Associations Between Pornography Consumption Patterns, Pornography Consumption Motives, and Social Wellbeing among U.S. College Students: A Latent Profile Analysis with a Primarily Female Sample

Abstract

This study identified profiles of pornography motivations and outcomes and assessed differences between profiles on three measures of social well-being: social support, fear of intimacy, and loneliness. Latent profile analysis and group comparisons were conducted using cross-sectional data from college students (N = 389). Results indicated four profiles: low motivation/average distress, porn for enjoyment, high motivation/average guilt, low motivation/high distress. Those in the high motivation/average guilt profile reported more social well-being difficulties relative to the other profiles and non-pornography consumers. Results suggest that individuals who report varying pornography use motivations and negative outcomes may report difficulties with social well-being, with implications for intimate relationships.

In the United States, it is estimated that 46% of men and 16% of women between the ages of 18–39 have intentionally consumed pornography in a given week (Regnerus et al., 2016). When expanded to include pornography consumption in the last month, 91.5% of men and 60.2% of women in a large internet sample reported using pornography (Solano et al., 2020), including an estimated 15% of 10- to 17-year-olds who intentionally view pornography (Ybarra & Mitchell, 2005). The number of individuals who reported watching an X-rated movie in the last year increased across genders and ages from 1973–1980 to 2008–2012 with few exceptions (Price et al., 2015). Thus, access to and viewing pornography has changed over time, and pornography consumption is developmentally typical among adolescents (including early teenagers) and young adults (Binnie & Reavey, 2020; Ybarra & Mitchell, 2005).

Pornography Use, Relationships, and Loneliness

While some pornography consumption may be adaptive (e.g., facilitating sexual gratification without risk of STIs or pregnancy, promoting female sexual quality among couples who view pornography together; Bőthe et al., 2022; Poulsen et al., 2013), ongoing research suggests negative aspects of extensive pornography use on sexual and non-sexual relationships. One specific concern regarding adolescent and young adult’s consumption of pornography is the negative effects on sexual relationships, including the development of expectations for sexual relationships and sexual satisfaction (Dawson et al., 2020; Löfgren-Mårtenson & Månsson, 2010; Miller et al., 2019). Further, pornography consumption has been shown to be associated with increased rates of divorce (Wright et al., 2017) and loss of trust (Kohut et al., 2017). Pornography consumption has been inversely linked to experienced affection, relational satisfaction, and closeness (Hesse & Floyd, 2019). Similarly, emotional intimacy, referring to a willingness to be open about personal thoughts and feelings and having them received well by another (including non-sexual/romantic partners), has also predicted pornography consumption (Waggoner, 2021). Building upon a seminal investigation in the mid 2000’s (Yoder et al., 2005), research exploring how unsatisfactory social relationships and loneliness impact pornography consumption has become more prevalent in recent years. Indeed, numerous investigations across countries and age groups have linked pornography use to loneliness (Efrati & Amichai-Hamburger, 2019; Grubbs et al., 2018; Ostrander, 2021; Tian et al., 2018; Weber et al., 2018; Wéry et al., 2020).

The relationship between loneliness and pornography use is bi-directional. Specifically, loneliness increases pornography consumption via maladaptive coping and avoidance and pornography consumption increases loneliness (Butler et al., 2018). This bidirectional relationship between loneliness and pornography consumption has also been alluded to in qualitative research, where researchers thematically analyzed posts in an online pornography abstinence forum. Users of the forum reported loneliness as a negative outcome of no longer engaging in pornography consumption [described as a possible pornography “withdrawal symptom” (Fernandez et al., 2021)]. Notably, forums users reported social support, feeling as though people are there for them, was a key factor in remaining abstinent from pornography consumption.

Pornography Use Motives

While pornography consumption may be used to avoid feelings such loneliness, the most common motive for pornography consumption is sexual arousal, and pornography consumption almost always involves masturbation (for review, Binnie & Reavey, 2020). In addition to sexual arousal, there are likely myriad motives for pornography use. A summary of well over a decade of research by Grubbs, Wright, et al., (2019) identified several broad categories that may motivate the consumption of pornography. The first major motivator identified was individual differences, encompassing endorsement along the ertophobia-erotophilia spectrum (Paul, 2009), sensation seeking (Weisskirch & Murphy, 2004), and narcissistic traits (Kasper et al., 2015). In studies that assessed self-reported reasons for pornography consumption, participants frequently identified arousal, physical pleasure, masturbation, and sexual enhancement (Brown, Durtschi et al., 2017) as their primary motive. Researchers also found that individuals frequently use pornography as a form of sex education wherein participants report hoping to gain education, information, or address curiosity (Emmers-Sommer, 2018). The final category identified for pornography consumption was coping motives. Research suggests that coping with stress (Paul & Shim, 2008), psychological distress (Rissel et al., 2017), boredom (Rothman et al., 2015), and loneliness (Weber et al., 2018) are all meaningful reasons for pornography consumption.

