Operationalizing premature or rapid ejaculation.

Premature or rapid ejaculation (RE) is widely believed to be the most common male sexual dysfunction and has been shown to have a substantial negative impact on men's sexual relationships (Metz, Pryor, Nesvacil, Abuzzahab, & Koznar, 1997; Rowland, Cooper, Slob, & Houtsmuller, 1997). Nonetheless, methodological difficulties in RE research, most notably the inconsistency in operational definitions, have prevented a full understanding of this phenomenon (Grenier & Byers, 1995, 1997). RE research has, for the most part, made the (untested) assumption that the various operational definitions that have been used (e.g., ejaculatory latency, perceived ejaculatory control, number of intravaginal thrusts) all measure the same phenomenon and thus are interchangeable. Further, researchers have largely used these criteria to dichotomize men on the basis of whether they do or do not have RE, creating additional methodological difficulties related to arbitrary and inconsistent cut-off points.

Grenier and Byers (1997) surveyed the ejaculatory behavior of 110 heterosexual college men using the five most commonly employed measures of RE: ejaculatory latency, perceived ejaculatory control, concern over ejaculating more rapidly than desired, satisfaction with ejaculatory control, and the percentage of antiportal ejaculation. In order to determine the extent to which they are related, they assessed these criteria as continuous rather than as dichotomous predictors. They found that the RE criteria were only modestly related, sharing between 2% and 46% (M = 10%) of their variance, challenging the assumption that the various RE criteria are interchangeable. However, these results were based on a sample of university men who were fairly homogeneous in age, sexual experience, and frequency of intercourse. This was unfortunate as each of these variables has been etiologically linked with RE. For example, a number of theorists have suggested that younger, less sexually experienced men ejaculate faster than older, more sexually experienced men (Reinisch & Beasely, 1990; Ruff & St. Lawrence, 1985). However, these associations remain to be empirically demonstrated. Similarly, although a lower frequency of intercourse has been theoretically linked with RE (Gospodinoff, 1989; Williams, 1984), the empirical support for these associations is inconsistent. Spiess, Geer, and O'Donohue (1984) found that greater periods of sexual abstinence were positively correlated with shorter ejaculatory latencies, whereas other researchers have found no such association (Rowland, Strassberg, de Gouveia Brazao, & Slob, 2000; Strassberg, Kelly, Carroll, & Kircher, 1987).

The present study replicated and extended the findings of Grenier and Byers (1997) by investigating the relationships between various RE criteria in a more heterogeneous sample of heterosexual men, using a much larger sample than is typical of RE investigations. We assessed the five RE criteria that are most frequently found in the literature, albeit usually as dichotomous criteria, and that were also assessed by Grenier and Byers (1997): (a) latency to ejaculation after intromission, (b) control over the occurrence of ejaculation, (c) concern about ejaculating too soon, (d) (dis)satisfaction with the ability to select the moment of ejaculation, and (e) the occurrence of antiportal ejaculation. In addition we assessed two other continuous RE criteria which are used less frequently in the RE literature: (f) the percentage of intercourse experiences in which ejaculation occurs sooner or faster than desired, and (g) the percentage of intercourse experience in which specific attempts are made to delay ejaculation and prolong intercourse.

Research on RE has used objective criteria to place men into RE and non-RE groups but has failed to simultaneously query men about their perceptions of their RE status. Consequently, little is known about the aspect(s) of ejaculatory behavior that men use to identify themselves as experiencing a problem or difficulty with RE and/or to seek help for RE. Nor has the impact of having a self-identified RE problem been studied in men who have not sought therapy. Therefore, we also examined (h) men's self-identification as having an RE problem.

The following research questions were addressed in this study:

1. How do estimates of the prevalence of RE based on different dichotomous definitions of RE found in the literature compare to each other?

2. How do these estimates compare to estimates of RE based on having a self-identified RE problem?

3. What is the degree of overlap among the RE criteria when they are used as continuous variables?

4. What are men's perceptions of the impact of RE on their self-esteem, sexual pleasure, sexual relationship quality, overall relationship quality, and avoidance of sexual intercourse.

Although erectile dysfunction (ED) was not a focus of this study, we also examined the prevalence of ED, the overlap between ED and RE, and the impact of ED on men's self-esteem, sexual pleasure, sexual relationship quality, and overall relationship quality.

Etiology of Rapid Ejaculation

A number of hypothesis have been advanced in the literature regarding the etiology of RE. The classical view first advanced by Masters and Johnson (1970) and still endorsed by some authors as "the" cause of RE (e.g. Crooks & Baur, 1993; McCarthy, 1988a, 1988b) is that men who have initial intercourse experiences that are rushed and/or clandestine, either because their first partner(s) wished intercourse to end quickly or because there was a danger of being caught or discovered while having intercourse, become conditioned to ejaculate rapidly. However, this theory has never been tested. If Masters and Johnson are correct, men reporting more RE characteristics should also report that they felt more rushed, and experienced more negative affect during initial intercourse experiences than do men who report fewer RE characteristics. However, Masters and Johnson did not compare men with RE to a control group in concluding that RE is conditioned by anxious or rushed early intercourse experiences. Thus, it is possible that most men's initial intercourse experiences are characterized by a degree of anxiety, not just men with RE, and that early conditioning is not etiologically related to RE (Coles & Stokes, 1985). If this is the case, negative feelings during initial intercourse would not predict RE. Therefore, we examined whether the RE criteria are related to initial intercourse experiences characterized by being rushed or by negative emotions. As age, amount of sexual experience, and the frequency of intercourse have also been proposed to be linked etiologically with RE (Gospodinoff, 1989; Reinisch & Beasely, 1990; Ruff & St. Lawrence, 1985; Williams, 1984), we also examined the relationships between these factors and RE.

