Problematic pornography consumption attenuates approach tendencies for nudes

Problematic pornography consumption attenuates approach tendencies for nudes Sercan Kahveci

Supervisors: Bram van Bockstaele & Reinout Wiers Universiteit van Amsterdam, 2017

Abstract

Individuals displaying addictive behaviors like alcoholism and smoking were previously shown to have an automatic approach tendency to images of their addictive substance (approach bias). Approach-avoidance tasks were administered in which participants decided on approaching or avoiding based on whether the stimulus was an addiction cue, and another in which approach-avoid decisions were made based on addiction-irrelevant features. Individuals who use more pornography show a stronger approach bias for erotic stimuli. However, this relationship is attenuated by symptoms of problematic porn use. Despite its lower reliability, the irrelevant-feature approach/avoidance task relates to many more measures regarding sexuality and internet usage. It is speculated that self-initiated stopping attempts in problematic users may reduce approach bias.

Problematic pornography consumption attenuates approach tendencies for nudes Eighty-seven percent of young adult American men (Carroll et al., 2008) are pornography consumers (porn users). Many users report that it improves their sex life (Albright, 2008), and for years, the effects of porn on its users had not been investigated in detail. This has been changing in the last decade, as researchers have come to see that vulnerable porn users can grow dependent (Young, 2008). In the current study, it will be investigated whether porn use is driven by automatic implicit tendencies to approach addiction stimuli (approach biases), particularly in those who have difficulty controlling their use.

Compulsive use of porn may be a form of addiction. Compulsive porn users display tolerance for erotic stimuli: the striatum shows less activation, and event-related late positive potentials are less pronounced when compulsive users are presented with porn (Kühn & Gallinat, 2014; Prause, Steele, Staley, Sabatinelli, & Hajcak, 2015). Compulsive users experience more craving and anticipatory brain activity (Gola et al., 2017) but less liking when presented with porn (Voon et al., 2014), consistent with the incentive-sensitization theory of addiction (Robinson & Berridge, 1993). Additionally, frequent use in adolescents is correlated with sensation-seeking (Beyens, Vandenbosch, & Eggermont, 2015; Peter & Valkenburg, 2016) low self-control (Holt, Bossler, & May, 2012) and impulsivity (Carroll et al., 2008), traits previously linked to substance dependence (Wills, Vaccaro, & McNamara, 1994; Jentsch & Taylor, 1999). Like in substance dependency, compulsive porn users also display attentional biases for

addiction cues. While nudes catch the eye even in healthy people (Jiang, Costello, Fang, Huang, & He, 2006), individuals with compulsive sexual behaviors display an even stronger attentional bias for sexually explicit content (Mechelmans et al., 2014; Kagerer et al., 2014). The same attentional biases for addiction cues were also found in individuals dependent on alcohol, nicotine, heroin, and cannabis (reviewed in Field & Cox, 2008).

Those who attempt to cut back on porn face a challenge. Addiction cues increase craving (Cooney, Gillespie, Baker, & Kaplan, 1987; Carter & Tiffany, 1999). One cannot live their life completely avoiding porn cues like one can avoid illicit drugs like heroin. As such, those who try to stop are still exposed to erotic imagery daily through billboards, newspapers, online

thumbnails, spam, adverts and popular video games. It is crucial for those who are addicted, to not only avoid porn cues, but also to resist approaching porn when those cues are encountered by accident. This may prove difficult, as recent research has shown the existence of automatic

approach tendencies in individuals dependent on cannabis (Field, Eastwood, Bradley, & Mogg, 2006), tobacco (Bradley, Field, Mogg, & De Houwer, 2004), alcohol (Field, Mogg, & Bradley, 2005), and excessive videogaming (Rabinowitz & Nagar, 2015). The aim of the current study is to extrapolate these findings to porn users.

In approach-avoidance tasks (AAT) the participant is instructed to move pictures towards or away from themselves or a stick figure. Whether to approach or avoid depends on features of the picture. One variant is the relevant-feature AAT (rAAT), where these features are directly relevant to the measured bias. For example, a relevant-feature AAT will instruct porn users to approach erotic images during some blocks by pulling a joystick towards them, simulating an approach movement. During other blocks, they may be instructed explicitly to avoid erotic stimuli, pushing the joystick away when encountered. An approach bias is indicated when porn users are faster when they have to approach erotic images, than when they must avoid them.

Another variant is the irrelevant-feature AAT (iAAT), where participants approach or avoid pictures depending on a feature that is not related to the bias being measured. For instance, heavy porn users may be instructed to pull images featuring blonde-haired individuals, regardless if these images are pornographic or not; simultaneously they may be instructed to avoid images featuring brown-haired individuals, the “irrelevant feature” being hair color. It is assumed that approach biases will nevertheless affect reaction times to erotic versus non-erotic content, so porn-dependent individuals are faster to approach an erotic picture than a control picture, even if the approach or avoidance response depends on features like hair color, image background color, or image tilt.