In recent studies, researchers have focused on examining the relationship between different pornography consumption motives and excessive pornography consumption. Emerging evidence suggests using pornography for stress reduction, emotional distraction, boredom avoidance, fantasy, and sexual pleasure are all associated with increased pornography usage (Bőthe et al., 2020). These findings are consistent with assessments looking at if experiential avoidance – avoiding thoughts, feelings, and emotions -- predicting problematic pornography consumption (Levin et al., 2019; Wetterneck et al., 2012). Studies examining the frequency of pornography use based on motivations suggest two groups of pornography users. There are those that use occasionally only for their own sexual arousal or gratification (i.e. masturbation and physical pleasure) and those that use pornography frequently for sexual arousal as well as for excitement and knowledge (Brown, Durtschi et al., 2017). A similar two group pattern has been found when profiles are analyzed based on biological sex, although for women, the group that viewed pornography less frequently were viewed as pornography abstainers suggesting some difference between men and women in porn use (Brown, Conner et al., 2017). While these studies advanced understanding of how motives relate to use, they are limited in that frequency of pornography consumption does not always mean that the use is problematic, that is, associated with negative outcomes or distress.

Researchers have attempted to identify profiles of problematic pornography users. Through use of latent profile analysis, accumulated evidence suggests three profiles of pornography users may exist. One three-profile solution suggests non-problematic users, low-risk users, and at-risk (of problematic use) users (Bőthe et al., 2018). A second three profile solution indicates profiles for those who view pornography frequently and experience their use as problematic, those who view pornography frequently and it’s not problematic, and those who don’t view pornography frequently and don’t engage in problematic behaviors (Bőthe et al., 2020). Profiles examining loneliness and problematic pornography use are rare, but findings suggest that higher levels of loneliness are associated with more problematic pornography viewing (Bőthe et al., 2018). However, research on problematic pornography use frequently does not assess individual’s motives for consuming pornography (Bőthe et al., 2018; Vaillancourt-Morel et al., 2017) or assesses potential motives without asking if they were the reason for pornography viewing (Bőthe et al., 2020), such as asking about the tendency to become bored but not if pornography was viewed to reduce boredom. Further, no study to date has examined variables contributing to social intimacy (e.g., fear of intimacy) as they relate to profiles of pornography consumption, nor have past studies explicitly integrated motives for pornography use with problematic pornography use behaviors into profiles of pornography use.

The current study sought to replicate and expand upon previous studies. The first goal of the current study was to identify profiles of pornography consumption motives and problematic behaviors associated with pornography use. Given findings from previous studies examining pornography use profiles detailed above, it was hypothesized that three profiles would emerge. A second aim of the current study was to assess how the extracted profiles differed in their social connection based on measures of loneliness, fear of intimacy, and social support. It was hypothesized that the profiles with the highest levels of problematic use would have the highest level of loneliness, fear of intimacy, and social support.

Method

Procedure and Participants

Prior to the initialization of the study, the current study was reviewed and approved by the Human Subjects Institutional Review Board at the institutes at which data collection took place. Recruitment took place at a medium sized Hispanic-Serving Institute (Institute 1) and a medium sized rural institute (Institute 2). To be eligible for the study, individuals had to be 18 years of age or older and enrolled in courses at the university where data collection took place. Participants were recruited through classroom announcements and e-mail announcements detailing the study. Potential participants were informed that the study consisted of completing questionnaires online, the nature of the questionnaires, and the anticipated time it would take to complete the questionnaires prior to agreeing to participate in the study. Attention checks were inserted throughout the questionnaires to ensure that responding was associated with the questions that were presented.

The final sample analyzed consisted of 389 participants. Informed consent was provided by 428 individuals; 5% (n = 22) did not pass data validity checks (operationalized as missing any of eight attention check questions) and were not included in analyses. Seventeen participants did not respond to items assessing pornography motivations and behaviors (e.g., PCI and CPUI) and were excluded from data analysis, with ultimately 389 participants included in the final analysis. The demographics of the analyzed sample are presented in Table 1.

