Critique of "Perceived Addiction to Internet Pornography and Psychological Distress: Examining Relationships Concurrently and Over Time" (2015)
- Update: A much more comprehensive analysis of Joshua Grubbs's claims and studies can be found here - Is Grubbs pulling the wool over our eyes with his "perceived porn addiction" research? (2016)
Here are a few of the headlines birthed from this new study by Joshua B. Grubbs, Nicholas Stauner, Julie J. Exline, Kenneth I. Pargament, and Matthew J. Lindberg (Grubbs et al., 2015):
- Psychology Research Links Distress to Perceived Internet Pornography Addiction
- Watching Porn Is OK. Believing In Porn Addiction Is Not
- Perceived Addiction To Porn Is More Harmful Than Porn Use Itself
- Believing You Have Porn Addiction Is the Cause of Your Porn Problem, Study Finds
In essence the study's main claim is reported as: "perceived addiction" to pornography is more related to psychological distress than are current daily hours of porn viewing. An excerpt from one of the above articles:
A new study in the journal Psychology of Addictive Behaviors has found that perceived addiction to pornography—that is, “feeling addicted to Internet pornography irrespective of actual pornography use”—is associated with forms of psychological distress including depression, anxiety, anger, and stress. Pornography use itself, the authors found, was “relatively unrelated to psychological distress.”
While the above quote contains inaccuracies which we will explore, let's take it at face value. The reader is left with the impression that actual porn use is no big deal, but "believing" you are addicted to porn will cause you psychological distress. The take away: It's perfectly healthy to use porn as long as you don't believe you are addicted.
Grubbs et al.'s claim, and all the resulting headlines, are built upon this finding: Subjects' current hours of porn use did not correlate strongly enough (in researchers' subjective view) with scores on Grubbs's own porn use questionnaire (the Cyber Pornography Use Inventory "CPUI"). To put it another way, if porn addiction really existed there "should" be, in the authors' view, a one-to-one relationship between current hours of use and scores on the CPUI. Grubbs et al. also reported that "psychological distress" was related to scores on the CPUI, but not as strongly related to current hours of use.
Here's the thing: There's absolutely no scientific basis for declaring the CPUI a measure of "perceived addiction," and yet that's what all the inflated headlines rest on! The CPUI was never validated for "perceived" as opposed to "real" addiction.
For Grubbs et al.'s claims and interpretations to be valid, BOTH of the following must be true and supported by actual research:
1) The Cyber Pornography Use Inventory (CPUI) must assess "perceived addiction" to porn but not actual porn addiction.
- Grubbs himself developed the 9-item CPUI as an inventory of online porn problems, not a "perceived addiction" test. Here he chose to use it in lieu of other validated addiction tests, precisely to create the illusion that he could measure "perceived addiction" rather than addiction. In fact, the CPUI measures the same signs, symptoms and indications of addiction as do standard addiction tests.
- In the current study, Grubbs et al. use the phrase "perceived porn addiction" synonymously with subjects' total score on the CPUI, without scientific justification.
2) Internet porn addiction must equal hours of porn viewing.
- This is refuted by the scientific literature. Internet porn addiction ≠ hours of porn viewing.
- Shockingly, the Grubbs et al. study reveals there actually was a strong correlation between hours of use and the CPUI! From p. 6 of the study:
"Additionally, average daily pornography use in hours was significantly and positively associated with depression, anxiety, and anger, as well as with perceived addiction."
With respect to the first point, Grubbs developed his own porn addiction questionnaire (CPUI), and then later capriciously declared that it measures only "perceived addiction to porn" - without demonstrating any justification for his recharacterization. (Really!)
With respect to the second point, previous research teams have found that the variable "hours of use" is not correlated with cybersex addiction (or video-gaming addiction). That is, addiction is more accurately predicted by other variables than "hours of use." That said, as you can see from the above excerpt, Grubbs actually found a significant correlation between hours of use and psychological distress.
We'll look at details about why Grubbs et al.'s assumptions are neither true nor supportable below, but here's how the researchers could have described their actual findings without misleading the public:
"Study finds that porn addiction is strongly related to psychological distress and less strongly (but still) related to current hours of use."
The cliff notes version: Addiction is related to psychological distress, and so are hours of use. So much for the attention-grabbing, misleading headlines spawned by the study.
