You are on page 1of 26

Sexual and Relationship Therapy

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/csmt20

The Consumption of Pornography Scale – General


(COPS – G)

S. Gabe Hatch , Charlotte R. Esplin , H. Dorian Hatch , Aeriel Halstead ,


Joseph Olsen & Scott R. Braithwaite

To cite this article: S. Gabe Hatch , Charlotte R. Esplin , H. Dorian Hatch , Aeriel Halstead ,
Joseph Olsen & Scott R. Braithwaite (2020): The Consumption of Pornography Scale – General
(COPS – G), Sexual and Relationship Therapy, DOI: 10.1080/14681994.2020.1813885

To link to this article: https://doi.org/10.1080/14681994.2020.1813885

Published online: 14 Sep 2020.

Submit your article to this journal

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at


https://www.tandfonline.com/action/journalInformation?journalCode=csmt20
SEXUAL AND RELATIONSHIP THERAPY
https://doi.org/10.1080/14681994.2020.1813885

The Consumption of Pornography Scale – General


(COPS – G)
S. Gabe Hatcha, Charlotte R. Esplinb, H. Dorian Hatchb, Aeriel Halsteadb,
Joseph Olsenb and Scott R. Braithwaiteb
a
Department of Psychology, The University of Miami, Coral Gables, FL, USA; bDepartment of
Psychology, Brigham Young University, Provo, UT, USA

ABSTRACT ARTICLE HISTORY


Pornography is widely used, distributed, and researched. Received 11 December 2019
Nevertheless, few psychometric scales have yet to be developed Accepted 18 August 2020
that assesses this complex construct without pathologizing the
KEYWORDS
behavior or labeling the behavior as problematic. The current study
Pornography; pornography
sought to develop a 34-item, purely behavioral and topographic consumption; measurement;
measure of pornography consumption without assigning value to psychometrics
the behavior. In the current study (N ¼ 317), we examine the over-
arching factor structure and provide initial evidence for the scale’s
psychometric characteristics. In line with our hypotheses, we found
a bifactor model with one general factor and four subfactors fit the
data best; the scale was reliable, and scores were stable over time.
Further, the current study provides initial evidence for the scale’s
convergent validity in its relation to sexual sensation seeking, exces-
sive pornography use, and gender as well as evidence for divergent
validity by possessing small correlations with depressive symptoms
and sexual satisfaction. Finally, the current study also provides initial
evidence for the scale’s incremental validity and discusses limita-
tions and ideas for future research.

Pornography is widely used by Americans (Cooper, 1998; Pornhub.com, 2018;


Ropelato, 2006). Because of its ubiquity, it is necessary to have a well-validated and
reliable psychometric scale that measures the consumption of pornography to
adequately examine its potential impact. Some have called into question whether
existing scales assessing the consumption of pornography have acceptable content val-
idity (B}
othe et al., 2018; Duffy et al., 2016; Short et al., 2012). Although these claims
are difficult to disentangle, one interpretation is that the current literature lacks a
purely behavioral measure of pornography consumption that does not pathologize
the behavior or label it as problematic. Indeed, the lack of a psychometrically vali-
dated behavioral measure of pornography consumption may be to blame for some of
the conflicting results in the existing literature (e.g. pornography use and sexual
aggression; see Fisher & Barak, 2001; Fisher & Grenier, 1994; Malamuth et al., 2000 ).

CONTACT S. Gabe Hatch sgh49@miami.edu Department of Psychology, The University of Miami, 5665 Ponce
De Leon Blvd., Coral Gables, FL 33124, USA
ß 2020 College of Sexual and Relationship Therapists
2 S. G. HATCH ET AL.

Within this area of research, this is perhaps best illustrated in the use of single-
item measures. It is currently common practice to use a single-item measure to assess
pornography consumption (Goldsmith et al., 2017; Kohut et al., 2020; Short et al.,
2012). Some authors have estimated that approximately 62% of the studies investigat-
ing the effects of pornography consumption only use a single-item measure, 95% of
studies use varying definitions of pornography, and only 5% of studies use psycho-
metrically validated measures to assess pornography consumption (Short et al., 2012).
A reliable and valid single-item measure may be useful in some contexts if it is reli-
able and valid (see Fuchs & Diamantopoulos, 2009); however, single-item measures
do not appear to have served this area of research well. The primary concern with
idiosyncratic single-item measures for each study is that they may differ in the con-
tent being assessed (Kohut et al., 2020). This is perhaps best illustrated when trying
to estimate the prevalence of pornography consumption – something that should be a
simple task. Ropelato (2006), for instance, suggests that approximately 13% of U.S.
adults regularly view pornography. Wright (2013) asked participants whether they
had viewed a pornographic film within the past year and 30.8% of U.S. male partici-
pants answered that they had. Grubbs et al. (2015) asked their college participants
how often they had viewed pornography in the past month, 50.6% of participants
indicated that had. Zheng and Zheng (2014), reported that approximately 90% of
their participants reported engaging in online sexual activities and Donevan and
Mattebo (2017) reported that nearly 98% of their respondents reported watching
pornography. Although these estimates come from different samples, which could
explain some of the differences, idiosyncratic single-item measures appear to explain
some of the variance in these divergent estimates. Indeed, Kohut et al. (2020) recently
argued that “if such basic issues around measurement remain unresolved in this field,
how confident can we be about the reported impact of pornography” (p. 1)?
In addition to using single-item idiosyncratic questions across studies, another
limitation in the measurement of pornography consumption is the clear harm-
focused narrative that most of the current measures employ (Kohut et al., 2020). Two
examples of such scales include The Cyber-Pornography Use Inventory (CPUI;
Grubbs et al., 2010), and The Problematic Pornography Consumption Scale (PPCS;
B}othe et al., 2018). The CPUI broadly measures the addictive patterns of pornography
use, guilt regarding online pornography use, and online sexual activities. As the indi-
vidual names of each subscale within CPUI suggests, this scale appears to measure
addiction to pornography use and negative emotions accompanying use. Additionally,
as its name implies, the PPCS measures problematic consumption and is theoretically
based on an addiction model (B} othe et al., 2018). Each of these measures frames
pornography consumption within a harm-focused narrative and assess reasons for
consuming pornography as opposed to consumption patterns. If pornography con-
sumption is framed in this theoretical context, investigators are no longer assessing
the behavior of consuming pornography but an individual’s negative perceptions of
their own pornography use. Indeed
If we are uncritical we shall always find what we want: we shall look for, and find,
confirmations, and we shall look away from, and not see, whatever might be dangerous
to our pet theories. In this way, it is only too easy to obtain what appears to be
SEXUAL AND RELATIONSHIP THERAPY 3

overwhelming evidence in favor of a theory which, if approached critically, would have


been refuted (Popper, 1945, p. 79).
Past research has generally found that measuring perceptions of use yields different
outcomes than when assessing behavior (Hald & Malamuth, 2008) and if scale devel-
opment in this area of research continues to focus on creating perception- and val-
ues-based measures of pornography consumption, how can we begin to investigate
the full range of effects of consuming pornography (positive or negative)?
The Pornography Consumption Questionnaire has attempted to accomplish this
task (PCQ; Hald, 2006). The PCQ defines pornography, assesses sociodemographic
characteristics, exposure to pornographic content, pornography preferences, money
spent on pornography, and the positive and negative effects of pornography (Hald,
2006). The PCQ is one of the only measures that appears to assess the behavioral
aspects of pornography use (although it fails to inquire about arousal to pornographic
content). Although the PCQ has many excellent features, it is long, multifaceted, and
lacks theoretical development. With 139 items, the PCQ cannot be used in research
settings where investigators are concerned about excessive item load. To our know-
ledge, the PCQ as a whole has never undergone psychometric development – only
sections of the PCQ have been psychometrically validated. These validation efforts
have 1) focused on perceived effects of pornography rather than behaviors, 2) have
been psychometrically validated only after they have been proposed (Miller et al.,
2019), and 3) lack item content to create a purely behavioral measure of pornography
consumption. Thus, it is important that if we intend to investigate the full range of
effects stemming from pornography consumption, that we measure the behaviors
associated with pornography use rather than perceptions of use.