Research suggests that although some cases of RE have psychological causes, others have biological causes. Kameya, Deguchi, and Yokota (1997) have demonstrated that there are individual differences in ejaculatory latency even given the same sexual stimulation under the same conditions. In keeping with this view, greater penile sensitivity and a shorter bulbocavernosus reflex nerve response latency have both been linked to some cases of primary RE (Metz et al., 1997; Rowland & Slob, 1997). Unfortunately, it was impossible to conduct the required physiological assessment necessary to assess biogenic contributors to RE within the scope of this community-based research. However, even if there is a somatic component to some cases of RE, it is still important to determine whether other factors such as early conditioning, age, sexual experience, and frequency of intercourse also play a role in RE.

METHOD

Participants

A questionnaire packet was mailed to 1989 randomly selected male university alumni. Most (69%) had received an undergraduate degree, and the remainder had received either a Masters or Doctorate. Two hundred sixty (14.1%) men returned completed and usable questionnaires. Although lower than desirable, this rate is very similar to the response rate from two other mailed sexuality surveys conducted by our laboratory. Incomplete and thus unusable questionnaires were returned by 21 additional men. Respondents ranged in age from 23 to 78 years (M = 41.6 years, SD = 12.2). Most (82%) were married or cohabiting and the length of their married/cohabiting relationship ranged from 3 months to 50 years (M = 15.4 years, SD = 11.7 years). Half (50%) had one or more children living with them. On average, the men began having intercourse when they were 19 years old (SD = 3.2) and reported an average of nine intercourse partners over their lives (SD = 10.1). However, the median number of lifetime intercourse partners was five and the mode was one. Twelve men (5% of the sample) reported more than 40 intercourse partners, with three reporting 100 or more partners. The men reported an average intercourse frequency of 6.1 times in a 4-week period (SD = 4.7, range = 0-25 times.)

Procedure

A packet containing a cover letter and two questionnaires (a male and a female version, each inside a separate privacy envelope (1)), a postage-paid response card, and a postage-paid return envelope was mailed to a random sample of male university alumni who were living in Canada. A reminder letter was sent 10 days after the packet was mailed. Those interested in participating in the study returned their completed questionnaire (with or without a partner questionnaire) via prepaid post. Return of a completed questionnaire was interpreted as indicating informed consent. After the data had been analyzed, a summary of the results was sent to participants who had expressed an interest in receiving the results by returning the response card. This procedure was approved by the Ethical Review Board of the University of New Brunswick.

Materials

The questionnaire consisted of a large number of items divided into seven sections and printed in booklet form. Only the questions used in the present study are described here. A copy of the questionnaire may be obtained from the authors.

Participants completed a Background Questionnaire that assessed basic demographic information such as age, level of education, income level, and relationship status.

The Personal Sexual History section assessed various aspects of men's early sexual experience. Men indicated the age at which they first had sexual intercourse as well as their number of lifetime intercourse partners. After examination of the estimated number of partners for univariate outliers, men reporting more than 40 partners were reassigned a value of 41 on this variable (Tabachnick & Fidell, 1989). Next, the men were asked, "Thinking back to the first few times you ever had sexual intercourse, what were your feelings at those times?" The participants reported their feelings on the following 10 bipolar scales, each ranging from not at all (1) to very (7): excited, anxious, frightened, relaxed, guilty, rushed, worried, happy, nervous, and sexually aroused. Ratings of feeling rushed were used as one variable. Ratings for how anxious, frightened, guilty, worried, and nervous the men had felt were added to create the Negative Affect Scale, with a possible range of scores from 5 to 35 (this procedure is similar to the one used for computing scores on the Feelings Scale which assess positive and negative affect; Smeaton & Byrne, 1998). In order to provide some empirical support for the reliability of the Negative Affect Scale, the internal consistency and item-total correlations were examined. The internal consistency was good (Chronbach's alpha = .79.) All five items correlated above .42 with the total score, and the internal consistency was not significantly improved by dropping any of the items. Finally, the men were asked if they had made any specific attempts to prolong or to hasten intercourse during those early intercourse experiences by checking one of three response options: "I tried to make it last as long as possible"; "I tried to get it over as quickly as possible"; and "I made no specific attempt to either prolong or hasten intercourse."

The Current Sexual Behavior section contained questions about the men's current sexual and ejaculatory behavior, including their frequency of intercourse over a typical 4-week period. Eight separate RE measures were incorporated into this section, five of which were adapted from Grenier and Byers (1997). The men reported (a) their latency to ejaculate following intromission in minutes (LATENCY); (b) their level of concern about ejaculating earlier or sooner than they want to, with anchors of no concern (1) and extreme concern (8) (CONCERN); (c) their level of satisfaction with their ability to select the moment when they first ejaculate, with anchors of extremely satisfied (1) and extremely dissatisfied (8) (SATISFACTION); (d) the degree of control they feel they have over the occurrence of their first ejaculation, with anchors of no control (1) and absolute control (8) (CONTROL); (e) the percent of intercourse experiences during which they involuntarily ejaculate before intercourse starts, on an 11-point scale ranging from 0% to 100% (% ANTIPORTAL). Four other RE measures were developed specifically for the current investigation. The men reported, on 11-point scales ranging from 0% to 100%, the percentage of intercourse experiences in which they (f) specifically attempt to prolong intercourse by delaying the occurrence of ejaculation (% DELAY); (g) feel they have control over when they ejaculate for the first time (% CONTROL); and, (h) ejaculate sooner or faster than they would like (% RAPID). As CONTROL and % CONTROL, were found to be relatively strongly correlated, r = .71 p < .001, the latter measure was eliminated from further analyses in order to reduce potential problems of multicollinearity.

The Specific Sexual Issues section assessed the men's self-definition as having a current or past problem with RE. Participants indicated whether they currently had a sexual difficulty or problem [due to] early ejaculation. Early ejaculation was defined as "your ejaculation occurs before you want it to, despite your desire to delay it". These data were used to create a dichotomous variable assessing whether men self-identified a current RE problem (RE PROBLEM/NO RE PROBLEM). In a separate question, the men indicated whether early ejaculation had been a continuing or recurring problem for them in the past. They also rated the impact of early ejaculation on four different aspects of their life using 7-point scales ranging from very negative impact (1) to very positive impact (7): their feelings of self-esteem, their sexual pleasure, their sexual relationship, and their overall relationship. Finally, all the men were asked if they had ever avoided or declined a sexual intercourse opportunity because of concerns or worries about early ejaculation, and if so, how frequently they had done so on a scale ranging from almost never (1) to almost all the time (7). The men also indicated whether they had sought information or help to resolve their problem with early ejaculation.