The CPUI Assesses Porn Addiction, Not "Perceived Addiction"
In Grubbs's initial 2010 paper he validated the Cyber-Pornography Use Inventory (CPUI) as a questionnaire assessing actual porn addiction. Nowhere in the 2010 paper does he use the phrase "perceived addiction". Excerpts from Grubbs's original paper confirming his CPUI assesses only real porn addiction:
The previously described models proposed for understanding behavioral addictions were the primary theoretical assumptions used to derive the instrument for this study, the Cyber-Pornography Use Inventory (CPUI), patterned after the Internet Sex Screening Test developed by Delmonico (Delmonico & Griffin, 2008). The CPUI design was based on the principle that addictive behavior is characterized by an inability to stop the behavior, significant negative effects as a result of the behavior, and a generalized obsession with the behavior (Delmonico & Miller, 2003).
The CPUI does indeed show promise as an instrument assessing Internet pornography addiction. Whereas previous instruments, such as the ISST, had assessed only broad-spectrum online sexual addiction, this scale did demonstrate promise in specifically assessing Internet pornography addiction. Furthermore, the items on the previously explained Addictive Patterns scale seem to find some level of theoretical support and potential construct validity when compared with the diagnostic criteria for both Substance Dependence and Pathological Gambling, an ICD.
Finally, five of the items on the Addictive Patterns scale from the original Compulsivity scale seem to directly tap into the individual’s perceived or actual inability to stop the behavior in which they are engaging. Inability to stop a problematic behavior under any circumstances is not only an important diagnostic criterion for both SD and PG, but it can also can be thought of as one of the core elements of both addiction, as manifested in SD, and ICDs (Dixon et al., 2007; Pontenza, 2006). It seems that it is this inability that creates the disorder.
In a 2013 study Grubbs reduced the number of CPUI questions from 32 to 9, removed the guilt and shame section of the test, and re-labled his actual porn addiction test a "perceived porn addiction" test. He did so without and explanation, while using the phrase "perceived addiction" 80 times in the 2013 paper. Let's be very clear - Grubbs did not validate his CPUI as an assessment tool differentiating actual porn addiction from "perceived porn addiction".
In reality, Grubbs's Cyber Pornography Use Inventory (CPUI) questionnaire very similar to many other drug and behavioral addiction questionnaires. Like other addiction tests, the CPUI assesses behaviors and symptoms common to all addictions, such as: the inability to control use; compulsion to use, cravings to use, negative psychological, social and emotional effects; and preoccupation with using. In fact, only 1 of the 9 questions below even hints at "perceived addiction." Yet we are told that a person's total score for all 9 questions is synonymous with "perceived addiction" rather than addiction itself. Very misleading, very clever, and without any scientific basis. Agnotology fodder, anyone? (Agnotology is the study of culturally induced ignorance or doubt, particularly the publication of inaccurate or misleading scientific data.) Reality check: other researchers describe the CPUI as an actual porn addiction assessment questionnaire, and use it as such in their published studies:
- An Examination of Internet Pornography Usage among Male Students at Evangelical Christian Colleges (2011)
- Questionnaires and scales for the evaluation of the online sexual activities: A review of 20 years of research (2014)
- Problematic cybersex: Conceptualization, assessment, and treatment (2015)
- Clarifying the Links Among Online Gaming, Internet Use, Drinking Motives, and Online Pornography Use (2015)
- Examining Correlates of Problematic Internet Pornography Use Among University Student (2016)
The last study above used a longer version of the Grubbs CPUI and an Internet pornography addiction questionnaire derived from the DSM-5 Internet video-gaming addiction criteria. The graphs show the same subjects' scores on the two different porn addiction questionnaires:
No surprise. Very similar results and distribution for the Grubbs CPUI and the "actual" porn addiction questionnaire. That's because the CPUI is an addiction test, invalidating Grubbs' attempt to relabel it as a "perceived addiction test."
Note that decades of established addiction assessment tests for both chemical and behavioral addictions rely on similar questions to assess actual, not merely perceived, addiction. Let's compare the CPUI to a commonly used addiction assessment tool known as the "4 Cs." The CPUI questions that correlate with the four Cs are noted as well.