Defining pornography
Before developing the theoretical underpinnings of the behavior of pornography con-
sumption, it is important to have a conceptual definition and the current study
sought to capture breadth rather than specificity (Kohut et al., 2020). Yet, heterogen-
ous conceptual definitions and the inability to conceptually define what pornography
is, and is not, plague the current literature (Busby et al., 2017; Kohut et al., 2019;
Marshall & Miller, 2019; McKee et al., 2020). A popular definition is that “sexually
explicit materials intended to arouse” (McKee et al., 2020). The problem with this
definition includes having to guess the intention of the material and what is consid-
ered to be sexually explicit – something that may change over time or never be
understood (McKee et al., 2020). Other definitions include “media content depicting
nudity and explicit sexual acts” (Wright et al., 2016, p. 954). Sexual education text-
books would fall under this definition of pornography but would fail to capture what
some with non-nudity-related fetishistic disorders may consider being pornographic.
These idiosyncratic definitions in the research community are not uncommon
(Kohut, 2014; Marshall & Miller, 2019), and appear to reflect a larger issue with the
pornography consumption literature – that the study and individual interpretations
(i.e. by scientists and the participants they study) of this inherently self-identified
4 S. G. HATCH ET AL.

fuzzy concept are phenomenological, falling inline within the definition of an “open
concept” (see Lilienfeld et al., 2015; Meehl, 1986).
This difficulty is perhaps best illustrated in a paper by Kohut (2014) who, using
qualitative and quantitative methodologies, sought to conceptually refine the defin-
ition of pornography. Of interest in his paper was one theme that emerged among
previous definitions in the research community: defining pornography as art, or lack
thereof. Interestingly, when college students were asked to endorse whether this “art”
definition was a defining characteristic of pornography, it appeared to be the most
divisively endorsed trait. This division is due, at least in part, to define an open con-
cept (pornography) with another open concept (art). If an artistic definition were
endorsed, a picture of a shoe might be considered pornography because it is “artistic,”
and to someone with a fetishistic disorder, it also falls in line with the two highest
endorsed elements of pornography: something that promotes sexual arousal and grati-
fication. Many of these elements are obscure and, and when distilling these data into
a uniform definition, require necessary value judgments by the researchers and value
judgments that can interrupt unbiased investigation. The conceptual definition we
most prefer (see Kohut et al., 2020) has many great qualities; however, is still quite
vague. In sum, we believe the more accurate the definition of pornography, the more
ambiguous it becomes.
To this end, we prefer a different tactic when defining pornography, that is, we prefer
a “bottom-up” approach. Instead of demanding participants conform to our definition of
pornography, thereby measuring a small fraction of what pornography might or might
not be, we prefer to leave that within the agency of the participant. Indeed, the lay popu-
lation may know something about pornography that the research community may not
(Grubbs et al., 2019), in this case, they know what pornography is. This falls in line with
the logic of Supreme Court Justice Potter Stewart who, when attempting to define porn-
ography, said: “I know it when I see it.” Though we do acknowledge that this approach
might be seen as a weakness, and comes with its share difficulties, we believe that leaving
this concept “open” is in the best interest of the research as we are not disregarding an
individual’s personal experiences with pornography because they do not align with phe-
nomenological definitions of a scientist.
We posit that the impact of pornography is phenomenological—a function of indi-
vidual interpretation and felt experience, rather than an ill-fitting definition foisted
on a construct that has inherently fuzzy boundaries. Thus, our operationalization of
pornography consumption does not impose a definition but allows participants to
indicate whether they viewed “pornography” according to their judgment and mean-
ing of that word. When certain definitions are required, to maintain precision in our
language and not tax the working memory of our participants, this strict definition is
introduced at the item- rather than the scale-level.

Dimensions of behavior
Considering these limitations, we sought to develop a purely behavioral measure of
pornography consumption through a topographical lens. Four dimensional quantities
are important when assessing behavior: Repeatability, Temporal Extent, Topography,
SEXUAL AND RELATIONSHIP THERAPY 5

and Magnitude (Cooper et al., 2014; Johnston & Pennypacker, 1993). Repeatability
can be measured by how frequently the behavior occurs within a period of time
(Cooper et al., 2014). The current literature has typically utilized frequency to meas-
ure the behavior of pornography consumption (see Hald & Stulhofer, 2016; Perry,
2015, 2016; Short et al., 2012). The predominant nature of repeatability or frequency
measures within the literature is understandable because the frequency of a behavior
is the fundamental datum when measuring that behavior (Skinner, 1938). However,
simple measures of frequency are insufficient as they do not assess how long a par-
ticipant viewed pornography, the type of pornography being consumed, or physio-
logical responses to pornographic content.
These are deficiencies best captured by the other three behavioral dimensions:
Temporal Extent, Topography, and Magnitude. Temporal Extent has been defined as
the duration of the behavior, or the amount of time in which the behavior occurs
(Cooper et al.,2014; Johnston & Pennypacker, 1993). In the current literature, this
could be best conceptualized as duration or the amount of time spent viewing porn-
ography. Topography refers to the physical form of the behavior, or the type of porn-
ography being consumed, as well as whether pornography was consumed deliberately
or accidentally. However, the importance of delineating between deliberate or acci-
dental exposure may be best illustrated by the final behavioral dimension: Magnitude
(Cooper et al., 2014). Magnitude is defined as the intensity with which a response is
emitted or arousal to pornographic content; in other words, the more aroused the
greater the intensity. Different arousal levels may be a function of whether an individ-
ual was deliberately or accidentally exposed (Flood, 2007). Also, because sexual
arousal may be a function of the type of pornography being consumed rather than a
sexual act, we conceptualize that exposure and arousal belong to the same dimension.
In line with the above model, we theorize a four-factor behavioral measure of porn-
ography consumption: frequency of exposure, duration of exposure, accidental expos-
ure and arousal to pornographic content, and deliberate exposure and arousal to
pornographic content. Together these dimensions inform behavioral measurement in
ways a single-dimensional quantity cannot; measure the behavior of pornography
consumption with theoretical consistency; and take a topographical view of pornog-
raphy consumption thereby not pathologizing the behavior (Cooper et al., 2014).

The current study


In light of our behavioral model of pornography consumption, the current study has
three aims. The first aim is to examine the factor structure of pornography consump-
tion. Specifically, we intend to investigate whether these behaviors are best under-
stood as being related to one another, belonging to a single general factor, or
understood as both a general factor and individual subscales. Because past research
has generally understood the frequency of pornography consumption, the duration of
pornography consumption, and accidental/deliberate exposure and arousal to porno-
graphic material to be interrelated but distinct constructs, we hypothesize that the
four-factor associative model will provide the best model fit. Also, these models will
prove to be reliable and stable over time.
6 S. G. HATCH ET AL.

Second, after determining the best-fitting model, convergent validity will be tested
by examining the relationships between pornography consumption and measures of
sexual satisfaction, excessive pornography use, depressive symptoms, sexual sensation
seeking, and gender. Past research has indicated that males view pornography more
frequently than females (Hald, 2006), that use is negatively associated with sexual sat-
isfaction and depressive symptoms (Nelson et al., 2010; Wolak et al., 2007; Wright
et al., 2017), and positively associated and sexual sensation seeking (Beyens et al.,
2015; Zheng & Zheng, 2014). We theorize that similar patterns will exist with our
behavioral measure of pornography consumption.
In the final aim of the study, we seek to determine whether our newly developed
measure adds incremental validity above and beyond preexisting measures in the cur-
rent literature. We hypothesize that our measure will account for more variance than
a single-item measure (i.e. what is employed in the current literature). Incremental
validity will be tested with sexual satisfaction, excessive pornography use, depressive
symptoms, sexual sensation seeking, and gender each being specified as outcomes.

Method
Before collecting data, permission was sought and granted from the local IRB. Data
come from a larger study examining whether pornography consumption changes over
time. Exactly 321 responses were collected from Amazon’s Mechanical Turk using
TurkPrime.com. Four responses were deleted from the analysis because they were
duplicates. In total, 317 individuals (172 males, 141 females, and 4 that did not indi-
cate their gender) were recruited via TurkPrime.com (Litman et al., 2017). Eligible
individual participants were restricted to the United States and were allowed to have
any range of HIT completion percentage. The average age of participants was 36.32
(SD ¼ 10.20). A majority of participants were White (73.19%), 10.41% were Black,
5.36% were Asian, 5.36% were Latina(o), 3.15% reported being Biracial, and 1.26%
reported belonging to a race not listed. Participants were asked to complete a 20-
minute survey each month for six months. Surveys were published at the end of each
month and left online for 14 days. In exchange for their responses, participants were
compensated $0.50 per survey and entered into monthly randomized drawings for
Amazon.com gift certificates.

