Issues in the Laboratory Study of Human Sexual Response: A Synthesis for the Nontechnical Sexologist.

Laboratory-based studies on sexual response represent a significant part of the current volume of research on sexuality. Such studies share the common goal of measuring sexual arousal and response as the dependent variable to any number of possible independent, predictor, or correlated variables within a controlled laboratory setting. Because these studies are often limited in scope and utilize procedures unfamiliar to many students of sexology, findings may sometimes seem tangential to mainstream issues of sexuality. Yet, these studies have enabled the testing of various theoretical perspectives pertaining to sexual dysfunctions and deviancies, and it behooves the contemporary sexologist not only to understand the value of such research, but also to become a critical consumer of it. In the following pages, I present an overview of advantages and drawbacks of laboratory based sex research so that nonlab sexologists can develop an appreciation of it. Special attention is paid to some issues currently faced by laboratory researchers of sexuality.

ROOTS OF THE LABORATORY STUDY OF HUMAN SEXUALITY

Although many believe that the laboratory study of human sexuality began with the groundbreaking work of Masters and Johnson (1966), historical documents suggest that isolated studies on this topic were carried out as early as the beginning of the twentieth century. These studies were aimed at describing physiological correlates of sexual response, and most took advantage of the developing technology that enabled measurement of autonomic responses such as heart rate and EEG during sexual stimulation (e.g., Mosovich & Tallafero, 1954). Among these early investigations was the covert work of the behaviorist John B. Watson (Magoun, 1981). Watson's work on sexuality mired him in deep trouble: As if the topic of his research was not shocking enough, there were hints that Watson carried out the research with the involvement of his female assistant (McConnell, 1974). Not surprisingly, Watson's wife filed for divorce, and Watson himself was forced to resign his academic position and move into the world of advertising. From the perspective of sexology, data that might have helped usher in a new era of sex research was lost to the scandal.

It took the pioneering efforts of Masters and Johnson (1966) to lend legitimacy to the idea that sexual response could and should be subjected to laboratory analysis. Their bold studies, begun in the 1950s and first published in 1966, provided basic information not only about the sexual response cycle but also about creative ways to measure it. Equally important, their studies paved the way for a growing number of scholars interested in studying sexuality from the perspective of a natural science, one that emphasizes precise measurement of physiological responses and control over independent and/or extraneous variables. From those seminal studies emerged the field of sexual psychophysiology, a field that has since undergone a number of conceptual and methodological transformations. In this paper, I provide an assessment of the current state of the laboratory investigation of sexual response.

SCOPE OF LABORATORY INVESTIGATIONS

Description of Functional and Dysfunctional Sexual Response

Early laboratory studies of human sexual response were primarily descriptive in nature and attempted to characterize physiological changes associated with self-induced sexual arousal or orgasm. In these studies, measurement of nongenital autonomic responses that generated bioelectric signals (Davis & Buchwald, 1957: EEG, electrodermal response, heart rate, EMG) was sometimes supplemented with the direct observation of the subject by the investigator. Following the comprehensive work of Masters and Johnson (1966), precise measurement of genital response became more feasible as instrumentation was developed that could respond to the shape and mechanical changes of the genitals during arousal. In fact, because remote genital-measuring devices eliminate the need for direct observation of the subject by the researcher--a situation that might influence sexual response--there is strong interest in continued refinement of monitoring devices which afford privacy, precision, and ease.

Long after the publication of Masters and Johnson's (1966) initial research, basic descriptive analysis of sexual arousal and response still continues to be an important goal of sexual psychophysiological research (e.g., Meston & Gorzalka, 1996; Rowland, Cooper, Houtsmuller, & Slob, 1997). Improved instrumentation for assessing physiological response, in conjunction with psychological scales that measure the subject's perceptions and affects during arousal, has expanded the view of sexual response to include multiple and interrelated dimensions. For example, a number of studies have demonstrated that in normal (as opposed to deviant) sexually-functional male and female subjects, increased genital response and sexual arousal is typically associated with increased positive affect (Barlow, 1986; Heiman, 1980; Koukounas & McCabe, 1997; Morokoff & Heiman, 1980; Rowland, Cooper, & Slob, 1996).

As the understanding of normal sexual response in men and women has increased, the interest in understanding how it contrasts with dysjunctional sexual response (i.e., lack of ability to perform adequately) has also grown. As a result, quasi-experimental studies in which the response patterns of functional subjects are compared with those of dysfunctional subjects have become increasingly common. We now know, for example, that dysfunctional women show less subjective and autonomic arousal to erotic cues, have greater difficulty labeling their genital cues, and give self-ratings of arousal tied less to bodily cues than their functional counterparts (Morokoff & Heiman, 1980; Palace & Gorzalka, 1992). Men with erection problems not only show inhibited genital response to sexual stimulation but also respond with higher negative affect, less sensuality, and more self-monitoring of their genital response than sexually functional men (Rowland, Cooper, & Heiman, 1995; Rowland & Heiman, 1991).

Experimental Analysis of Sexual Response

Important to understanding any behavioral response is the specification of factors that can account for its variation. Both within and among sexually functional subjects, wide variation in sexual response is typical. Some individuals are more erotophilic than others; some stimulus conditions enhance arousal, others interfere with it; and various conditions may alter the relationships among the physiological, perceptive, and affective components of the sexual response. Therefore, a second goal of the laboratory investigation of sexual response has been to identify factors that influence sexual arousal and to explain how such factors might operate. For example, mental tasks that distract attention from the sexual stimulus and, consequently, from genital response, appear to attenuate sexual arousal (Elliott & O'Donohue, 1997; Geer & Fuhr, 1976), while the demand to perform sexually, induced by a set of instructions to the subject, seems to enhance it (Heiman & Rowland, 1983). Although such findings may not be surprising, they represent tests of hypothetical factors (e.g., excessive self-monitoring and performance demand) presumed to account for inhibited sexual response in men with psychogenic impotence (Kaplan, 1974; Masters & Johnson, 1970). When these same stimulus conditions are manipulated in men with erectile dysfunction, opposite patterns appear--moderate distraction from the immediate situation facilitates arousal, whereas demand for successful sexual performance disrupts it (see Abrahamson, Barlow, & Abrahamson, 1989; Cranston-Cuebas & Barlow, 1990; Heiman & Rowland, 1983). Because response patterns may differ between functional and dysfunctional subjects, many investigators have included subjects from both populations as they attempt to specify conditions that influence sexual arousal and response. Such investigations are, of course, not restricted to the study of psychological factors, but may explore the biological domain as well (e.g., using pharmacological agents, altering the sensitivity of sensory receptors, etc.).