- Compulsion to use (2, 3)
- Inability to Control use (2, 3, maybe 4-6)
- Cravings to use (3 especially, but 1-6 could be interpreted as cravings)
- Continued use despite negative consequences (4-6, perhaps 7-9)
The Cyber Pornography Use Inventory (CPUI) (developed by Grubbs)
1. I believe I am addicted to Internet pornography.
2. I feel unable to stop my use of online pornography.
3. Even when I do not want to view pornography online, I feel drawn to it
4. At times, I try to arrange my schedule so that I will be able to be alone in order to view pornography.
5. I have refused to go out with friends or attend certain social functions to have the opportunity to view pornography.
6. I have put off important priorities to view pornography.
7. I feel ashamed after viewing pornography online.
8. I feel depressed after viewing pornography online.
9. I feel sick after viewing pornography online.
Addiction experts rely on assessment tools like the 4Cs as indicating addiction because neuroscientists have correlated the symptoms those questions address with underlying addiction-related brain changes in decades of basic-research studies. As a medical matter, addiction is a disorder of the brain. It shows up in very specific behaviors, but can't be assessed from superficial indications such as hours of use. See the public policy statement of the American Society of Addiction Medicine.
In short, the CPUI does a better job at actually identifying that which it purports to differentially diagnose against (actual addiction) than at identifying "perceived addiction," as Grubbs claims it does.
Finally, we must ask why Grubbs found it necessary to create his own internet porn addiction test. Others, well established and thoroughly validated, were available to him. Might it be that as the author of the CPUI he assumes he has the power to re-label it as a "perceived addiction to porn" test in order to fool readers into believing that all porn addicts are misdiagnosing themselves? Sorry, if it walks like a duck and talks like a duck....
Bottom Line: The CPUI assesses actual porn addiction, not "perceived" porn addiction. Delete the word "perceived" from the study, and every article about it, and you are left with an unsurprising finding: psychological distress is related to
perceived porn addiction.
Current Hours of Use Are Not Related To Porn Addiction
Grubbs et al.'s conclusion is largely based on a faulty premise: The extent of a porn addiction is best assessed simply by hours of internet porn viewing. As Grubbs et al. did not find a tight enough correlation (in their view) in their subjects, they concluded their subjects merely had "perceived addiction" instead. Two huge holes in the story render Grubbs et al.'s claim highly suspect.
As described earlier, the first gaping hole is that Grubbs et al. actually found a pretty strong correlation between hours of use and the CPUI! From p. 6 of the study:
"Additionally, average daily pornography use in hours was significantly and positively associated with depression, anxiety, and anger, as well as with perceived addiction."
Stop the presses! This excerpt directly contradicts all the headlines, which claim that pornography use was NOT strongly correlated with psychological distress or "perceived addiction." Again, whenever you see the phrase "perceived addiction" it actually denotes the subjects' total score on the CPUI (which is a porn addiction test).
To say all of this another way: Both psychological distress and CPUI scores were significantly correlated with hours of use. Does any journalist or blogger ever read an actual study?
The second hole in this study's underpinnings, which you could drive a truck through, is that research on internet porn and videogame use (1, 2, 3, 4, 5, 6, 7, 8) has reported that neither addiction correlates with hours of use. The variable 'hours of use' is an unreliable measure of addiction, and established addiction assessment tools evaluate addiction using multiple other factors (such as those listed in the CPUI). The following cybersex addiction studies, which Grubbs omitted, report little relationship between hours and indications of addiction:
"Results indicate that self-reported problems in daily life linked to online sexual activities were predicted by subjective sexual arousal ratings of the pornographic material, global severity of psychological symptoms, and the number of sex applications used when being on Internet sex sites in daily life, while the time spent on Internet sex sites (minutes per day) did not significantly contribute to explanation of variance in Internet Addiction Test sex score (IATsex). We see some parallels between cognitive and brain mechanisms potentially contributing to the maintenance of excessive cybersex and those described for individuals with substance dependence."
"Recent findings have demonstrated an association between CyberSex Addiction (CA) severity and indicators of sexual excitability, and that coping by sexual behaviors mediated the relationship between sexual excitability and CA symptoms. Results showed strong correlations between CA symptoms and indicators of sexual arousal and sexual excitability, coping by sexual behaviors, and psychological symptoms. CyberSex Addiction was not associated with offline sexual behaviors and weekly cybersex use time."
According to our best knowledge this study is the first direct examination of associations between the frequency of porn use and actual behavior of treatment-seeking for problematic porn use (measured as visiting the psychologist, psychiatrist or sexologist for this purpose). Our results indicate that the future studies, and treatment, in this field should focus more on impact of porn use on the life of an individual (quality) rather than its mere frequency (quantity), as the negative symptoms associated with porn use (rather than porn use frequency ) are the most significant predictor of treatment-seeking behavior.
Relation between PU and negative symptoms was significant and mediated by self-reported, subjective religiosity (weak, partial mediation) among non-treatment seekers. Among treatment-seekers religiosity is not related to negative symptoms.