Creating items for the Consumption of Pornography Scale – General


Although we borrowed iterations of some items previously used in academic papers
(e.g. “Please indicate how often you viewed pornography in the past 30-days?”; Hatch
et al., 2020; Negash et al., 2016) most of the questions that were developed for the
current scale were developed by our research team. All items that were created were
intended to represent the four theoretical constructs commonly measured in the lit-
erature (i.e. frequency of exposure, duration of exposure, and accidental arousal to
pornographic content, and deliberate exposure and arousal to pornographic content)
using the logical-content approach to test development (Kaplan & Saccuzzo, 2017).
SEXUAL AND RELATIONSHIP THERAPY 7

Items measuring the frequency of exposure


Items assessing the frequency of exposure were variants of a Likert-scale item fre-
quently used in research on pornography consumption (i.e. “Please indicate how
often you viewed pornography in the past 30-days?”). We both expanded and nar-
rowed the scope of this question. In addition to asking how often respondents viewed
pornography in the past 30-days, we asked participants how often they had viewed
pornography in the past year and how many times they viewed pornography this
week (see Appendix A).

Items measuring the duration of exposure


Motivations for those who consume pornography frequently but not for long periods
of time may be different than those who consume pornography infrequently but for
long periods of time. Those who view pornography infrequently but for long amounts
of time may do so for educational purposes (Esplin et al., 2020), do so while edging,
or when the risk of getting caught is low. Thus, we will be able to parse apart two
very different groups of pornography users and study their outcomes separately. We
developed three questions intended to measure the amount of time spent consuming
pornography during a viewing session (see Appendix A).

Items measuring exposure and arousal to pornographic content


The type of pornography consumed, the intention behind consumption, and the
arousal that occurs during consumption could play a large role in reinforcing the fre-
quency and duration of exposure. Past studies investigating exposure to pornographic
content have utilized more thorough methods than the one we will employ in this
study (see Hald, 2006; Hald & Stulhofer, 2016). However, gathering genres directly
from pornography websites may be problematic for a few reasons. First, asking poten-
tial consumers about every single genre of pornography that they may have viewed
would likely increase the variation within the scale – it would make the scale more
time-consuming. Second, it would be useful to make this scale as generalizable as
possible, for as long as possible. It is unreasonable to assume that every single genre
of pornography has already been created – more are likely to be produced.
To cover as much conceptual ground as possible with the fewest items, participants
were, in an alternating fashion, first asked to indicate whether they had been acciden-
tally exposed to different types of pornography e.g. pornography depicting heterosex-
ual couples, same-sex couples, solitary individuals, three partners, group sex with
more than three partners, and depictions of S&M, they were then asked whether they
had been deliberately exposed to the aforementioned pornography genres (see
Appendix A). On each of the seven general types of pornography, participants were
asked to indicate either “Yes” (coded as 1) or “No” (0). If participants endorsed that
they had been exposed to one of these types of pornography (either on purpose or by
accident), they were asked to indicate how aroused they had become on an eight-
point Likert scale. Those that answered “No” to the exposure question were coded as
1 on the arousal question (not at all aroused) and then multiplied by the associated
arousal question creating a single-item and getting rid of any missingness problems.
8 S. G. HATCH ET AL.

Measures
The Global Measure of Sexual Satisfaction (GMSEX)
The GMSEX is a five-item scale that measures sexual satisfaction using a seven-point
semantic differential scale, e.g. good-bad (Lawrance & Byers, 1995). Composite reli-
ability of the GMSEX in the current sample was excellent (x ¼ 0.97).

The Problematic Pornography Use Scale – Excessive Use Subscale


The Excessive Use subscale of the Pornography Consumption Use Scale is a three-
item test on a six-point Likert scale (Kor et al., 2014). Scores on the Excessive Use
subscale are significantly correlated with the compulsive use subscale of the CPUI (r
¼ .56; Grubbs et al., 2015), providing evidence of convergent validity. Composite reli-
ability in the current sample was excellent (x ¼ 0.90).

The Center for Epidemiological Studies Depression Scale (CES-D-10)


The CES-D-10 is a screener for depression (Andresen et al., 1994). The CES-D-10
was highly correlated with the BASIS-24 Depression and Functioning subscale (r ¼
.85) and had a low correlation with the BASIS-24 psychosis (r ¼ .18) and substance
abuse (r ¼ .09) subscales, providing evidence of convergent and divergent validity,
respectively. Composite reliability in the current sample was good (x ¼ 0.89).

The Sexual Sensation Seeking Scale (SSSS)


Sexual sensation seeking is the propensity to seek out new and exciting sexual stimu-
lation or arousal (Kalichman et al., 1994). The nine-item test is measured on a four-
point Likert scale (e.g. “I like wild uninhibited sexual encounters”). In an attempt to
keep our survey as short as possible, we used the shortest version for this scale, one
that had been validated among men who have sex with men. The SSSS was also cor-
related with unprotected anal sex (r ¼ .30, p < .05) and number of sexual partners (r
¼ .34, p < .05), providing evidence for the scale’s validity. Composite reliability for
this scale in the current sample was good (x ¼ 0.85).

Gender
Participants were also asked to indicate their gender. Those who identified as male
were coded as 0 and those that identified as female were coded as 1.

Data analysis
In the first aim of the study, we intend to examine the factor structure of our newly
proposed measure and its reliability. To do this, we conducted three confirmatory
factor analyses: an associated factors model, a hierarchical model, and a bifactor
model. These three models suited our research question well, as we will be able to
determine whether the unique subscales are related but best understood as independ-
ent factors (an associated factors model), caused by a higher-order factor (a hierarch-
ical model), or whether unique subscales co-exist with a general factor (a bifactor
model). All confirmatory factor analyses were computed using the diagonally
weighted least square estimator in Mplus (WLSMV; which is utilized when categorical
SEXUAL AND RELATIONSHIP THERAPY 9

variables are specified as indicators, see Rosseel, 2012) and the model that displays
the best fit to the data will be selected for the proceeding reliability and validity anal-
yses. Similar to other studies in this area of research (B}
othe et al., 2018), the goodness
of fit was determined according to the following criteria: CFI ( 0.95 for good fit),
TLI ( 0.95 for good fit), RMSEA (0.05 for good fit, 0.05-0.08 adequate fit, 0.08-
0.10 mediocre fit), and SRMR ( 0.10 for acceptable fit,  0.05 for good fit; see
Schermelleh-Engel et al., 2003). Additionally, all coefficients reported in the current
article were standardized.
Next, instead of reporting Cronbach’s alpha as our metric for reliability, we report
McDonald’s Omega and Omega Hierarchical coefficients. These indices outperform
Cronbach’s alpha when congeneric models are being estimated – i.e. models where
each parameter is theorized to contribute uniquely to the confirmatory factor analysis.
Lastly, test-retest reliability will also be computed using correlations, allowing us to
determine whether the construct(s) are stable over time.
If the bifactor model fits the data best, variance added ratios (VAR) will be com-
puted to see whether the addition of the subscales offers value above and beyond the
general factor. Past research recommended that VARs greater than 1.1 provide valu-
able information otherwise not captured by the general factor. Thus, subscales will be
reported if their VAR is greater than 1.1 (Feinberg & Jurich, 2017; Feinberg &
Wainer, 2014).
The second aim of the current study seeks to build upon the first by offering ini-
tial evidence for scale validation by providing evidence of convergent validity. To
accomplish this, we compute correlations between the latent variables of interest and
the measured variables in the SEM framework. This suits our research question well
in that we are able to examine whether our scale has convergent validity with com-
monly assessed measures in the current literature.
In addressing the final aim of the current study, which answers whether the cur-
rent measure adds incremental validity above and beyond other measures employed
in the current literature, we compare the fit of regression models that only assess for
the frequency of pornography use in the past 30 days to regression models that
include other factors supported by Aim 1 using summed scores because factor scores
are not currently supported within the WLSMV framework in Mplus (Muthen &
Muthen, 1998–2017).

Missing data
Missing data were handled by generating five multiple imputed datasets using the
mice package in RStudio (Buuren & Groothuis-Oudshoorn, 2010). Specifically, data
for the confirmatory factor analyses were imputed using the proportional odds model
for ordered categorical variables. As debates in the field continue regarding how
model fit indices in structural equation models should be computed across imputa-
tions, we attempt to address this concern by reporting mean indices across imputa-
tions. Structural and path coefficients, however, are only reported from the first
imputed dataset as no substantial differences between models were noted. Data for
10 S. G. HATCH ET AL.

the second and third aims of the study, on the other hand, were imputed using pre-
dictive mean matching using five imputed datasets.