A lot of variations exist of the AAT, yet we do not know the effects of a lot of these manipulations: earlier studies show that relevant- and irrelevant-feature AATs measure different forms of approach bias (Wiers, Gladwin, & Rinck, 2013). Additionally, researchers have been using arrow keys and joysticks to simulate pulling or pushing away the relevant stimulus, without investigating whether keys carry the same implicit meaning of pushing and pulling as moving the joystick does.

The main aim of the current study was to assess whether there is an approach bias to pornography-related cues, and whether this bias is related to frequency of porn use and

symptoms of problematic porn use. The effects of religiosity, porn craving and sexual excitation and inhibition were also explored, as were the reliability of the irfeature and

relevant-feature AAT. Half of the participants did the tasks using keyboard input and half with joystick input. It was hypothesized that there is approach bias to porn cues in both irrelevant-feature and relevant-feature approach/avoidance tasks and that this bias relates to frequency of porn use. It was also predicted that AAT measures would predict porn use and addiction symptoms over and above attitudes about porn, sexual excitability, sexual inhibition and relationship status. Lastly, it was predicted that the joystick-variant of the AAT would elicit larger approach or avoidance bias scores overall, and that it is a more reliable predictor of porn use.

Method Participants

62 adult heterosexual or bisexual males (Mage = 24.47) participated for course requirements or financial compensation at the University of Amsterdam. Data of one additional participant was lost due to a computer crash. A power analysis for simple regression with power of .80 and a medium effect size (f ² = .15), indicated that at least 55 participants would be needed to find a significant medium effect for the main hypothesis, Therefore, testing was continued until at least this number was reached.

Questionnaires

Gender, age and sexual orientation (Kinsey scale) were measured to check if inclusion criteria were met. Questionnaires regarding expectancy of porn use, computer use, hair color preference, religiosity, pornography craving and erectile function were included for exploratory and descriptive analysis. Questions per questionnaire were presented in random order.

The Problematic Pornography Use Inventory (PPUS: Kor et al., 2014) consists of 12 items with a six-point Likert scale. The PPUS has great internal consistency in this study, α = .92. The scale consists of four factors measuring continued use despite harm, excessive use, control difficulties and use for avoiding negative emotions. The scale has been shown to correlate with frequency of use and other measures of problematic porn use and psychopathology.

Erectile dysfunction was assessed using the International Index of Erectile Function (Rosen et al., 1997). The literature reports excellent internal consistency (α = .91) and construct validity. It consists of 15 items, answered on five-point and six-point scales.

The Sexual Inhibition and Sexual Excitation scales (Janssen, Vorst, Finn, & Bancroft, 2002) were used to measure sexual excitability and inhibition. The full questionnaire consists of

44 items answered on a four-point true-to-false scale. The three scales are Sexual Excitability (SES, α = .80), Sexual Inhibition due to performance concerns (SIS1, α = .76) and Sexual Inhibition due to consequences of sex (SIS2, α = .73). Besides performance concerns, SIS1 measures self-reported tendency to quickly lose one‟s erection when stimulation is absent.

The Pornography Craving Questionnaire (Kraus & Rosenberg, 2014) was used to measure cravings for porn both before and after exposure to erotic cues in the AAT. The scale consists of twelve items answered on a seven-point Likert scale and has good reliability in this study, α = .85. It correlates with measures of compulsive porn and internet use.

Participants rated their attitude about porn on a -100 to 100 visual analogue scale. They also indicated whether they expect positive and negative effects of porn use on a questionnaire loosely based on Wiers, van Woerden, Smulders and de Jong (2002). Positive expectations included reduction of anxiety and feeling ecstatic, whereas negative expectations included feeling dirty and estranged from oneself. Porn use frequency was measured by asking how many separate times and for was used on each day of the week preceding the study, as well as the duration of each of these sessions.

Materials

For the AAT, pictures of 40 blonde nude models, 40 brown-haired nude models, 40 blonde clothed models and 40 brown-haired clothed models were downloaded from the internet. Effort was made to match the pictures in all conditions on location, body position, and attractiveness. Images were cropped and resized to 1024 x 768. Additionally, hair color was sometimes artificially lightened or darkened to facilitate distinguishing between blonde and brown hair. Pictures that featured more than one individual, dark skin color, ambiguous hair color or lack of visible hair were excluded to facilitate the task and to ensure that other features in the image did not give away the correct response.

Stimuli were presented on an Asus VG236H 23” monitor. Input was provided through the Gigabyte Aivia Osmium keyboard or through the Logitech Attack 3 joystick. The experiment was programmed in Inquisit 5 (2016) and questionnaires were presented in Qualtrics software (2017).