Laboratory Investigation as a Diagnostic Aid for Dysfunctional Sexual Response

While much of the effort of psychophysiological investigations has focused on a basic understanding of sexual response, within clinical circles psychophysiological procedures have gained moderate acceptance as an aid in the differential diagnosis of sexual dysfunction in men (Rowland & Slob, 1995; Slob, Blom, & van der Werff ten Bosch, 1990) for two reasons. First, as part of a sexual diagnosis the laboratory approach may be helpful in establishing whether penile erection can be provoked under any stimulus conditions. Second, the patient's self-assessment of somatic (sexual) response through an interview or questionnaire may be biased, or sometimes overly optimistic or pessimistic (for a discussion of this more general issue, see Moss-Morris & Petrie, 1997). Therefore, laboratory procedures that permit control over the type and intensity of sexual stimulation and firsthand monitoring of sexual response may serve as additional strategies for assessing maximal potency of the patient (see Rowland & Slob, 1995). Although a lack of genital response in these situations yields no helpful information for the clinician, a positive response reduces the likelihood of using an unnecessary or inappropriate treatment strategy for that individual (e.g., use of a biological approach when a psychogenic dysfunction stems from the dynamics of the dyadic relationship).

Although the clinical use of lab procedures emphasizes diagnostic efficacy, this "applied" endpoint is usually combined with a research orientation. Laboratory protocols may be designed with the goal of collecting information that improves the understanding not only of diagnostic procedures but also of the dysfunctional sexual response (e.g., Incrocci, Hop, & Slob, 1996). Unfortunately, attempts to apply this strategy to differentiating types of sexual dysfunction in women (e.g., low sex desire or vaginismus) have been limited. The reasons may be many, ranging from a general lack of attention to the woman's sexual response to a reduced need to rely on nonverbal measures for adequate assessment of women's sexual problems.

Assessment and Treatment of Sexual Deviancies

A sizable body of laboratory research has been carried out with the purpose of understanding men's and women's reactions to sexually inappropriate stimuli (e.g., children, coercive sex: see Malamuth & Check, 1980). Research has also identified and studied men with deviant sexual behaviors, or paraphilias. These include problems such as fetishism, pedophilia, and sadomasochistic behaviors. Central to their assessment is the assumption that these deviations represent responses (perhaps conditioned) evoked by specific kinds of stimuli. Furthermore, emphasis is often placed on genital response, the assumption being that because this is a reflexive response the unconstrained penis will not lie. As a hypothetical example, the pedophile shows sexual interest in children or adolescents, so thoughts or images of youngsters would presumably elicit a different genital response in pedophiles than in controls. Substantial evidence argues that these laboratory procedures can indeed differentiate men with certain disorders from controls (Abel, Barlow, Blanchard, & Guild, 1977; Barbaree, Marshall, & Lanthier, 1979) as well as offenders with differing magnitudes of the disorder (e.g., violent rapists from other rapists: see Quinsey & Chaplin, 1984).

Given the stimulus-response (S-R) model that presumably forms the basis of sexual deviations, some laboratory treatments have been aimed at altering this S-R connection. For example, aversive and counter conditioning of the undesirable response, satiation of the response to the deviant stimuli, and reinforcement of desirable alternative responses have all been attempted in the laboratory with limited degrees of success (Abel, Osborn, Anthony, & Gardos, 1992; Rosen & Beck, 1988). However, because sexual deviancy entails complex medical, legal, and moral questions which are unresolved, and because the degree to which laboratory results mirror real-world events has not been firmly established, laboratory-based assessment and treatment of these problems have met with significant controversy (see McAnulty & Adams, 1992; McConaghy, 1989).

THE LABORATORY ADVANTAGE

In a typical sexual psychophysiological study, a subject is presented with an erotic stimulus in a quiet, private setting. During stimulation, physiological responses (including genital) are monitored using remote-measuring devices, and afterwards the individual's perceptions, cognitions, and feelings about the experience may be assessed. There are multiple reasons for applying this paradigm to the study of sexual response. Some are general and pertain to virtually all types of laboratory research on sexual response, whereas others are limited to studies of sexual dysfunction or deviancies as described above. Four advantages of this approach stand out, and each is discussed briefly.

First, laboratory analysis enables rigorous control over sexual stimulus parameters, those intended for the study of functional sexual response as well as those used in the study of dysfunctional or deviant sexual response. By using controlled audio/visual sexual stimulation, the subject's attention can be focused so that variation in cognitive content during sexual arousal is limited. In addition, stimuli can be tailored to match the particular population or response under investigation. For example, in a study where the experimenter wants to identify situational variables that enhance sexual response, stimuli that induce submaximal genital arousal and subjective arousal (which allow for an increase in response) may be chosen (Heiman & Rowland, 1983). In contrast, when a strong response is desirable, as in the clinical diagnosis of erectile dysfunction, intense visual stimuli supplemented with genital stimulation may be more appropriate (Rowland, den Ouden, & Slob, 1994). Stimulus conditions may be further honed to reflect the differing values that men and women place on various kinds of sexual interactions. For example, women generally find scenes involving mutual attraction between sexual partners and prolonged foreplay more subjectively arousing than ones involving anonymous sex with strong emphasis on intercourse (Laan, Everaerd, van Bellen, & Hanewald, 1994). In the case of deviant sexual response, the content of the erotica may be focused on those specific stimulus characteristics most associated with the deviancy and therefore most likely to elicit arousal.

Beyond the actual sexual stimuli, the role of other stimuli relevant to sexual response can be investigated by introducing new variables into the situation in a controlled manner. For example, the effect of a presumed therapeutic drug might be evaluated, or the relationship between sexual arousal and psychological factors such as cognitive distraction (Elliott & O'Donohue, 1997) or a particular emotional state (anxiety or mood: Beck, Barlow, Sakheim, & Abrahamson, 1987; Beggs, Calhoun, & Wolchik, 1987; Palace & Gorzalka, 1992; Mitchell, DiBartolo, Brown, & Barlow, 1998) might be determined. In each case, the experimenter can manipulate the variable of interest, verify the efficacy of the manipulation, and measure changes in the response. If, for example, the goal is to assess the role of coercion on sexual arousal (as might be used in assessing arousal in men convicted of rape), stimulation involving a nonconsensual activity might be superimposed on the sexual scenario (Lohr, Adams, & Davis, 1997).

