Role of attribution in affective responses to a partnered sexual situation among sexually dysfunctional men


David L. Rowland, Valparaiso University, 1700 Chapel Drive, Valparaiso, IN 46383, USA. e-mail:


Study Type – Therapy (qualitative)

Level of Evidence 4

What's known on the subject? and What does the study add?

Sexually dysfunctional men report higher negative affect and lower positive affect than sexually functional men. Furthermore, men with sexual problems tend to make internal, self-blaming, attributions for negative sexual events, which can result in a diminished sense of self-efficacy and cause men to expect similar negative outcomes across future sexual situations. This pattern may sustain and actually intensify the sexual problem.

This study shows that causal attributions for dysfunctional response influence emotional response to a partnered sexual situation. Specifically, sexually dysfunctional men who attribute their problem to a medical condition do better emotionally than those who attribute the problem to unknown or psychological factors – and this attribution process is a more powerful and reliable predictor of emotional response than the man's actual diagnosed somatic risk for a sexual problem. Furthermore, the study strongly suggests that those men who attribute their problem to a physical/medical issue are able to ‘externalize’ it – needing to assume less responsibility for and/or control over it, thereby reducing their psychological burden. These findings indicate not only that communication between physician and patient regarding the aetiology of a sexual problem may be critically important but also that, at least for some patients, pharmacological treatment should be combined with strategies that promote the development of internal positive self-serving attributions surrounding sexual activity.


  • • To investigate factors that influence sexually dysfunctional men's emotional response within a partnered sexual situation, and, most specifically, whether they attribute their dysfunctional response to a specific biomedical cause vs a psychological or unknown cause.


  • • Based on a sample of 59 sexually dysfunctional men visiting a urology clinic, linear regression was used to determine the relationship between patients' attributions and five global affective factors derived through principal components analysis: apprehension, insecure, arousable, affection, and pleasant.
  • • Two other covariates were included: actual (diagnosed) somatic risk as determined by the patient's medical history and clinician's notes, and the patient's self-reported importance of sexual intimacy.


  • • Attribution (biomedical vs psychological/unknown) had significant effects on three psycho-affective factors (insecure, arousable, affection); men who attributed their problem to a biomedical cause had higher positive affect and lower insecurity.
  • • Diagnosed somatic risk was significant for insecure and marginally for affection.
  • • Importance of sex was significant on four psycho-affective variables (insecure, arousable, pleasant, affection), with higher endorsement associated with higher insecurity and lower positive affect.


  • • Men who can attribute their sexual dysfunction to a medical condition do much better emotionally (higher positive affect and lower negative affect) than those who attribute the problem to unknown or psychological factors – and this factor is more predictive of their emotionality than the patient's actual somatic risk for sexual dysfunction.
  • • As demonstrated in a previous analysis, higher ratings of the importance of sex independently predicted stronger negative affect for men experiencing sexual problems.


Psycho-affective responses to sexual situations constitute one of the most notable and reliable psychological differences between sexually dysfunctional and functional men [1–4]. Not surprisingly, dysfunctional men tend to report higher negative affect and lower positive affect than functional men. However, even within samples of sexually dysfunctional men, significant variation in emotional response occurs: some dysfunctional men show high levels of distress and negative affect while others show little or none [4,5]. Some such variation can be accounted for by the importance the patient ascribes to his sex life, in fact more so this factor than the severity of the dysfunction itself [6].

Men's emotional experience, along with sexual response itself, is clearly affected by the cognitive framework and expectations they bring to a sexual situation. Specifically, how they use the information about the situation to understand their actions and attribute cause to them affects their interpretation of and response to the sexual situation [7–9].

Broadly speaking, studies on cognitive attribution of cause have found that people tend to make internal-self attributions for personal positive behaviours or outcomes, and external-other attributions for negative behaviours or outcomes [10]. Such attribution processes might apply to sexual situations as well: for example, externalized, self-serving attributions may be adaptive when dealing with failed sexual experiences such as premature ejaculation, erectile dysfunction or lack of orgasm, as they would help ensure positive feelings toward future sexual experiences. In contrast, internal, self-serving attributions may be more typical for successful sexual experiences.

