Study invalidates the CPUI-9 as an instrument to assess either “perceived pornography addiction” or actual pornography addiction

course.corr_.jpg

SECTION 1: Introduction

A new study (Fernandez et al., 2017) tested and analyzed the CPUI-9, a purported “perceived pornography addiction” questionnaire developed by Joshua Grubbs, and found that it couldn’t accurately assess “actual porn addiction” or “perceived porn addiction” (Do Cyber Pornography Use Inventory-9 Scores Reflect Actual Compulsivity in Internet Pornography Use? Exploring the Role of Abstinence Effort). It also found that 1/3 of the CPUI-9 questions should be omitted to return valid results related to “moral disapproval”, “religiosity”, and “hours of porn use.” The findings raise significant doubts about conclusions drawn from any study that has employed the CPUI-9 or relied on studies that employed it. Many of the new study’s concerns and criticisms mirror those outlined in this extensive YBOP critique.

In simple terms the CPUI-9 studies and the headlines they spawned contributed to the following questionable assertions:

  1. “Belief in porn addiction” or “perceived pornography addiction” can be distinguished from “actual pornography addiction” by the CPUI-9.
  2. “Current levels of porn use” is the one valid proxy for actual porn addiction, not scores on porn addiction assessment questionnaires.
  3. In some subjects “current levels of porn use” did not correlate linearly with total CPUI-9 scores. Grubbs asserts these individuals falsely “believe” they are addicted to porn.
  4. In the CPUI-9 studies, “religiosity” correlates with Total CPUI-9 scores. Because of this Grubbs suggests that most religious porn users only believe they are addicted, and do not have an actual porn addiction.
  5. In some of these studies both “religiosity” & “moral disapproval” correlate with Total CPUI-9 scores. Because of this Grubbs and his teams claim that religious porn users have shame-induced “belief in pornography addiction,” not actual pornography addiction.

Articles based on various CPUI-9 studies sum up these findings as:

  • Believing in porn addiction is the source of your problems, not porn use itself.
  • Religious porn users are not really addicted to porn (even if they score high on the CPUI-9) – they are simply experiencing shame & guilt surrounding their porn use.

In this extraordinary 2016 Psychology Today article, Joshua Grubbs sums up his views, claiming that porn addiction is nothing more than religious shame:

Being labeled “porn addict” by a partner, or even by oneself, has nothing to do with the amount of porn a man views, says Joshua Grubbs, assistant professor of psychology at Bowling Green University. Instead, it has everything to do with religiosity and moral attitudes toward sex. In short, he says, “It’s shame-motivated.”

Contrary to Grubbs’s above statement, his studies actually found that “the amount of porn a man views” is very much related to porn addiction (scores on the CPUI-9).

Grubbs continues:

….Grubbs calls it “perceived pornography addiction.” “It functions very differently from other addictions.”

As Fernandez et al., 2017 reveals, the CPUI-9 has, in fact, failed to assess “perceived porn addiction.” And actual porn addiction functions very much like other addictions.

Bottom line: The results of Fernandez et al., 2017 place all assertions based on CPUI-9 results, and all of the resulting headlines, in serious doubt.

The problems with the “perceived pornography addiction” questionnaire (CPUI-9)

To understand the new study’s importance we need to first examine the Cyber Pornography Use Inventory (CPUI-9). Important to note:

  • The CPUI-9 is divided into 3 named sections with 3 questions each (take special note of the “Emotional Distress” questions).
  • Each question is scored using a Likert scale of 1 to 7, with 1 being “not at all,” and 7 being “extremely.”
  • Whenever Grubbs uses the phrase “perceived addiction” he really means nothing more than the total score on his CPUI-9 test, yet the test cannot actually distinguish “perceived” addiction from real addiction.

Perceived Compulsivity Section

  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

Access Efforts Section

  1. At times, I try to arrange my schedule so that I will be able to be alone in order to view pornography.
  2. I have refused to go out with friends or attend certain social functions to have the opportunity to view pornography.
  3. I have put off important priorities to view pornography.

Emotional Distress Section

  1. I feel ashamed after viewing pornography online.
  2. I feel depressed after viewing pornography online.
  3. I feel sick after viewing pornography online.

Examining the CPUI-9 reveals three glaring truths exposed by the authors of Fernandez et al., 2017 (and in the YBOP critique):

  • The CPUI-9 cannot differentiate between an actual porn addiction and a mere belief in porn addiction (“perceived addiction”).
  • The first two sections (questions 1-6) assess the signs and symptoms of an actual pornography addiction (not “perceived pornography addiction”).
  • The “Emotional Distress” questions (7-9) assess levels of shame and guilt, and are not found in any other type of addiction assessment (i.e., they don’t belong).