In order to broaden the content of the questionnaire the men were then asked the following question with respect to erection difficulties (ED): "Do you currently consider yourself to have a sexual difficulty or problem because you regularly have difficulty getting or maintaining an erection?" Men who reported a self-identified ED difficulty then rated the impact of ED on their self-esteem, sexual pleasure, sexual relationship, and overall relationship, and indicated whether they had sought help to resolve their erectile difficulties.

RESULTS

Prevalence of Rapid Ejaculation

We assessed rapid ejaculation using seven separate continuous measures. The means and standard deviations for each of these measures are presented in Table 1. The average latency to ejaculation reported by the men was approximately eight minutes and typically they reported a moderate degree of CONTROL, CONCERN, and SATISFACTION. On average, the men reported that during at least one third of sexual intercourse opportunities, ejaculation occurred sooner or faster than desired (% RAPID), and during half of all intercourse experiences they specifically attempt to delay the occurrence of their ejaculation (% DELAY). Although the men reported that on average antiportal ejaculation occurred during 1 out of 20 intercourse opportunities, almost three quarters of men reported that they had never experienced antiportal ejaculation. Thus, as would be expected from a community sample, most of them men did not experience rapid ejaculation on any of the criteria.

Participants indicated whether early ejaculation was a current sexual difficulty/problem for them, or if it had been a problem for them in the past. Two thirds of the men responded negatively to both questions. The remaining one third (32%) indicated that RE had been a problem for them at some time in their lives. These latter men could be grouped according to three different patterns of RE: 20 men (7.8%) reported that they had experienced RE as a sexual problem for some discrete period of time in the past but no longer felt that RE was a problem for them; 46 men (18.1%) reported that RE had been a life-long problem for them; and 14 men (5.5%) reported currently experiencing RE but not having had RE at any other time in their past. Thus, allowing men to classify themselves, 60 respondents (23.6%) reported a current difficulty with RE (RE PROBLEM group). Six other respondents either failed to complete these questions or provided inconsistent responses, and were excluded from analyses involving the RE PROBLEM variable.

Only 15 men (6%; 9 men reporting a current difficulty with RE and 6 men reporting that they no longer had a problem with RE) reported that they (or their partner) had sought help to resolve their RE problem. The most commonly reported source of help was books on sexual functioning or sex manuals (reported by 11 men) followed by the use of other published material (e.g., magazine or newspaper articles, reported by 9 men). Only four men had sought assistance from a professional. Similarly, few men (4.8%) had sought help with their ED difficulty.

Twenty-eight men (11.2%) reported a current ED difficulty. There was no difference in the percentages of men with and without ED reporting an RE problem (35.7% and 22.0%, respectively), [X.sup.2] (1) = 2.61, p > .10. Overall, 10 of the men in the sample (4%) self-identified as currently having both RE and ED problems.

To assess the consequences of adopting the various common approaches to identifying men as having RE, the various RE criteria were dichotomized (see Table 2). As can be seen, using single criteria, RE prevalence rates varied between 4% and 61% depending upon which RE criterion was examined and which cutoff value was applied. Fully 52% of the men reported some concern over ejaculating too early (score of 5 or greater). Frequently within the RE literature, two or more criteria are used in combination to determine the presence of RE. When the two most commonly used RE criteria, ejaculatory latency and ejaculatory control, are considered simultaneously using the most commonly used latency of 2 minutes or less along with the most stringent criterion of little or no control (1 on the CONTROL scale), only 2% of the sample would meet these rapid ejaculation criteria. If a LATENCY of 2 minutes or less is combined with a cut-off of 4 or less on the CONTROL scale, 10% of the sample would meet the rapid ejaculation criteria. Finally, using the most inclusive cutoff values found in the literature (i.e., a LATENCY of 7 minutes or less and 4 or less on the CONTROL scale), 32% would be classified as rapid ejaculators.

Fully 100% of the men with a LATENCY of 1 minute or less considered themselves to have an RE PROBLEM. However, as LATENCY increased, the percentage of men who self-identified as having an RE PROBLEM decreased: 42% of the men with a LATENCY of 2 minutes; 32% with a LATENCY of 3 minutes; and 15% of men with a LATENCY of 4 minutes.

Relationships Among Rapid Ejaculation Criteria

The Pearson zero-order correlations among the eight RE criteria are reported in Table 3 (nonparametric tests yielded correlations of very similar magnitudes). All but two of these correlations were significant. However, the correlations were small to moderate with the various criteria that were significantly correlated sharing between 2% and 36% of their variance. In fact, LATENCY and CONTROL, the two most commonly used criteria, shared only 12% of their variance suggesting that these variables are relatively independent. Further, LATENCY was not significantly related to CONCERN.

To investigate whether the RE criteria are related to self-identification as having an RE problem, a one-way MANOVA was conducted using the seven continuous RE criteria (LATENCY, CONTROL, CONCERN, SATISFACTION, % DELAY, % ANTIPORTAL, % RAPID) as dependent measures and RE PROBLEM as a two-level independent measure. The MANOVA was significant, F (7, 245) = 34.89, p < .001. Follow-up univariate ANOVAs, reported in Table 4, indicated that the RE PROBLEM and NO RE PROBLEM groups were significantly different on all of the assessed RE criteria in the expected directions. Examination of the zero-order correlations (see Table 3) indicates that the strength of association of these significant differences were small to moderate. The strongest association with RE PROBLEM was seen in % RAPID, sharing 45% of the variance. In contrast, CONTROL and LATENCY shared 9% and 18%, respectively, of their variance with RE PROBLEM.

A discriminant function analysis using a stepwise procedure was conducted to identify the most parsimonious combination of RE criteria which would predict self-identified RE as a problem. Three of the RE criteria, % RAPID, CONCERN, and CONTROL, were retained in the final model, [X.sup.2] (3) = 166.85, p <.001. Examination of the standardized coefficients revealed that, compared to the NO RE PROBLEM group, the RE PROBLEM group reported a greater percentage of intercourse occasions during which they feel their ejaculation occurs too soon (.73), greater concern over ejaculating rapidly (.37), and less ejaculatory control (-.21).