Yet, the attribution process surrounding successful (or unsuccessful) sexual responses, first, appears to be quite different for sexually functional vs sexually dysfunctional men and women, and second, is not fully consistent with general predictions made by attribution theory. Men with erectile disorder, for example, tend to make internal (self-blaming) attributions for negative sexual events, whereas men without erectile disorder do not [11]; women with low orgasm rates show a similar pattern, adopting a self-handicapping interpretation of placing blame on themselves for their lack of orgasm and giving credit for orgasm to an outside source such as a partner [12]. As noted, such patterns contrast with attribution theory expectation: one might anticipate that sexually dysfunctional individuals would not blame themselves but rather make a self-serving attribution by blaming their partner or outside circumstances [9,13].

Internal attribution for the cause of sexual inadequacy has the potential to precipitate a cycle of failure and create a self-fulfilling prophesy, thus making attribution processes relevant to understanding men with sexual problems. Men who attribute their failure to internal causes (i.e. to something about themselves rather than to the partner or situation) typically report a diminished sense of self-efficacy, and as such come to expect similar negative outcomes across future sexual situations. This predisposition may become a major factor in maintaining or intensifying the sexual problem [14–18].

The above studies suggest that men with sexual problems are more likely to assume an internal attribution (i.e. ‘something is wrong with me’), assigning blame (and by implication ‘control’ over the problem) to themselves, thereby setting up a negative orientation regarding future sexual situations. Indeed, men with sexual problems are known to approach sexual situations more negatively than their functional counterparts, giving support to this assumption [19]. However, not all internal attributions are the same. For example, internal attributions – ones where men feel both personally responsible and presumably in control – are likely to be different from those that, although internal (i.e. directed toward the self), absolve men from assumptions regarding personal responsibility and control.

In this study, we were interested in investigating the relationship between the psycho-affective response of men with a sexual dysfunction and their attributions/assumptions about the cause of their dysfunctional sexual response. Specifically, we hypothesized that men who attributed their sexual failure to a physical/medical condition (although an internal attribution, not one for which the man must assume responsibility and control) would respond more positively emotionally than men who attributed their problem to a psychological/relationship problem or to an unknown cause (internal attributions over which an individual might assume greater responsibility and control). Our analysis included two additional factors: (i) an index of clinician-diagnosed ‘somatic risk’, a variable that assessed the likelihood that the man's sexual problem was actually due to a physical/medical problem; and (ii) the patient's self-reported ‘importance of sex’, a factor which, when strongly endorsed, has been shown to exacerbate dysfunctional men's negative emotional response within the sexual situation [6].



Study participants were 59 heterosexual men (mean age 56.8 years, sd (standard deviation) 10.6) visiting the James Buchanan Brady Urological Institute at the Johns Hopkins Medical Institutions for consultation and possible treatment of a sexual problem. Eighteen potential participants were eliminated from a larger, initial pool due to incomplete data, language barriers, referral to a different clinic, or self-exclusion due to lack of partner or unwillingness to attempt intercourse with the partner. All final participants had been in a sexual relationship with their partner for at least 6 months prior to participation in the study.


Upon scheduling an appointment with the clinic, participants were sent a packet of forms to complete prior to the office consultation. The packet gathered information about (i) demographics and educational attainment; (ii) general medical history; (iii) self-assessment of sexual functioning; (iv) patient self-efficacy regarding sexual functioning; and (v) affective response to a specific ‘partnered’ sexual situation. In addition, insurance information and detailed informed consent as approved by the Medical Ethics Committee of the Johns Hopkins Hospital were included in the packet.