Approach-avoidance task

During the approach-avoidance task, participants saw erotic and non-erotic pictures of women. They approached or avoided the stimulus by pressing the downward and upward arrow keys or by moving the joystick backward and forward. Input device was counterbalanced between subjects. Approach was accompanied by zooming into the picture, avoidance by zooming out. Participants received 8 blocks of 40 trials each, in fixed order. In block 1 they had to approach nude models and avoid clothed models, the opposite was required in block 2. In block 3, they had to approach blonde-haired models and avoid brown-haired models, the

opposite responses were required in block 4. The same order blocks was repeated for blocks 5 to 8. All pictures were shown once during the first half of the study and once during the second half. Pictures were presented in semi-random order with no more than 2 images of the same category appearing after each other.

Procedure

Upon arrival, participants signed an informed consent form and filled in the porn craving questionnaire. Subsequently they performed the approach-avoidance task. After task completion, they filled in the questionnaires as described in the Questionnaires section. Lastly participants were debriefed and rewarded for participating.

Analysis

Approach scores per picture type were calculated by subtracting the median of approach trials from the median of avoid trials. Porn approach bias was then calculated by subtracting approach scores for erotic pictures from bias for control pictures (as done by Cousijn, Goudriaan, & Wiers, 2011). This produced approach bias scores separately for the irrelevant-feature and relevant-feature trials, controlling for overall reaction time and for overall differences between approach and avoid responses. A craving change score was calculated by subtracting the participants‟ Porn Craving Questionnaire scores taken before the experiment from those taken after the experiment.

The presence of overall approach bias itself was pre-specified to be analyzed by comparing comparing bias scores to zero, separately for the relevant-feature and irrelevant feature AAT with one-sample t-tests. To see whether approach bias predicted porn use, a hierarchical regression was pre-specified where porn use was predicted first by relational status, attitude about porn, sexual excitability and inhibition scales, and in the second step with approach bias

scores for both tasks. To see whether the AAT with joystick was more sensitive to approach bias than the AAT with keys, an independent samples t-test was pre-specified comparing keys and joystick conditions on approach bias magnitude.

To see whether the AAT with joystick was more valid than AAT with, regression analysis was pre-specified with frequency and addiction symptoms as predictors, interacting with dummy variable input condition and predicting approach bias in both tasks. This analysis was not run due to lack of power. Finally, bias scores for the two tasks were pre-specified to be predicted by all collected questionnaire scores to explore differences between relevant-feature and irrelevant-feature AATs.

Results Confirmatory Results

Five participants were excluded from the analyses: one met exclusion criteria (more homosexual than heterosexual on the Kinsey scale), one had extremely slow reaction times, one had an excessive error rate, and two had outlying scores on the relevant-feature and irrelevant-feature AATs, respectively, which could skew regression findings. The latter four participants, all deviating more than 3 standard deviations from the sample mean on their specific variables, were not excluded for analyses that involved only questionnaire data.

There was no overall relevant-feature AAT approach bias, t (56) = 1.74, p = .088, and there was no overall irrelevant-feature AAT approach bias either, t (56) = 1.39, p = .170. Instead, many individuals showed an approach bias, but some did not, instead avoiding erotic stimuli.

A hierarchical regression analysis was conducted where sexual excitability (SES), sexual inhibition due to performance threat (SIS1), sexual inhibition due to performance consequences (SIS2), attitudes on porn and relational status were used to predict weekly minutes spent on porn in step 1, and relevant- and irrelevant-feature approach bias scores were added in step 2. As can be seen in Table 1, no predictors were significant in step 1. In step 2, AAT scores did not improve prediction. Both models were not significant.

Table 1. Hypothesis-driven regression analysis predicting minutes of porn use per week.

B SE β t p

Step 1 R² = .14, F (5, 51) = 1.64, p = .166

Sexual excitability 2.62 1.35 .27 1.94 .058

Sex. Inhibition 1 (performance threat) 2.93 1.55 .25 1.89 .064 Sex. inhibition 2 (performance consequences) .35 1.77 .03 .20 .844

Porn Attitude -.27 .28 -.13 .95 .349

Relational status 16.16 18.14 .13 .89 .377

Step 2 ΔR² = 0, F (2, 49) = .08, p = .925

Relevant-feature AAT bias .06 .14 .05 .39 .700

Irrelevant-feature AAT bias .01 .21 .01 .03 .973

The same hierarchical regression was repeated with Problematic Porn Use Inventory scores as dependent variable. One participant with an outlying residual (+3.5 SDs) after Step 1 was excluded before initiating analysis. Exclusion of this participant made the effect of irrelevant-feature AAT bias on problematic porn use non-significant. As can be seen in Table 2, SES, as well as SIS1 scores, representing rapid loss of arousal and sexual performance concerns, significantly predicted problematic use in step 1. Relevant- and irrelevant-feature AAT bias scores did not help with prediction in step 2. Both models were not significant.