The Impact of Rapid Ejaculation

On average, participants in the RE PROBLEM group reported that rapid ejaculation had had a slightly negative impact on their self-esteem (M = 3.2, SD = 1.0), sexual pleasure (M = 3.0, SD = 1.1), and current sexual relationship (M = 3.3, SD = 1.2), but little impact on their overall relationship (M = 3.8, SD = 0.8). Only 16 (6%) of the 260 men in the study reported that they had ever avoided or declined a sexual intercourse opportunity due to concerns or worries about RE. However, even for these men, this had occurred infrequently (M = 2.7, SD = 1.1 on a scale ranging from almost never (1) to almost always (7)). Twelve of these 16 men were in the RE PROBLEM group.

Participants who reported a current ED problem reported that ED had a somewhat negative impact on their self-esteem (M = 2.9, SD = 1.1), sexual pleasure (M = 2.7, SD = 1.2), and current sexual relationship (M = 3.0, SD = 1.2), but little impact on their overall relationship (M = 3.5, SD = 1.2). Comparison of the men who reported RE only to those men reporting ED only using a one-way MANOVA indicated that the two types of sexual problems did not differ in their impact, F (4, 60) = 0.80, p >. 10.

Rapid Ejaculation Etiology

A canonical correlation analysis was conducted between the eight RE criteria and the five variables theorized to be etiologically associated with RE: age, number of lifetime sexual intercourse partners, frequency of intercourse, feeling "rushed" during early intercourse experiences, and experiencing negative affect during early intercourse experiences. One significant pair of variates was found which accounted for 17% of the variance, Rc = .41, F(40, 1185) = 1.97, p < .001 (see Table 5). Using a cut-off value of .30 (Tabachnick & Fidell, 1989), only frequency of intercourse was correlated with the etiology variate, while all the variables except % ANTIPORTAL were correlated with the RE set. Interpretation of the pair of variates suggests that an increased frequency of intercourse was associated with higher ejaculatory latency, control, and satisfaction as well as with lower concern over ejaculating too rapidly and self-identification as having an RE problem. In addition, an increased frequency of intercourse was associated with a lower percentage of intercourse experiences in which ejaculation occurs sooner than desired, and a lower percentage of attempts to delay ejaculation. Thus, men who had intercourse less often were more likely to report behaviours and reactions that characterize RE. For example, the RE PROBLEM group reported having sex 4.4 times in a 4-week period whereas the frequency for the NO RE PROBLEM group was 6.6 times.

DISCUSSION

Among human sexuality researchers and therapists, RE is usually described as the most common male sexual dysfunction. However, there is disagreement as to the actual prevalence of RE, with prevalence rates ranging from 10% to 20% of men (Ende, Rockwell & Glasgow, 1986; Masters, Johnson, & Kolodny, 1986) to more than 50% (Kinsey, Pomeroy, & Martin, 1948; Metz & Seifert, 1990). Discussions of RE directed primarily toward the lay public place RE prevalence at about a third of men (e.g., McCarthy, 1988a; Zilbergeld, 1999). Based on their probability sample, Laumann, Gagnon, Michael, and Michaels (1994) placed the RE prevalence rate at 28.5% of men. Their RE operationalization was based entirely upon male self-identification and is similar to the operationalization of self-identified RE PROBLEM used in the current study. Our finding that 23.6% of the men identified themselves as having a current problem with RE is similar to the rate found by Laumann et al. (1994); the finding that 11.2% of the men in this study reported a current problem with ED is similar to Laumann et al.'s (1994) estimate of 10.4% of men with erectile difficulties. Similarly, our average ejaculatory latency of 7.9 minutes was similar to the mean of 7.8 minutes found by Kameya et al. (1997) in their laboratory investigation of men's ejaculatory latency during manual stimulation by a woman. Based upon these comparisons, it seems likely that the current investigation adequately sampled men at least in terms of the erectile and ejaculatory functioning. However, the degree to which selection bias due to the low response rate and participant's high level of education may have affected the results is not known. For example, volunteers to participate in sexuality research tend to be more erotophilic and sexually active than are nonvolunteers (Bogaert, 1996; Strassberg & Lowe, 1995). In addition, Laumann, Paik, and Rosen (1999) found that college graduates were more likely to report problems with RE than were less educated men. Further, the men who self-identified as having an RE PROBLEM may differ substantially from the men seen in clinical practice who both self-identify and seek help as only 15% of the RE PROBLEM group had sought help of any kind of help for their problem. Nonetheless, a strength of this study is that it sampled men from the community and thus was not subject to the biases inherent in using a clinical sample.

When the criteria commonly used in previous RE research were dichotomized creating RE and non-RE groups, RE prevalence rates ranged from less than 1% of the sample to more than 60%. This is similarly to the prevalence range for RE (i.e., less than 1% to 59%) reported by Grenier and Byers (1997) in their relatively homogeneous sample of young men. The broad variation in RE prevalence supports the contention that a uniform but multidimensional approach to operationalizing RE is needed (Grenier & Byers, 1995, 1997). The finding that each of the assessed RE criteria produce a vastly different RE prevalence rate is not surprising considering the limited degree of relatedness found among them. Although the criteria were, for the most part, positively correlated, the amount of shared variance amongst them was relatively small. This calls into question the comparability of the results of the vast majority of RE research as few RE studies have operationalized RE in the same way. The fact that there were significant differences between men who did and who did not identify themselves as having a current RE problem on each of the seven RE criteria suggests that these dimensions all represent aspect of RE. Men who feel that they have a current problem with RE report shorter ejaculatory latencies, less perceived ejaculatory control, less satisfaction at the ability to choose the moment of ejaculation during intercourse, more concern over ejaculating sooner than desired, and so on.