As part of the pre-consultation information, patients provided extensive self-assessment on a number of measures regarding sexual response, including interest in sex, erectile quality and difficulties, general arousability, feelings of control over ejaculation, pain during sex, partner information and so on (see Table 1). In addition, men who were willing to attempt sexual intercourse with their partner were asked to do so prior to the office visit and, immediately after they had finished the attempt and no more than 24 h after, fill out a form that assessed their experience and feelings surrounding the sexual episode. Patients were instructed to keep that specific sexual episode in mind (and only that one) while they rated, in private from their partner, 28 affective descriptors (see reference [6]). Patients typically had 3–4 weeks to carry out this request, and most were highly motivated to participate in this aspect of the study as they were visiting the clinic to deal with a problem that interfered with sexual performance. In fact, only five from the initial pool of dysfunctional men declined to do so (several because they had no current partner). In some instances, when the time interval between scheduling the appointment and visiting the clinic was relatively short, patients may not have had an opportunity to attempt intercourse with their partner prior to the consultation and so were asked again during the clinic visit to do so and return the forms either through the mail or upon their next visit.

Table 1. Self-report sexual functioning in men visiting a urology clinic
CategoryMean (±sem)
  1. *1, attended high school; 2, completed high school; 3, trade school; 4, attended college; 5, completed college; 6, postgraduate; 7, advanced degree; 8, other. †1, always; 5, never. ‡1, very often; 7, almost never. §1, not at all; 7, very much.

 Age56.8 (1.31)
 Education*4.7 (0.19)
Metabolic disorder 
 Diagnosed with metabolic disorder10.1%
 Diagnosed with cardiovascular problems27.8%
Sexual problems 
 Difficulty getting erection2.3 (0.15)
 Difficulty keeping erection2.0 (0.13)
 Get erection when you want5.2 (0.24)
 Keep erection when you want5.4 (0.22)
 Ejaculate when you want4.0 (0.28)
 Ejaculate sooner than you want4.6 (0.29)
Other sexual parameters 
 Having sex important2.1 (0.10)
 Having sex important to partner2.4 (0.10)
 Desire for sex with partner2.5 (0.11)
 Frequency of thoughts about sex2.7 (0.17)
 Anxious about sex with partner2.9 (0.22)
 Feel good about sex with partner3.5 (0.24)
 Satisfied with last sexual activity§3.9 (0.26)

The information packet was handed in at the initial consultation with the urologist, and relevant information from the forms was used to aid in the consultation and to verify or refine a diagnosis, carried out by the same urologist to assure consistency and uniformity. After a comprehensive clinical interview with the urologist, individuals were classified according to their problem: erectile dysfunction (n= 48), both erectile dysfunction and premature ejaculation (n= 8) or primarily premature ejaculation (n= 3).


Participants rated their sexual experience with their partner on 28 affective descriptors that have been used previously in studies assessing the relationship between affective and sexual response. These descriptors comprised two broad sets of scales, one measuring positive affect (13 items) and the other negative affect (15 items). In generating this list, we selected and phrased items so that they would be meaningful to a sexual context; the procedure for doing so has been fully described elsewhere [4,19–21]. Each item was rated on a seven-point Likert-type scale (1 = not at all; 7 = very much). Approximately 6–8 weeks later, a subset of 36 men who had not yet begun treatment or for whom no treatment was recommended were administered the Psycho-affective Response Form for a second time to test the stability of such measures over time (test–retest reliability 0.86). Data used in this study came only from the initial questionnaire.


In order to reduce the number of items, principal components analysis was used to create omnibus variables/factors designed to better characterize the positive and negative emotional response of the patients in our study. A detailed description of this procedure is provided elsewhere [4], but briefly, principal components analysis was run separately on positive and negative item pools, using varimax rotation and retaining only components yielding eigenvalues greater than 1.0. This analysis yielded five distinct ‘affect’ factors, two related to negative affect and three related to positive affect. The two negative factors were named Apprehensive and Insecure; the three positive factors were named Affection, Pleasant and Arousability. Correlations among affective factors (Table 2) were generally moderate, with negative factors correlating negatively with positive factors and the higher correlations typically occurring among factors within the same general positive or negative domain; Pleasant was the exception, showing higher negative correlations with Apprehensive and Insecure.