We will first provide a brief summary of Fernandez et al., 2017 followed by excerpts from its findings with our comments.

SECTION 2: Fernandez et al., 2017 – Design & Findings

A brief description of Fernandez et al., 2017:

This was a unique study in that it asked participants to abstain from internet porn for 14 days. (Only a handful of studies have asked participants to abstain from porn, which is one of the most unambiguous ways to reveal its effects.) Participants took the CPUI-9 before and after their 14-day attempt at porn abstinence. (Note: They did not abstain from masturbation or sex, only porn.) The researchers’ main objective was to compare ‘before’ and ‘after’ scores of the 3 sections of the CPUI-9 to the following 3 variables:

1) Actual compulsivity. The fact that the participants were attempting to quit porn allowed the researchers to measure actual compulsivity (with respect to porn use). The researchers used a formula of “failed abstinence attempts X abstinence effort” to measure actual compulsivity. This is the first study to compare actual compulsivity to subjects’ scores on a porn addiction questionnaire (the CPUI-9).

2) Frequency of Internet porn use. Subjects’ frequency of internet porn use prior to the study.

3) Moral Disapproval questionnaire. In addition to taking the CPUI-9, Fernandez’s subjects took a Moral Disapproval questionnaire, so researchers could correlate its results with CPUI-9 questions. Moral disapproval of pornography was measured by four items rated on a 7-point Likert scale from 1 (not at all) to 7 (extremely):

  • “Viewing pornography online troubles my conscience,”
  • “Viewing pornography violates my religious beliefs,”
  • “I believe that viewing pornography is morally wrong,” and
  • “I believe that viewing pornography is a sin.”

Note that 3 of the 4 “moral disapproval” questions involve religiosity.

Let’s explore what Fernandez et al., 2017 reported and what it had to say about the CPUI-9 and the conclusions put forth in studies employing the CPUI-9.

What did Fernandez et al., 2017 report?

Findings #1: Higher frequency of porn use was related to: 1) Total CPUI-9 scores, 2) “Perceived Compulsivity” questions, and 3) actual compulsivity (failed abstinence attempts X abstinence effort). However, frequency of porn use was unrelated to scores on “Emotional Distress” questions 7-9 (which assess guilt & shame).

Translation: No matter how you measure it, actual porn addiction is strongly correlated with higher levels of porn use. However, guilt & shame questions 7-9 shouldn’t be part of a porn addiction (or even “perceived porn addiction”) assessment because they are unrelated to frequency of porn use. The 3 “Emotional Distress” questions do not belong. In fact, they skew CPUI-9 results.

Take away 1: The Grubbs studies (or any study that used the CPUI-9) did not assess “perceived porn addiction” or “belief in porn addiction” or “labeling themselves as addicted. It’s important to keep in mind that “perceived pornography addiction” indicates nothing more than the total score on the CPUI-9. A headline such as “Believing You Have Porn Addiction Is the Cause of Your Porn Problem, Study Finds” should now be re-interpreted as “Having a Porn Addiction Is the Cause of Your Porn Problem, Study Finds.” It’s important to note that there’s no scientific precedent for a “perceived addiction” assessment test, and the CPUI-9 has not been validated as such.

Take away 2: Guilt & shame questions 7-9 have no place in a porn addiction questionnaire because they skew Total CPUI-9 scores far lower for non-religious porn users, while elevating scores for religious porn users. For example, if an atheist and devout Christian have identical scores on CPUI-9 questions 1-6, it’s almost certain that the Christian will end up with far higher CPUI-9 scores, after questions 7-9 are added – regardless of the degree of addiction in either subject.

Take away 3: Omitting guilt & shame questions 7-9 results in “hours of porn use” (not religion) being the strongest predictor of porn addiction. To say it another way, “Emotional Distress” questions correlate strongly with “religiosity” but not with “hours of porn use.” Contrary to misleading articles, the CPUI-9 studies found that higher levels of porn use correlated with so-called “perceived pornography addiction.”

Findings #2: Failed abstinence attempts correlated with the 1) Total CPUI-9 scores, and 2) “Perceived Compulsivity” questions – but not with “Emotional Distress” questions 7-9.

Translation: The inability to control use correlated with CPUI-9 actual addiction questions 1-6, but not with the guilt & shame questions 7-9.

Take away: Once again, CPUI-9 questions 1-6 assess actual porn addiction, while guilt & shame questions 7-9 do not. Inclusion of the “Emotional Distress” questions leads to far lower CPUI-9 scores for porn addicts and far higher CPUI-9 scores for religious individuals, or just about anyone who would prefer not to be using porn.