The results of the discriminant function analysis indicate that % RAPID, CONCERN, and CONTROL are all uniquely related to men's self-identification as having an RE problem. This suggests that there are three components that affect men's evaluation of their RE status: a behavioural component (or at least the subjective report of a behavior, i.e., the regularity of RE), an affective component (i.e., worry or concern about RE), and an efficacy component (i.e., having little control over the timing of ejaculation). These three components are also contained in the DSM-IV diagnostic criteria for premature ejaculation (American Psychiatric Association [APA], 1994). Specifically, the DSM-IV identifies premature ejaculation as "persistent or recurrent ejaculation with minimal sexual stimulation before, on, or shortly after penetration and before the person wishes it" (APA; p. 511). This statement identifies the regularity of RE and ejaculatory control as essential features, corresponding to the variables % RAPID and CONTROL, respectively, assessed in this study. The DSM-IV also specifies that rapid ejaculation must "cause marked distress or interpersonal difficulty" (APA; p. 511). This statement corresponds to the CONCERN variable identified as a unique aspect of RE.

The notable difference between the DSM-IV definition and the current results is the role of ejaculatory latency. Ejaculatory latency is a central feature of the DSM-IV diagnostic criteria. However, ejaculatory latency did not appear as a unique predictor of a self-identified RE problem, and was only moderately correlated with the other indicators of RE. As we used retrospective evaluation of ejaculatory latency, this may be a result of poor reliability of this measure. However, there is evidence for the reliability of retrospective measures of ejaculatory latency. Rowland et al. (2000) found high consistency between retrospective and prospective measures of RE. Further, reports of ejaculatory latency by the female partners of 152 of the men in the present study were significantly correlated with the men's self-reports (r = .64).

Alternately, the failure of ejaculatory latency to appear as a unique predictor of having a self-identified RE problem may indicate that ejaculatory latency is not as important a contributor to men's distress about their ejaculatory behavior as has often been assumed. In fact, within the self-identified RE PROBLEM group, although the average reported latency was 4.5 minutes, there were three self-identified RE men who reported average latencies of 10 minutes or more, ejaculatory latencies that were longer than the average for the sample. These men would be unlikely to be identified as have an RE problem by a therapist or a partner (that is, meet objective or diagnostic criteria for RE); nonetheless they express significant concern over the timing of their ejaculation. These men may have unrealistic beliefs, perhaps media-induced, about the average length of time that most men can delay ejaculation.

Together, these finding suggest that a multifaceted approach is needed to identify RE. This multifaceted approach must include measures of behavior, affect, and self-efficacy, and must also assess the degree or severity of each dimension. This approach to operationalizing RE would be useful from a diagnostic perspective in that it would recognize individual patterns of RE characteristics rather than treating men with RE as a homogeneous group (Metz et al., 1997.)

There are two possible strategies that could be adopted. First, diagnosis of RE could be based on the three RE scales used in the current study that were found to uniquely identify men who self-identify as having an RE PROBLEM--% RAPID, CONTROL, and CONCERN--or some modification of them. To determine the utility of this strategy, we used these three measures to classifying men as to whether they self-identified as having an RE PROBLEM using the criteria of limited control (4 or less on the CONTROL scale), some concern (5 or greater on the CONCERN scale), and ejaculation occurring sooner than desired on at least 50% of occasions. Using these three criteria, 65% of men who had self-identified as having an RE PROBLEM were correctly classified, and only 8% of the NO RE PROBLEM were misclassified as having an RE problem (i.e., constituted false positives). Further, the average ejaculatory latency for the men who met these three criteria was 4.3 minutes (SD = 4.0) compared to a mean of 10.1 minutes (SD = 6.5) for the men who did not meet the three criteria. In contrast, although using CONCERN and only one of the other two RE criteria raised the hit rate to between 73% and 76%, it also resulted in between 22% and 32% of NO RE PROBLEM men being misclassified. These results support the utility of basing an operational definition of RE on the combined assessment of the regularity of RE, worry or concern about ejaculating soon than desired, and control over the timing of ejaculation. It may also be necessary to include the criterion that the man has a short ejaculatory latency in order to ensure that men who have unrealistic expectations about ejaculatory latency do not fit the diagnostic criteria for RE. This could be done by making a short ejaculatory latency a precondition of diagnosing RE. It is also possible that the % RAPID item could be made more objective and clearly tied to short latency while still retaining its predictive validity (for example, by assessing the percentage of time in which the man ejaculates in "less than 2 minutes" rather than "sooner or faster than you want to").

Alternately, although there is no such scale currently available, an RE scale that consists of items tapping a variety of RE facets and that yields a total score that quantifies the severity of RE could be used, particularly for empirical work. The current investigation has identified a number of the salient components that should be included in this type of scale. It should be noted that both of these assessment strategies assume that the couple follows the typical sexual script in which there is one intromission that leads to ejaculation, and thus would need to be modified for couples that do not follow this script.

Etiological Factors Related To Rapid Ejaculation

We investigated a number of factors that have been etiologically linked with RE within the literature: age, number of sexual partners, intercourse frequency, feeling rushed during early intercourse experiences, and experiencing negative affect during these early experiences. With the exception of frequency of intercourse, the results of the current investigation find little support for these presumed associations, lending support to a biological origin of RE in at least some cases. However, men who reported having more frequent intercourse did report longer ejaculatory latencies, greater ejaculatory control, less concern over ejaculating too soon, greater satisfaction with the ability to select the moment of ejaculation, and a lower percentage of intercourse occasions when they feel they ejaculated too quickly, and were less likely to identify themselves as having a problem with RE. Although we did not measure sexual abstinence directly, these findings are consistent with those of Gospodinoff (1989) and Spiess et al. (1984) who reported that greater periods of sexual abstinence were positively correlated with shorter ejaculatory latencies. Typically, the correlational nature of these data would leave the direction of this relationship unknown. However, in the current study men were also asked if they had ever avoided or declined an intercourse opportunity due to their concerns about RE. Only a very small minority of men (6%) reported they had ever done this, and the majority of these men reported that they had done so very rarely. Thus, it is unlikely that the association between a lower intercourse frequency and the increase in the characteristics of RE is due to RE men specifically reducing their intercourse opportunities due to RE concerns. The mechanism by which the frequency of intercourse may be related to RE remains to be delineated, however. It may be that this is due to a practice effect, where men who engage in intercourse more often become more familiar with their preejaculatory feelings, discover positions which afford them greater control, or learn or develop other mechanisms/techniques that allow them to prolong intercourse (Kaplan, 1989; McCarthy, 1988a, 1988b; Zilbergeld, 1992). Alternately, more frequent intercourse may lead to a blunting of a preexisting heightened somatic sensitivity (Colpi, Fanciullacci, Beretta, Negri, & Zanollo, 1986; Fanciullacci, Colpi, Beretta, & Zanollo, 1988; Strassberg, Mahoney, Schaugaard, & Hale, 1990). In either case, as the magnitude of the relationship between intercourse frequency and the RE criteria was small, accounting for 17% of the variance, there appear to be other more salient factors related to the etiology of RE than the ones investigated here, such as biogenic factors.

