care System
129 PUBLICATIONS   1,403 CITATIONS    52 PUBLICATIONS   1,443 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Identifying Innovations for Managing High-Cost Mental Health Patients View project

Assessment of Student Learning Outcomes in Psychology View project

All content following this page was uploaded by Jessica G. Irons on 12 March 2018.

The user has requested enhancement of the downloaded file.

JOURNAL OF CAFFEINE RESEARCH
Volume 6, Number 1, 2016
ª Mary Ann Liebert, Inc.
DOI: 10.1089/jcr.2015.0012

Development and Initial Validation of the Caffeine

Consumption Questionnaire-Revised

Jessica G. Irons, PhD,1 Drew T. Bassett, BS,2 Caroline O. Prendergast, BS,3

R. Eric Landrum, PhD,4 and Adrienne J. Heinz, PhD5,6

Objective: We sought to outline the rationale and procedures for revising the Caffeine Consumption Ques-
tionnaire (CCQ) and to examine the utility and validity of the CCQ-revised (CCQ-R). Given that caffeine is
a widely used drug with myriad potential health effects, it is important to measure and understand its use;
however, to date, there are no reliable and valid biological tests readily available and existing self-report
measures do not have standardized administration or scoring procedures.
Method: The CCQ-R incorporates pictures, visual cues of relative sizes, sample products, and details about
products to aid in responding. Additionally, a standardized scoring system is provided.
Results and Conclusion: Validity data are provided and suggest that the CCQ-R offers a valid approach for
measuring self-reported caffeine use.

Introduction include drowsiness, headaches, fatigue, and negative

mood.3,10 Only recently have caffeine intoxication and

C affeine is the most commonly consumed psycho-

active substance worldwide1,2 and occurs naturally
in a variety of foods and beverages (e.g., chocolate, cof-
fee, tea). Caffeine is also added to many consumables, in-
cluding sodas, energy drinks, some chewing gums, and
mints, as well as various over-the-counter and prescription
drugs. In the United States, *80% of adults3 and 75% of
adolescents4,5 report regular caffeine consumption. Low-
to-moderate dosages of caffeine (50–300 mg) can typi-
cally be consumed without adverse consequences6 and
short-term effects can include reduced fatigue, increased
wakefulness, improved performance on simple and com-
plex attention tasks, perceived thought clarity, and an-
algesic effects for headaches.3,7,8 Negative effects are
associated with higher doses of caffeine (500–1000 mg)
and can include hypertension, cardiac arrhythmia, anxiety,
tachycardia, insomnia, and detrimental effects on working
memory at high load levels.3,9
Habitual caffeine consumers may develop tolerance
and experience withdrawal symptoms upon cessation fol-
lowing prolonged repeated use. Withdrawal symptoms
withdrawal been recognized in the Diagnostic and Stat-
istical Manual-5 (DSM-5) following review of substan-
tial evidence from preclinical and clinical studies that
support the validity, reliability, and clinical significance
of caffeine withdrawal.11 Indeed, accumulating data on
the impact of caffeine withdrawal and other detrimental
outcomes associated with caffeine use (e.g., intoxication)
led to the inclusion of caffeine use disorder as a condi-
tion for further study in the DSM-5.11 In addition to
withdrawal and potential caffeine-related disorders,
long-term caffeine use may lead to increased risk of car-
diovascular disease12,13 and has been shown to be associ-
ated with depression among children and adolescents14
and with increased anxiety and depression in adults.15,16
Additionally, caffeine use has clinical implications re-
lated to sleep as well as drug interaction effects that are
not yet well characterized (e.g., alcohol)17,18 and are in
need of further study.
Despite estimated use prevalence among adults and
children and potential negative effects of use, no validated
measure of caffeine use exists to date. Furthermore, many

1
Department of Psychology, James Madison University, Harrisonburg, Virginia.
2
Department of Psychology, Auburn University, Auburn, Alabama.
3
Lynch School of Education, Boston College, Boston, Massachusetts.
4
Department of Psychology, Boise State University, Boise, Idaho.
5
National Center for Posttraumatic Stress Disorder, Palo Alto VA Healthcare System, Menlo Park, California.
6
Center for Innovation to Implementation, Palo Alto VA Healthcare System, Menlo Park, California.