Table 2. Hypothesis-driven regression analysis predicting symptoms of problematic porn use (PPUS).

B SE β t p

Step 1 ΔR² = .38, F (5, 50) = 6.20, p < .001

Sexual excitability .42 .17 .29 2.44 .018

Sex. Inhibition 1 (performance threat) .88 .20 .49 4.38 .000 Sex. inhibition 2 (performance consequences) .37 .23 .18 1.62 .112

Porn Attitude 1.68 2.30 .09 .73 .468

Relational Status -.04 .04 -.13 1.04 .303

Step 2 ΔR² = .01, F(2, 48) = .48, p = .624

Relevant-feature AAT bias .01 .02 .050 .42 .674

Irrelevant-feature AAT bias -.03 .03 -.112 .92 .362

Keyboard versus joystick. Independent samples t-tests failed to confirm that the

magnitude of approach bias recorded with the joystick-based relevant-feature AAT was different from that recorded with the keyboard-based relevant-feature AAT, t (55) = .89, p = .380, and no magnitude difference could be indicated with the two versions of the irrelevant-feature AAT either, t (55) = 1.20, p = .236. Analyses regarding differences in validity of acquired AAT bias scores were not conducted, because the current study did not have enough power to demonstrate

interaction effects. On top of that, participants in keyboard and joystick conditions had different rates of porn use, confounding any potential analyses of validity. Reliability of input methods is explored in the next section.

Exploratory Results

A different operationalization of porn use frequency. A regression analysis was run with the same predictors as in Table 1, but predicting number of porn-viewing sessions per week, rather than minutes per week. While nothing predicted minutes of porn use per week, weekly

sessions of porn use was predicted by sexual inhibition due to performance concerns in step 1,

and by relevant-feature AAT bias score in step 2, as can be seen in Table 3. Step 2 significantly improved on prediction over step 1. For the rest of the exploratory section, sessions per week rather than minutes per week will be used as the variable representing frequency of porn use. Table 3. Predicting number of porn-viewing sessions per week

with exploratory hierarchical regression

B SE β t p Step 1 R² = .15, F (5, 50) = 1.74, p = .143 Sexual excitability .05 .06 .12 .86 .393 Sex. Inhibition 1 .15 .07 .29 2.18 .034 Sex. Inhibition 2 .08 .08 .14 1.04 .303 Porn Attitude -.26 .79 -.05 -.33 .741 Relational status .01 .01 .16 1.13 .265 Step 2 ΔR² = .10, F (2, 48) = 3.28, p = .046 Relevant-feature AAT bias .01 .01 .28 2.12 .040 Irrelevant-feature AAT bias .01 .01 .16 1.20 .237

Exploration of questionnaire data. Correlations between questionnaire data are displayed in Table 4. The number of self-reported porn-viewing sessions per week is correlated with

problematic porn use, sexual inhibition due performance threat, minutes per week spent viewing porn, lower erectile functioning, and positive expectancy of porn use. Porn-viewing sessions per week is not correlated with the self-reported increase in arousal due to exposure to erotic images in the AAT, sexual excitability, overall use of the internet, or attitude towards porn. The amount of porn viewed weekly seemed to be more strongly linked to variables indicating dysfunctions (potentially) caused by porn, than to variables related to sexual appetite, a liking for porn, or general internet use.

Table 4. Correlation matrix for questionnaire data. 1 2 3 4 5 6 7 8 9 10 1. PPUS - 2. Craving gain -.04 - 3. SES .16 .08 - 4. SIS1 * .53 .03 .01 - 5. SIS2 .08 .01 .06 .05 -

6. Sessions per week * .56 .02 .14 * .29 .10 -

7. Internet use .10 .08 -.04 -.03 .05 .10 -

8. Porn attitude .09 * .33 * .33 -.11 .05 .19 .09 - 9. Erectile functioning *-.37 .15 -.07 -.24 -.18 *-.36 .08 .11 -

10. Pos. expectancy .22 .22 * .46 -.02 .14 * .26 -.04 * .48 .12 - 11. Neg. expectancy * .44 .04 .02 * .48 -.02 .16 -.07 -.19 -.06 .08 * indicates p < .05. The erectile functioning score was derived from the IIEF, excluding questions relating to sex.

Positive Expectancy and Negative Expectancy scores were derived from the porn expectancy questionnaire

using principal component analysis with oblimin rotation. PPUS = Problematic Porn Use Scale; SES = Sexual Excitability Scale; SIS1 = Sexual Inhibition due to performance concerns; SIS2 = Sexual Inhibition due to performance consequences.