Findings #3: “Moral disapproval” of pornography use was strongly correlated with 1) Total CPUI-9 scores, and 2) “Emotional Distress” questions. However, “moral disapproval” was only slightly related to CPUI-9 “Perceived Compulsivity” scores. In other words, the most addicted subjects did not score higher on religiosity.

Translation: “Moral disapproval” of porn correlated strongly with the CPUI-9 guilt & shame questions 7-9. Most importantly, questions 7-9 are the only reason “moral disapproval” correlated with Total CPUI-9 (“perceived porn addiction”). Inclusion of the “Emotional Distress” questions is what generates the misleading claim that “belief in porn addiction” is driven by moral disapproval.

Take away 1: Omitting the guilt & shame questions (7-9) results in “moral disapproval” having nothing to do with porn addiction. The “Emotional Distress” questions assessing guilt and shame cause just about anyone who would prefer not to be using porn (especially religious individuals) to have much higher CPUI-9 scores.

Take away 2: Inclusion of guilt & shame questions 7-9 leads to artificially strong correlations between “moral disapproval” and the Total CPUI-9 (perceived addiction). The fact that religious individuals score very high on both “moral disapproval” and the “Emotional Distress” questions has led to unsupported claims that religious people are far more likely to “perceive” themselves addicted to porn (remember “perceived addiction” is shorthand for “total CPUI-9 score”). However, this is simply not true, because the “extra” points religious people earn on questions 7-9 do not measure addiction, or even “perception” of addiction. They measure nothing but emotional distress due to conflicted values.

Take away 3: Religious individuals score very high on both the “moral disapproval” questions and the “Emotional Distress” questions. CPUI-9 based studies have adopted the correlation between “moral disapproval” and the 3 “Emotional Distress” questions to create a mythology that religious individuals only believe they are addicted to porn. However, these questions assess neither porn addiction nor “belief” nor “perception” of addiction, so they are out of place in this instrument.

In summary, the conclusions and claims spawned by the CPUI-9 are simply invalid. Joshua Grubbs created a questionnaire that cannot, and was never validated for, sorting “perceived” from actual addiction: the CPUI-9. With zero scientific justification he re-labeled his CPUI-9 as a “perceived pornography addiction” questionnaire.

Because the CPUI-9 included 3 extraneous questions assessing guilt and shame, religious porn users’ CPUI scores tend to be skewed upward. The existence of higher CPUI-9 scores for religious porn users was then fed to the media as a claim that, “religious people falsely believe they are addicted to porn.” This was followed by several studies correlating moral disapproval with CPUI-9 scores. Since religious people as a group score higher on moral disapproval, and (thus) the total CPUI-9, it was pronounced (without actual support) that religious-based moral disapproval is the true cause of pornography addiction. That’s quite a leap, and unjustified as a matter of science.

We will now present excerpts from Fernandez et al., 2017 accompanied by comments and clarifying images.


SECTION 3: Excerpts of Fernandez et al., 2017 (with comments)

The discussion section of Fernandez et al., 2017 contained three main findings, three theoretical implications, and two clinical implications. They follow.

First main finding: The CPUI-9 “Perceived Compulsivity” questions assess actual compulsivity not “belief” in porn addiction

Fernandez et al., 2017 discuss how the actual compulsivity scores align with scores on the CPUI-9 “Perceived Compulsivity” questions, but not with the “Emotional Distress” questions.

We found partial support for our second hypothesis, that failed abstinence attempts would interact with abstinence effort to predict higher CPUI-9 scores, controlling for moral disapproval. However, this relationship was limited to Perceived Compulsivity scores, and not Emotional Distress scores and CPUI-9 full scale scores. Specifically, when failed abstinence attempts are high and abstinence effort is high, higher scores on the Perceived Compulsivity subscale are predicted. This finding is consistent with our proposition that it is not merely frequency of pornography use which contributes to perceptions of compulsivity, but that this would also depend on an equally important variable, abstinence effort. Previously, studies have demonstrated that frequency of pornography use accounts for some variance in the CPUI-9 (Grubbs et al., 2015a; Grubbs et al., 2015c), but frequency of pornography use alone is not sufficient to infer the presence of compulsivity (Kor et al., 2014). The present study posits that some individuals may view IP frequently, but may not be exerting substantial effort in abstaining from IP. As such, they might have never felt that their use was compulsive in any way, because there was no intention to abstain. Accordingly, the present study’s introduction of abstinence effort as a new variable is an important contribution. As predicted, when individuals tried hard to abstain from pornography (i.e., high abstinence effort) but experienced many failures (i.e., high failed abstinence attempts), this aligned with greater scores on the Perceived Compulsivity subscale.