1
2 IRONS ET AL.
individuals who consume caffeinated products are un-
aware of the amount of caffeine they consume or the asso-
ciated effects of their use patterns.19 Reasons for a lack of
awareness regarding caffeine consumption levels may in-
clude a perception that caffeine is harmless because it is
legal or a lack of understanding of its potential effects20
and/or a lack of user-friendly, reliable, and valid methods
of assessment. In addition, the U.S. Food and Drug
Administration (FDA) does not currently require that caf-
feine content be included on product labels such that caf-
feine content information is not readily available for many
items.21 Because consumers are often unaware of their
own caffeine consumption habits (in terms of quantifiable
servings and caffeine levels) and no standard clinical as-
sessment exists for caffeine use, a sound measurement
tool is needed both for clinical and experimental purposes.
Biological testing is the standard and preferred method of
assessing substance use for most drugs (e.g., cocaine, mar-
ijuana, heroine).22,23 Just as with other drugs, there are
several biological methods of assessment of caffeine con-
sumption, including breath tests, urinary tests, and salivary
tests—each with varying levels of validity. Even among
the most accurate of these available tests (urinalysis and
salivary tests),23,24 biological tests only allow for determi-
nation of the amount of caffeine currently present in the
body at a given time. Given that the half-life of caffeine
is between 4 and 6 hours for healthy adults, the information
gained from a biological measurement at a single time
point is limited to recent acute use. As such, actual patterns
and amounts of caffeine use are virtually impossible to dis-
cern over longer time periods.17 Given the restricted data
available from biological tests and the costs typically asso-
ciated with such tests, it is important to consider alternative
methods for measuring caffeine consumption. Use of other
substances is commonly measured through self-reports in
both clinical and research settings; however, there are
few self-report options for caffeine use. The self-report
measures that exist are not ideal from the client/patient/
participant perspective or from the practitioner/researcher
perspective given that there are no standardized scoring
systems and the measures rely solely on respondents’
knowledge of caffeinated products.
Although self-report measures may be subject to report-
ing inaccuracies, researchers have shown that self-report
measures of substance use are as high as 95% consis-
tent with biological measures.25 For example, Timeline
Follow-Back (TLFB) is a widely used, calendar-based
self-report measure of substance use frequency and quan-
tity. Several researchers have found high percentages of
agreement between the TLFB and biological measures
for illicit substances, as high as 95.7% for cannabis,26
84.1% for cocaine, and 94.0% for opiates.25 Thus, self-
report measures can reduce the need for biological tests
if shown to be valid measures of behavior. As such, an ef-
fective self-report instrument is both possible and needed
for measuring caffeine use.
To date, there are only two self-report caffeine measures
of which we are aware. The first, the Caffeine Consump-
tion Questionnaire (CCQ),27 is a self-report questionnaire
designed to circumvent the need for knowledge of the
caffeine concentrations in each product. Patients or partic-
ipants report whether they have ingested specific products,
including coffee, tea, soda, cocoa, chocolate milk, and
over-the-counter medications. A common serving size is
designated for each type of product (e.g., one serving
size per item category; 12 oz soft drink, 5 oz coffee).
The practitioner or researcher then calculates how much
caffeine was ingested. Although the measure allows par-
ticipants to report caffeine ingestion in terms of the num-
ber of servings consumed, the CCQ is not intuitive for
consumers to complete. Consumers often do not know
the amount of a product they consumed (i.e., how many
ounces of soda or coffee consumed), thus rendering
their CCQ data a potentially unreliable assessment of caf-
feine use. In addition, the CCQ is laborious for practition-
ers and researchers to score and many newer caffeinated