In contrast, neither age nor number of sexual partners were related to the RE criteria. This is consistent with Grenier and Byers (1997) and Laumann et al. (1999), who also found little relationship between age or sexual experience and RE. Further, in comparing the ejaculatory latency and control reported by the young men in the Grenier and Byers study to the results in the current study, the college men reported longer ejaculatory latencies (13.6 vs 7.9 minutes) but similar ejaculatory control (4.2 vs 4.6). Thus, our results do not support what has sometimes become a cliche in the sexuality literature; that is, that problems with rapid ejaculation routinely decrease with age and experience. It is possible that decreases in RE with age that have been observed in some men by clinicians may be due to increased sexual frequency, not to increased age or sexual experience. Alternately, it maybe that although some men gain ejaculatory control as they age, anxiety about erectile failure adversely impacts the ejaculatory control of other men, thus obscuring the relationship between age and ejaculatory control. Finally, as ejaculatory latency and control may not be uniform across various cohorts, a longitudinal design would provide a better test of the relationship between age and RE.

Similarly, there was no support for Masters and Johnson's (1970) theory that RE is conditioned during early rushed and/or clandestine sexual experiences. In fact, on average the men reported a moderate level of negative affect (M = 19.6, SD = 6.3) and being moderately rushed (M = 3.7, SD = 1.9) during early intercourse experiences, suggesting that negative affect and feeling rushed during early intercourse experiences is a common event rather than a unusual conditioning event. It is also possible that problems associated with retrospective data such as faulty memory may have attenuated any relationship between early experiences and RE. Further, the role of early (or continuing) masturbatory experiences, which could also play a role in conditioning men to ejaculate rapidly, remain to be investigated.

Conclusion

Estimates of the prevalence of RE are sensitive to variations in the way rapid ejaculation is defined. Nonetheless, almost one quarter of men identified themselves as having a current RE problem, and reported experiencing considerable concern over ejaculating sooner than desired and dissatisfaction with their ability to select the moment of ejaculation. A strength of this study is that it included men who are concerned about their ejaculatory control but would not be included in a clinical sample. RE does not appear to have a major impact on most of these men's functioning. Men with self-identified RE perceived the problem as having only a slightly negative impact on their self-esteem, sexual pleasure, sexual relationship, or romantic relationship. Further, few men avoided or declined intercourse opportunities because of the concerns about rapid ejaculation. Similarly, among this community sample, ED had only a slightly negative impact on the men's sexual functioning. Given that few of the men had sought help from the literature or from a professional for their problem with rapid ejaculation or erectile dysfunction, it may be that RE and ED do not cause most men to experience significant emotional distress. It may be that only the relatively few men who experience emotional or relationship distress associated with a sexual dysfunction or who experience significant concern about their sexual functioning seek professional help. Thus, results from clinical samples may not generalize well to men in the community experiencing the same sexual dysfunctions.

Table 1. Descriptive Statistics for the Rapid Ejaculation
Criteria

Measure         M      SD    N     %

LATENCY         7.9    6.4
CONTROL         4.6    1.8
CONCERN         4.3    2.2
SATISFACTION    4.8    1.9
% RAPID        35.5   28.8
% DELAY        52.3   29.8
% ANTIPORTAL    5.6   14.0
RE PROBLEM                   60   23.6

Note. N = 254-260. LATENCY = ejaculatory latency following vaginal
intromission in minutes; CONTROL = control over the occurrence of
first ejaculation during sexual intercourse; CONCERN = concern over
ejaculating sooner than desired; SATISFACTION = satisfaction with
ability to select the moment of ejaculation; % RAPID = percentage of
intercourse experiences in which ejaculation occurs sooner than
desired; % DELAY = percentage of intercourse experiences in which
specific attempts are made to prolong intercourse by delaying
ejaculation; % ANTIPORTAL = percentage of intercourse experiences where
ejaculation occurs before intercourse has started; RE PROBLEM =
self-identification of having a current problem because ejaculation
occurs before desired.
Table 2. Percentage of Men Classified as Experiencing RE Using Various
Dichotomized Criteria

                                        LATENCY

                          [less than     [less than     [less than
                          or equal to]   or equal to]   or equal to]
RE criiterion               1 minute       2 minute       3 minute

LATENCY
  [less than or
    equal to] 1 minute
  [less than or
    equal to] 2 minutes
  [less than or
    equal to] 3 minutes
  [less than or
    equal to] 4 minutes
  [less than or
    equal to] 7 minutes
CONTROL
  Rating = 1                  0.4%           1.9%           2.3%
  Rating [less than or
    equal to] 2               3.1%           6.9%           8.8%
  Rating [less than or
    equal to] 3               3.8%           9.6%          13.8%
  Rating [less than or
    equal to] 4               5.0%          10.0%          16.2%
CONCERN
  Rating = 8                  1.2%           3.5%           3.5%
  Rating [greater than
    or equal to] 5            5.4%          10.8%          15.8%
RE PROBLEM                    5.5%           9.4%          12.6%

                                  LATENCY

                          [less than
                          or equal to]
RE criiterion               4 minute     Overall