SUMMARY: First, frequency of porn use was strongly related to the CPUI-9 “Perceived Compulsivity” questions and actual compulsivity (“failed abstinence attempts X abstinence effort”).

Second, porn users who tried really hard to stop, yet repeatedly failed, had the highest scores on the CPUI-9 “Perceived Compulsivity” questions. Put simply, CPUI-9 questions 1-3 assess actual compulsivity (cravings and inability to control use) rather than “belief in addiction.” That means they offer no support for the concept of “perceived addiction.”

Third, the “Emotional Distress” questions (assessing guilt & shame) are immaterial in assessing actual porn addiction, and only function to skew Total CPUI-9 scores higher for religious individuals and those who disapprove of porn use.

Let’s do visual stats. Here are some tips for understanding the numbers in the following tables and images: Zero means no correlation between two variables; 1.00 means a complete correlation between two variables. The bigger the number the stronger the correlation between the 2 variables. If a number has a minus sign, it means there’s a negative correlation between two things. (For example, there’s a negative correlation between exercise and heart disease. Thus, in normal language, exercise reduces the chances of heart disease. On the other hand, obesity has a positive correlation with heart disease.)

We start with the table of correlations from Fernandez et al., 2017. Number 1 is “frequency of internet porn use”, which correlates strongly the CPUI-9 “Perceived Compulsivity” questions (0.47), Abstinence Effort (0.28), and Failed Abstinence Attempts (0.47). Frequency of porn use was unrelated to “Emotional Distress” questions (0.05) and negatively correlated with “moral disapproval” (0.14).

The results without the 3 “Emotional Distress” questions skewing the results: “Frequency of porn use” is by far the strongest predictor of actual porn addiction – not religiosity! As Fernandez et al. pointed out, the above correlations are similar for all the CPUI-9 studies conducted by Grubbs’s teams.

The core premise of the “perceived porn addiction” studies rests upon the unfounded assertion that Total CPUI-9 scores should correlate perfectly with “current hours of porn use”. The researchers presume that – if a person’s CPUI-9 scores are relatively high, yet their “hours of porn use” only moderately high – the individual falsely “believes” they are addicted to porn. A graphic representation of this assertion:

However, as Fernandez et al. and many other studies point out, current level of porn use is an unreliable measure of addiction. More importantly, the 3 “Emotional Distress” questions greatly weaken correlations between frequency of use and the Total CPUI-9 scores.

Bottom line: There is no such thing as “perceived compulsivity” or “perceived porn addiction.” If a porn user scores high on a porn addiction test, it means he is experiencing the signs and symptoms of an actual addiction. In addition, it is scientifically unsound to presume that current levels of porn consumption can be used as a proxy for actual porn addiction (as many studies have concluded).


Second main finding: Needing greater effort to abstain correlated with CPUI-9 “Perceived Compulsivity” questions

Fernandez et al., 2017 point out that needing greater effort to abstain correlated strongly with the CPUI-9 “Perceived Compulsivity” questions and frequency of porn use, but not with the “Emotional Distress” questions:

Interestingly, abstinence effort as an individual predictor also demonstrated a significant positive predictive relationship with the Perceived Compulsivity subscale (but not the Emotional Distress subscale and the CPUI-9 full scale), controlling for failed abstinence attempts and moral disapproval, although this relationship was not hypothesized a priori. We predicted in the present study that only individuals who actually experienced failed abstinence attempts might infer compulsivity from their own behavior, leading to perceptions of compulsivity. However, we found that greater abstinence effort predicted higher scores on the Perceived Compulsivity subscale, and that this relationship was seen even independent of failed abstinence attempts. This finding has the important implication that trying to abstain from pornography in and of itself is related to perceptions of compulsivity in some individuals.

SUMMARY: Similar to the first finding, higher scores on the CPUI-9 “Perceived Compulsivity” questions strongly correlated with features of actual compulsivity (needing high levels of effort to abstain from porn). Put simply, the CPUI-9 “Perceived Compulsivity” questions assess actual compulsivity. However, needing greater effort to abstain from porn had little to do with guilt, shame or remorse (“Emotional Distress” questions). Guilt & shame surrounding porn use has little to do with actual porn addiction, let alone a “belief” in porn addiction.

Bottom line: There is no such thing as “perceived compulsivity” or “perceived porn addiction.” The “Emotional Distress” questions have no place in the CPUI-9, except to skew scores higher for religious porn users and create unsupported conclusions and headlines.