items (e.g., energy drinks, caffeinated foods) have been
made commercially available since its original concep-
tion. Finally, no data are available to provide evidence
for the validity of CCQ scores.
A more recent self-report measure, the Stanford Caf-
feine Questionnaire (SCQ), was developed based on
the CCQ for the purposes of studying the effects of chro-
notype and caffeine on sleep.28 As part of the study, the
researchers examined the validity of the SCQ using
saliva tests for detecting caffeine use. Similar to the
CCQ, using the SCQ, participants self-reported their con-
sumption of caffeinated items, including cola, diet cola,
pepper soda, citrus soda, other soda, tea (hot), iced tea,
instant coffee, brewed coffee, other coffee, store-bought
coffee, energy drink, other drink, food, and medications.
Although the SCQ measure included common conver-
sions for serving size (e.g., 1 can = 12 oz and Red Bull =
8.3 oz), it suffers from many of the same challenges
affecting the CCQ—neither measure accounts for con-
sumers not knowing the amount of caffeinated products
they have consumed (i.e., how many ounces of soda or
coffee). Furthermore, the SCQ requires participants to re-
port their daily caffeine consumption in 6-hour blocks for
7 days. The use of time blocks may be beneficial for dis-
cerning temporal patterns of use; however, time blocks
may pose a challenge with respect to reporting accuracy
given that caffeine use is often ongoing throughout the
day. Furthermore, the use of time blocks necessarily ren-
ders SCQ scoring more complex. Participants also pro-
vided saliva samples as a validity check for the SCQ;
however, saliva tests for caffeine use provide a poor stan-
dard for testing validity of self-report measures designed
to assess specific use details.17 Indeed, Nova et al. found
that only 41% of the variance in salivary caffeine con-
centration level accounted for participant responses on
the SCQ.28 Thus, it remains unclear if the SCQ is a
CAFFEINE CONSUMPTION QUESTIONNAIRE-REVISED
valid measure of caffeine use or if the saliva test is sim-
ply an insufficient tool for validity assessment.
The current study aims to determine the validity of a
modified CCQ-revised (CCQ-R), based on Landrum’s ini-
tial measure,27 designed to be more easily completed by
consumers and scored by practitioners and researchers.
The validity of caffeine self-report measures may be chal-
lenged because (1) typical consumers lack awareness of
caffeine content in consumables and (2) there is no stan-
dard unit of caffeine dosing (compared with alcohol stan-
dard drinks; 5 oz wine = 1.5 oz hard liquor = 12 oz beer)
given the variety of methods of ingestion across myriad
types of consumables. In an effort to improve self-report
of caffeine consumption, the CCQ-R includes the follow-
ing three advantages over extant measures of caffeine
consumption. First, it provides a visual presentation of
a variety of commonly consumed caffeinated products
across a wide range of commonly available serving sizes.
Second, it includes current commercially available items
that were not in existence when previous measures were
developed (i.e., energy drinks, caffeinated mints, and
caffeinated gum). Third, the CCQ-R simplifies how par-
ticipants report consumption of the caffeinated product
by asking about how many times the caffeinated product
is consumed, on average, per week. Previous measures
ask consumers to report how many times caffeinated
products are consumed on average daily, which is prob-
lematic because many consumers may not ingest caffeine
on a daily basis, use may vary widely across days, and
patterns of use may not be captured by examining aver-
age daily use. We aim to provide a new self-report mea-
sure of caffeine use as well as validity data to suggest its
practical utility and potential viability for clinical and re-
search purposes using a validity assessment plan that
does not include biological testing.
Materials and Methods
Participants
Participants were 96 college students (57% female;
84% Caucasian, 9% African American, 5% Asian, and
2% Hispanic) from a mid-Atlantic university in the
United States.
Measures
Demographic survey. A demographic survey was used
to assess gender, class standing, race, and marital status.
In addition to the previous items, participants were asked