Table 1:

Demographics

	N = 389
Age	22.93 (10.17)
Location
Institute 1 (medium-sized, Hispanic serving institute)	84.3%
Institute 2 (small, rural) institute)	15.7%
Gender
Male	16.3%
Female	83.7%
Racial identity
Asian or Asian American	2.2%
American Indian or Alaska Native	0.8%
Black or African American	6.1%
Multiracial	12.9%
Native Hawaiian or Pacific Islander	.6%
White	77.4%
Ethnicity
Hispanic or Latinx	49.3%
Not Hispanic or Latinx	50.7%
Employment status
Full-time	12.9%
Part-time	48.1%
Not employed	39.1%
Watches pornography
Does not watch pornography	20.8%
Watches pornography	79.2%

Measures

Cyber Pornography Consumption Inventory

The Cyber Pornography Use Inventory (CPUI; Grubbs et al., 2010) is a measure used to assess internet pornography usage as it relates to behavioral patterns of addiction. Based upon updated scoring recommendations (Grubbs & Gola, 2019; Grubbs et al., 2015), two subscales examining perceived compulsivity regarding pornography use and emotional distress related to pornography use were computed. Perceived compulsivity was comprised of four items capturing perceived loss of control related to pornography use and was scored according to the CPUI-4 (Grubbs & Gola, 2019). Of note, the updated version of the CPUI-4 contains an item “I have put off things I needed to do in order to view pornography” which is no included in the original CPUI, and the current analyses instead used the item “pornography has sometimes interfered with certain aspects of my life.” Emotional distress was scored according to the CPUI-9, with three items capturing distress regarding pornography use (α = .93; ex. “I feel depressed after viewing pornography online”). Items on the CPUI are scored on a 7-point Likert scale ranging from “Strongly Disagree” to “Strongly Agree.”

Fear of Intimacy Scale

The Fear of Intimacy Scale (FIS; Descutner & Thelen, 1991) is a 35-item measure designed to assess individual’s anxiety about close, dating relationships. The measure asks questions about willingness to engage in personal disclosure and experience feelings of closeness. Items are scored on a 5-point scale where high overall scores indicate higher fear of intimacy. The measure has good validity and high internal consistency (α = .93).

Medical Outcome Study Social Support Scale

The Medical Outcome Study Social Support Survey (Sherbourne & Stewart, 1991) is a 19-item measure that assess different types of social support available to individuals. The measure assesses four different types of social support including emotional support, tangible support, affectionate support, and positive social interactions. The measure has good validity and high internal consistency (α = .97).

Pornography Consumption Inventory

The Pornography Consumption Inventory (PCI; Reid et al., 2011) is a 15-item measure that assess motivations for pornography consumption. The PCI is composed of four subscales (Emotional Avoidance, Sexual Curiosity, Excitement Seeking, and Sexual Pleasure) based around motives for viewing pornography. Items are scored on a 5-point scale from 1 (Never Like Me) to 5 (Very Often Like Me). The overall measure has good internal consistency (α=.83) with each of the subscales having acceptable or good internal consistency (Emotional Avoidance, α=.85; Sexual Curiosity, α=.87; Excitement Seeking, α=.73; and Sexual Pleasure, α=.71).

UCLA Loneliness Inventory Version 3

The UCLA Loneliness Inventory Version 3 (UCLALI; Russell, 1996) is a 20-item assessment of loneliness. The measure is scored on a scale of 1 (Never) to 4 (Always). The UCLALI has high internal consistency (α = .92) and has demonstrated good validity.

Data analysis plan

Latent profile analysis (LPA) was used to identify profiles (e.g., subgroups) of participants based upon their 1) motivations for pornography use (ex. sexual pleasure, emotional avoidance assessed by the PCI,) and 2) perceived compulsion to view pornography and emotional distress assessed by the CPUI. Next, profile differences for social variables (i.e., loneliness, fear of intimacy, social support) were evaluated. LPA is an empirical, person-centered approach for identifying groups or profiles of individuals within heterogenous data who share a meaningful and identifiable pattern of responses on measures of interest (Collins & Lanza, 2010; Ferguson et al., 2020). LPA has an advantage over variable-centered approaches, which assume all individuals belong to a single population, with no differentiation between latent subgroups. In contrast, LPA approaches assume individual differences are present and important, occur in a way that may be examined through patterns, and in which a small number of patterns are meaningful (for review, Ferguson et al., 2020). Within LPA, the profiles are conceptualized as unobserved latent mixture components, in which the observed variables are continuous (Ferguson et al., 2020; Oberski, 2016). Identification of latent profiles is done using joint and marginal probabilities in within-class and between-class models.

Guided by the six steps of LPA as outlined by Ferguson et al. (2020), prior to conducting statistical analyses, data were cleaned for analysis and observed for missing values, normality, and data validity procedures. Of the included participants, 20.8% (n = 81) indicated the lowest possible response for all items on the PCI and CPUI and were conceptualized as non-pornography users. The remainder of participants (n = 308) were included in the LPA. Notably, there were no differences across data collection locations regarding if participants were classified as pornography users [X²(1) = .86, p > .35]. The required sample size of an LPA is dependent on the number of profiles and distance between profiles (Ferguson et al., 2020). Simulation studies investigating power within LPA analyses suggest samples of 300 to 500 qualify as minimum sample (for review, Ferguson et al., 2020), indicating that while the current sample may be smaller than other samples, it falls within the realm of acceptable statistical power. To determine power for our specific research questions, we conducted a post hoc power analysis in MPlus using a Monte Carlo simulation with a sample size of 300 and 1000 replications (Nylund et al., 2007; Nylund-Gibson & Choi, 2018). The specified population was comprised of the six indicators with values based upon past empirical studies (Bőthe et al., 2018, 2020, 2022; Wilt et al., 2016). Based upon existing recommendations for determining power (Muthén & Muthén, 2002; Tein et al., 2013), power was determined to be adequate for up to a six profiles solution, with coverage ranging from .92 to .97.