Second, laboratory analysis permits measurement of covert (autonomic) responses during arousal which would otherwise go undetected by either the subject or experimenter These autonomic measures of arousal signal activation of a reflexive response system that may not be apparent from the subject's overt response or verbal report. Since these autonomic responses are difficult to detect with the unaided senses, psychophysiological instrumentation is used to amplify and modify the biological signals that underlie them (e.g., heart rate, pulse amplitude, electrodermal response, and EMG). They then can be recorded and expressed in a standard unit of measurement that allows comparison across subjects, or within subjects across situations. Even when the subject may be aware of aspects of their physiological sexual response (e.g., vaginal wetness or penile erection), quantification of these responses without some genital yardstick is difficult. The use of measurement devices that monitor and amplify these responses ensures precise quantification of genital response.

Third, laboratory-based research provides a very reliable way of measuring the physiological component of sexual response. Reliability is improved in several ways through lab-based research. Within the lab setting, instrumentation designed to measure covert or subtle changes in responsivity allows for precise, sensitive, and standardized measurement of the response variable. Thus, for a genital response of a given magnitude, repeated samplings within or across individuals will yield consistent data.

Equally important, lab-generated information about sexual response is potentially more reliable than information obtained through self-report strategies. In the laboratory, the experimenter knows exactly what parameters are being measured, how they are measured, and what other factors might impact upon those measures. Compare this with a subject's verbal response to the question posed within an interview asking "How aroused were you during your last sexual activity?" Here the experimenter must not only assume the subject's motivation to answer honestly, but also must rely on the subject's ability to recall the events clearly and to assess those events accurately. Assessment of objective (genital) sexual response in the laboratory is influenced less by these factors. Even subjective assessment of sexual response made by the participant under lab conditions may be more accurate, because issues of recall are minimized and distortion arising from specific motivations (conscious or not) or intentions is likely to be detected. A subject that reports an absence of "subjective" arousal in the face of a strong genital response (or vice versa) is likely to be questioned further by the researcher. This latter consequence may be particularly salient in persons exhibiting sexual dysfunctions or deviancies, since these individuals sometimes have cause to misrepresent or misperceive aspects of their sexual response.

Finally, laboratory investigations create an in situ sexual situation which provides the opportunity to integrate the multiple components of sexual response. In addition to the physical component, sexual response includes perceptual, affective, and cognitive dimensions; the interactions among these dimensions may not be clear from the subject's anamnestic reports of past sexual experiences. They may, however, become apparent when these dimensions are tapped simultaneously during an actual sexual experience. The psychophysiological laboratory provides just such an opportunity to integrate various dimensions of sexual response (e.g., Beck & Barlow, 1986; Bozman & Beck, 1991)--this laboratory "snapshot" of sexual response may well surpass any insight that could be garnered through the subject's own verbal reconstruction of the experience.

In sum, laboratory analysis offers important advantages in the study of sexual arousal, dysfunction, and deviancy. It can address questions about sexuality that cannot be readily tested using alternate research strategies. Nevertheless, laboratory investigations should not substitute for other research procedures investigating sexual response (diagnostic procedures, self-report, etc.); rather, laboratory investigations should complement other procedures.

ISSUES SURROUNDING THE LABORATORY STUDY OF SEXUAL RESPONSE

The study of sexual response in the laboratory has its drawbacks. For convenience, these have been grouped into three broad categories: factors external to the laboratory which influence the efficacy of the studies, conceptual and methodological problems arising from the laboratory studies themselves, and issues regarding the predictive and external validity of results obtained from the laboratory.

External Factors

Political and moral considerations. Although the legitimacy of sexual psychophysiological studies has been accepted by sexologists and much of the general scientific community, the judgment of the public-at-large appears ambivalent. On the one hand, lab studies may project a more scientific aura than survey or interview studies and therefore may be afforded more credibility by the press and the public. On the other hand, among some religions there are prohibitions against nonprocreative sexual activity, and therefore, unless psychophysiological procedures can be justified as part of a diagnostic routine to aid successful reproduction, they may be viewed with skepticism or rejected as immoral. While few religions would contest the noble goals of this research, some would clearly find fault with methodology involving erotic stimulation and arousal to orgasm outside a procreative marital relationship.

Issues of religion aside, lab-based studies of sexuality are viewed by many individuals within the civic or academic community as unnecessary or unwarranted (e.g., Holden, 1976; Laumann, Gagnon, Michael, & Michaels, 1994). Although objections are varied, comments culled from letters to this author include the idea that bedroom activities should not be subjected to scientific scrutiny, or that such studies are prurient, distasteful, or immoral. In the United States, such attitudes do not go unnoticed by politicians, who must spend significant effort appeasing interest groups with a specific political or moral agenda. Because of its potential liability, sex research that incorporates pornography and actual sexual acts as part of its research protocol will receive greater scrutiny by politicians than research of a less controversial nature, even though the design of the sex research may be sound and the health objectives both important and achievable. Of course, attitudes toward sexuality differ across cultural and national boundaries, so obstacles encountered in one locale may be quite different from those encountered in another.

Although the impact of various religious/attitudinal views on lab-based sex research has not been investigated in any large scale study, several consequences are probable: (a) Where research topics or methodologies conflict with the ideologies of significant segments of society, research support is likely to be limited and research findings met with skepticism by a substantial portion of the public and (b) because men and women holding an opposing moral or attitudinal view are unlikely to participate in research studies of this type, a subject selection bias is inevitable. This bias, which limits the impact of sexual psychophysiology, may not be readily overcome and creates the larger problem of generalization of results to the population at large.