Are the bias scores reliable? To compute the reliability of the tasks, bias scores were calculated separately for even and odd trials, and correlated after exclusion of outliers. There was no relationship between the irrelevant approach bias score calculated from even trials and the same score from odd trials, r = -.11, p = .402, but a medium correlation could be demonstrated for the relevant-feature approach bias, r = .40, p = .002. The relevant-feature AAT was thus more reliable than the irrelevant-feature AAT, z = 2.85, p = .004. The relevant-feature AAT reliability did not differ for joystick, r = .46, p = .012, and keyboard, r = .28, p = .149, comparison: z = .76,

p = .446. The same was true for the irrelevant-feature AAT; joystick: r = .05, p = .780, keyboard: r = -.35, p = .069, comparison: z = 1.51, p = .131. The fact that a marginally significant negative

correlation could emerge from a test of reliability is somewhat alarming, however. Approach bias scores for the relevant-feature and irrelevant-feature AAT did not correlate, r = .15, p = .259. These two tasks may measure different mechanisms, or simply correlate less due to their low reliability.

What predicts approach bias? To understand the differences of what is measured by relevant-feature AAT and irrelevant-feature AAT, regression analyses were run predicting relevant-feature AAT and irrelevant-feature AAT bias from craving change score, SES, SIS1, porn-viewing sessions per week, daily internet use in hours, relational status, attitude towards porn, positive porn expectancy, negative porn expectancy, and problematic porn use. Religiosity

was not used due to very little variance in the sample, perhaps due to self-selection. Sessions per week, rather than minutes per week, was used as predictor because it has a much stronger relationship with approach bias scores. The outlier with extreme residuals in the confirmatory analyses was also excluded. Results are displayed in Table 5.

Relevant-feature AAT bias was only predicted by the number of porn viewing sessions per week. Irrelevant-feature AAT bias, however, was predicted by sexual excitability, sexual

inhibition due to performance threat, porn viewing sessions per week, daily hours of internet use, and negatively by problematic porn use. Symptoms of problematic porn use attenuate irrelevant-feature approach bias. Irrelevant-irrelevant-feature AAT bias was more related to explicit behavioral measures than relevant-feature AAT bias (34% versus 5% of explained variance).

Table 5. Exploratory regression analysis predicting bias scores from questionnaire data.

Relevant-feature AAT Irrelevant-feature AAT

B SE β t p B SE β t p Problematic use -1.25 1.30 -.19 -.96 .344 -1.78 .81 -.41 -2.21 .032 Weekly sessions 7.89 3.58 .35 2.20 .033 4.91 2.22 .33 2.21 .032 SES 1.99 1.48 .21 1.34 .186 2.50 .92 .40 2.73 .009 SIS1 .41 1.92 .04 .21 .833 2.63 1.18 .34 2.22 .031 Internet use 4.54 3.41 .19 1.33 .189 5.35 2.11 .33 2.54 .015 Relational status 1.01 18.35 .01 .06 .956 17.27 11.34 .21 1.52 .135 Porn attitude -.44 .33 -.23 -1.34 .188 -.07 .20 -.06 -.35 .729 + expectancy 10.36 10.43 .17 .99 .326 -10.49 6.45 -.25 -1.63 .111 - expectancy -10.40 10.44 -.16 -1.00 .325 6.03 6.46 .14 .93 .355 Craving gain -.27 .76 -.05 -.36 .722 -.13 .47 -.04 -.27 .786 F (10, 45) = 1.29, p = .266, R² = .05 F (10, 45) = 2.32, p = .27, R² = .34

Could there be approach bias to one’s preferred hair color? Relevant-feature AAT and irrelevant-feature AAT approach bias scores were calculated for images of blonde and brunette women in men who prefer either blonde or brunette women. Three participants were excluded with extreme scores on either the relevant- or irrelevant-feature AATs for hair-color preference (±3SD). Blonde or brown hair-preferring men did not differ in their relative approach bias for blonde or brown hair on the relevant-feature AAT, t (57) = .61, p = .542, or irrelevant-feature AATs, t (57) = 1.65, p = .105. The same analysis was repeated for erotic images only, with two ouliers excluded (±3SD). Brown hair-preferring men had larger approach bias for blonde women than blonde-hair preferring men in the relevant-feature AAT, t (58) = 2.22, p = .030; no

Discussion

The current study was initiated to investigate whether automatic approach tendencies to erotic content can be discerned, and whether they are related to one‟s degree of porn use and symptoms of problematic porn use. Participants who use more porn were shown to have stronger approach bias. This relationship is complicated by the fact that symptoms of problematic porn use attenuate approach bias on the irrelevant-feature AAT, contrary to predictions. Relevant- and irrelevant-feature AATs did not predict (problematic) porn use over and above explicit

questionnaire scores. The irfeature AAT scores were less reliable than the relevant-feature AAT scores, but nevertheless were more strongly related to measures of sexual behavior.