LPA was conducted utilizing MPlus (Muthen & Muthen, 1998–2020). We engaged in an iterative model estimation process, starting with a model of one profile and subsequently adding estimating more profiles until six were reached, which was guided by past published research suggesting that LPA studies typically find their best fitting model theoretically and statistically after comparing five to six models (Masyn, 2013). The four subscales of the PCI (Reid et al., 2011) and two subscales of the CPUI (Grubbs et al., 2010) were standardized to facilitate ease of interpretation and entered as indicators for the models. The composite scores of these subscales were used as indicated rather than item-level data to reduce complexity and increase model convergence (Ferguson et al., 2020). Each model was compared against the previous model with model retention determined by Bayesian Information Criterion (BIC), sample-sized adjusted BIC (SSBIC), Akaike’s Information Criterion (AIC), Lo-Mendell Ruben (LMR), and bootstrap likelihood ratio test (BLRT). Both BIC and AIC are relative fit indices, where lower values suggest improved model fit. The BLRT compares the fit of the current model to a comparison model with one fewer class, and a non-significant p-value indicating the model with fewer classes is a better fit (Arminger et al., 1999). Entropy values, which is a measure of variability in the stochastic system (Larose et al., 2016), were also examined, with higher entropy values (closer to 1.0) indicating improved fit (Ferguson et al., 2020). Existing recommendations recommend that fit indices should be contextualized by theory and model utility, including the “conceptual meaningfulness and plausibility of each [profile] solution” (Nylund-Gibson & Choi, 2018, p. 12) and their degree of differentiation. Guided by recommendations, we also inspected the item probability plots and the qualitative differences among the profiles. Each model was evaluated based on its adherence to past theoretical and empirical literature, and the profiles were examined to determine if they were interpretable.

After the best fitting model was determined, the BCH approach was used to identify associations between profile membership and loneliness, fear of intimacy, and social support. The BHC approach is examines means across classes for continuous auxiliary variables and is recommended as a more robust analytic approach to examining covariates relative to other auxiliary approaches (Asparouhov & Muthén, 2014; Ferguson et al., 2020).

Comparisons to non-pornography users

Because 20% of the sample were conceptualized as non-users of pornography, comparisons between participants in the latent classes and non-pornography users were conducted. Participants in the LPA were assigned to their most likely membership group and a categorical variable was created with individuals who reported not watching porn included as their own group. Differences across groups for loneliness, fear of intimacy, and social support were conducted through one-way ANOVAs and post-hoc t-tests with a Benjamini-Hochberg false discovery correction (FDR; Benjamini & Hochberg, 2000; Keselman et al., 2002). Chi-square analyses were also conducted to ascertain any differences in group members by gender and location of data collection. Post hoc methods are not generally recommended because such analyses treat the latent profiles as fixed, categorical variables and may place an individual into a profile without justification. However, such analyses may be conducted if entropy values are greater than .80 (Clark & Muthen, 2009; for review, Ferguson et al., 2020). As the final model had an entropy value above .80 (see Results), post hoc comparisons were determined to be appropriate.

Results

Bivariate correlations, means, and standard deviations are presented in Table 2. Missingness at the composite level was rare, ranging from .1% to 2.22%, and was estimated in MPlus using full information maximum likelihood estimation. Solutions for the LPA models are summarized in Table 3. Profiles containing less than 5% of the sample (Models 5 and 6) were rejected not only for their lack of interpretability (Masyn, 2013), but also their nonsignificant LMR. Fit indices (e.g., BIC, AIC, SSBIC) supported that the four profile model demonstrated the best fit to the data. Notably, entropy levels for a three profile model were slightly higher than a four-factor model, although both were in an acceptable range. LMR and BLRT tests were significant for a four-profile model, suggesting that a more parsimonious model was not necessarily indicated. Examination of each profile and probability plots indicated that four distinct profiles were reflected in the data.