The need for interdisciplinary collaborative efforts. Psychophysiological research aimed at understanding sexual response and its disorders requires, by definition, an interdisciplinary framework. Lab-based research on both men and women has origins in two disciplines: psychology (psychophysiology) and the biomedical sciences, and there is need for collaboration between these two schools at both the conceptual and practical levels. Not only does each discipline provide an important perspective in defining the relevant questions, but in order to include the type of subjects that will yield the most valid results, a research program must also have access to individuals with sexual problems. Subjects are most easily recruited not just from psychology clinics, but from clinics located in psychiatry, urology, endocrinology, and gynecology departments of medical institutions. However, for a variety of reasons these cooperative arrangements seem difficult to establish and maintain. Departmental economics, physical distance, and divergent objectives often discourage cooperation. In addition, the research goals of the psychophysiologist and the medical staff are often quite disparate. Whereas the psychophysiologist may be interested in a variety of basic questions about sexual response, the physician is typically concerned with developing procedures that will improve treatment efficacy. As a result, physicians are frequently reluctant to press their patients to participate in psychophysiological studies, particularly when the procedure has the potential for psychological discomfort and when it offers little or no guarantee of an immediate benefit to the patient. Although such concerns may stand contrary to the immediate pressures on the medical profession (Tiefer, 1996) and thus may impede collaborative efforts, they do not always present insurmountable problems. A number of cooperative efforts have emerged over the past decades, and these may serve as models for other attempts at studying sexual response from an interdisciplinary perspective.

Without such collaborations, lab-based research loses much of its efficacy. Studies restricted to healthy, sexually functional men and women (as might occur in a program housed in a psychology department) may not provide much insight into clinical issues of diagnosis and treatment. And, vice versa, studies carried out by clinicians in medical settings without the benefit of a psychophysiologist may suffer from a lack of attention to the methodological and conceptual issues relevant to psychophysiological and psychological measurement.

Time and cost. Compared with (or in addition to) interview techniques or questionnaires, lab procedures designed to obtain information about sexual response require significant labor, time, and space. Lab research places a heavy demand on research or clinical personnel, stemming from the fact that multiple subjects cannot be run concomitantly and, further, that each subject requires the dedicated attention of the researcher. There is also a substantial time investment required of the research participant, and because participants with significant sexual experience or with sexual dysfunctions or deviations will yield more valid tests of hypotheses regarding problematic sexual response, extra effort and expense must be devoted to the recruitment and compensation of subjects drawn from these populations. A practical consequence of the high time and cost investment is that sample sizes for this type of research end up being relatively small, typically averaging between 10 and 20 subjects per group. Although this by itself may not be a problem since fewer subjects are usually required for these kinds of studies, within dysfunctional or deviant groups there may already exist considerable heterogeneity. For example, a group of 15 premature ejaculators may consist of several subsets of men, some with lifelong premature ejaculation, some with coexisting erection problems, and some with situational premature ejaculation. Sample sizes that are initially small, in conjunction with attenuated effect sizes from heterogeneity within groups, reduce statistical power and increase the odds of missing real effects, a genuine liability given the sizable investment of resources.

Beyond these practical concerns, however, there is a more pressing issue that needs to be addressed. As the support for research funds continues to be highly competitive, premium is placed on the efficiency of investigatory procedures. Given the resources required for lab research, it is appropriate to ask whether lab-based research on sexuality is worth the funds spent. That is, might these same ends be achieved with less costly and, according to some, less objectionable research procedures?

While a cost-benefit analysis is not among the goals of this paper, examples of obvious benefits can be (and have been) identified (see Rosen & Beck, 1988). The demand of laboratory research for rigorous procedures, tight definitions, and precise measurements has undoubtedly hastened progress in a field that lags behind many other subfields of psychology and physiology. Aside from enabling a basic understanding of the mind-body mechanics of sexual response, sexual psychophysiology has generally sharpened the conceptualization of sexual response, and sexual dysfunction and deviancy in particular. For example, anxiety has long been assumed to play a causal role in psychogenic erection failure and premature ejaculation (Kaplan, 1974). But lab-based research has been instrumental in recognizing the multiple ways anxiety can be conceptualized, and for generating operational definitions of this construct. Furthermore, such lab studies have shown that different sexual dysfunctions may be characterized by quite different affective profiles (Barlow, 1986; Rowland, Cooper, & Heiman, 1995; Rowland, Cooper, & Slob, 1996) and have addressed the multiple and complex ways an emotional state may interact with the sexual experience to either enhance or inhibit arousal (Rowland, 1996). As a result, lab studies have spurred the conceptualization of sexual response through several stages of evolution: These conceptual refinements would probably not have resulted solely from clinically based research and/or interview research.

From a more clinical perspective, lab-based studies aimed at diagnosis have demonstrated the critical role that psychological factors play in the patient's response to penile injection of vasoactive substances (e.g., papaverine, PG2) for inducing erection (Slob, Rowland, Blom, & van der Werff ten Bosch, 1991). These clinical procedures have undoubtedly spared a number of patients from more invasive diagnostic procedures aimed at pinpointing causes of erectile failure (Slob et al., 1990). And procedures aimed at identifying deviant patterns of arousal have been useful in specifying stimulus factors associated with paraphilias as well as testing treatment procedures targeted at this population (see Simon & Schouten, 1991, for an overview). In sum, given that the psychophysiological study of sex is yet in its infancy, its contribution to the understanding of sexuality appears both promising and unique.

Conceptual and Methodological Issues From the Laboratory Session

Overcoming extraneous variation generated by the laboratory situation. For most people, sexual response is a very private matter, and potential lab participants seem to respond quite differently to the idea that their response will be monitored by the experimenter. For some, the clinical setting of the laboratory, along with the "white coat" staff, seems to have minimal effect on sexual response. In fact, the permissive clinical atmosphere of the lab may actually diminish feelings of guilt or depravity associated with erotica and thus may enhance sexual response. For other participants, the knowledge of being observed undoubtedly inhibits sexual arousal. Unfortunately, it is not yet possible to specify the extent to which the laboratory setting enhances or inhibits arousal, or why individual subjects respond differently. I know of no studies that have systematically queried subjects about perceived differences between their laboratory sexual response and their bedroom response, or that have attempted to identify factors that differentiate individuals who feel inhibited in the lab from those who are either unaffected or even disinhibited.

Participants also react differently to various types of sexual stimulation. Based on our own research involving hundreds of men, some individuals have little or no experience with the kinds of erotic stimuli used in sexual psychophysiological research. For these people, the novelty of the sexual stimuli may induce strong arousal. On the other hand, these subjects might just as easily react with repulsion or disdain. In contrast, persons having positive experiences with erotica usually show some interest and enjoyment in the stimuli and are likely to find them sexually arousing. However, even these subjects may show surprising results. Based on exit interviews with participants, we have found that a subset of volunteers, because of their extensive histories with erotica, exhibit little or no arousal due to the purportedly mild effects of the lab stimuli.