Table 2. Correlations among outcome (affective) variables
  • **

    P < 0.001(two-tailed).



Table 3. Correlations among predictor variables
 Somatic riskSex importance
  1. *P < 0.05 (two-tailed). **P < 0.001 (two-tailed).

Somatic risk  
Sex importance0.211* 
Table 4. Description of input variables
Variable codeDescriptionScale
  • *

    Reverse scored for purposes of the analysis.

Importance of sexPatient's self-reported importance1–5: 1, not at all; 5, very much*
AttributionPatient attribution of cause of problem1, includes physical/medical; 2, psychological/relational/unknown and not including ‘1’
Somatic riskDiagnosed likelihood of a somatic explanation for the sexual dysfunction1, high somatic risk; 2, moderate somatic risk; 3, low somatic risk

Participants' causal attribution for their sexual problem was specified using an item that asked participants to identify their suspected reason for the sexual problem, with the response selected from one or more of three primary categories: (i) physical/medical (such as injury, illness, specific medical conditions, chronic disease, surgery, medication, or other physical/medical problems not otherwise specified); (ii) psychological/relational (such as stress, stressful life event or trauma, relationship problems, psychological/emotional issues or problems, anxiety, mental illness etc.); or (iii) unknown. For the analysis, two categories were created from patients' responses. The first category represented those patients (n= 34) that attributed cause to a physical/medical condition (even if the patient indicated additional psychological or unknown causes). The second category represented those patients (n= 25) who did not identify a physical condition as the presumed cause; thus these patients indicated a psychological/relational cause, an unknown cause, or both.

A measure of ‘diagnosed’ somatic risk was also included as a covariate. This index was established through medical records and the clinician's notes and diagnosis. For patients having one or more of the following, somatic risk was designated as ‘high’: longer term history of diabetes, pelvic surgery, metabolic disease, cardiovascular compromise, medication use, or other conditions known to affect sexual response. For patients having recent onset (e.g. within the past year) of any one (but not more) of the above diseases, having a condition sometimes associated with sexual dysfunction (e.g. Peyronie's disease) or taking medication that purportedly affects sexual function (e.g. selective serotonin re-uptake inhibitors), somatic risk was designated as ‘moderate’. For patients with no medical or physical symptoms or no medication use generally associated with sexual dysfunction, somatic risk was designated as ‘low’. The correlation between diagnosed somatic risk and patient attribution of cause was 0.57 (32% shared variance; Table 3).

Finally, the participant's rating of ‘the importance of sex with a partner’, as determined by a single questionnaire item on a five-point Likert scale (1 = very important; 5 = not at all important), was included as an independent covariate because previous research has shown this factor to explain significant variation in psycho-affective response in men with sexual dysfunction [6]. Our goal in including this covariate was to determine whether its effect on psycho-affective response held up when these additional variables were included.



Responses of self-reported sexual functioning were generally consistent with the diagnosis given by the urologist (Table 1), indicating problems in either erectile response or ejaculatory control or both. Overall sexual functioning in this sample was low relative to a sexually functional clinical group compared in a previous study [4].


Linear regression was run for each of the five affective factors (Apprehensive, Insecure, Arousability, Affection and Pleasant) as the dependent variable and Attribution, Diagnosed Somatic Risk and Importance of Sex as covariates (Table 5). In general, the combination of these three predictor variables generated multiple R values from 0.21 to 0.49. To facilitate understanding of the results, the findings have been organized and presented around each predictor variable (rather than the various positive and negative affective response factors).