LATENCY
  [less than or
    equal to] 1 minute                     6.6%
  [less than or
    equal to] 2 minutes                   16.6%
  [less than or
    equal to] 3 minutes                   26.6%
  [less than or
    equal to] 4 minutes                   31.9%
  [less than or
    equal to] 7 minutes                   60.9%
CONTROL
  Rating = 1                  3.5%         3.8%
  Rating [less than or
    equal to] 2              14.6%        16.1%
  Rating [less than or
    equal to] 3              23.5%        30.0%
  Rating [less than or
    equal to] 4              31.5%        43.8%
CONCERN
  Rating = 8                  5.0%        5.8%
  Rating [greater than
    or equal to] 5           34.2%        51.5%
RE PROBLEM                   20.1%        23.6%

Note. N = 254-260. LATENCY = ejaculatory latency following vaginal
intromission in minutes; CONTROL = control over the occurrence of first
ejaculation during sexual intercourse (1 = no control, 8 = absolute
control); CONCERN = concern over ejaculating sooner than desired
(1 = no concern, 8 = extreme concern); RE PROBLEM = self-identification
of having a current problem because ejaculation occurs before desired.
Table 3. Zero-Order Correlations Amongst Rapid Ejaculation Criteria

RE criterion         1          2          3          4          5

1. LATENCY
2. CONTROL         .35 ***
3. CONCERN        -.09       -.26 ***
4. SATISFACTION    .32 ***    .58 ***   -.53 ***
5. % RAPID        -.38 ***   -.52 ***    .55 ***   -.60 ***
6. % DELAY         .16 **     .20 ***    .35 ***    .04       .15 *
7. % ANTIPORTAL   -.13 *     -.13 *      .21 ***   -.13 *     .25 ***
8. RE PROBLEM     -.30 ***   -.42 ***    .51 ***   -.49 ***   .67 ***

RE criterion        6         7

1. LATENCY
2. CONTROL
3. CONCERN
4. SATISFACTION
5. % RAPID
6. % DELAY
7. % ANTIPORTAL   .18 **
8. RE PROBLEM     .19 **   .21 ***

Note. N = 254 - 260. LATENCY = ejaculatory latency following vaginal
intromission; CONTROL = control over the occurrence of first
ejaculation during sexual intercourse; CONCERN = concern over
ejaculating sooner or faster than desired; SATISFACTION = satisfaction
with ability to select the moment of ejaculation during sexual
intercourse; % RAPID = percentage of intercourse experiences in which
ejaculation occurs sooner than desired; % DELAY = Percentage of
intercourse experiences where specific attempts are made to prolong
intercourse by delaying ejaculation; % ANTIPORTAL = Percentage of
intercourse experiences in which ejaculation occurs before intercourse
has started; RE PROBLEM = self-identification of having a problem
because ejaculation occurs before desired.

* p < .05.

** p < .01.

*** p < .001.
Table 4. Comparisons of Men With and Without a
Self-identified RE Problem on Seven Continuous
RE Criteria

Criterion      RE PROBLEM   NO RE PROBLEM   [F.sub.univ]

LATENCY            4.6           9.1          25.6 ***
CONTROL            3.3           5.0          53.5 ***
CONCERN            6.4           3.7          89.4 ***
SATISFACTION       3.1           5.3          80.2 ***
% RAPID           69.7          24.7         202.9 ***
% DELAY           62.7          49.5           9.3 **
% ANTIPORTAL      10.9           3.9          11.9 ***

Note. [F.sub.mult] (7,245) = 828.38, p < .001. N = 253. LATENCY =
ejaculatory latency following vaginal intromission in minutes;
CONTROL = control over the occurrence of first ejaculation during
sexual intercourse; CONCERN = concern over ejaculating sooner than
desired; SATISFACTION = satisfaction with ability to select the moment
of ejaculation during sexual intercourse; % RAPID = percentage of
intercourse experiences in which ejaculation occurs sooner than
desired; % DELAY = percentage of intercourse experiences in which
specific attempts are made to prolong intercourse by delaying
ejaculation; % ANTIPORTAL = percentage of intercourse experiences in
which ejaculation occurs before intercourse has started; RE
PROBLEM = self-identification of having a current problem because
ejaculation occurs before desired.

** p < .01.

*** p < .001.
Table 5. Canonical Correlation Analysis Between Variables
Etiologically Linked With RE and the RE Criteria

Variable set                                 r (a)

Etiological variable set
  Age                                         .16
  Frequency of intercourse                    .83
  Number of sexual partners                   .25
  Feeling rushed during early intercourse    -.29
  Negative affect during early intercourse   -.13
RE criteria set
  LATENCY                                     .38
  CONTROL                                     .51
  CONCERN                                    -.83
  SATISFACTION                                .60
  % RAPID                                    -.72
  % DELAY                                    -.44
  % ANTIPORTAL                                .02
  RE PROBLEM                                 -.57

Note. N = 251. [R.sub.c] = .41, F(40, 1185) = 1.97, p < .001.
LATENCY = ejaculatory latency following vaginal intromission in
minutes; CONTROL = control over the occurrence of first ejaculation
during sexual intercourse; CONCERN = concern over ejaculating sooner
or faster than desired; SATISFACTION = satisfaction with ability to
select the moment of ejaculation during sexual intercourse; %
RAPID = percentage of intercourse experiences in which ejaculation
occurs sooner than desired; % DELAY = percentage of intercourse
experiences in which specific attempts are made to prolong intercourse
by delaying ejaculation; % ANTIPORTAL = percentage of intercourse
experiences in which ejaculation occurs before intercourse has
started; RE PROBLEM = self-identification of having a current problem
because ejaculation occurs before desired.

(a) Correlation with the canonical variate.

(1) The privacy envelopes were used to ensure that the members of the couple completed their questionnaires independently. Only the results of the questionnaires completed by the men are reported here.

REFERENCES

American Psychiatric Association. (1994). Diagnostic and statistical manual of mental disorders (4th ed.). Washington, DC: Author.

Bogaert, A. F. (1996). Volunteer bias in human sexuality research: Evidence for both sexuality and personality differences in males. Archives of Sexual Behavior, 25, 125-140.