Add to the above extraneous factors the strong demand characteristics of participation in a laboratory study, the use of invasive devices and the physical restraint they necessitate, concerns about a hygienic environment (e.g., sterilized measuring devices), and unfamiliar staff, it is a wonder that the effects of any manipulated variable can be detected. Furthermore, such extraneous factors may generate opposite effects on different types of subjects. Sexually functional men, for example, tend to respond with greater arousal under demand conditions for arousal, whereas dysfunctional men respond with inhibition (Heiman & Rowland, 1983); in women, response patterns to demands for sexual arousal and/or suppression also appear to be related to specific strategies that can be predicted by identifiable group characteristics (Beck & Baldwin, 1994).

Researchers attempt to deal with lab factors, including demand characteristics, in a variety of ways. To moderate anxiety, subjects are carefully briefed as to what to expect and are allowed prior inspection of the lab and measuring devices. The experiments are conducted in a setting where the sense of privacy is maximized and confidentiality is guaranteed. Subjects are usually assured that a wide range of responses to the stimuli is normal. In some experiments, a trial session is incorporated into the research protocol so participants can adapt to lab procedures, measuring devices, and erotic stimulus materials. Such procedures help reduce lab anxiety that might interfere with sexual arousal. To deal with individual reactions to erotica, subjects may be questioned about their prior experience with sexual materials and, after the recording session, may be asked to react to the specific erotic stimuli used in the study (e.g., rated on a pleasant-unpleasant continuum). Variation from these measures can then be partialled out when comparisons are made between various treatment conditions or groups.

The limitations of sexual stimulation in the lab. Sexual response, including dysfunctional and deviant response, usually represents a series of dynamic interactions between two people, interactions which act to sustain and enhance the response. This dyadic aspect of sexuality is not readily explored in the psychophysiological lab, and this is the major limitation of most laboratory-based sex research today. Sexual response to a partner is really a composite of many psychological and physiological responses. These include the person's responses to the sensory stimuli, including body position, smell, touch, and visual and auditory input. The responses involve knowledge about one's sexual partner, the previous emotions (positive or negative) associated with that partner or the sexual situation, and the imaginative (or fantasy) cognitions that accompany the act. In this respect, the kinds of one-dimensional stimuli offered in the lab represent a poor substitute for the normally rich experience of sexuality with one's sexual partner. Nevertheless, even in this isolated context the erotic stimuli used in laboratory studies are surprisingly effective in their capacity to elicit sexual arousal.

Erotic stimulation used in psychophysiological research is most often audio-visual, consisting of segments drawn from videotapes produced commercially for entertainment or educational purposes. In some studies, vibrotactile stimulation around the clitoris or glans penis may supplement the visual erotica. Audiotaped vignettes describing sexual encounters or situations are used less often. These stimuli serve to focus the participant's attention and assure that variation in cognitive content is limited. Interestingly, recent research has suggested that despite the use of such controlled stimuli, men and women recall different aspects of identical stimuli, with males attending more to sexual features and females to the romantic aspects (e.g., Geer & McClone, 1990).

A number of studies have also utilized self-generated sexual fantasies and thoughts as an erotic stimulus condition as well. Arousal power varies both within and across stimulus modalities. We have found, for example, that visual erotica combined with tactile genital stimulation is highly effective in its capacity to induce strong arousal, often to the point of orgasm (Rowland, den Ouden, & Slob, 1994); in contrast, self-generated fantasies tend to produce lower levels of arousal (Rowland & Heiman, 1991).

Stronger stimulation is not always better stimulation. The choice of the type of stimulation is often suggested by the goals of the study or by the population under investigation. For example, the types of stimuli appropriate for studying low sex desire in women are considerably different from those used to study deviant arousal in men convicted of child molestation. No matter what the situation, these stimuli are more imaginative than real--in current lab studies an actual sexual partner does not participate in the lab session, so the response under investigation may be more akin to masturbation than coitus. Early lab studies on sexual response using sexual surrogates or couples (e.g., Masters & Johnson, 1966) had greater potential to reveal aspects of sexual response during coitus. However, the self-selection of participants in these studies (which involve on-camera sexual intercourse), along with the strong focus on physiology, would limit the efficacy of this strategy in most current psychophysiological research, which aims to delineate more than the basic physical parameters of sexual response.

A limited number of lab studies have relied on self-stimulation to induce arousal, a particularly useful strategy when the participant needs to reach orgasm (e.g., studies on premature ejaculation: Strassberg, Mahoney, Schaugaard, & Hale, 1990; or on oxytocin response in men and women: Carmichael et al., 1987). However, in such studies the experimenter surrenders control over the stimuli, and because genital measuring devices can interfere with self-stimulation, the ability to collect objective data on genital response may be diminished. On rare occasions, manual stimulation has been supplied by the experimenter. Notable examples include Alzate and Londono's (1983) study to determine the locus of vaginal erotic sensitivity (women were manually stimulated to orgasm by the male or female experimenter), and Kameya, Deguchi, and Yokota's (1997) investigation of premature ejaculation (men were manually stimulated to orgasm by a female masseuse). Such techniques offer the slight advantage of greater consistency in the type of stimulation (rhythm, intensity). But they also introduce so many experimenter-related variables and concerns about the subject's privacy that their efficacy seems highly limited.

In summary, sexual stimulation in laboratory research lacks the richness and interactive aspects of coitus with one's partner. This disadvantage is partly offset by the greater control that can be exerted over the content, length, intensity, and type of stimulation.

Defining sexual response. Precise and standardized measurement of genital and nongenital physical response is the prime advantage of the sexual psychophysiological method. Yet, physical response constitutes only the most overt index of sexual response--it does not necessarily represent the best or most valid one, and under some conditions or for some populations it may not be as reliable as presumed (Worminth, 1986). As implied earlier, sexual response is a proxy for a constellation of physical, perceptual, affective, and cognitive events occurring during arousal. Despite the multifaceted nature of the response, laboratory studies have focused mainly on genital response and single-item indices of subjective sexual arousal. Except where the interest has been the subject's response to deviant stimuli (e.g., coercive sex), only a handful of studies have attempted to capture in any detail the perceptions and affects associated with the experience.