Third main finding: Moral Disapproval was related to the “Emotional Distress” questions, but not to actual compulsivity or the CPUI-9 addiction questions (1-6)

Keep in mind that “moral disapproval of pornography” is the sum of 4 non-CPUI-9 questions, while the 3 CPUI-9 “Emotional Distress” questions assess guilt and shame. Fernandez et al., 2017 (and the other CPUI-9 studies) found that “moral disapproval of pornography” had little to do with actual porn addiction. The excerpt:

We found that when the CPUI-9 was taken as a whole, moral disapproval was the only significant predictor. However, when broken down, moral disapproval predicted only a specific domain of the CPUI-9, the Emotional Distress subscale (e.g., “I feel ashamed after viewing pornography online”) and had no influence on the Perceived Compulsivity subscale. This is consistent with previous research showing moral disapproval of pornography to be related only to the Emotional Distress subscale and not the Perceived Compulsivity or Access Efforts subscales (Wilt et al., 2016). This also lends support to Wilt and colleagues’ finding that moral disapproval accounts for a unique aspect of the CPUI-9, which is the emotional aspect (Emotional Distress), rather than the cognitive aspect (Perceived Compulsivity). Thus, although the Emotional Distress and Perceived Compulsivity subscales are related, our findings suggest that they need to be treated separately as they seem to be formed via different underlying psychological processes.

SUMMARY: Moral disapproval was strongly related to the 3 “Emotional Distress” questions, but only slightly related to the CPUI-9 “Perceived Compulsivity” questions. This means that “moral disapproval” is not related to porn addiction, but only to guilt & shame. Below are the correlations from the study cited in the excerpt (Wilt et al., 2016). Correlations between “moral disapproval” and the three CPUI-9 sections are highlighted:

As with the other CPUI-9 studies, believing porn is morally wrong or sinful correlated strongly with the CPUI-9 “Emotional Distress” section (#4). Yet there’s very little (or a negative) correlation between “moral disapproval” and the legitimate CPUI-9 porn addiction questions (“Access Efforts”, “Perceived Compulsivity”). Fernandez et al. says shame & guilt (questions 7-9) needs to be examined separately from actual porn addiction (questions 1-6). They do not assess addiction or “perceived” addiction.

Bottom line: The “Emotional Distress” questions have no place in the CPUI-9, except to skew scores higher for religious porn users. Researchers have exploited the natural correlation between “moral disapproval of porn” and the “Emotional Distress” questions to claim that moral objections causes the “belief in porn addiction” (Total CPUI-9 score). Since religious individuals score high on both “moral disapproval” and “Emotional Distress,” researchers incorrectly claim religion causes porn addiction, but study results supply little evidence that this is so.


Theoretical implications #1: “Perceived” pornography addiction is a myth. Moral disapproval plays no part in actual porn addiction.

Fernandez et al., 2017 found that the CPUI-9 “Perceived Compulsivity” questions assess actual compulsivity, and that moral disapproval plays no role in actual porn addiction.

Our findings have three important theoretical implications. First, the present study elucidates the previously unexplored relationship between perceived addiction to IP, as measured by the CPUI-9, and actual compulsivity. In our sample, we found that perceptions of compulsivity were indeed reflective of reality. It appears that an actual compulsive pattern (failed abstinence attempts x abstinence effort), and abstinence effort on its own, predict scores on the CPUI-9 Perceived Compulsivity subscale. We found that this relationship held even after holding moral disapproval constant. Thus, our findings suggest that regardless of whether an individual morally disapproves of pornography, the individual’s Perceived Compulsivity scores may be reflective of actual compulsivity, or the experience of difficulty in abstaining from IP. We propose that while actual compulsivity does not equate to actual addiction, compulsivity is a key component of addiction and its presence in an IP user might be an indication of actual addiction to IP. Therefore, the current study’s findings raise questions about whether research on the CPUI-9 to date can to some extent be accounted for by actual addiction, beyond mere perception of addiction.

SUMMARY: When Fernandez et al. says “perceptions of compulsivity” it means the CPUI-9 “Perceived Compulsivity” questions. Scores on “Perceived Compulsivity” aligned with actual compulsivity (failed abstinence attempts x abstinence effort). Put simply, CPUI-9 questions 1-3 assess actual compulsivity (cravings and inability to control use) rather than “belief in porn addiction.” The authors express serious reservations about using the phrase “perceived addiction” interchangeably with CPUI-9 test scores. Finally, assessing moral disapproval tells us nothing about actual porn addiction.

Next we use the data from another CPUI-9 paper co-authored by Grubbs (“Transgression as Addiction: Religiosity and Moral Disapproval as Predictors of Perceived Addiction to Pornography“), as its provocative title suggests that religious-based moral disapproval causes porn addiction.