Table 5. Regression output tables
 Unstandardized coefficientsStandardized β t Significance R 2
B Standard error
(A) Apprehensive      
 Constant14.7745.328 2.7730.0080.046
 Importance of sex1.9281.5700.1661.2290.224 
 Somatic risk2.0162.7840.1140.7240.472 
(B) Insecure      
 Constant6.6554.658 1.4290.1590.140
 Attribution−2.1061.005−0.327−2.096 0.041  
 Importance of sex2.3321.3690.2261.7040.094 
 Somatic risk5.2192.3830.3372.190 0.033  
(C) Affectionate      
 Constant30.2862.655 11.4050.0000.230
 Attribution1.9500.5790.4983.369 0.001  
 Importance of sex−1.8760.773−0.303−2.429 0.019  
 Somatic risk−2.3571.373−0.251−1.7160.092 
(D) Pleasant      
 Constant20.7222.958 7.0050.0000.097
 Importance of sex−1.8720.888−0.282−2.108 0.040  
 Somatic risk−0.3371.562−0.034−0.2160.830 
(E) Arousability      
 Constant16.0122.049 7.8150.0000.235
 Attribution1.0960.4250.3542.576 0.013  
 Importance of sex−2.0500.596−0.415−3.438 0.001  
 Somatic risk−0.5811.022−0.078−0.5690.572 

Attribution of the cause of the sexual problem was significant for the negative affect Insecure (P= 0.041) and the positive affects Arousability (P= 0.013) and Affection (P= 0.001), such that attribution to a clear physical/medical cause led to lower feelings of insecurity and higher feelings of arousability and affection during a partnered sexual experience.

Diagnosed somatic risk (low, moderate, high) was significant for the negative affect Insecure (P= 0.033) and suggestive for Affection (P= 0.090): the greater the diagnosed somatic risk, the lower the feelings of insecurity and to a lesser degree the higher the feelings of affection. Finally, self-rated Importance of Sex was significant for the positive affects Pleasant (P= 0.040), Affection (P= 0.019) and Arousability (P= 0.001) and suggestive for the negative affect Insecure (P= 0.094). Generally, the more important sex was to the man, the less pleasant, affectionate and arousable he felt, and the more insecure he felt, in response to the partnered sexual experience.


This study is, to our knowledge, the first to investigate the relationship between causal attribution of sexual dysfunction in men and how it affects their emotional response in a sexual situation with their partner. Three broad findings are apparent. First, men that have a clear physical/medical cause to which they can attribute their problem experience less negative and greater positive emotion in the sexual situation. Second, ‘diagnosed’ somatic risk for sexual dysfunction, although having some predictive potential regarding emotional response, was less relevant than the participant's beliefs regarding the cause of his sexual problem. Third, the greater the importance of sex to the participant, the lower his positive affect and the higher his negative affect – an effect identified in a previous study and, from this analysis, shown to affect emotional response independent of causal attributions or diagnosed somatic risk.

It is not surprising that men with sexual dysfunction report overall more negative affective responses than functional men, a finding that has been demonstrated repeatedly [4,5]. More interesting is the effect that the man's attribution of cause of the sexual problem had on his emotional response. Specifically, men who can attribute their sexual dysfunction to a medical or physical condition do much better emotionally (higher positive affect and lower negative affect) than those who attribute the problem to unknown or psychological factors.

Previous research has shown that men with a sexual dysfunction who experience sexual failure are more likely to internalize the cause (self-blame) than men without sexual dysfunction [11]. Furthermore, these men have greater difficulty with sexual functioning in the future than men who externalize the problem [22]. Our study, however, clearly differentiates between types of ‘internal’ attributions, suggesting that the critical factor with respect to attribution is not merely its reference to something about ‘oneself’ (vs something about an external factor), but rather its reference to situations where the individual can assume responsibility and control. That is, the key distinction is not external vs internal, but rather control/efficacy vs lack of control/efficacy. Specifically, those men who can attribute their problem to a physical/medical issue are able to ‘externalize’ the problem – needing to assume less responsibility for and/or control over the problem. In doing so, they reduce the burden of responsibility on themselves and thus, even though sexual performance itself may be unaffected, tend to have a more positive affective experience – lower insecurity and stronger feelings of arousability and affection – during sex with their partner. In contrast, men with a sexual dysfunction who attribute their problem to an internal factor for which they feel responsible respond more negatively emotionally (and less positively), a pattern of responding that is likely to maintain or even intensify the problem. Such men, for example, less confident in their ability to perform and possibly more distracted during sexual experiences [4], may be less mindful of the moment with their partner, a situation that in turn is likely to further affect performance. Furthermore, because feelings of insecurity (i.e. low confidence and trust), arousability and affection are grounded in the dyadic relationship, this attributional pattern may well affect overall relationship functioning and quality. By implication, such men are probably less likely to overcome the effects of sexual dysfunction through medication alone, a conclusion similar to one that we reported previously based on a sample of men with premature ejaculation [19].