Table 2:

Bivariate correlations, Means, and Standard Deviations

	1	2	3	4	5	6	8	9	10
1. Emotional Avoidance	--
2. Sexual Curiosity	.47**	--
3. Excitement Seeking	.69**	.68**	--
4. Sexual Pleasure	.54**	.63**	.71**	--
5. Perceived Compulsivity to Pornography Use	.36**	.21**	.28**	.24**	--
6. Emotional Distress	-.11*	-.32**	-.24**	-.36**	.07	--
7. UCLA	.27**	.15**	.21**	.21**	.16**	-.07	--
8. Fear of Intimacy	.23**	.17**	.12**	.05	.13**	.04	.50**	--
9. Social Support Scale	-.23**	-.08	-.15**	-.10	-.12**	.04	-.65**	-.41**	--
Mean (SD)	8.40 (4.47)	9.05 (4.52)	5.96 (3.15)	8.25 (4.42)	1.76 (1.16)	2.20 (1.67)	2.37 (.58)	80.52 (23.37)	72.80 (17.19)

Table 3.

Fit statistics for LPA models of pornography motivations and compulsions

Number of classes	BIC	AIC	SSBIC	LMR	BLRT	Entropy
2 classes	4997.554	4926.682	4937.294	.35	<.001	.82
3 classes	4848.376	4751.394	4765.915	.04	<.001	.866
4 classes	4620.313	4743.406	4638.744	<.02	<.001	.860
5 classes	4709.425	4560.221	4582.562	.60	<.01	.876
6 classes	4669.351	4494.036	4520.286	.37	<.01	.877
Number of classesProbs2	BIC	AIC	SSBIC	LMR	BLRT	Entropy
2 classes	4964.489	4893.679	4904.230	.00	<.001	.82
3 classes	4811.257	4714.359	4728.797	.02	<.001	.88
4 classes	4706.782	4583.796	4602.121	.01	<.001	.86
5 classes	4660.399	4511.325	4533.537	.21	<.01	.86
6 classes	4622.047	4446.885	4472.983	.17	<.01	.87

Profiles

Means and standard deviations of the standardized composite subscales among the participants in the three classes are presented in Figure 1. Profile 1 (32.79%) “unmotivated, unperturbed” had low scores (approximately .5 SD below the mean) across all PCI subscales and perceived compulsivity and distress regarding pornography use approached the mean scores for the sample. Profile 2 (44.48%) “porn for enjoyment” group scored within the mean on all measures of pornography motives, compulsivity, and distress, the exception of sexual pleasure subscale of the PCI, where they were half of a standard deviation above the mean. Profile 3 (9.09%) “high pornography distress” reported very high levels of distress regarding pornography use, contextualized by low scores on all pornography motivation items and perceived compulsivity to use pornography, although their emotional avoidance motivates and perceived compulsivity scores fell within .5 standard deviations of the mean. Their level of distress regarding pornography use (~2.5 standard deviations above the sample mean) distinguishes them from every other profile identified. Profile 4 (13.36%) “high motivations and average distress” reported scores ranging from one to one and half standard deviations above the mean on all pornography use motivations, between one and one and a half standard deviations above the mean regarding perceived compulsivity and average distress regarding pornography use. Notably, this group reported emotional avoidance to be their highest endorsed motive for pornography use, followed by sexual curiosity, and while they reported the highest perceived compulsions to view pornography of any group, their distress was similar to the sample mean. To aid in comparison to the extant literature, unstandardized means and standard deviations for each profile are presented in Table 4.

Figure 1.

Results of LPA analysis of pornography motivations and compulsions

Profile 1 has low scores; approximately .5 SD below the mean across all PCI subscales and perceived compulsivity and distress regarding pornography use approaching the mean scores for the sample. Profile 2 scores are within the mean on all measures of pornography motives, compulsivity, and distress, the exception of sexual pleasure subscale of the PCI, where they were half of a standard deviation above the mean. Profile 3 reported very high levels of distress regarding pornography use, approximately 2.5 standard deviations above the sample mean, contextualized by low scores on all pornography motivation items and perceived compulsivity to use pornography, although their emotional avoidance motivations and perceived compulsivity scores fall within .5 standard deviations of the mean. Profile 4 reported scores ranging from one to one and half standard deviations above the mean on all pornography use motivations, between one and one and a half standard deviations above the mean regarding perceived compulsivity and average distress regarding pornography use.