Note that the “Emotional Distress” questions produce the strong correlations between “moral disapproval” and Total CPUI-9 scores. Note: “Access Efforts” questions 4-6 assess core addiction behaviors (inability to control use despite severe negative consequences), yet are largely unrelated to moral disapproval and religiosity.

Bottom line: There is no such thing as “perceived porn addiction.” If a porn user scores high on a valid porn addiction test, it means he is experiencing the signs and symptoms of an actual addiction. If you believe you are addicted, you are addicted. How one feels morally about pornography has virtually nothing to do with actual pornography addiction. To be accurate, spin laden phrases such as “perceived pornography addiction” or “belief in porn addiction” should more accurately be replaced with “pornography addiction.”


Theoretical implications #2: The 3 “Emotional Distress” questions inflate Total CPUI-9 scores for religious individuals while deflating Total CPUI-9 scores for actual porn addicts.

Fernandez et al., 2017 discuss how the 3 “Emotional Distress” questions skew all results from any study that employed the CPUI-9.

Second, our findings cast doubts on the suitability of the inclusion of the Emotional Distress subscale as part of the CPUI-9. As consistently found across multiple studies (e.g., Grubbs et al., 2015a,c), our findings also showed that frequency of IP use had no relationship with Emotional Distress scores. More importantly, actual compulsivity as conceptualized in the present study (failed abstinence attempts x abstinence effort) had no relationship with Emotional Distress scores. This suggests that individuals who experience actual compulsivity in their pornography use do not necessarily experience emotional distress associated with their pornography use.

Rather, Emotional Distress scores were significantly predicted by moral disapproval, in line with previous studies which also found a substantial overlap between the two (Grubbs et al., 2015a; Wilt et al., 2016). This indicates that emotional distress as measured by the CPUI-9 is accounted for mainly by dissonance felt due to engaging in a behavior that one morally disapproves of, and is unrelated to actual compulsivity. As such, the inclusion of the Emotional Distress subscale as part of the CPUI-9 might skew results in such a way that it inflates the total perceived addiction scores of IP users who morally disapprove of pornography, and deflates the total perceived addiction scores of IP users who have high Perceived Compulsivity scores, but low moral disapproval of pornography.

This may be because the Emotional Distress subscale was based on an original “Guilt” scale which was developed for use particularly with religious populations (Grubbs et al., 2010), and its utility with non-religious populations remains uncertain in light of subsequent findings related to this scale. “Clinically significant distress” is an important component in the diagnostic criteria proposed for Hypersexual Disorder for the DSM-5, where diagnostic criterion B states that “there is clinically significant personal distress … associated with the frequency and intensity of these sexual fantasies, urges, or behaviors” (Kafka 2010, p. 379). It is doubtful that the Emotional Distress subscale taps into this particular sort of clinically significant distress. The way the items are phrased (i.e., “I feel ashamed/depressed/sick after viewing pornography online”) suggests that distress need not be associated with the frequency and intensity of the sexual fantasies, urges, or behavior, but could be brought about merely from engaging in the behavior even in a non-compulsive way.

SUMMARY: This is the core finding: The 3 “Emotional Distress” questions have no place in the CPUI-9, or any porn addiction questionnaire. These guilt & shame questions do not assess distress surrounding addictive porn use or “perception of addiction.” These 3 questions artificially inflate Total CPUI-9 scores for religious individuals while deflating Total CPUI-9 scores for nonreligious porn addicts.

It’s important to note that assessment questionnaires for other types of addiction typically do not have questions about guilt & shame. Certainly, none make one third of their questionnaires about guilt and shame. For example, the DSM-5 criteria from Alcohol Use Disorder contain 11 questions. Yet none of the questions assess remorse or guilt after a drinking binge. Nor does the DSM-5 Gambling Addiction questionnaire contain a single question about remorse, guilt or shame.

Bottom line: Eliminate the 3 “Emotional Distress” questions and all the claims and correlations they were based upon disappear. Let’s examine how the  3 “Emotional Distress” questions skew CPUI-9 results.

Claim #1: First, it has been claimed over and over that “hours of porn use” were unrelated to “perceived porn addiction” (Total CPUI-9 scores). That’s not true as correlations taken from Grubbs’s “Transgression” study reveal:

In fact, hours of porn use is a stronger predictor of porn addiction (Total CPUI-9) than is religiosity. This alone debunks most of the headlines spawned by the the CPUI-9 “perceived addiction” studies.

While there’s still a correlation between religiosity and Total CPUI-9 scores, it’s largely produced by the 3 “Emotional Distress” questions. This data (taken from Grubbs’s “Transgression” study #2) reveals how the 3 “Emotional Distress” questions drastically lower correlations between hours of porn use and Total CPUI-9 scores:

As you can see actual porn addiction (as assessed by questions 1-6) is powerfully related to levels of porn use.