Higher scores on addictive measures of internet porn use were correlated with daily or more frequent use of internet porn. However, the results indicate that there was no direct link between the amount and frequency of an individual’s pornography use and struggles with anxiety, depression, and life and relationship satisfaction. Significant correlations to high internet porn addiction scores included an early first exposure to internet porn, addiction to video games, and being male. While some positive effects of internet porn use have been documented in previous literature our results do not indicate that psychosocial functioning improves with moderate or casual use of internet porn.
Thus, from the outset this study and its assertions collapse because its conclusions rest upon equating current hours of use with the level of addiction/problems/distress reported by subjects as a valid measure of addiction.
Why don't addiction specialists rely on hours of use? Imagine trying to assess addictions by simply asking, "How many hours do you currently spend eating (food addiction)?" or "How many hours do you spend gambling (gambling addition)?" or "How many hours do you spend drinking (alcoholism)?" To demonstrate how problematic hours of use would be, consider alcohol as an example:
- A 45-year old Italian man has a tradition of drinking 2 glasses of wine every night with dinner. His meal is with his extended family and it takes 3 hours to complete (lots of yakking). So he drinks for 3 hours a night, 21 hour per week.
- A 25 year-old factory worker only drinks on the weekends, but binge drinks both Friday and Saturday night to the point of passing out or getting sick. He regrets his actions and wants to stop, but can't, drives drunk, gets in fights, is sexually aggressive, etc. He then spends all of Sunday recovering, and feels like crap until Wednesday. However, he spent only 8 hours a week drinking.
Which drinker has a problem? This is why "current hours of use" alone cannot inform us as to who is addicted and who is not.
Finally, we must ask why Grubbs et al. chose to create the CPUI when other, thoroughly validated addiction tests were readily available.
Bottom line: The study's claims depend upon "current hours of use" being a valid criterion for true addiction. They are not. Moreover, once you get past the abstract, the full study reveals that "current hours of use" is actually related to both psychological distress and CPUI scores!
"Current Hours of Use" Omits Many Variables
A secondary methodological problem is that Grubbs et al. assessed porn use by asking subjects about their "current hours of porn use." That question is troublingly vague. Over what period? One subject may be thinking "How much did I use yesterday?" another "over the last week?" or "on average since I decided to quit viewing because of unwanted effects?" The result is data that are not comparable can't be analyzed for the purpose of drawing reliable conclusions.
More important, the "current porn use" question, on which the study's conclusions rest, fails to ask about key variables of porn use: age use began, years of use, whether the user escalated to novel genres of porn or developed unexpected porn fetishes, the ratio of ejaculation with porn to ejaculation without it, amount of sex with a real partner, and so forth. Those questions would likely enlighten us more about who really has a problem with porn use than simply "current hours of use."
Deeper Analysis: Addiction Correlates With Addiction, not Emotional Distress
We have established that the CPUI is in no way a "perceived addiction" test. Instead, the CPUI is an actual porn addiction test assessing 3 categories of compulsive pornography use:
- Compulsion to use & Inability to control use
- Efforts to use (negative consequences)
- Emotional distress after using (shame, depression, feel sick)
What do the conspicuous headlines claim and the authors assert? That psychological distress about one's porn addiction is the real problem, rather than the addiction itself. If that were true, we would expect "porn addicts" to score especially high in Emotional Distress category of the CPUI. Not so. From p. 9 of the study:
"In all SEM analyses, emotional distress consistently had the weakest loadings on the latent factor of perceived addiction...... Such a consistent pattern across studies strongly indicates that emotional distress regarding use is not the primary driving factor in the link between perceived addiction and psychological distress."
Translation: Scores on the CPUI Emotional Distress section were the least related to scores on separate Psychological Distress questionnaires (which measure such things as stress, depression and anger). Put simply, the psychological distress of porn addiction does NOT arise from shame or guilt. Instead, psychological distress arises from the inability to control use despite negative consequences, as measured by CPUI sections 1 and 2. From pg 9 of the study:
By contrast, perceived compulsivity—the direct acknowledgment of feelings of addiction to pornography—was consistently the primary factor driving perceived addiction. Additionally, access efforts [negative consequences] loaded more strongly on perceived addiction than emotional distress regarding use.
Translation: The inability to control use (questions 1-3) coupled with negative consequences (questions 4-6), was more strongly related to "psychological distress." Put simply, guilt and shame weren't such a big deal, but inability to control use, the compulsion to watch porn when subjects don't want to, refusing to go out with friends or attend social functions in order to view pornography, and putting off important priorities to view pornography, are what really upset subjects.