Table 4:

Three-Profile Model Unstandardized Means and Standard Deviations

Variable	Profile 1 Unmotivated, Unperturbed (n = 101)	Profile 2 Porn for Enjoyment (n = 137)	Profile 3 High Distress (n = 28)	Profile 4 High motivations/ Average Distress (n = 42)	Non-Users of Pornography (n = 81)
Pornography Consumption Inventory
Emotional Avoidance	6.38 (2.46)	9.54 (3.67)	7.64 (3.55)	16.64 (3.69)	NA
Sexual Curiosity	7.99 (3.20)	11.34 (3.61)	7.25 (2.58)	15.02 (3.39)	NA
Excitement Seeking	4.10 (1.26)	7.51 (1.93)	4.67 (1.66)	11.98 (1.94)	NA
Sexual Pleasure	6.36 (2.58)	11.91 (2.44)	4.82 (2.21)	13.26 (1.85)	NA
Cyber Pornography Use Inventory
Perceived Compulsivity	1.48 (0.78)	1.85 (1.10)	2.26 (1.65)	2.74 (1.68)	NA
Emotional Distress	1.44 (0.60)	1.59 (.79)	5.20 (.97)	1.83 (1.07)	NA
UCLA Loneliness Inventory	48.31 (11.90)a+	48.38 (10.51)a+	47.29 (10.45) +	52.76 (11.84)a+	42.86 (11.25)a
Fear of Intimacy Scale	83.30 (24.02)b	80.30 (23.24)	85.44 (24.43)	86.46 (24.40)b	75.11 (20.90)b
Social Support Scale	72.16 (17.32)	72.77 (17.97)	72.36 (17.67)	66.86 (16.48)c	77.13 (16.64)c

^aSignificant difference between non-users of pornography and Profile 1, Profile 2, and Profile 4 via posthoc analysis

⁺Significant difference between Profile 1, Profile 2 and Profile 3 compared to Profile 4 via BCH and post hoc analyses

^bSignificant difference between non-users of pornography and Profile 1 and Profile 4

^cSignificant difference between non-users of pornography and Profile 4

BCH analyses suggested significant differences in loneliness between Profile 4 (M = 52.76) and Profiles 1, 2, and 3 (M = 48.31, M = 48.38, and M = 47.29, respectively), however overall chi-square results were not significant, so this finding should be interpreted with caution.

Comparison to non-pornography users

All three one-way ANOVAs were significant regarding group differences in loneliness, fear of intimacy, and social support when including non-pornography users and all three profiles (all F’s > 2.46, all p’s < .05). After FDR correction, when comparing the non-pornography users to the profiles, results indicated non-users of pornography reported lower levels of loneliness relative to Profiles 1, 2, and 4 [t(180) = −3.52, p = .001; t(216) = −4.05, p < .001; t(121) = −4.81, p < .001; respectively]. That is, members of these three pornography profiles reported, on average, more loneliness than non-users of pornography. Non-users of pornography reported lower levels of fear of intimacy than individuals in Profile 1 (“unmotivated, unperturbed”) and 4 (“high motivations and average distress”) [t(172) = −3.20, p = .001; t(115) = −3.37, p = .001]. Non-users of pornography reported higher levels of social support than individuals in Profile 4 “high motivations and average distress” (t(120) = 3.128, p = .002) There were no differences in social support between non-users of pornography and any other pornography profiles.

Chi-square analyses indicated there were no differences in pornography class based upon location (X²(4) = 2.30, p > .60). However, there were significant differences in pornography class based upon gender (X²(4) = 28.00, p < .001). A higher proportion of female participants indicated they did not watch pornography relative to male participants (24.5% vs. 4.8%) and a higher proportion of female participants were classified as Profile 3 (“high pornography distress”) relative to male participants (8.2% vs 3.2%). A lower proportion of female participants were classified as Profile 4 (“high motivations and average distress”) relative to male participants (7.5% vs. 25.8%). Notably, approximately equal proportion of the female and male participants were grouped into Profile 1 (“unmotivated, unperturbed”; 25.4% vs. 29.0%) and Profile 2 (“porn for enjoyment”; 34.5% vs. 37.10%).

Discussion

The present study assessed different profiles of pornography use motives and problematic pornography usage through a latent profile analysis and compared social wellness variables across profiles. Researchers have posited that deficiency in meaningful relationships may increase the probability of problematic pornography consumption. As such, the current study sought to expand upon findings linking loneliness to pornography use by not only assessing loneliness across pornography profiles, but also examining potential underlying causes of loneliness in social support and fear of intimacy. Our hypotheses were partially supported; four profiles of pornography users and problematic use were identified reflecting a range of pornography use motives and problematic pornography use. Findings also demonstrated that the profile characterized by highest endorsement of motives and perceived compulsive use, reported more loneliness, more fear of intimacy, and less social support than the other three pornography profiles and non-pornography users.