But even attempts at measuring simple subjective arousal may be inherently problematic. Although no standard definition of subjective arousal has been offered, it is most frequently assessed with questions such as "How sexually aroused are you or how sexually aroused do you feel?" asked during or after the stimulus presentation. While no one can say for certain how an individual assesses their state of arousal, it is probably based on multiple aspects of the situation--the stimulus context, feelings about the stimulus, past responses to similar stimuli, expectations and desire to respond, and actual physical response--and represents a mental summary of these conditions. In most individuals in most situations, internal consistency exists among these conditions: The person views the situation as a sexual one, is attracted to the stimulus, wants and expects to respond sexually, and shows a genital response consistent with expectations and desire. But what happens when inconsistency occurs, for example, in sexually dysfunctional subjects who may want to respond physically, but in fact show no genital response. Do subjects base their subjective arousal on their psychological or mental state, on their genital response, or on other criteria'? We have shown, for example, that when there is substantial disparity between ratings of subjective arousal, mental arousal, and physical arousal in men with erectile dysfunction (Rowland & Heiman, 1991), subjective arousal is more closely aligned with mental arousal than physical arousal. In contrast, sexually functional men rely more heavily on their physical arousal to define their subjective arousal. Such findings suggest that constructs like subjective arousal may be determined differently when subjects experience internally inconsistent information, as might occur in dysfunctional or paraphilic participants. It would seem that analysis of this construct into more specific components is worth considering, particularly in view of the fact that the two genders may use or weigh the individual components differently when rating their overall subjective arousal (e.g., evidence suggests that women in general may rely less on genital feedback to assess their state of arousal than do men: Laan & Everaerd, 1995).

Given the problem in defining subjective arousal, it is not surprising that the correspondence between people's self-reported and genital arousal is moderate at best in men, even less in women, and generally different for sexually functional and dysfunctional subjects. Beyond the reason discussed above, other factors may underlie this disparity, and these factors may again be different for men and women. Some factors may emanate from the way genital response is measured: Genital devices fall short of capturing all the relevant physical changes related to arousal, so these devices may under- or overestimate, or miss completely, various stages of genital arousal. Other factors may be related to the ceiling effect on genital response: Subjective arousal may continue to increase well after maximum genital response has been achieved. And finally, there may be theoretical reason to expect only moderate correspondence between genital and subjective arousal: Attentional focus on the partner during sexual activity may preclude strong attention on one's own genital response. Strong self-monitoring of one's genital response is thought to interfere with sexual arousal (Sakhiem, Barlow, Beck, & Abrahamson, 1984; Wincze, Venditti, Barlow, & Mavissakalian, 1980).

Given the less-than-perfect correlation between genital and subjective arousal, the issue ultimately centers around which measure constitutes the more valid index of arousal. When there is strong consistency between measures there is little need for concern, but in the event of inconsistency, the experimenter may follow the strategy of reporting both measures or choosing between the two. To some extent the choice depends on the goals of the study, but the temptation to favor genital measures, because they are precise and presumably less subject to falsification, has generally been resisted. However, problems occur because, in the absence of other strong and reliable measures, genital response often acquires more significance than it should. Psychophysiological studies could benefit by developing comprehensive (i.e., multivariate) indices that capture the multiple dimensions of sexual response. Such a strategy would help maintain the balance between the physical and the psychological aspects of sexual response, and could lead to a better or more valid measure of sexual response, the components of which could be analyzed both separately and as a composite. Indeed, one of the promises of sexual psychophysiological research is its potential for integrating the multiple dimensions of sexual response, a process that is only now beginning to take shape.

The problem of repeated stimulation. Due to the large effort invested in each participant, relatively small sample sizes are typically utilized in laboratory-based research on sexual response. To increase efficiency, studies often use a within-subject design where each participant is exposed to several different treatment conditions. Since each treatment condition typically requires exposure to erotic stimulation, the effects of repeated exposures to erotic stimulation need to be considered, particularly when multiple treatment conditions are incorporated into a single lab session. As with any within-subject design, order effects may arise from any number of sources. Specifically, exposure to an initial erotic stimulus has the potential to alter one's physiological state, which then can affect one's response to a subsequent erotic stimulus. For example, because of changing hemodynamics in the genital regions, the second vaginal engorgement or erection within a session may be easier or more difficult than the first. An initial response, for example, could act to "prime" a later response. It has been demonstrated that preliminary exposure to erotic materials in men stimulates LH and androgen release, which in turn may account for enhanced arousal to an erotic stimulus presented an hour later (Rowland, Heiman, Hatch, Doering, & Weiler, 1987). A similar enhancement of arousal has been described in women, though it could not be tied to specific endocrine changes (Heiman, Rowland, Hatch, & Gladue, 1991).

Other order-related effects also require attention. Repeated exposure to the same erotic stimulus eventually results in habituation of central arousal, although initial presentations may actually facilitate arousal (see Over & Koukounas, 1995, for review). By introducing novel stimuli (e.g., a different videotape segment) with each exposure, the habituation response is diminished, but this strategy adds another important variable requiring control, namely whether each stimulus (e.g., erotic videotape) is equally arousing. Such effects can be handled through counterbalanced designs in which different subsets of participants experience the various treatment conditions and videotapes in different temporal sequences. In theory, this procedure is straightforward, but in practice it has significant implications. Even in the simplest possible design (e.g., one having two treatment conditions within a single lab session), both the treatment (independent variable) and the videotapes must be counterbalanced. Although this strategy is necessary for establishing sources of variance in responding, it further subdivides each group, perhaps already limited in size, into four. Statistically, this procedure has the potential to decrease power, that is, the ability to detect real effects. Thus, the problem of low sample size and consequent statistical power may be compounded.

Finally, studies involving women present a special challenge. The most commonly used measure of genital response, vaginal pulse amplitude (VPA), has no absolute unit of measure as a reference. Consequently, not only are comparisons across studies nearly impossible, but comparisons across groups within a single study are fraught with problems. One way to circumvent this problem is to rely on multiple treatment conditions within a single lab session. However, both VPA (a measure of the strength of the pulse of the vessels in the vaginal epithelium) and thermal readings from the labium minus may show residual effects long after the stimulus has been removed. In some cases, baseline recovery may never occur within the time frame of the experimental session, and so it becomes difficult to interpret responses that only have meaning when they are expressed as a change from baseline. One creative approach to shortening the recovery period for labial temperature, which may require 30 minutes or more, involved fanning the genital area after stimulation (Slob, Bax, Hop, Rowland, & van der Werff ten Bosch, 1996)--not very elegant, but functional. The challenge not only to devise instrumentation that captures critical phasic changes associated with functional vaginal response (i.e., becoming lubricated) on a standardized scale, but also to find ways to match experimental protocols to minimize measurement error, is one of the major tasks facing the study of female sexual psychophysiology today. In this regard, the use of the heated oxygen electrode, a device that measures both heat dissipation and oxygen perfusion in the vagina, may hold promise (Amberson & Hoon, 1985).