Results
Aim 1: examining the factor structure of pornography consumption
To examine the factor structure of the general behavior of pornography consumption,
we examined a four-factor associative model, a hierarchical model, and a bifactor
model which has been recommended in other measurement studies (Morin et al.,
2016). We first examined the four-factor associative model which implies that the
unique subscales may be related but are best understood as independent factors. The
four-factor solution was fit and the mean indices across imputations indicated that
the model displayed good fit according to the CFI and TLI, mediocre fit according to
the RMSEA, and acceptable fit for the SRMR (v2 (164) ¼ 685.39, p < 0.001; CFI ¼
0.962; TLI ¼ 0.956; RMSEA ¼ 0.100, 90% CI [0.092, 0.108]; SRMR ¼ 0.065; see
Figure 1).
Next, we fit a hierarchical model. The hierarchical model suggests that while acci-
dental exposure to pornographic content, deliberate exposure to pornographic con-
tent, the frequency of exposure to pornography, and the duration of exposure to
pornography are all important in understanding pornography consumption, these fac-
tors are best understood as being caused by a higher-order behavior – i.e. general
pornography consumption. The mean indices on the hierarchical model provided
good fit according to the CFI and TLI, mediocre fit according to the RMSEA, and
acceptable fit according to the SRMR (v2 (166) ¼ 614.93, p < 0.001; CFI ¼ 0.968; TLI
¼ 0.962; RMSEA ¼ 0.092, 90% CI [0.085, 0.100]; SRMR ¼ 0.066; see Figure 2).
Indeed, our findings indicated acceptable model fit for both the associative factors
model as well as the hierarchical model. One useful way of assessing whether the
model is determined to reflect multiple related factors, a single higher-order factor, or
if these factors co-exist, is to fit a bifactor model. This is done by examining goodness
of fit, assessing factor loadings, as well as computing reliability coefficients (see
Morin et al., 2016; Rodriguez et al., 2016). Unsurprisingly, the bifactor model pro-
vided the best fit. According to the mean estimates of the CFI and TLI the model fit
the data well, the mean estimate of RMSEA suggested mediocre fit, and the mean
estimate of SRMR was acceptable (v2 (150) ¼ 496.80, p < 0.001; CFI ¼ 0.974; TLI ¼
0.968; RMSEA ¼ 0.085, 90% CI [0.077, 0.094]; SRMR ¼ 0.060; see Figure 3). The
general factor of pornography consumption was well-defined with factor loadings fall-
ing between 0.243 and 0.784 (M ¼ 0.523, all p-values < 0.001). Above and beyond
this, three of the specific factors also appeared to be well-defined including the fre-
quency of pornography consumption (jkj ¼ 0.661-0.773, M ¼ 0.723, all p-values <
0.001), the duration of pornography consumption (jkj ¼ 0.771-0.904, M ¼ 0.849, all
p-values < 0.001), and accidental exposure and arousal to pornography consumption
(jkj ¼ 0.252-0.832, M ¼ 0.580, all p-values < 0.001) each offering unique residual spe-
cificity even after taking into account the general behavior of pornography consump-
tion. Deliberate exposure and arousal to pornography consumption, however,
appeared to be the least well-defined (jkj ¼ 0.074-0.587, M ¼ 0.384), with one of the
SEXUAL AND RELATIONSHIP THERAPY 11

Figure 1. Four-factor associative model solution. Note. Item numbers in the figure correspond to
the item numbers in the Appendix. Items in the accidental exposure and arousal as well as the
deliberate exposure and arousal subscales represent the interaction between the exposure and cor-
responding arousal item.

Figure 2. Hierarchical model solution. Note. Item numbers in the figure correspond to the item
numbers in the Appendix. Items in the accidental exposure and arousal as well as the deliberate
exposure and arousal subscales represent the interaction between the exposure and corresponding
arousal item.

factor loadings not being statistically significant and offering lower levels of residual
specificity when the general factor is considered.
Next, the omega reliability coefficient for the total scale was good (x ¼ 0.843) and
64% of the variance of unit-weighted total scores could be attributed to individual
differences on the general factor (xH ¼ 0.639). The reliability of the frequency of the
pornography consumption subscale was good (x ¼ 0.879) and it explained 4% (xHS
¼ 0.036) of the variance after accounting for the general factor. This was a similar
finding for each of the subscales in that each possessed good reliability, and explained
up to 12% of the variance after accounting for the general factor (duration x ¼ 0.885,
xHS ¼ 0.050; accidental exposure and arousal x ¼ 0.871, xHS ¼ 0.115; deliberate
exposure and arousal x ¼ 0.853, xHS ¼ 0.053).
Because the bifactor model had the best model fit, value-added ratios (VAR) were
calculated to see whether the addition of the subscales offers additional value and are
worth reporting. Computation of each VAR indicated that each subscale (frequency
VAR ¼ 3.42; duration VAR ¼ 6.18; accidental exposure and arousal VAR ¼ 1.15)
provided valuable information not entirely captured by the general factor. Although
the VAR was slightly below the recommended cutoff for the deliberate exposure and
arousal subscale (VAR ¼ 1.05), we nevertheless continued to include the subscale in
12 S. G. HATCH ET AL.

Figure 3. Bifactor model solution. Note. Item numbers in the figure correspond to the item num-
bers in the Appendix. Items in the accidental exposure and arousal as well as the deliberate expos-
ure and arousal subscales represent the interaction between the exposure and corresponding
arousal item.

the proceeding analysis because it did not appear to fall within the misleading range
as specified by Feinberg and Jurich (2017) and to help maintain uniformity through-
out the proceeding analyses.
So far, our findings have indicated that the associative model and the hierarchical
model provide acceptable fit to the data; however, the bifactor model appeared to
offer the best fit. Upon closer examination of the bifactor model, we found that 1)
factor loadings appeared to load onto both the general factor as well as the individual
subscales under most scenarios, 2) reliability appeared to be high for the entire scale
as well as each of the subscales, and 3) although the general factor accounts for a
large part of the variance, the variance accounted for by some of the factors (e.g. the
accidental exposure and arousal subscale) should not be ignored and offer valuable
information.
Test-retest reliability fell between r ¼ 0.71-0.93 for one month, r ¼ 0.71-0.81 for
two months, r ¼ 0.67-0.82 for three months, r ¼ 0.65-0.81 for four months, r ¼ 0.70-
0.81 for five months, and r ¼ 0.74 for six months between assessments suggesting that
general pornography consumption is stable over time. The frequency subscale
(r ¼ 0.72  0.92, M ¼ 0.83), duration subscale (r ¼ 0.59-0.84, M ¼ 0.72), accidental
exposure and arousal (r ¼ 0.48-0.87, M ¼ 0.70), and deliberate exposure and arousal
(r ¼ 0.70-0.88, M ¼ 0.80) subscales possessed similar reliability over the six-month
period suggesting that the scale is stable over time.

Aim 2: convergent validity of the Consumption of Pornography Scale – General


The general pornography factor was significantly related to sexual sensation seeking
(r ¼ 0.64, 95% CI [0.56, 0.72], p < 0.001), depressive symptoms (r ¼ 0.16, 95% CI
[0.04, 0.28], p ¼ 0.009), excessive pornography use (r ¼ 0.41, 95% CI [0.31, 0.51],
SEXUAL AND RELATIONSHIP THERAPY 13