Our findings build upon previous studies that utilized profiles of pornography use by incorporating motives for use with potentially problematic use. Profile 1 identified individuals who reported few motives for using pornography and experienced minimal distress after using pornography. Profile 2 included individuals who indicated more frequently using pornography for sexual pleasure relative to other motives and scored close to average on perceived problematic behaviors. Profile 3 included individuals who reported low motivations for consuming pornography but reported high levels of distress and elevated feelings of compulsivity related to pornography use. The final group of individuals fell into Profile 4, which included individuals who viewed pornography for many reasons, reported perceived compulsive patterns of behavior, but reported only average distress from pornography consumption. When compared to non-pornography users, pornography consumers typically reported higher levels of loneliness however, there was no significant differences between non-pornography users and Profile 3. Additionally, compared to Profile 4, those who fit best with the other Profiles reported significantly less loneliness. Relative to non-pornography users, users who fit the profile reporting the most motives for use and greater perceived compulsivity regarding pornography use (Profile 4) reported higher levels of loneliness, higher levels of fear of intimacy and lower levels of social support than non-users. Interestingly, there were no significant differences in loneliness, fear of intimacy, and social support between non-pornography users and two pornography profiles. While overall, non-pornography users reported lower scores on loneliness and fear of intimacy, and higher scores on social support than any pornography profile, these scores are statistically indistinguishable from the fear of intimacy reported by individuals classified within Profile 2 and Profile 3. Differences in social support for non-users were only statistically significant from those classified by Profile 4, the high motivations, average distress profile.

The current results are mostly consistent with past literature suggesting a spectrum of motivations, attitudes, and behaviors regarding pornography among college and college-aged individuals (Brown, Conner, et al., 2017; Brown, Durtschi, et al., 2017; Tóth-Király et al., 2019). A prior LPA that examined motivations and attitudes regarding pornography, frequency of pornography use, and religiosity found three classes of pornography users (Brown, Durtschi, et al., 2017). Similar to our findings, classes were distinguished by 1) abstainers and individuals who rarely used pornography and primarily used pornography due to sexual curiosity and to educate themselves about sex; 2) pornography users who primarily used pornography for personal enjoyment and masturbation, but did not use pornography to avoid uncomfortable emotions, and 3) pornography users who reported the highest reports of pornography acceptance, use, and benefits who reported using pornography for sexual enjoyment, as well as expanding their knowledge of sexual possibilities and sexual excitement (Brown, Durtschi, et al., 2017). Notably this third group also reported higher levels of using pornography to change negative mood relative to the other two groups. The current study suggests the possibility of a fourth profile where individuals have low motivation for use but experience a significant amount of distress associated with use.

The current findings also replicate and expand upon previous research linking pornography usage to loneliness. Previous analyses (Bőthe et al., 2018) found that higher risk (of problematic use) users of pornography experienced more loneliness than lower risk users. Our results reflect this past finding. Individuals in Profile 4, which reported higher pornography motivations and compulsive pornography use, reported higher loneliness than the other three pornography profiles. Our results suggest that this finding extends to individuals who report no motives for pornography use and no problematic behaviors (i.e., the individuals who did not fit into one of the four profiles in the current study) as they also reported less loneliness than those in Profile 4. Additionally, our findings suggest fear of intimacy and social support, potential precursors to loneliness (Maitland, 2020), are higher in individuals with potentially problematic pornography use than non-pornography users and in the case of fear of intimacy, also higher in those who report minimal distress and low motivation for pornography consumption (Profile 1), which may have meaningful implications for understanding how the social environment is associated with pornography consumption. While these findings are consistent with the theorized bi-directional relationship between loneliness and pornography use (Butler et al., 2018), they indicate that there may be isolation or relationship quality issues that underly the relationship between loneliness and pornography use.

The derived profiles suggest that isolated reasons for pornography use are not sufficient for identifying individuals who may have more problems related to their pornography use. The reports of problematic pornography use were highest among those that endorsed many reasons for viewing pornography (e.g., Profile 4). Within this profile, perceived compulsive use was endorsed most frequently, and participants endorsed every motive for pornography consumption at a higher level than any other profile. This finding introduces questions about why pornography is the chosen outlet for satisfying these motives, such as emotional avoidance or sexual excitement. It is possible that these individuals do not have the environmental availability or behavioral repertoires to get these needs met through topographically different behaviors. As the focus of pornography consumption may be related to relief from negative emotions or sensation seeking, participants within this group are more focused on achieving these outcomes, rather than distress associated with pornography use. Those in Profile 2 report using pornography for multiple reasons at an average level for the sample, but there is an increase in reported viewing of pornography for sexual pleasure. Interestingly, this group reports low levels of distress regarding pornography use despite having average reports of compulsive patterns. This profile may best reflect permissive attitudes towards pornography, wherein pornography use is viewed as relatively normative and an aspect of sexual enjoyment. Profile 1 reported few reasons for viewing pornography but reported higher levels of emotional avoidance and sexual curiosity compared to excitement seeking or sexual pleasure although these were all low relative to the other profiles. Finally, those in Profile 3 report low motives for pornography viewing, especially in the domain of sexual pleasure, but report high levels of emotional distress upon viewing.