This is a very different reality than readers get from either the misleading headlines or study abstract.
Grubbs Introduction Distorts Current State of the Research
In the introduction and discussion sections Grubbs et al. toss aside decades of neuropsychological and other addiction research (and related assessment tools) to attempt to persuade readers that the scientific literature shows that internet porn addiction doesn't exist (and that therefore that all evidence of porn addiction must be "perceived," not real). A new review shows just how farfetched this contention is. See Neuroscience of Internet Pornography Addiction: A Review and Update, which aligns decades of addiction neuroscience research with recent neuroscience and neuropsych studies on porn users themselves. It concludes (logically and scientifically) that internet pornography addiction is quite real, and in fact a subset of internet addiction (based on more than 100 brain studies, as well as hundreds of other relevant studies).
In their opening paragraphs, Grubbs et al. demonstrate their profound bias by basing their claim about the nonexistence of internet porn addiction on the papers of two self-proclaimed "internet porn addiction debunkers": David Ley, author of The Myth of Sex Addiction, and former UCLA researcher Nicole Prause, whose work has been formally criticized in the medical literature for weak methodology and unsupported conclusions.
For example, Grubbs et al. rely on a one-sided paper by Ley, Prause and their colleague Peter Finn, which claimed to be a review (that is, an impartial analysis of the existing literature). However, it omitted or misrepresented nearly every study that found negative effects of internet porn use, while also ignoring the dozens of recent internet addiction studies demonstrating addiction-related structural brain changes in internet addicts' brains. (Line-by-line critique can be found here.)
Equaling telling is Grubbs et al.'s omission of every brain scan and neuropsychological study that found evidence in support of the porn addiction model (over a dozen collected here). Instead of hard science from the many omitted studies, the reader is given an overreaching conclusion:
In sum, there is a fair amount of evidence suggesting that many individuals feel addicted to Internet pornography, even in the absence of a clinically verified diagnosis to subsume such a disorder.
Finally, the only neurological study cited by Grubbs as refuting porn addiction (Steele et al.) actually supports the porn addiction model. Steele et al. reported higher EEG readings (P300) when subjects were exposed to porn photos. Studies consistently show that an elevated P300 occurs when addicts are exposed to cues (such as images) related to their addiction. In addition, the study reported that greater cue-reactivity to porn correlated with less desire for partnered sex. As neither result matched the headlines, Grubbs perpetuated the flawed conclusions of the original authors (the "debunkers of porn addiction").
Given its unsupported conclusions and biased claims about the non-existence of porn addiction, it seems likely that Grubbs et al. designed this study to meet a specific agenda - to re-label porn addiction as "perceived addiction" and persuade readers that porn use is harmless and they need only worry about believing it can harm. Agnotology mission accomplished!
This saying comes to mind: What the abstract giveth, the full study taketh away. The headlines and claims spawned by Grubbs et al. are not even supported by the underlying study. Again,
- Both internet gaming and internet porn addiction studies have already established that hours of use do not correlate well with addiction. This fact alone guts this entire study and its misleading headlines.
- The Cyber Pornography Use Inventory (CPUI) assesses the signs, symptoms and behaviors of an addiction, not "perceived addiction". Don't be fooled; the CPUI was never validated for "perceived" as opposed to "real" addiction.
- Grubbs et al.'s study reveals that the "average daily pornography use in hours was significantly and positively associated with depression, anxiety, and anger, as well as with perceived addiction (the CPUI)." This directly contradicts the press claims that say hours of use were not related to CPUI scores or psychological distress.
Any one of the above dismantles this study, but all three mean that this study should be ignored as the work of agnotology that it is.
Research such as this contributes to the ongoing campaign to confuse the public about the reality of internet porn addiction. For example, one frequently sees attempts by Grubbs's colleagues to conflate internet porn addiction with sex addiction and then sweep both away as "unsupported," even though the neuropsychological and medical evidence demonstrating internet addiction is already overwhelming. Another tactic is to conflate internet porn addiction with "Hypersexual disorder" and then claim that the DSM-5, by rejecting the latter, has rejected the former. In fact, internet porn addiction was never formally proposed, or evaluated, for inclusion in the DSM-5. It's time it was, given the mounting evidence that both internet addiction and its subtype internet porn addiction are true addictions.
In the same tradition, Grubbs et al., without justification, now attempt to sweep aside decades of addiction research and assessment tests developed for all kinds of addictions, and substitute their own worldview (that internet porn addiction doesn't exist and should be recast as "perceived addiction"). Should society and its headline-hungry journalists allow this? You be the judge.