Individuals without problematic porn use are likely driven to use porn by their implicit desire to approach it, while the situation is more complex for individuals displaying symptoms of problematic porn use as well. After all, symptoms of problematic porn use, which tend to

correlate with porn use, predict a reduction in approach bias rather than an increase. Schiebener, Laier and Brand (2015) found that individuals with problematic porn use showed either approach or avoidance of porn using a multitasking paradigm. Some participants with problematic porn use may show an avoidance bias to porn because they have trained themselves to avoid porn, or have developed an aversion to porn. Even though trained avoidance bias was shown to reduce drinking in problematic drinkers (Wiers, Rinck, Kordts, Houben, & Strack, 2010), it remains to be shown whether training an avoidance bias actually helps to reduce porn use. On the other hand, some problematic porn users may have become desensitized to the relatively mild erotic still images used in this task, or conversely have come to see regular images of women as pornographic cues. Lastly, research within the attentional bias realm (Albery et al., 2017) suggests that years of sexual activity may have a protective effect against development of (attentional) biases in porn users, allowing sexually experienced men to use porn without developing implicit biases.

Are these findings consistent with what is found in other addictions? A direct comparison cannot be made, as most studies compare clinical populations with controls, failing to dissociate the amount of use from the degree of self-perceived maladaptivity. However, the correlation between amount of use and approach bias is also found for casual cannabis users (Cousijn, Goudriaan, & Wiers, 2011). For this reason, it is critical to investigate what causes some heavy

porn or drug users to report low levels of problematic use, whereas others report more problems. For porn users, one such factor may be erectile dysfunction.

Participants who used more porn showed more signs of erectile dysfunction and sexual inhibition due to performance threat, consistent with recent reports stating that porn use can cause erectile dysfunction (Porto, 2016; Park et al., 2016). Recent epidemiological research did not find this link (Landripet & Štulhofer, 2015), although a link was found between internet sex addiction symptoms and erectile dysfunction elsewhere (Wéry & Billieux, 2016). The current study used more accurate methods to measure both erectile functioning and amount of porn use, by using a full questionnaire to measure erectile functioning and by requesting participants to report the number of porn-viewing sessions for every day separately, rather than requesting a single estimate from the participant. This may have revealed a relationship that was lost in measurement inaccuracy in the aforementioned epidemiological study. Additionally, the sample used here was much more restricted than the one used by Landripet and Štulhofer: the

relationship between erectile dysfunction and porn use is likely to be much stronger for young and healthy students than for men up to age forty of variable health and occupation, in whom other factors are also likely to cause erectile dysfunction. Physiological research is needed to confirm that large amounts of porn use do indeed translate to measurable erectile dysfunction, and not just subjective experience.

Perhaps the greatest limitation of the study is the low reliability of the irrelevant-feature approach-avoidance task. Although this is undesirable, it is not a surprising occurrence: implicit tasks were previously found to lack reliability but have predictive power nonetheless (Van Bockstaele et al., 2011). Future research within the realm of approach bias should aim to develop more reliable implicit measures of approach bias, to ensure that these paradigms can eventually be used in clinical practice.

Additionally, no ratings of attractiveness or distinguishability of hair color were collected for the used stimuli. This makes it impossible to verify whether erotic and non-erotic stimuli were successfully matched on attractiveness and whether hair color was easily distinguishable.

Current results also should not yet be generalized to clinical practice, due to the relatively low number of participants using porn in a manner that can be considered problematic.

results regarding erectile dysfunction will only generalize for males of the same age bracket as in this study.

Future research should be directed towards determinants of porn use and the role approach bias plays in porn use. Attempted reduction of porn use may mediate the relationship between self-reported problematic porn use and irrelevant-feature AAT porn avoidance bias. It is worth investigating if self-reported attempts at abstinence cause an avoidance bias towards porn, and if the presence or emergence of avoidance bias predicts successful abstinence. Approach bias modification could jump-start successful reduction of porn use in individuals who actually have a bias, leading to a computer-based treatment for an addiction that predominantly unfolds online behind a computer.

References

Albery, I.P., Lowry, J., Frings, D., Johnson, H.L., Hogan, C., & Moss, A.C. (2017). Exploring the relationship between sexual compulsivity and attentional bias to sex-related words in a cohort of sexually active individuals. European Addiction Research, 23, 1-6.

https://doi.org/10.1159/000448732

Albright, J. M. (2008). Sex in america online: An exploration of sex, marital status, and sexual identity in internet sex seeking and its impacts. The Journal of Sex Research, 45, 175-186. https://doi.org/10.1080/00224490801987481

Beyens, I., Vandenbosch, L., & Eggermont, S. (2015). Early adolescent boys‟ exposure to Internet pornography: Relationships to pubertal timing, sensation seeking, and academic performance. Journal of Early Adolescence, 35, 1045–1068.

https://doi.org/10.1177/0272431614548069

Bradley, B.P., Field, M., Mogg, K., & De Houwer, J. (2004). Attentional and evaluative biases for smoking cues in nicotine dependence: component processes of biases in visual orienting. Behavioural Pharmacology, 15, 29–36

Carroll, J. S., Padilla-Walker, L. M., Nelson, L. J., Olson, C. D., Barry, C. M., & Madsen, S. D. (2008). Generation XXX: Pornography acceptance and use among emerging adults.