p < 0.001), and gender (r ¼ 0.59, 95% CI [-0.56, 0.54], p < 0.001) corroborating
our initial predictions. However, general pornography consumption was unrelated
with sexual satisfaction (r ¼ 0.02, 95% CI [-0.14, 0.08], p ¼ 0.70).
The frequency of pornography consumption was significantly related to sexual sen-
sation seeking (r ¼ 0.63, 95% CI [0.49, 0.78], p < 0.001), excessive pornography use
(r ¼ 0.46, 95% CI [0.39, 0.54], p < 0.001), sexual satisfaction (r ¼ 0.16, 95% CI
[-0.29, 0.02], p ¼ 0.02), and females reported viewing pornography less frequently
than males (r ¼ 0.65, 95% CI [-0.77, 0.54], p < 0.001) but not related to depres-
sive symptoms (r ¼ 0.11, 95% CI [-0.03, 0.25], p ¼ 0.13).
The duration of pornography consumption was significantly related to sexual sen-
sation seeking (r ¼ 0.23, 95% CI [0.12, 0.35], p < 0.001), women appeared to consume
pornography for less amounts of time than men (r ¼ 0.06, 95% CI [-0.11, 0.01],
p ¼ 0.02), and the duration of use was related to excessive pornography use (r ¼ 0.44,
95% CI [0.34, 0.55], p < 0.001). However, the duration of consumption was unrelated
with sexual satisfaction (r ¼ 0.02, 95% CI [-0.09, 0.14], p ¼ 0.70) and there was insuffi-
cient evidence to conclude that there was a relationship between the duration of
pornography consumption and depressive symptoms (r ¼ 0.01, 95% CI [-0.11,
0.12], p ¼ 0.92).
Accidental exposure and arousal to pornography was significantly related to sexual
sensation seeking (r ¼ 0.50, 95% CI [0.36, 0.64], p < 0.001), depressive symptoms
(r ¼ 0.19, 95% CI [0.05, 0.34], p ¼ 0.009), excessive pornography use (r ¼ 0.43, 95% CI
[0.26, 0.59], p < 0.001), and females reported less accidental exposure and arousal to
pornography than men (r ¼ 0.25, 95% CI [-0.40, 0.12], p ¼ 0.002). However, acci-
dental exposure and arousal to pornography was not significantly related to sexual
satisfaction (r ¼ 0.05, 95% CI [-0.20, 0.09], p ¼ 0.49).
Deliberate exposure and arousal to pornography was significantly related to sexual
sensation seeking (r ¼ 0.86, 95% CI [0.63, 0.99], p < 0.001), excessive pornography use
(r ¼ 0.26, 95% CI [0.12, 0.40], p < 0.001), and females reported being less deliberately
exposed and aroused to pornography than men (r ¼ 0.47, 95% CI [-0.64, 0.30],
p < 0.001). However, deliberate exposure and arousal to pornography was unrelated
to depressive symptoms (r ¼ 0.07, 95% CI [-0.10, 0.22], p ¼ 0.43) or sexual satisfaction
(r ¼ 0.06, 95% CI [-0.08, 0.20], p ¼ 0.40).
Although our initial prediction was that each of the different facets of pornography
consumption would have statistically reliable relationships with gender, sexual satis-
faction, sexual sensation seeking, depressive symptoms, and excessive pornography
use, this pattern of findings is not surprising. Given that the different scales have a
different pattern of statistically significant relationships, and that the strength of each
of these relationships varies, this pattern of findings serves to validate the newly pro-
posed scale. Each of these behavioral constructs has the propensity to measure differ-
ent types, patterns, and strengths of relationships.

Aim 3: incremental validity


Finally, to establish incremental validity, we compared a multiple regression model
that included single-item measure of frequency (i.e. how often have you viewed
14 S. G. HATCH ET AL.

pornography in the past 30 days?) to 1) a model that added the newly developed gen-
eral factor, and 2) a model that added the newly developed subfactors. The newly
developed frequency of exposure subscale was not included as a predictor due to
anticipated multicollinearity issues with the single-item measure.
After controlling for the frequency of exposure within the last 30 days, neither the
addition of the general factor (v2 (1, 721.56) ¼ 0.24, p ¼ 0.62, D R2 ¼ 0.000), nor the
subfactors (v2 (3, 2,938.56) ¼ 0.44, p ¼ 0.73, D R2 ¼ 0.003) added incremental valid-
ity when predicting gender. This pattern of findings was consistent when sexual satis-
faction was specified as an outcome (General Comparison: F(1, 603.93) ¼ 0.97,
p ¼ 0.32, D R2 ¼ 0.003; Subscale Comparison: F(3, 345.42) ¼ 2.23, p ¼ 0.08, D R2 ¼
0.021). When depressive symptoms were specified as the outcome of interest,
although the subfactors did not appear to account for more variance than a single-
item measure of frequency (Subscale Comparison: F(3, 467.86) ¼ 1.18, p ¼ 0.32, D R2
¼ 0.017) the general factor accounted for 2% more variance compared to measures
typically employed in the current literature (General Comparison: F(1, 613.70) ¼
4.98, p ¼ 0.03, D R2 ¼ 0.017). However, when sexual sensation seeking was specified
as the outcome, both the general factor (F(1, 403) ¼ 52.62, p < 0.001, D R2 ¼ 0.124)
and the subfactors (F(3, 1,944.47) ¼ 10.63, p < 0.001,D R2 ¼ 0.121) accounted for
12% more variance than the single-item measure. Indeed, a similar pattern emerged
when excessive use was the specified outcome, the addition of the general factor
accounted for 5% more variance (F(1, 486.70) ¼ 17.14, p < 0.001, DR2 ¼ 0.049) and
the subfactors accounted for 14% more variance (F(3, 784.52) ¼ 17.61, p < 0.001, D
R2 ¼ 0.137) than the single-item measure.
Similar to the previous aim, this pattern of findings was unexpected. We theorized
that the addition of the newly developed measures would add incremental validity in
predicting gender, sexual satisfaction, depressive symptoms, sexual sensation seeking,
and excessive pornography use. Although statistically significant differences in incre-
mental validity were not obtained when gender and sexual satisfaction were specified
as outcomes, it was apparent in others. These findings suggest that the newly devel-
oped scale may not provide incremental prediction in all scenarios; however, it 1) can
provide incremental prediction in some scenarios, and 2) when it does not, it seems
to specify which behavioral facet of pornography consumption accounts for the vari-
ance. This may highlight the potential capability this scale has to provide a more
nuanced understanding of a complex behavioral construct.

Discussion
The overarching aim of the current study was to develop a reliable and valid meas-
urement of general pornography consumption. Analyses confirmed initial assump-
tions that pornography consumption lends itself to a bifactor solution where specific
factors, as well as a unidimensional factor, co-exists. Each of these factors was found
to be reliable and stable over time. Further, we found significant relationships
between our newly developed measures and sexual sensation seeking, depressive
symptoms, excessive pornography use, and gender, providing evidence of convergent
and divergent validity. Finally, we sought to establish incremental validity and found
SEXUAL AND RELATIONSHIP THERAPY 15

that the inclusion of measures related to pornography consumption, beyond a single-


item measure of frequency, accounted for more variance for predicted outcomes of
sexual sensation seeking, excessive pornography use, and depressive symptoms.
However, in scenarios where incremental validity was not established, the scale may
offer a nuanced view of the behavioral construct.
Regarding convergent validity, our findings were consistent with previous litera-
ture. For example, we found that the general subfactors had significant, strong, and
positive relationships with sexual sensation seeking. This is consistent with previous
research that has found that frequent pornography use is highly correlated with sex-
ual sensation seeking (Zheng & Zheng, 2014) and that sensation seeking predicted
pornography use (Beyens et al., 2015). In addition, women were also found to con-
sume pornography than males, which is consistent with at least one study that has
found that men consume pornography for a greater aggregate amount of time than
women (Hald, 2006).
Regarding divergent validity, we also found that the general pornography con-
sumption factor and the accidental exposure and arousal to pornographic content
each possessed a small but significant relationship with depressive symptoms (Nelson
et al., 2010; Wolak et al., 2007) showing that these two constructs have small relation-
ships. Finally, a small statistically significant relationship emerged between the fre-
quency of pornography consumption and sexual satisfaction, overlapping with the
confidence interval reported by Wright and colleagues (2017) and displaying further
evidence of divergent validity.
One surprising finding was that, despite accounting for addition variance above
and beyond the general factor, neither general factor nor subfactors accounted for
more variance than a single-item frequency measure of pornography in predicting
gender. One theoretical explanation falls in line with the Theory of Erotic Plasticity
(Baumeister, 2000). This theory suggests that a female’s sexuality is more socially mal-
leable than males. Several studies investigating opposite-sex couples have found that
when females view pornography, they often view it with a male partner (Bridges &
Morokoff, 2011; Maddox et al., 2011). It could be that although males consume porn-
ography more frequently than females, females quickly adopt the viewing preferences
of their male partners. This would suggest that after controlling for the frequency of
exposure, few differences between the sexes would exist.
In addition to providing the first psychometrically reliable and validated topo-
graphical measure of the behavior of pornography consumption, this study possesses
three other notable strengths. First, this is the first general-behavioral measure of
pornography consumption that was created without value-based assumptions and
does not label the behavior as problematic. Second, the foundation for individual sub-
scales was based on four dimensional quantities of behavior: Repeatability, Temporal
Extent, Topography, and Magnitude (Cooper et al., 2014; Johnston & Pennypacker,
1993). Each of these dimensional quantities can inform the research, both together
and separately, in ways a single-dimensional quantity cannot. Third, therapists assist-
ing in the treatment of those who seek help for problematic pornography use may
find it beneficial to know which behavior (i.e. frequency, duration, deliberate and
accidental exposure and arousal levels) would be best to target. While this may seem
16 S. G. HATCH ET AL.