So, using Total CPUI-9 incorrectly leads to Claim #2: that being religious is strongly related to “perceived pornography addiction.” This correlation is reinterpreted as “religious people falsely believe they are addicted to porn.” Nether is true as actual porn addiction is, in fact, powerfully related to levels of porn use, and not related to religiosity. Comparing correlations between the CPUI-9 core addiction behaviors (“Access Efforts’) and Religiosity or Hours of porn use shows that religion has nothing to do with porn addiction:

The above correlation is the most important take away from this entire article: Religiosity has virtually nothing to do with actual porn addiction! Again, “Access Efforts” questions 4-6 assess core addiction behaviors (the inability to control despite severe negative consequences). In this section we provide four possible reasons why religious porn users may score higher on CPUI-9 actual addiction questions 1-6.

If religious subjects were more likely to “feel addicted” to porn, religiosity should correlate very strongly with actual porn addiction. It doesn’t. To say it another way, those subjects who are most addicted do not score higher in religiosity.


Theoretical implications #3: Actual compulsivity (failed abstinence attempts x abstinence effort) aligns with so-called “perceived compulsivity”

Fernandez et al., 2017 points out what is obvious to porn addicts: trying really hard to quit, yet continually failing, reveals the depth of your compulsion.

Third, this study introduced abstinence effort as an important variable in relation to understanding how perceptions of compulsivity might develop. It is noted that in the literature, frequency of IP use has been investigated without taking into account participants’ varying levels of abstinence effort. The present study’s findings demonstrate that abstinence effort on its own, and when interacting with failed abstinence attempts, predicts greater perceived compulsivity. We have discussed the experience of difficulty at abstaining or craving for pornography as a possible explanation of how the abstinence effort on its own may predict greater perceived compulsivity, in that the difficulty experienced may reveal to the individual that there may be compulsivity in their pornography use. However, at present, the exact mechanism by which abstinence effort relates to perceived compulsivity remains uncertain and is an avenue for further research.

SUMMARY: higher scores on the CPUI-9 “Perceived Compulsivity” were strongly related to features of actual compulsivity (needing greater effort to abstain from porn, yet being unable to do so). Put simply, so-called “perceived compulsivity” equates with actual compulsivity.

Bottom line: If you believe you are addicted to porn (because you are using it compulsively), you are addicted. All future studies should stop employing inaccurate and spin laden phrases such as “perceived pornography addiction” or “belief in porn addiction” as a proxy for CPUI-9 scores.

As an exercise in accuracy we remove the spin laden terms from a few “perceived addiction” studies, so the reader can understand the findings accurately:

Leonhardt et al., 2017 said:

“it appears that pornography users feel relationship anxiety surrounding their use only insofar as they believe themselves to have a compulsive, distressing pattern of use.”

Leonhardt et al., 2017 with accurate terminology:

Pornography addicts feel relationship anxiety surrounding their porn use.

Grubbs et al., 2015 said:

“These findings strongly underscore the claim that perceived addiction to Internet pornography likely contributes to the experience of  psychological distress for some individuals.”

Grubbs et al., 2015 with accurate terminology:

Addiction to Internet pornography is correlated with psychological distress.


Clinical implications #1:

Fernandez et al., 2017 suggests that clinicians can believe patients when they say they are addicted to pornography.

Finally, our findings provide important implications for the treatment of individuals who report being addicted to Internet pornography. There has been evidence in the literature to suggest that there have been an increasing number of individuals reporting being addicted to pornography (Cavaglion, 2008, 2009; Kalman, 2008; Mitchell, Becker-Blease, & Finkelhor, 2005; Mitchell & Wells, 2007). Clinicians working with individuals who report being addicted to pornography need to take these self-perceptions seriously, instead of being skeptical about the accuracy of these self-perceptions. Our findings suggest that if an individual perceives compulsivity in their own IP use, it is likely that these perceptions might be indeed reflective of reality.

In the same way, clinicians should realize that “perceived compulsivity” could be seen as a useful perception to have, if the perception is reflective of reality. Individuals who experience compulsivity in their IP use might benefit from gaining self-awareness that they are compulsive, and can use this insight into their own behavior to decide whether they need to take steps toward changing their behavior. Individuals who are unsure about whether their IP use is compulsive or not can subject themselves to a behavioral experiment such as the one employed in this study, with abstinence as the goal (for a 14-day period or otherwise). Such behavioral experiments might offer a useful way to ensure that perceptions are grounded in reality, through experiential learning.