to report whether they had ever been advised to avoid
caffeine, whether they were particularly sensitive to the
effects of caffeine, and to provide their best estimates
of how many milligrams of caffeine they believe they
consume in an average day.
Caffeine Consumption Questionnaire-revised. The CCQ-R
is a web-based self-report measure of caffeine consump-
3
tion adapted from the CCQ. The measure instructs partic-
ipants to report their average weekly consumption of
several caffeinated beverages (e.g., coffee, tea, soda, en-
ergy drinks), caffeinated foods (e.g., chocolate, candy
bars, baked goods), and over-the-counter caffeinated
drugs (e.g., weight-loss drugs, headache powder). If an
item is not consumed, participants may enter 0 or leave
the item blank. Most items pair the caffeinated product
with pictures of a variety of commonly available serving
sizes of the products. For example, when asked how
much coffee participants consume in an average week,
they are presented with four pictures of cups of varying
sizes that are typically available for purchase (e.g., 8, 12,
16, and 20 oz) and asked to indicate how many of each
size of the beverage they consume in an average week.
Over-the-counter drugs and caffeinated food products, ex-
cluding candy and chocolate bars, are the only products not
paired with pictures. The reported number of each serving
size is used to determine the total number of units (e.g., oz)
of each item type consumed in an average week (to create
item-type subscores). This value is then multiplied by the
caffeine content for that item type to produce an item-
type subscore. Subscores are then summed to reflect the
total amount of caffeine consumed in an average week.
Procedures
Participants were enrolled in introductory psychology
classes and they received credit for their participation in
the study. The university Institutional Review Board ap-
proved the study. All participants reported to a computer
laboratory to complete the measures (see Fig. 1). After
providing written informed consent, participants were
asked to read directions and complete the demographic
survey and the CCQ-R (self-report of their own personal
caffeine use to be used for descriptive purposes: admin-
istration 1) using a web-based survey (Qualtrics). Upon
completing the demographic survey and the CCQ-R,
each participant was presented with a box of products
containing both caffeinated items (e.g., Coca-Cola, choc-
olate bar, Starbucks Coffee) and noncaffeinated foil
items (e.g., Sprite, crackers). Each box of products con-
tained 13 or 14 items; up to 12 items in each box con-
tained caffeine and the others were noncaffeinated foils
(not specifically labeled as caffeine-free). Participants
were instructed to imagine that the items inside the pro-
vided box were what they personally consumed in an av-
erage week. Participants were then asked to complete the
CCQ-R accordingly, as if the content of the provided box
was what they consumed (box content; administration 2).
The minimum/mean/maximum box totals (total amount
of caffeine present in the box based on scoring values)
were calculated to compare with the box scores repor-
ted by participants. Items that were reported as other
were hand scored and manually added to the box totals
(administration 2).
4 IRONS ET AL.
FIG. 1. Flowchart of procedures.
Data analyses
A total of 96 participants completed the measures; of
those, 15 (females = 9) cases were excluded from analyses.
Data exclusion occurred if (1) <75% of the total number of
caffeinated items in the box were reported, (2) more items
than were present in the box were reported, or (3) multiple
items were entered that were not present in any of the
boxes. Most cases of exclusion occurred because partici-
pants entered serving sizes in lieu of number of servings.
Analyses were conducted with remaining participants’
data (N = 81). Those excluded from analyses were not sta-
tistically different ( ps > 0.05) from the remainder of the
sample with respect to any demographic variables. All an-
alyses were conducted using the mean value estimate
score (see Supplementary Data: CCQ-R scoring; Appen-
dix A for scoring details, Appendix B for SPSS scoring
syntax, and Appendix C for scoring codes; Supplementary
Data are available online at www.liebertpub.com/jcr). For
an example of box contents and administration 2 re-