Journal of Adolescent Research, 23, 6-30. https://doi.org/10.1177/0743558407306348

Carter, B. L., & Tiffany, S. T. (1999). Meta-analysis of cue-reactivity in addiction research.

Addiction, 94, 327-340.

Cooney, N. L., Gillespie, R. A., Baker, L. H., & Kaplan, R. F. (1987). Cognitive changes after alcohol cue exposure. Journal of Consulting and Clinical Psychology, 55, 150-155. Cousijn, J., Goudriaan, A.E., & Wiers, R.W. (2011). Reaching out towards cannabis:

approach-bias in heavy cannabis users predicts changes in cannabis use. Addiction, 106, 1667-1674. https://doi.org/10.1111/j.1360-0443.2011.03475.x

Field, M., & Cox, W.M. (2008). Attentional bias in addictive behaviors: a review of its development, causes, and consequences. Drug and Alcohol Dependence, 97(1-2), 1-20. https://doi.org/10.1016/j.drugalcdep.2008.03.030

Field, M., Eastwood, B., Bradley, B. P., & Mogg, K. (2006). Selective processing of cannabis cues in regular cannabis users. Drug and Alcohol Dependence, 85, 75-82.

Field, M., Mogg, K., Bradley, B.P. (2005). Craving and cognitive biases for alcohol cues in social drinkers. Alcohol and Alcoholism, 40, 504–510.

https://doi.org/10.1093/alcalc/agh213

Gola, M., Wordecha, M., Sescousse, G., Lew-Starowicz, M., Kossowski, B., Wypych, M., . . . Marchewka, A. (2017). Can pornography be addictive? An fMRI study of men seeking treatment for problematic pornography use. Neuropsychopharmacology,

https://doi.org/10.1038/npp.2017.78

Holt, T. J., Bossler, A. M., & May, D. C. (2012). Low self-control, deviant peer associations, and juvenile cyberdeviance. American Journal of Criminal Justice, 37, 378–395.

https://doi.org/10.1007/s12103-011-9117-3

Inquisit 5 [Computer software]. (2016). Retrieved from http://www.millisecond.com.

Janssen, E., Vorst, H., Finn, P., & Bancroft, J. (2002). The sexual inhibition (SIS) and sexual excitation (SES) scales: I. Measuring sexual inhibition and excitation proneness in men.

The Journal of Sex Research, 39, 114-126.

Jentsch, J. D., & Taylor, J. R. (1999). Impulsivity resulting from frontostriatal dysfunction in drug abuse: Implications for the control of behavior by reward-related stimuli.

Psychopharmacology, 146, 373-390.

Jiang, Y., Costello, P., Fang, F., Huang, M., & He, S. (2006). A gender- and sexual orientation-dependent spatial attentional effect of invisible images. Proceedings of the National

Academy of Sciences of the United States of America, 103, 17048-17052.

https://doi.org/10.1073/pnas.0605678103

Kagerer, S., Wehrum, S., Klucken, T., Walter, B., Vaitl, D., & Stark, R. (2014). Sex attracts: Investigating individual differences in attentional bias to sexual stimuli. PLoS ONE, 9(9) https://doi.org/10.1371/journal.pone.0107795

Kor, A., Zilcha-Mano, S., Fogel, Y.A., Mikulincer, M., Reid, R.C., & Potenza, M.N. (2014). Psychometric development of the Problematic Pornography Use Scale. Addictive

Behaviors, 39, 861-868. http:// doi.org/10.1016/j.addbeh.2014.01.027

Kraus, S. & Rosenberg, H. (2014). The Pornography Craving Questionnaire: Psychometric Properties. Archives of Sexual Behavior, 43, 451-462. https://doi.org/10.1007/s10508-013-0229-3

Kühn, S., & Gallinat, J. (2014). Brain structure and functional connectivity associated with pornography consumption: the brain on porn. JAMA Psychiatry, 71, 827–34.

https://doi.org/10.1001/jamapsychiatry.2014.93

Landripet, I., & Štulhofer, A. (2015). Is Pornography Use Associated with Sexual Difficulties and Dysfunctions among Younger Heterosexual Men? Journal of Sexual Medicine, 12, 1136-1139. https://doi.org/10.1111/jsm.12853