SUMMARY: Since so-called “perceived compulsivity” equates with actual compulsivity in Fernandez et al., 2017, patients who claim to be addicted to porn, are in fact likely to be addicted to porn. If there is any doubt about the presence of actual addiction, clinicians should have the client try to abstain from porn for an extended period of time.

Bottom line: “Perceived addiction” doesn’t exist and its use should not be tolerated in scientific circles. Patients should be believed, regardless of the clinician’s personal bias or CPUI-9 score. Organizations such as AASECT, which has officially proclaimed that porn addiction doesn’t exist, may be causing harm to patients and the public.


Clinical implications #2:

From the Fernandez et al., 2017 discussion:

Importantly, our findings suggest that cognitive self-evaluations of compulsivity are likely to be accurate even if the individual morally disapproves of pornography. Clinicians should not be too quick to dismiss cognitive self-evaluations of individuals who morally disapprove of pornography as overly pathological interpretations due to their moralistic beliefs.

On the other hand, clinicians need to keep in mind that the emotional distress associated with pornography use experienced by clients, especially ones who morally disapprove of pornography, appears to be separate from the cognitive self-evaluation of compulsivity. Emotional distress, at least in the way it is measured by the CPUI-9, is not necessarily the result of compulsive IP use, and needs to be treated as a separate issue.

Conversely, clinicians need to also be aware that an individual could be experiencing actual compulsivity in their IP use without necessarily feeling emotions such as shame or depression associated with their IP use.

SUMMARY: First, clinicians should respect (even religious) patients’ self assessments when they feel addicted to porn in the absence of strong evidence to the contrary. Clinicians shouldn’t allow their own biases or a patient’s moral views to influence their assessments. Second, the “emotional distress” assessed by the three CPUI-9 guilt & shame questions have nothing to do with actual porn, or perceived, addiction. Clinicians are urged to avoid conflating actual or perceived porn addiction with guilt & shame – as the CPUI-9 studies have done.

Bottom line: Moral disapproval has nothing to with actual or perceived porn addiction. Claims that morality plays a role porn addiction arise from the CPUI-9’s use of inappropriate shame & guilt questions (“Emotional Distress”) to assess addiction. Clinicians harm patients by suggesting their porn-related difficulties arise from moral disapproval, shame or guilt when they in fact arise from actual compulsion.


SECTION 4: Final Thoughts

It is important to ponder how a flawed instrument like the CPUI-9 became elevated to such a position of influence in the sexology field and related articles in the mainstream. As Fernandez et al. shows, the CPUI-9 body of research is not solid science. Nor was the CPUI-9 ever validated as being able to distinguish real from “perceived” addiction. Yet the claims based on CPUI-9 findings have been enshrined as infallible, influential truths in some circles (whose preconceptions these claims appear to support).

What is really going on? As Fernandez et al. points out, the CPUI-9 appears to be aimed at producing claims about religious people – specifically, aimed at distorting “perceived addiction” results with respect to religious subjects and drawing far-reaching conclusions. Whether or not the teams employing the CPUI-9 intended this result, the “perceived addiction” claims have very effectively achieved this end, and it is not surprising that those who delight in such an outcome find the conclusions appealing and worthy of ongoing publicity.

The developer of the CPUI-9 is ex-religious, and it is not inconceivable that he has set out, consciously or unconsciously, to bring into disrepute strict religious upbringings like his own by means of his research. Some mainstream accounts, quoting him extensively, have gone even further, suggesting that his “perceived addiction” findings are evidence that any concern about porn use contributes to (or even generates) a belief in porn addiction. This unsupported assertion does a great disservice to porn users (whether religious or nonreligious) who are suffering from a wide range of severe symptoms, and trying to understand porn’s effects. Many of today’s nonreligious users have no shame whatsoever about their porn use, apart from their distress about their inability to control their porn use when they attempt to do so.

Sadly, few critics seem willing to examine the premises on which CPUI-9 study claims and mainstream interpretations are based. Instead, most psychologists and journalists take at face value assertions that scores on this highly distorted instrument are, in fact, evidence of shame-based “perceived addiction.” Yet upon even the slightest reflection, it becomes evident that no single score (and certainly not the score on a profoundly distorted questionnaire like the CPUI-9) could possibly reveal a distinction between “perceived” and actual addiction, let alone justify the far-reaching claims for which it is being cited.

All of this means that work such as Fernandez et al. is vital. Highly publicized claims like those about CPUI-9 data are unwarranted unless the validity of the instrument on which they rest is tested and results carefully evaluated for other, more plausible explanations. Thanks to Fernandez et al. it is now evident that, as a research instrument, the CPUI-9 is flawed and unreliable. As a responsible scientist and academic, its creator himself no doubt sees this.