sponses, see Supplementary Data: Appendix D. Percent
correct scores were determined by using the following
formula:
100 (the absolute value of the difference between
reported mg of caffeine and actual calculated mg of caf-
feine in the box)/(the actual calculated mg of caffeine in
each box)
As such, higher values (i.e., percent correct scores) in-
dicated a greater agreement between participant-reported
values and actual calculated values of caffeine. The abso-
lute value was employed for tests of inferential statistics
to ease interpretation regarding the magnitude of the es-
timate difference. Roughly 40% of participants achieved
a percent correct score above 90%; 59% of participants
achieved a percent correct score above 85%. Approxi-
mately 43% of the sample under-reported caffeine con-
tent of items in boxes.
Results
Average caffeine consumption among the sample (as
reported by the CCQ-R) was 1327.15 mg (SD = 1071.82;
range = 0–4681 mg) per week or 189.59 mg per day
(items reported as other were not included in CCQ-R per-
sonal use scoring). Average number of days per week con-
suming caffeine was 4.43 (SD = 2.09) and five participants
reported no caffeine use. A one-way analysis of variance
revealed no differences in participant accuracy with re-
spect to the various boxes assigned ( p > 0.05)—regardless
of the box provided, accuracy was similar. Similarly, there
were no differences across races or genders with respect
to accuracy ( ps > 0.05). Furthermore, personal caffeine
use (measured by administration 1; see Procedures
section) was not correlated with percent correct scores
(r = 0.03, p = 0.811). Participants’ estimates of their
own consumption (in mg/day) were not correlated with
CCQ-R scores of their personal use (administration 1;
r = 0.13, p = 0.264) or percent correct scores (r = 0.20,
p = 0.071).
An independent t-test revealed a significant difference
between participants who reported having been advised
to avoid caffeine (n = 19; M = 89.25, SD = 10.08) and
those who had not (M = 81.94, SD = 13.26; t(79) = 2.11,
p = 0.04) with respect to CCQ-R accuracy. The average
percent correct score was 83.47% (SD = 12.95). The range
of percent correct scores was 57.03 (42.96–99.99%). A de-
pendent t-test revealed no significant difference between
participants’ CCQ-R scores for the boxed items and the ac-
tual amount of caffeine in the boxes, t(79) = 0.96, p = 0.340.
Discussion
We sought to examine the practical utility and poten-
tial validity of the CCQ-R for estimating caffeine use.
The CCQ-R is a more flexible and comprehensive self-
report measure of caffeine use than its predecessors
that may be administered and scored electronically.
Although participants were unable to accurately estimate
CAFFEINE CONSUMPTION QUESTIONNAIRE-REVISED
their own caffeine use (compared with their personal
CCQ-R scores; administration 1), participants’ CCQ-R
scores for the items included in the provided boxes were,
on average, *84% correct. We suggest that the CCQ-R
may provide a valid estimation of caffeine consumption.
CCQ-R accuracy was not impacted by any demograph-
ics except whether participants had been advised to avoid
caffeine. We believe these data provide evidence for con-
struct validity, in that we would expect no influence of
most demographics on the ability to complete the measure
accurately; however, we might expect that those instructed
to avoid caffeine would have better knowledge of what
products may contain caffeine (and thus make fewer
reporting errors) than those who have not been instructed
to avoid caffeine. Percent correct scores yielded nearly
84% average accuracy on the CCQ-R and *60% of the
sample scored 85% correct or higher; 85% agreement
is an accepted standard of inter-rater agreement.29 The
CCQ-R demonstrated criterion validity, in that a major-
ity of percent correct scores reached and exceeded 85%
correct inter-rater agreement (i.e., researcher-calculated
scores compared with participant-reported scores).
In addition to evidence for validity, the CCQ-R has a
number of strengths. The measure is easily administered
using a web-based survey or it can be printed for paper/
pencil completion. The CCQ-R is readily scored using
Statistical Package for the Social Sciences (SPSS) scor-
ing syntax that includes caffeine concentration values