Issues About the Meaning and Validity of Lab-Based Findings

One of the most important issues raised about any response studied under laboratory conditions is that of external validity--the extent to which the findings can be generalized to situations beyond the laboratory. The establishment of external validity in psychophysiological research needs to be addressed more aggressively if this approach is to deepen its credibility within clinical and applied circles. The reasons for slow progress in this area are perhaps understandable if one considers the origins of sexual psychophysiology.

With most experimental research in psychology conducted through the mid-1980s, little attention was given to establishing the external validity of lab-based research, and sexual psychophysiology was no exception. Second, research on physiological response systems has traditionally been under less pressure to establish external validity than research on psychological/behavioral systems which, by definition, are more dependent on situational and contextual factors. Early lab studies on sexual response were conceptualized mainly as attempts to elucidate the physiological aspects of the response, and in this respect models of sexual response were not viewed much differently from the stimulus-response (S-R) models of other sensorymotor systems, such as pain or emotion. Third, as with much of the research grounded in a behavioristic perspective, the specific characteristics of the stimulus, or even the context in which the stimulus occurred, were viewed as somewhat arbitrary. In fact, the nature of the stimulus was often defined by the response it evoked rather than by its inherent or a priori characteristics. Fourth, to establish external validity researchers must have access to the situations and populations to which the findings are being generalized. Populations of sexually dysfunctional and deviant subjects, and the situations in which their behaviors are expressed, have been difficult to probe.

The alliance of psychophysiological research and clinical assessment of sexual response has evolved over the years, with the need for establishing external validity paralleling the applied potential of the findings. In the study of sexual deviancy, the potential for lab-based findings to predict behavior outside the lab (predictive validity) was identified as an important issue early on (Freund, Chan, & Coulthart, 1979; Quinsey, 1973). Although recently questioned, the development of the rape index (Abel et al., 1977), a propensity to respond to deviant stimuli, represented a clear attempt to relate information about sexual response in the laboratory to a real world problem. The more recent use of laboratory research in the 1980s to appraise sexual dysfunctions (erectile failure, hypoactive sexual desire, etc.) in individual men and women has only now begun to generate attempts to establish greater external validity for findings from these types of studies (Incrocci & Slob, 1994; Janssen, Everaerd, van Lunsen, & Oerlemans, 1994). Despite increasing attempts toward establishing external validity, the relevance of research based on solitary sexual experiences in the laboratory to the understanding of sexual experiences with one's partner needs to be demonstrated more convincingly and with greater regularity. Ongoing concerns about validity are typically related to the conceptual and methodological issues raised in previous sections of this paper, and include such factors as the selection of subjects, the artificial nature of the stimuli, and the meaning of the response.

Subject selection bias. A number of studies indicate that volunteers for sex research are different from nonvolunteers, and so study samples may not represent the variation in responding typical of the population at large. Furthermore, these differences are magnified in studies using invasive techniques, as occurs in the laboratory study of sexual response (Morokoff, 1986). This selection bias may not be as significant a problem in the recruitment of patients from dysfunctional populations; at least in some types of studies (e.g., testing prosexual drugs), dysfunctional subjects may be highly motivated to participate. But it is a significant concern in the establishment of sexually functional comparison groups that are tested against dysfunctional or deviant groups in quasi-experimental studies. Specifically, control-group volunteers generally have more sexual experience, are more erotophilic, and report lower sex guilt than nonvolunteers. Although evidence for differences in personality and social variables (e.g., authoritarianism, religiosity, social class, etc.) between volunteers and nonvolunteers is mixed (Bogaert, 1996), it is safest to assume that these do affect the samples of sex studies (Strassberg & Lowe, 1995: Trivedi & Sabini, 1998). Yet, even studies investigating this issue are limited in the conclusions they can offer. For example, many studies demonstrating differences between volunteer and nonvolunteer subjects have relied on college students, already atypical of the general population in terms of age, social class, and education. And whether findings from such studies generalize to men and women aged 25 to 60, the more appropriate cohort group for comparison in a study on sexual deviation or dysfunction, is not known. To add to the complexity, volunteer men may exhibit a different profile than volunteer women--the volunteer men tend to be more gender-role stereotyped than nonvolunteer men, whereas the women tend to be less gender-role stereotyped (Wolchik, Braver, & Jensen, 1985). The extent to which such biases might influence the outcome of psychophysiological studies is unclear. At least one study (Saunders, Fisher, Hewitt, & Clayton, 1985) suggests that volunteer bias has only minimal effects on arousal measures in psychophysiological sex research. Nonetheless, given the aforementioned heightened erotophilia of volunteers, response differences between control and sexually dysfunctional men may be overestimated in laboratory studies, and differences between control and deviant men may be underestimated. Similar kinds of projections might apply to control and dysfunctional female groups. Furthermore, other aspects of sexual response (affective, cognitive, perceptive), as yet unexplored in studies on volunteerism, might well differ between participants and nonparticipants in sexual psychophysiological studies.

Despite limitations from volunteer bias, statistical strategies could be used to partly offset this problem in studies which compare effects across functional/normal and dysfunctional/deviant groups. Most studies on sexual response include data on the participant's sexual experience, one factor that differentiates volunteers from nonvolunteers. Once other salient factors that characterize volunteer groups are identified, they too could be assessed during a presession interview. The variance from these factors could then be partialled out as comparisons are made across dysfunctional/deviant and volunteer comparison groups. However, until the relevant variables are specified and such statistical controls are routinely implemented, sex researchers and sex research-consumers alike should exercise caution in generalizing findings from study samples to the larger population.