Coles, R., & Stokes, G. (1985). Sex and the American teenager. New York: Harper.

Colpi, G. M., Fanciullacci, F., Beretta, G,, Negri, L., & Zanollo, A. (1986). Evoked sacral potentials in subjects with true premature ejaculation. Andrologia, 18, 583-586.

Crooks, R., & Baur, K. (1993). Our sexuality (5th ed.). Don Mills, Canada: Benjamin/Cummings.

Ende, J., Rockwell, S., & Glasgow, M. (1986). The sexual history in general medicine practice. Archives of Internal Medicine, 144, 558-561.

Fanciullacci, F., Colpi, G. M., Beretta, G., & Zanollo, A. (1988). Cortical evoked potentials in subjects with true premature ejaculation. Andrologyia, 20, 326-330.

Gospodinoff, M. L. (1989). Premature ejaculation: Clinical subgroups and etiology. Journal of Sex & Marital Therapy, 15, 130-134.

Grenier, G., & Byers, E. S. (1995). Rapid ejaculation: A review of conceptual, etiological and treatment issues. The Archives of Sexual Behavior, 24, 447-472.

Grenier, G., & Byers, E. S. (1997). The relationship among ejaculatory control, ejaculatory latency, and attempts to prolong heterosexual intercourse. Archives of Sexual Behavior, 26, 27-47.

Kameya, Y., Deguchi, A., & Yokota, Y., (1997). Analysis of measured values of ejaculation time in healthy males. Journal of Sex and Marital Therapy, 23, 25-28.

Kaplan, H. S. (1989). RE: How to overcome premature ejaculation. New York: Brunner/Mazel.

Kinsey, A. C., Pomeroy, W. B., & Martin, C. E. (1948). Sexual behaviour in the human male. Philadelphia: W. B. Saunders Co.

Laumann, E. O., Gagnon, J. H., Michael, R. T., & Michaels, S. (1994). The social organization of sexuality: Sexual practices in the United States. Chicago: University of Chicago Press.

Laumann, E. O., Paik, A., & Rosen, R. C. (1999). Sexual dysfunction in the United States: Prevalence and predictors. Journal of the American Medical Association, 281, 537-544.

Masters, W. H., & Johnson, V. E. (1970). Human sexual inadequacy. Boston: Little, Brown & Co.

Masters, W. H., Johnson, V. E., & Kolodny, R. C. (1986). Masters and Johnson on sex and human loving. Boston: Little, Brown & Co.

McCarthy, B. (1988a). Male sexual awareness: Increasing sexual satisfaction. New York: Carroll & Graf.

McCarthy, B. (1988b).Cognitive-behavioral strategies and techniques in the treatment of early ejaculation. In S. R Leiblum & R. C. Rosen (Eds.), Principles and practices of sex therapy: Update for the 1990s (pp. 141-167). New York: Gilford Press.

Metz, M. E., Pryor, J. L., Nesvacil, L. J., Abuzzhab, F., & Koznor, J. (1997). Premature ejaculation: A psychophysiological review. Journal of Sex and Marital Therapy, 23, 3-23.

Metz, M, E., & Seifert, M. (1990). Men's expectations of physicians in sexual health concerns. Journal of Sex and Marital Therapy, 16, 79-89.

Reinisch, J. M., & Beasley, R. (1990). The Kinsey Institute new report on sex: What you must know to be sexually literate. New York: St. Martin's Press.

Rowland, D. L., Cooper, S. E., Slob, A. K., & Houtsmuller, E. J. (1997). The study of ejaculatory response in the psychophysiology laboratory. The Journal of Sex Research, 34, 161-166.

Rowland, D. L., & Slob, A. K. (1997). Premature ejaculation: Psychophysiological considerations in theory, research and treatment. The Annual Review of Sex Research, 8, 224-253.

Rowland, D. L., Strassberg, D. S., de Gouveia Brazao, C. A., & Slob, A. K. (2000). Ejaculatory latency and control in men with premature ejaculation: An analysis across sexual activities using multiple sources of information. Journal of Psychosomatic Research, 48, 69-77.

Ruff, G. A., & St. Lawrence, J. S. (1985). Premature ejaculation: Past research progress, future directions. Clinical Psychology Review, 5, 627-639.

Smeaton, G., & Byrne, D. (1998). The feelings scale: Positive and negative affective responses. In C. M. Davis, W. L. Yarber, R. Bauserman, G. Schreer, & S. L. Davis (Eds). Handbook of sexuality-related measures (pp. 50-52). Thousand Oaks, CA: Sage.

Spiess, W. F., Geer, J. H., & O'Donohue, W. T (1984). Premature ejaculation: Investigation of factors in ejaculatory latency. Journal of Abnormal Psychology, 93, 242-245.

Strassberg, D. S., Kelly, M. P., Carroll, C., & Kircher, J. C. (1987). The psychophysiological nature of premature ejaculation. Archives of Sexual Behavior, 16, 327-336.

Strassberg, D. S., & Lowe, K. (1995). Volunteer bias in sexuality research. Archives of Sexual Behavior, 24, 369-382.

Strassberg, D. S., Mahoney, J. M., Schaugaard, M., & Hale, V. E. (1990). The role of anxiety in premature ejaculation: A psychophysiological model. Archives of Sexual Behavior, 19, 251-258.

Tabachnick, B., & Fidell, L. (1989). Using multivariate statistics (2nd ed.). New York: Harper and Row.

Williams, W. (1984). Secondary premature ejaculation. Australian and New Zealand Journal of Psychiatry, 18, 333-340.

Zilbergeld, B. (1999). The new male sexuality (rev. ed.). New York: Bantam.

Manuscript accepted October 3, 2001

This research was the conducted as the doctoral dissertation of the first author under the supervision of the second author. The authors would like to thank Suzanne Lemieux and Lisa Price for their help in the design of this study and David Rowland for his helpful comments on an earlier draft of the manuscript.

Address correspondence to E. Sandra Byers, Ph.D., Department of Psychology, University of New Brunswick, Fredericton, New Brunswick, Canada E3B 6E4; e-mail: [e-mail].