from the US Food and Drug Administration (FDA),
Mayo Clinic, manufacturers’ company websites, and prod-
uct packaging; these values can be readily updated in the
scoring syntax should new concentration data become
available for any items. Additionally, three different scores
may be garnered from our scoring syntax: a conservative
estimate using minimum caffeine concentration values
available in calculations, a mean estimate using average
caffeine concentration values, and a liberal estimate
using maximum caffeine concentration values. Further-
more, researchers and clinicians may choose to be
more or less inclusive with respect to item categories
for which data are gathered and/or scored. For example,
if only coffee consumption is of interest, respondents may
complete only the coffee items category. Items not explic-
itly included on the CCQ-R may be captured using other
options (e.g., caffeinated gum) in each item category;
however, other items will need to be hand scored or man-
ually added to scoring syntax.
The CCQ-R has several implications for research and
clinical practice. As discussed, individuals often under-
estimate their caffeine consumption because the caffeine
content of caffeinated products is not provided on nutri-
tional labels. Indeed, this was evidenced in the current
study such that participants’ estimated use (i.e., an esti-
mate provided before completing the CCQ-R) was not
correlated with their use as measured by the CCQ-R. A
tool that yields a more accurate estimate of caffeine con-
5
sumption allows for researchers to invest more confi-
dence in their data and resultant research findings and
for clinicians to gain a better understanding of consump-
tion patterns that may be maladaptive.
Although we suggest the CCQ-R may be a viable self-
report measure for estimating caffeine use, several cave-
ats are necessary to consider. Given that there are myriad
caffeinated products and items available and little ability
to accurately quantify caffeine levels in many cases (e.g.,
differences in coffee roasting and brewing can result in
variability), it is important to consider the CCQ-R to be
an estimate of caffeine consumption. Not all possible
serving sizes are included on the CCQ-R and researchers
should standardize instructions for how best to estimate
and enter serving sizes for serving sizes not represented.
Furthermore, not all possible products are represented on
the CCQ-R and researchers should specify procedures
for reporting and scoring other items.
A variety of strategies may be helpful for improving
quality and accuracy of CCQ-R responses. Simple re-
spondent errors such as entering serving size where num-
ber of servings should be entered may be easily avoided
with standard instructions and a priori data exclusion cri-
teria. Respondents may also need guidance with respect
to identifying items that contain caffeine versus those
that do not. In the current study, participants who had
been advised to avoid caffeine yielded significantly
higher percent correct scores than those who had not,
presumably because avoiding caffeine requires a rela-
tively complete knowledge of which items contain caf-
feine. Providing a comprehensive list of specific items
containing caffeine may be helpful for improving accu-
racy of reporting. Finally, in some cases, it may be appro-
priate to ask respondents to keep a daily caffeine
consumption log to aid in completion of the CCQ-R.
Conclusions
In summary, the CCQ-R is the first self-report measure
for caffeine use for which there are data to provide evi-
dence for its validity. The CCQ-R is also user-friendly
for both researchers and clinicians, as well as for their re-
spective participants and patients. Specifically, it can be
administered using paper and pencil (see Supplementary
Data: Appendix A) or as a web-based questionnaire
(www.caffeineconsumptionmeasure.com). Scoring is
automated and total consumption along with consump-
tion in item-specific categories can be automatically
calculated with syntax in SPSS (see Supplementary
Data: Appendix B). Use of the CCQ-R outside of a re-
search and clinical context (i.e., self-monitoring pur-
poses) is an additional possible indication in need of
further study.
Author Disclosure Statement
No competing financial interests exist.
 6 IRONS ET AL.