One possible explanation for low pornography viewing motives but high levels of distress may be religiosity, as data was collected within geographic regions with high levels of religious involvement and identification (Garcia & Kruger, 2010). While pornography use and religiosity are typically negatively associated (Nelson et al., 2010), this association is likely complex. Specifically, it may not simply be belief in religiosity, but past and current practices that are negatively associated with pornography use (Nelson et al., 2010). Brown and colleagues three-class solution indicated that all profiles exhibited fairly high religiosity (Brown, Durtschi, et al., 2017). The authors hypothesize that individuals may report religious beliefs that are important, but these may not translate into specific religious practices (Brown, Durtschi, et al., 2017). While religiosity was not assessed within this sample and thus all possible interpretations are speculative, it is possible that the report of distress, particularly within Profile 3 is reflective of past research wherein individuals who are religious may still believe pornography to be wrong, but still may engage in the behavior (Nelson et al., 2010), a phenomena known as moral incongruence (Grubbs, Perry et al., 2019). Researchers have found that the relationship between pornography use, and self-reported addiction is moderated by religiosity (Grubbs et al., 2020). This finding suggest that it is possible that the same behaviors, including motives for use, are interpreted as problematic as a function of a person’s values.

Limitations and future directions

The results of current study must be contextualized by a number of limitations. The individuals who participated in the study were all college students who self-selected into the cross-sectional questionnaire-based study investigating pornography. Individuals who volunteer for sexuality research have shown to be more sexually experienced, hold less traditional sexual attitudes, report more sexual esteem, and are more sexual sensation seeking (Wiederman, 1999). As such, it is possible that the profiles created are not representative of all pornography consumers, so much as it is reflective of those who consume pornography and are also willing to participate in sexuality research. Further, given that all the individuals who participated in the study were college students, the generalizability of the findings is limited. While the sample did have roughly equal distribution of Hispanic and Non-Hispanic participants, the sample was overwhelming comprised of white women. It is important to recognize that women not only watch less pornography than men, but also may watch different types of pornography for different reasons and develop unique problems associated with use. Further, no analysis was conducted on any other demographic variables. As such, future studies seeking to expand on the current findings should consider recruiting a representative sample based on the population of interest. The cross-sectional nature of the study limits the ability to assess causality, thus we cannot draw any conclusions regarding the casual effects of motives, fear of intimacy, social support, or loneliness on pornography addiction and related behaviors or vice versa. Finally, the present study was based around self-report data which may only approximate true patterns of usage and may not account for all motives or outcomes of pornography usage. Other pornography use metrics may be beneficial in elucidating the behaviors that are under investigation.

Findings from the current study suggest several potential pathways to assess causality between the variables that were investigated including longitudinal studies. It is also possible to manipulate the variables being studied through use of therapy in a research context with the population the current sample was drawn from. Given the relatively elevated levels of fear of intimacy and loneliness of those in Profile 3, the efficacy of interventions targeting these constructs for problematic pornography users should be investigated. Given the lack of efficacy of loneliness interventions (Mann et al., 2017) and that fear of intimacy may function as a precursor to loneliness (Maitland, 2020) it is our suggestion that fear of intimacy be targeted. Previous research has indicated that Functional Analytic Psychotherapy prototypically targets and is an effective intervention for fear of intimacy (Maitland et al., 2016) and as such has the potential to be a useful intervention modality. Further, given the variety of motives endorsed by those who report the most problematic pornography consumption, researchers are encouraged to assess other behaviors that these individuals engage in that have the same function. For example, a study could recruit individuals who report viewing pornography in a problematic fashion as an experiential avoidance technique and assess if those individuals have other experiential avoidance techniques or if the repertoire limited to pornography consumption.

Conclusions

The present study presents findings in a latent profile analysis of college pornography consumers and social correlates of pornography consumption. Our findings suggest four profiles of users based on motives and problematic viewing tendencies. The profile with the largest proportion of participants (e.g, Profile 2) was characterized by average levels of motivation for consuming pornography but slightly higher levels of motivation in viewing pornography for sexual pleasure and average levels of compulsivity and distress. The second largest profile was characterized by low levels of motivation for viewing pornography but average levels of distress. The profile with the smallest proportion of participants (9.09% of the sample) included individuals who reported low levels of motivation for consuming pornography across potential motivators. Despite low levels of motivation, those included in Profile 3 reported average levels of compulsivity and high levels of distress. Finally, Profile 4 reported high motivation levels, above average levels of compulsivity but only average amounts of emotional distress. These profiles suggest that there is no single motive that is associated with problematic pornography consumption, but rather an abundance of motives. When users did report high levels of motivation for consuming pornography (e.g., Profile 4), they report more loneliness, higher fear of intimacy, and less social support. These findings when viewed together provide the support for more intensive clinical investigations and intense longitudinal studies.

Data availability statement

The data that support the findings of this study are available from the corresponding author, DWMM, upon reasonable request.