Men having greater ‘diagnosed’ somatic risk for developing a sexual dysfunction reported higher feelings of insecurity and, to a lesser extent, lower feelings of affection. Two points are notable. First, the correlation between attribution of cause and diagnosed somatic risk was lower than anticipated; we had actually expected near redundancy in these two variables. Clearly, what men believe to be the cause of their sexual problem, and what the actual probable cause is, are quite different. Hence, a differential diagnosis may continue to have value, even in an era of patient-centred approaches where patients' goals and ways to achieve them are given priority and where the establishment of aetiology is often de-emphasized. Second, the importance of clear communication between the clinician and patient about probable cause is obvious, as dysfunctions due to a medical or physical problem (assuming such a diagnosis is valid) weigh less heavily emotionally on the patient – the patient's perception of his situation is more important to his emotional response to sex than the actual situation.

Finally, the higher the dysfunctional man rated the importance of sex with his partner, the less positive his emotional experience, specifically in the areas of pleasantness, affection and arousability. Previous research has demonstrated the importance of ‘soft’ variables – including individual expectations, needs and desires – on a man's emotional response within the context of a partnered sexual experience [6]. This relationship persisted when variables on attribution and somatic risk were introduced into the analysis.

The findings of this study have specific implications for the treatment of men with sexual dysfunctions. Men who place a high value on their sex lives but who cannot attribute sexual failure to a physical/medical aetiology represent a particularly vulnerable group. Such men need to be given a clear route to understanding their physical/medical cause (assuming one actually exists), or they may be considered for more in-depth treatment that enables them to adopt a less self-blaming perspective and a more self-protective attribution process in future sexual experiences. One approach to treatment, for example, might include a biopsychosocial model that emphasizes individual vulnerability in any one of the major domains that affect sexual functioning – biological, psychological and relational – with the concomitant message that some individuals, while healthy biologically, may have an inherent psychological vulnerability within sexual situations. Such vulnerability might, however, stem from desirable personality characteristics such as sensitivity to the partner, valuing sex and intimacy with the partner, taking responsibility for one's actions and so on. That is, the same attributes that are important to a well functioning relationship and valued by society in other situations may increase psychological vulnerability in situations involving sexual ‘performance’.


Several sampling and other methodological issues may have affected the conclusions of this study. First, our sample consisted of men with a mix of sexual dysfunctions; although most had erectile dysfunction or combined erectile dysfunction with premature ejaculation, several had just premature ejaculation. Second, the study lacks an index of the quality of relationship functioning and an assessment of the partner's emotional response. Both measures would probably contribute to our understanding of variation in men's emotional response during sex with their partner. Finally, we did not distinguish between types of depression, whether clinical, event-related, sex-related and so on. Such a distinction might affect attribution categories – e.g. clinical (true psychopathological) depression may be considered a biological/medical cause rather than a psychological cause. Unless the patient's medical history indicated otherwise, the absence of this level of detailed information led us to classify any mention of depression as a psychological cause. On the positive side, few participants constituted an ‘undifferentiated’ depression attribution.

This study shows that causal attributions for dysfunctional response affect psycho-affective response to a partnered sexual situation and may, in turn, contribute to the development, maintenance and intensification of a sexual dysfunction. These findings further suggest that treatment focused on medication alone may be inadequate, and that a comprehensive strategy should include the development of internal positive self-serving attributions and feelings surrounding sexual activity.


The helpful assistance of Cody Neal, MA, in the preparation of this paper is appreciated.


None of the authors has a conflict of interest to declare. No commercial, foundation or government funding was provided.