Mechelmans, D.J., Irvine, M., Banca, P., Porter, L., Mitchell, S., Mole, T.B. et al. (2014). Enhanced attentional bias towards sexually explicit cues in individuals with and without compulsive sexual behaviours. PloS One, 9, e105476.

https://doi.org/10.1371/journal.pone.0105476

Park, B.Y., Wilson, G., Berger, J., Christman, M., Reina, B., Bishop, F., Klam, W.P., & Doan, A.P. (2016). Is Internet Pornography Causing Sexual Dysfunctions? A Review with Clinical Reports. Behavioral Sciences, 6(3), 17. https://doi.org/10.3390/bs6030017 Peter, J., & Valkenburg, P. M. (2016). Adolescents and pornography: A review of 20 years of

research. Journal of Sex Research, 53, 509-531. https://doi.org/10.1080/00224499.2016.1143441

Porto, R. (2016). Male masturbation habits and sexual dysfunctions. [Habitudes masturbatoires et dysfonctions sexuelles masculines] Sexologies, 25(4), 160-165.

https://doi.org/10.1016/j.sexol.2016.03.004

Prause, N., Steele, V.R., Staley, C., Sabatinelli, D., Hajcak, G. (2015). Modulation of late positive potentials by sexual images in problem users and controls inconsistent with „porn addiction‟. Biological Psychology, 109, 192–199.

https://doi.org/10.1016/j.biopsycho.2015.06.005.

Qualtrics 3.7 [computer software] (2017). Retrieved from www.qualtrics.com. Rabinovitz, S., & Nagar, M. (2015). Possible end to an endless quest? Cognitive bias

modification for excessive multiplayer online gamers. Cyberpsychology, Behavior, and

Social Networking, 18, 581-587. https://doi.org/10.1089/cyber.2015.0173

Robinson, T.E., & Berridge, K.C. (1993). The neural basis of drug craving: an incentive-sensitization theory of addiction. Brain Research Reviews, 18, 247-291.

Rosen, R.C., Riley, A., Wagner, G., Osterloh, I.H., Kirkpatrick, J., & Mishra, A. (1997). The international index of erectile function (IIEF): a multidimensional scale for assessment of erectile dysfunction. Urology, 49, 822-830.

Schiebener, J., Laier, C., & Brand, M. (2015). Getting stuck with pornography? overuse or neglect of cybersex cues in a multitasking situation is related to symptoms of cybersex addiction. Journal of Behavioral Addictions, 4, 14-21.

https://doi.org/10.1556/JBA.4.2015.1.5

Van Bockstaele, B., Verschuere, B., Koster, E.H.W., Tibboel, H., De Houwer, J., & Crombez, G. (2011). Differential predictive power of self report and implicit measures on behavioural and physiological fear responses to spiders. International Journal of Psychophysiology, 79, 166-174. https://doi.org/10.1016/j.ijpsycho.2010.10.003

Voon, V., Mole, T. B., Banca, P., Porter, L., Morris, L., Mitchell, S., Lapa, T. R., Karr, J., Harrison, N. A., Potenza, M. N., & Irvine, M. (2014). Neural correlates of sexual cue reactivity in individuals with and without compulsive sexual behaviours. PloS One, 9(7), e102419. https://doi.org/10.1371/journal.pone.0102419

Wéry, A., & Billieux, J. (2016). Online sexual activities: An exploratory study of problematic and non-problematic usage patterns in a sample of men. Computers in Human Behavior,

56, 257-266. https://doi.org/10.1016/j.chb.2015.11.046

Wiers, R. W., Rinck, M., Kordts, R., Houben, K., & Strack, F. (2010). Retraining automatic action-tendencies to approach alcohol in hazardous drinkers. Addiction, 105, 279-287. https://doi.org/10.1111/j.1360-0443.2009.02775.x

Wiers, R.W., Gladwin, T.E., & Rinck, M. (2013). Should we train alcohol-dependent patients to avoid alcohol? Frontiers in Psychiatry, 22. https://doi.org/10.3389/fpsyt.2013.00033 Wiers, R.W., van Woerden, N., Smulders, F.T.Y., & de Jong, P.J. (2002). Implicit and Explicit

Alcohol-Related Cognitions in Heavy and Light Drinkers. Journal of Abnormal

Psychology, 111, 648-658. https://doi.org/10.1037//0021-843X.111.4.648

Wills, T. A., Vaccaro, D., & McNamara, G. (1994). Novelty seeking, risk taking, and related constructs as predictors of adolescent substance use: An application of cloninger's theory.

Journal of Substance Abuse, 6, 1-20. https://doi.org/10.1016/S0899-3289(94)90039-6

Young, K.S. (2008). Internet sex addiction: Risk factors, stages of development, and treatment.