References intoxication: a systematic review and meta-analysis.

Neurosci Biobehav Rev. 2014;47:16–21.
1. Frary CD, Johnson RK, Wang MQ. Food sources and in-
19. Johnson-Greene D, Fatis M, Sonnek D, Shawchuck C. A
takes of caffeine in the diets of persons in the United
survey of caffeine use and associated side effects in a
States. J Am Diet Assoc. 2005;105:110–113.
college population. J Drug Educ. 1988;18:211–220.
2. Gilbert RM. Caffeine consumption. Prog Clin Biol Res.
20. Addicott MA. Caffeine use disorder: a review of the
1984;158:185–213.
evidence and future implications. Curr Addict Rep.
3. Julien RM, Advokat CD, Comaty JE. A Primer of Drug
2014;1:186–192.
Action, 12th ed. United States: Worth Publishers; 2011.
21. United States Food and Drug Administration. Why Isn’t
4. Anderson BL, Juliano LM. Behavior, sleep, and prob-
the Amount of Caffeine a Product Contains Required on
lematic caffeine consumption in a college-aged sample.
a Food Label?; 2015. Available at www.fda.gov/
J Caffeine Res. 2012;2:38–44.
AboutFDA/Transparency/Basics/ucm194317.htm
5. Ludden AB, Wolfson AR. Understanding adolescent caf-
22. Cone EJ. Saliva testing for drugs of abuse. Ann N Y
feine use: connecting use patterns with expectancies, rea-
Acad Sci. 1993;20:91–127.
sons, and sleep. Health Educ Behav. 2010;37:330–342.
23. Wolff K, Farrell M, Mardsen J, Monteiro MG, Ali R,
6. Nehlig A. Are we dependent upon coffee and caffeine? A
Welch S, Strang J. A review of biological indicators of
review on human and animal data. Neurosci Biobehav
illicit drug use, practical considerations and clinical use-
Rev. 1999;23:563–576.
fulness. Addiction. 1999;94:1279–1298.
7. Diamond S, Balm TK, Freitag FG. Ibuprofen plus caf-
24. Haeckel R. Relationship between intraindividual vari-
feine in the treatment of tension-type headache. Clin
ation of the saliva/plasma- and of the arteriovenous
Pharmacol Ther. 2000;68:312–319.
concentration ratio as demonstrated by the administra-
8. Einöther S, Giesbrect T. Caffeine as an attention enhanc-
tion of caffeine. J Clin Chem Clin Biochem. 1990;28:
er: revising existing assumptions. Psychopharmacology
279–284.
(Berl). 2013;225:251–274.
25. Hjorthøj CR, Hjorthøj AR, Nordentoft M. Validity of
9. Klaassen EB, de Groot RHM, Evers EAT, et al. The ef-
timeline follow-back for self-reported use of cannabis
fect of caffeine on working memory load-related brain
and other illicit substances: systematic review and
activation in middle-aged males. Neuropharmacology.
meta-analysis. Addict Behav. 2012;37:225–233.
2013;64:160–167.
26. Hjorthøj CR, Fohlmann A, Larsen AM, Arendt M,
10. Juliano LM, Griffiths RR. A critical review of caffeine
Nordentoft M. Correlations and agreement between
withdrawal: empirical validation of symptoms and
delta-9-tetrahydrocannabinol (THC) in blood plasma
signs, incidence, severity, and associated features. Psy-
and timeline follow-back (TLFB)-assisted self-reported
chopharmacology (Berl). 2004;76:1–29.
use of cannabis of patients with cannabis use disorder
11. Hasin DS, O’Brien CP, Auriacombe M, et al. DSM-5 cri-
and psychotic illness attending the CapOpus random-
teria for substance use disorders: recommendations and
ized clinical trial. Addiction. 2012;107:1123–1131.
rationale. Am J Psychiatry. 2013;170:834–851.
27. Landrum RE. College students’ use of caffeine and
12. Hartley TR, Sung BH, Pincomb GA, Whitsett TL, Wilson
its relationship to personality. Coll Stud J. 1992;26:
MF, Lovallo WR. Hypertension risk status and effect of caf-
151–155.
feine on blood pressure. Hypertension. 2000;36:137–141.
28. Nova P, Hernandez B, Ptolemy AS, Zeitzer JM. Model-
13. Klatsky AL, Friedman GD, Armstrong MA. Coffee use
ing caffeine concentrations with the Stanford Caffeine
prior to myocardial infarction restudied: heavier intake
Questionnaire: preliminary evidence for an interaction
may increase the risk. Am J Epidemiol. 1990;132:479–488.
of chronotype with the effects of caffeine on sleep.
14. Luebbe AM, Bell DJ. Mountain Dew or Mountain
Sleep Med. 2012;13:362–367.
Don’t?: a pilot investigation of caffeine use parameters
29. Smith RA, Davis SF. Psychologist as Detective: An
and relations to depression and anxiety symptoms in 5th-
Introduction to Conducting Research in Psychology.
and 10th-grade students. J Sch Health. 2009;79:380–387.
Upper Saddle River, NJ: Pearson; 2013.
15. Broderick P, Benjamin AB. Caffeine and psychiatric symp-
toms: a review. J Okla State Med Assoc. 2004;97:538–542.
16. James JE, Crosbie J. Somatic and psychological health
implications of heavy caffeine use. Br J Addict. 1987; Address correspondence to:
82:503–509. Jessica G. Irons, PhD
17. Advokat CD, Comaty JE, Julien RM. Julien’s Primer of Department of Psychology
Drug Action, 13th ed. United States: Worth Publishers; James Madison University
2014. Harrisonburg, VA 22807
18. Benson S, Verster JC, Alford C, Scholey A. Effects of
mixing alcohol with caffeinated beverages on subjective E-mail: ironsjg@gmail.com

View publication stats