contrary to the stated purposes and strengths of a scale said to measure behavior
without assigning pathology, it is not inherently so. The scale, while retaining its
topographical view, may provide the ideal scope to gain a more nuanced understand-
ing of the specific aspects of pornography consumption that are problematic and/
or addictive.
For better or worse, the current study took a theoretical risk by not supplying par-
ticipants with a blanket conceptual definition on the term pornography at the scale-
level but chose to define pornography at the item-level. This decision goes contrary
to several recent recommendations put forth by experts in the field (Kohut et al.,
2020; Marshall & Miller, 2019; Short et al., 2012) and we recognize that this may be a
critical limitation of the current study. However, we continue to wonder the degree
to which providing a strict scale-level conceptual definition of the term to participants
impacts the generalizability of the findings. A crucial direction for future research
could include supplying participants with varying definitions of “pornography” and
“use,” as some recommend (Kohut et al., 2020), and test for measurement invariance
between definitions. At the end of the study, investigators might consider asking par-
ticipants to write down the definition they originally received to see how well they
were able to maintain the definition in their working memory.
One primary limitation of the current study is that we did not assess for criterion val-
idity. Indeed, we could be measuring the perceptions of behavior surrounding pornog-
raphy consumption rather than the behavior of pornography consumption per se. We
find the recommendation put forth by Kohut et al. (2020) to be relevant to the current
study – a crucial validation step in the current study is comparing these self-report find-
ings to actual pornography consumption using smartphone data. Further limitations
include not testing for gender or time invariance, which serves as a restriction to estab-
lishing measurement quality. A final limitation was our inability to calculate factor scores
for our latent constructs. We recognize that estimating these scores would be in line with
best practice in scale development (McNeish & Wolf, 2020); however, limitations in
Mplus prevented us from computing factor scores with the WLSMV estimator (Muthen
& Muthen, 1998–2017). In spite of these faults, this was the first attempt at developing a
psychometric scale combining these theoretical constructs.
Those seeking future directions may look to four avenues: a further investigation
into the relationship between accidental pornography exposure and depression, fur-
ther validation of the current scale, item reduction, and exploring the factor structure
of the current scale with different definitions of pornography. One surprising finding
was the replicated result from Wolak et al.’s (2007) study that depressive symptoms
were associated with accidental or unwanted exposure to pornographic material.
Future studies should continue to explore the link between accidental exposure and
arousal to pornographic content and depressive symptoms. Another direction for
future research is to reduce the item load for the current scale. Some of our items are
likely to provide more information, have redundant information, or to discriminate
more than others. Uncovering the best items to ask participants would minimize the
amount of time required to measure pornography consumption. Next, further valid-
ation of the current scale is warranted. Though the sample was more representative
than a college sample, an increase in sample size would be valuable in establishing
SEXUAL AND RELATIONSHIP THERAPY 17

external validity. Finally, though this study chose to avoid imposing a strict definition
of pornography, future studies could compare different factor structures with differ-
ing definitions.
This paper attempted to respond to a call for a reliable and valid measure of porn-
ography consumption that does not have an underlying problematic behavior or
addiction assumption (Short et al., 2012). We found that pornography consumption
can be best understood as four unique but related factors or one general factor. We
found that this model was reliable, stable over time, and provided initial evidence for
convergent, divergent, and incremental validity.

Disclosure statement
No potential conflict of interest was reported by the authors.

Funding
This work was supported by an ORCA grant awarded to S. Gabe Hatch (grant number 5542).

Notes on contributors
S. Gabe Hatch is currently a Clinical Psychology graduate student at the University of Miami.
His program of research lies at the intersection of improving relationship well-being using
technology, improving sexual functioning, and applied quantitative methods.
Charlotte R. Esplin is a graduate student in Brigham Young University’s Clinical Psychology
PhD program. Her research interests are focused on measurement, personality, and successful
marital functioning.
H. Dorian Hatch is currently an undergraduate at Brigham Young University. He has interest
in psychopathology, psychopathy, and externalizing behaviors. He is also interested in the
measurement of psychopathology generally, and more specifically in the measurement and
implications of Cluster B disorders.
Aeriel Halstead is currently a graduate student at Brigham Young University. Her broad
research interests include the nature and prevention of intimate partner violence and methods
for increasing treatment accessibility for marginalized groups.
Joseph A. Olsen is Assistant Dean and Director of the Research Support Center in the College
of Family, Home, and Social Sciences at Brigham Young University. He teaches courses and
does research in advanced statistical methods for social and behavioral sciences.
Scott R. Braithwaite is currently an Associate Professor in the Clinical Psychology Department
at Brigham Young University and the developer of ePREP. Scott’s research interests involve
the dissemination of computerized and web-based relationship education and how to promote
relationship well-being.

References
Andresen, E. M., Malmgren, J. A., Carter, W. B., & Patrick, D. L. (1994). Screening for depres-
sion in well older adults: Evaluation of a short form of the CES-D. American Journal of
Preventive Medicine, 10(2), 77–84. https://doi.org/10.1016/S0749-3797(18)30622-6
18 S. G. HATCH ET AL.

Baumeister, R. F. (2000). Gender differences in erotic plasticity: The female sex drive as
socially flexible and responsive. Psychological Bulletin, 126(3), 347–374. https://doi.org/10.
1037/0033-2909.126.3.347
Beyens, I., Vandenbosch, L., & Eggermont, S. (2015). Early adolescent boys’ exposure to
Internet pornography: Relationships to pubertal timing, sensation seeking, and academic
performance. The Journal of Early Adolescence, 35(8), 1045–1068. https://doi.org/10.1177/
0272431614548069
B}
othe, B., T  Griffiths, M. D., Demetrovics, Z., & Orosz, G. (2018). The
oth-Kiraly, I., Zsila, A.,
development of the problematic pornography consumption scale (PPCS). The Journal of Sex
Research, 55(3), 395–406. https://doi.org/10.1080/00224499.2017.1291798
Bridges, A. J., & Morokoff, P. J. (2011). Sexual media use and relational satisfaction in hetero-
sexual couples. Personal Relationships, 18(4), 562–585. https://doi.org/10.1111/j.1475-6811.
2010.01328.x
Busby, D. M., Chiu, H. Y., Olsen, J. A., & Willoughby, B. J. (2017). Evaluating the dimension-
ality of pornography. Archives of Sexual Behavior, 46(6), 1723–1731. https://doi.org/10.1007/
s10508-017-0983-8
Buuren, S. V., & Groothuis-Oudshoorn, K. (2010). Mice: Multivariate imputation by chained
equations in R. Journal of Statistical Software, 12, 1–68.
Cooper, A. (1998). Sexuality and the Internet: Surfing into the new millennium.
CyberPsychology & Behavior, 1(2), 187–193. https://doi.org/10.1089/cpb.1998.1.187
Cooper, J. O., Heron, T. E., & Heward, W. L. (2014). Applied behavior analysis (2nd ed.).
Pearson.
Donevan, M., & Mattebo, M. (2017). The relationship between frequent pornography con-
sumption, behaviours, and sexual preoccupancy among male adolescents in Sweden. Sexual
& Reproductive Healthcare, 12, 82–87. https://doi.org/10.1016/j.srhc.2017.03.002
Duffy, A., Dawson, D. L., & Das Nair, R. (2016). Pornography addiction in adults: A system-
atic review of definitions and reported impact. The Journal of Sexual Medicine, 13(5),
760–777. https://doi.org/10.1016/j.jsxm.2016.03.002
Esplin, C. R., Hatch, S. G., Hatch, H. D., Diechman, C., & Braithwaite, S. R. (2020). What
motives drive pornography use?. The Family Journal.
Feinberg, R. A., & Jurich, D. P. (2017). Guidelines for interpreting and reporting subscores.
Educational Measurement: Issues and Practice, 36(1), 5–13. https://doi.org/10.1111/emip.
12142
Feinberg, R. A., & Wainer, H. (2014). A simple equation to predict a subscore’s value.
Educational Measurement: Issues and Practice, 33(3), 55–56. https://doi.org/10.1111/emip.
12035
Fisher, W. A., & Barak, A. (2001). Internet pornography: A social psychological perspective on
Internet sexuality. Journal of Sex Research, 38(4), 312–323. https://doi.org/10.1080/
00224490109552102
Fisher, W. A., & Grenier, G. (1994). Violent pornography, antiwoman thoughts, and anti-
woman acts: In search of reliable effects. Journal of Sex Research, 31(1), 23–38. https://doi.
org/10.1080/00224499409551727
Flood, M. (2007). Exposure to pornography among youth in Australia. Journal of Sociology,
43(1), 45–60. https://doi.org/10.1177/1440783307073934
Fuchs, C., & Diamantopoulos, A. (2009). Using single-item measures for construct measure-
ment in management research: Conceptual issues and application guidelines. Die
Betriebswirtschaft, 69(2), 195–210.
Goldsmith, K., Dunkley, C. R., Dang, S. S., & Gorzalka, B. B. (2017). Pornography consump-
tion and its association with sexual concerns and expectations among young men and
women. The Canadian Journal of Human Sexuality, 26(2), 151–162. https://doi.org/10.3138/
cjhs.262-a2
Grubbs, J. B., Perry, S. L., Wilt, J. A., & Reid, R. C. (2019). Pornography problems due to
moral incongruence: An integrative model with a systematic review and meta-analysis.
Archives of Sexual Behavior, 48(2), 397–415. https://doi.org/10.1007/s10508-018-1248-x
SEXUAL AND RELATIONSHIP THERAPY 19

Grubbs, J. B., Sessoms, J., Wheeler, D. M., & Volk, F. (2010). The Cyber-Pornography Use
Inventory: The development of a new assessment instrument. Sexual Addiction &
Compulsivity, 17(2), 106–126. https://doi.org/10.1080/10720161003776166
Grubbs, J. B., Volk, F., Exline, J. J., & Pargament, K. I. (2015). Internet pornography use:
Perceived addiction, psychological distress, and the validation of a brief measure. Journal of
Sex & Marital Therapy, 41(1), 83–106. https://doi.org/10.1080/0092623X.2013.842192
Hald, G. M. (2006). Gender differences in pornography consumption among young heterosex-
ual Danish adults. Archives of Sexual Behavior, 35(5), 577–585. https://doi.org/10.1007/
s10508-006-9064-0
Hald, G. M., & Malamuth, N. M. (2008). Self-perceived effects of pornography consumption.
Archives of Sexual Behavior, 37(4), 614–625. https://doi.org/10.1007/s10508-007-9212-1
Hald, G. M., & Stulhofer, A. (2016). What types of pornography do people use and do they
cluster? Assessing types and categories of pornography consumption in a large-scale online
sample. The Journal of Sex Research, 53(7), 849–859. https://doi.org/10.1080/00224499.2015.
1065953
Hatch, S. G., Esplin, C. R., Aaron, S. C., Dowdle, K. K., Fincham, F. D., Hatch, H. D., &
Braithwaite, S. R. (2020). Does pornography consumption lead to intimate partner violence
perpetration? Little evidence for temporal precedence. The Canadian Journal of Human
Sexuality. 1–8. https://doi.org/10.3138/cjhs.2019-0065
Johnston, J. M., & Pennypacker, H. S. (1993). Readings for strategies and tactics of behavioral
research. Lawrence Erlbaum Associates, Inc.
Kalichman, S. C., Johnson, J. R., Adair, V., Rompa, D., Multhauf, K., & Kelly, J. A. (1994).
Sexual sensation seeking: Scale development and predicting AIDS-risk behavior among
homosexually active men. Journal of Personality Assessment, 62(3), 385–397. https://doi.org/
10.1207/s15327752jpa6203_1
Kaplan, R. M., & Saccuzzo, D. P. (2017). Psychological testing: Principles, applications, and
issues. Nelson Education.
Kohut, T. A. (2014). An empirical investigation of the concept of “pornography Electronic Thesis
and Dissertation Repository. 2063. https://ir.lib.uwo.ca/etd/2063
Kohut, T., Balzarini, R. N., Fisher, W. A., Grubbs, J. B., Campbell, L., & Prause, N. (2020).
Surveying pornography use: A shaky science resting on poor measurement foundations. The
Journal of Sex Research, 57(6), 1–21. https://doi.org/10.1080/00224499.2019.1695244
Kor, A., Zilcha-Mano, S., Fogel, Y. A., Mikulincer, M., Reid, R. C., & Potenza, M. N. (2014).
Psychometric development of the Problematic Pornography Use scale. Addictive Behaviors,
39(5), 861–868. https://doi.org/10.1016/j.addbeh.2014.01.027
Lawrance, K. A., & Byers, E. S. (1995). Sexual satisfaction in long-term heterosexual relation-
ships: The interpersonal exchange model of sexual satisfaction. Personal Relationships, 2(4),
267–285. https://doi.org/10.1111/j.1475-6811.1995.tb00092.x
Lilienfeld, S. O., Sauvigne, K. C., Lynn, S. J., Cautin, R. L., Latzman, R. D., & Waldman, I. D.
(2015). Fifty psychological and psychiatric terms to avoid: A list of inaccurate, misleading,
misused, ambiguous, and logically confused words and phrases. Frontiers in Psychology, 6,
1100. https://doi.org/10.3389/fpsyg.2015.01100
Litman, L., Robinson, J., & Abberbock, T. (2017). TurkPrime.com: A versatile crowdsourcing
data acquisition platform for the behavioral sciences. Behavior Research Methods, 49(2),
433–442. https://doi.org/10.3758/s1342 https://doi.org/10.3758/s13428-016-0727-z
Maddox, A. M., Rhoades, G. K., & Markman, H. J. (2011). Viewing sexually-explicit materials
alone or together: Associations with relationship quality. Archives of Sexual Behavior, 40(2),
441–448. https://doi.org/10.1007/s10508-009-9585-4
Malamuth, N. M., Addison, T., & Koss, M. (2000). Pornography and sexual aggression: Are
there reliable effects and can we understand them? Annual Review of Sex Research, 11(1),
26–91. https://doi.org/10.1080/10532528.2000.10559784
Marshall, E. A., & Miller, H. A. (2019). Consistently inconsistent: A systematic review of the
measurement of pornography use. Aggression and Violent Behavior, 48, 169–179. https://doi.
org/10.1016/j.avb.2019.08.019
20 S. G. HATCH ET AL.

McKee, A., Byron, B., Litsou, K., & Ingham, R. (2020). An Interdisciplinary Definition of
Pornography: Results from a Global Delphi Panel. Archives of Sexual Behavior, 49(3),
1085–1091 https://doi.org/10.1007/s10508-019-01554-4
McNeish, D., & Wolf, M. G. (2020). Thinking twice about sum scores. Behavior Research
Methods, 1–19. https://doi.org/10.3758/s13428-020-01398-0
Meehl, P. E. (1986). Diagnostic taxa as open concepts: Metatheoretical and statistical questions
about reliability and construct validity in the grand strategy of nosological revision. In T.
Millon & G. L. Klerman (Eds.), Contemporary directions in psychopathology: Toward the
DSM-IV (pp. 215–231). Guilford Press.
Miller, D. J., Kidd, G., & Hald, G. M. (2019). Measuring self-perceived effects of pornography:
A short-form version of the Pornography Consumption Effects Scale. Archives of Sexual
Behavior, 48(3), 753–761. https://doi.org/10.1007/s10508-018-1327-z
Morin, A. J., Arens, A. K., & Marsh, H. W. (2016). A bifactor exploratory structural equation
modeling framework for the identification of distinct sources of construct-relevant psycho-
metric multidimensionality. Structural Equation Modeling: A Multidisciplinary Journal,
23(1), 116–139. https://doi.org/10.1080/10705511.2014.961800
Muthen, L. K., & Muthen, B. O. (1998–2017). Mplus user’s guide. 8th ed. Muthen & Muthen.
Negash, S., Sheppard, N. V. N., Lambert, N. M., & Fincham, F. D. (2016). Trading later
rewards for current pleasure: Pornography consumption and delay discounting. The Journal
of Sex Research, 53(6), 689–700. https://doi.org/10.1080/00224499.2015.1025123
Nelson, L. J., Padilla-Walker, L. M., & Carroll, J. S. (2010). “I believe it is wrong but I still do
it”: A comparison of religious young men who do versus do not use pornography.
Psychology of Religion and Spirituality, 2(3), 136–147. https://doi.org/10.1037/a0019127
Perry, S. L. (2015). Pornography consumption as a threat to religious socialization. Sociology of
Religion, 76(4), 436–458. https://doi.org/10.1093/socrel/srv043
Perry, S. L. (2016). From bad to worse? Pornography consumption, spousal religiosity, gender,
and marital quality. Sociological Forum, 31(2), 441–464. https://doi.org/10.1111/socf.12252
Popper, K. (1945). The poverty of historicism, III. Economica, 12(46), 69–89. https://doi.org/
10.2307/2549898
Pornhub.com. (2018, April 7). 2014 year in review [Web log post]. https://www.pornhub.com/
insights/2017-year-in-review
Rodriguez, A., Reise, S. P., & Haviland, M. G. (2016). Evaluating bifactor models: Calculating
and interpreting statistical indices. Psychological Methods, 21(2), 137–150. https://doi.org/10.
1037/met0000045
Ropelato, J. (2006). Internet pornography statistics. Retrieved on 25 Aug, 2020 from http://min-
istryoftruth.me.uk/wp-content/uploads/2014/03/IFR2013.pdf.
Rosseel, Y. (2012). Lavaan: An R package for structural equation modeling and more. Version
0.5–12 (BETA). Journal of Statistical Software, 48(2), 1–36. https://doi.org/10.18637/jss.v048.i02
Schermelleh-Engel, K., Moosbrugger, H., & M€ uller, H. (2003). Evaluating the fit of structural
equation models: Tests of significance and descriptive goodness-of-fit measures. Methods of
Psychological Research Online, 8(2), 23–74.
Short, M. B., Black, L., Smith, A. H., Wetterneck, C. T., & Wells, D. E. (2012). A review of
Internet pornography use research: Methodology and content from the past 10 years.
Cyberpsychology, Behavior, and Social Networking, 15(1), 13–23. https://doi.org/10.1089/
cyber.2010.0477
Skinner, B. F. (1938). The behavior of organisms: an experimental analysis. Appleton-Century.
Wolak, J., Mitchell, K., & Finkelhor, D. (2007). Unwanted and wanted exposure to online
pornography in a national sample of youth Internet users. Pediatrics, 119(2), 247–257.
https://doi.org/10.1542/peds.2006-1891
Wright, P. J. (2013). US males and pornography, 1973–2010: Consumption, predictors, corre-
lates. Journal of Sex Research, 50(1), 60–71. https://doi.org/10.1080/00224499.2011.628132
Wright, P. J., Tokunaga, R. S., & Kraus, A. (2016). Consumption of pornography, perceived
peer norms, and condomless sex. Health Communication, 31(8), 954–963. https://doi.org/10.
1080/10410236.2015.1022936
SEXUAL AND RELATIONSHIP THERAPY 21

Wright, P. J., Tokunaga, R. S., Kraus, A., & Klann, E. (2017). Pornography consumption and
satisfaction: A meta-analysis. Human Communication Research, 43(3), 315–343. https://doi.
org/10.1111/hcre.12108
Zheng, L., & Zheng, Y. (2014). Online sexual activity in Mainland China: Relationship to sex-
ual sensation seeking and sociosexuality. Computers in Human Behavior, 36, 323–329.
https://doi.org/10.1016/j.chb.2014.03.062

Appendix A
Frequency of exposure subscale
1. How often have you viewed pornography in the past year?
a. I viewed pornography once in the past year (1)
b. I viewed pornography about once every six months (2)
c. I viewed pornography about monthly (3)
d. I viewed pornography weekly (4)
e. I viewed pornography multiple times a week (5)
f. I viewed pornography daily (6)
g. I haven’t viewed pornography in the past year (0)
2. How often have you viewed pornography is the past 30-days?
a. I viewed pornography once in the past 30-days (1)
b. I viewed pornography twice in the past 30-days (2)
c. I viewed pornography weekly (3)
d. I viewed pornography multiple times a week (4)
e. I viewed pornography daily (5)
f. I viewed pornography multiple times a day (6)
g. I haven’t viewed pornography in the past 30-days (0)
3. How often have you viewed pornography in the past 7-days?
a. I viewed pornography once in the past 7-days (1)
b. I viewed pornography twice in the past 7-days (2)
c. I viewed pornography every day of the past 7-days (3)
d. I viewed pornography multiple times a day in the past 7-days (4)
e. I haven’t viewed pornography in the past 7-days (0)

Duration of exposure subscale

4. When I view pornography, I view it for …


a. Less than 5 minutes (1)
b. 6-15 minutes (2)
c. 16-25 minutes (3)
d. 26-35 minutes (4)
e. 36-45 minutes (5)
f. 46þ minutes (6)
5. When I visit a pornographic website, I visit for …
a. Less than 5 minutes. (1)
b. 6-15 minutes (2)
c. 16-25 minutes (3)
d. 26-35 minutes (4)
e. 36-45 minutes (5)
f. 46þ minutes (6)
22 S. G. HATCH ET AL.

6. The last time I viewed pornography, I viewed it for …


a. Less than 5 minutes (1)
b. 6-15 minutes (2)
c. 16-25 minutes (3)
d. 26-35 minutes (4)
e. 36-45 minutes (5)
f. 46þ minutes (6)

Accidental exposure and arousal – items 7, 8, 11, 12, 15, 16, 19, 20, 23, 24, 27,
28, 31, 32
Deliberate exposure and arousal – items 9, 10, 13, 14, 17, 18, 21, 22, 25, 26, 29,
30, 33, 34
7. Have you EVER viewed heterosexual pornography (where 1 male and 1 female are por-
trayed having sex) by ACCIDENT?
a. Yes (1)
b. No (0)
8. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
9. Have you EVER viewed heterosexual pornography (where 1 male and 1 female are por-
trayed having sex) on PURPOSE?
a. Yes (1)
b. No (0)
10. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
11. Have you EVER viewed gay men pornography (where 2 males are portrayed having sex)
by ACCIDENT?
a. Yes (1)
b. No (0)
12. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
SEXUAL AND RELATIONSHIP THERAPY 23

f. 5 (5)
g. 6 (6)
h. Very aroused (7)
13. Have you EVER viewed gay men pornography (where 2 males are portrayed having sex)
on PURPOSE?
a. Yes (1)
b. No (0)
14. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
15. Have you EVER viewed lesbian pornography (where 2 females are portrayed having sex)
by ACCIDENT?
a. Yes (1)
b. No (0)
16. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
17. Have you EVER viewed lesbian pornography (where 2 females are portrayed having sex)
on PURPOSE?
a. Yes (1)
b. No (0)
18. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
19. Have you EVER viewed solo pornography (where 1 individual are portrayed showing
genitals or masturbating) by ACCIDENT?
a. Yes (1)
b. No (0)
20. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
24 S. G. HATCH ET AL.

d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
21. Have you EVER viewed solo pornography (where 1 individual are portrayed showing
genitals or masturbating) on PURPOSE?
a. Yes (1)
b. No (0)
22. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
23. Have you EVER viewed threesome pornography (where EXACTLY 3 individuals are por-
trayed having sex) by ACCIDENT?
a. Yes (1)
b. No (0)
24. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
25. Have you EVER viewed threesome pornography (where EXACTLY 3 individuals are por-
trayed having sex) on PURPOSE?
a. Yes (1)
b. No (0)
26. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
27. Have you EVER viewed group sex pornography (where 4 or more individuals are por-
trayed having sex) by ACCIDENT?
a. Yes (1)
b. No (0)
28. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
SEXUAL AND RELATIONSHIP THERAPY 25

b. A little aroused (1)


c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
29. Have you EVER viewed group sex pornography (where 4 or more individuals are por-
trayed having sex) on PURPOSE?
a. Yes (1)
b. No (0)
30. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
31. Have you EVER viewed S&M pornography by ACCIDENT?
a. Yes (1)
b. No (0)
32. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)
33. Have you EVER viewed S&M pornography on PURPOSE?
a. Yes (1)
b. No (0)
34. If “Yes” was selected, how sexually aroused did you become?
a. Not at all aroused (0)
b. A little aroused (1)
c. 2 (2)
d. 3 (3)
e. 4 (4)
f. 5 (5)
g. 6 (6)
h. Very aroused (7)

You might also like