The problem of a single observation. A second consideration pertinent to external validity is whether the sexual response in the lab is typical of the sexual response in general. The fact remains that the laboratory provides only a single "snapshot" of an individual's sexuality. There is little guarantee that this picture is a representative one, as many factors, even beyond those associated with the laboratory itself, might have influenced the person's response during that particular session. Although we assume such factors are randomized out when viewing response patterns over groups of subjects, a conclusion about an individual's sexual response based upon a single observation is a gamble. Yet, assessment procedures aimed at persons with sexual dysfunctions or deviancies sometimes rely on such limited observations. In such instances, multiple sessions might be considered, although these add to cost and time, with only limited possibilities for useful research data.

The meaning of sexual response in the laboratory. Critical to sexual psychophysiological research is whether findings derived from the laboratory have meaning for sexual behavior under other conditions--for example, with one's sexual partner or under conditions which normally manifest a sexual deviancy. In order to generalize beyond the laboratory environment, the experimenter must first be able to assume that the data collected in the lab are accurate representations of the subject's sexual response. In most instances, the participant has no reason to generate nonrepresentative or false data; however, in the study of sexual deviancies, subjects may have strong motivation to falsity responses due to various legal or social consequences (Freund, 1963). Although lab studies improve the ability to detect inconsistencies in responses resulting from the participant's own attempts to control their response, they are not immune to the process. Self-report ratings of arousal are easily subject to falsification, and by using cognitive strategies that distract from the erotic stimulation, genital response can be suppressed as well. Manual manipulation of the genital measuring device to fake arousal during nonpreferred stimuli has also been reported (see McConaghy, 1989, for discussion of some of these issues). As a result, the truthfulness of genital response may sometimes be suspect, and steps should be implemented to minimize such possibilities. Accordingly, one of the first demands of the researcher in demonstrating external validity is that of assuring the internal accuracy and validity of the data.

Assuming internal validity, the next concern is the level at which the generalization is to occur. Psychophysiological studies generate data at two levels: the group (or treatment condition) and the individual. Experimental research has been most successful at delineating differences at the group level (e.g., functional vs. dysfunctional). Diagnostic and assessment procedures, on the other hand, are aimed at understanding and predicting variation among individuals. When research methodologies aimed at delineating group differences are used as diagnostic procedures to understand individual responses, firm conclusions are just not possible. The very principle that enables experimental research to detect group effects, namely randomization of extraneous variables within and across groups, is not operative at the individual level. Therefore, these laboratory-generated data have only limited value in understanding individual responses. For example, in the diagnosis of a man complaining of erectile dysfunction, a strong response in the laboratory is solid evidence that the man is capable of achieving an erection. Any response less than that may provide information of only limited use to the clinician.

Inability to interpret a subject's response has mild consequences compared with misinterpretation of the response. In research on sexually deviant responses (sexual behavior deemed inappropriate, as in nonconsensual sex), there exists significant potential to draw inappropriate conclusions about individual behavior. For example, male rapists tend to become aroused to sexually deviant stimuli and therefore score high on the rape index. From this, however, one cannot logically assume that men who generally score high on the rape index have a propensity toward coercive sexual behavior (these correlations are actually rather low; see Malamuth, 1986). Specifically, rapists may constitute only a small subset of men who, for whatever reason, become aroused by such kinds of stimuli. For the majority of such men, translation of sexual arousal in the lab into sexually deviant behavior involving a victim simply may not occur. This line of thinking is consistent with broader arguments against claims of causal relationships among pornography, violence, and sexual assault behavior, where a base-rate effect is evident. That is, for a low-frequency behavior overall predictive accuracy is highest if one predicts that all subjects will not engage in the behavior (e.g., sexual assault); even a reasonably valid predictor will produce too many false positives to be useful in recommending intervention. Given such problems of interpretation, sexual psychophysiological research continues to be challenged with identifying lab-based response patterns that truly discriminate deviant from normal populations and thus enable generalizability to situations and behaviors outside the laboratory.

The stimulus-response connection in context. The applicability of lab findings to real-world sexual situations is influenced by two factors: (a) the extent to which understanding variation of the response is relevant, and (b) the ability of the laboratory to simulate the real world when understanding response variation is relevant. In their initial research, Masters and Johnson (1966) were interested in describing the typical response associated with functional sexual arousal rather than in finding factors that accounted for variation in response. This typical response represents a fairly accurate description of the sequence of physiological events under most conditions of sexual arousal, not just those in the laboratory. But, when understanding response variation is important, the stimulus relevance or context also becomes important. Early lab studies on premature ejaculation (PE), for example, employed only visual erotica to test for differences in genital response patterns between PE and non-PE men. Not surprisingly, these studies failed to discriminate between PE and non-PE men. When we combined visual erotica with a stimulus more relevant to this disorder, namely penile stimulation, and added measures of ejaculatory threshold to the outcome, clear differences emerged (Rowland et al., 1997).

In general, it might be argued that the more important the stimulus context is to understanding variation in responding, the greater the burden on the researcher to simulate that context in the lab. Alternatively, the less the researcher is able to simulate that context, the less the response in the lab will predict responding in other situations. Thus, lab research might be weakest in its ability to understand how sociocultural and partner-generated variables (partner's attractiveness, performance evaluation, etc.) affect sexual arousal and performance. Although some of these variables can be manipulated in the partner's absence (e.g., a performance demand can be induced by an instructional set from the experimenter), it is not clear to what extent, if any, they replicate salient conditions with the sexual partner, where failure to respond has consequences for both the individual and the relationship.

Sexual response in and outside the lab: Is it different? Greater simulation of the outside world within the laboratory will undoubtedly increase the prospect for generalizability. Nevertheless, researchers are still faced with the task of actually demonstrating that responses exhibited in the lab do, indeed, correlate with those outside the lab. Specifically, psychophysiological studies need to demonstrate that parameters of the response (type, frequency, duration, etc.) under investigation in the lab correlate strongly with parameters of that same response in the more natural, nonlab setting. Less convincing but also of value is the demonstration that lab responses can predict group membership (e.g., functional vs. dysfunctional) based on some independently-validated criterion, or that sexual response at the group level (versus the individual level) shows consistency between lab and nonlab settings.

A handful of psychophysiological studies, conducted mainly on men, have begun to provide such data. For example, one recent study found that the subject's genital response to combined erotic visual and penile stimulation was highly accurate in predicting membership in impotent and nonimpotent groups (Janssen et al., 1994). Several of our own studies investigating pharmacological effects on premature ejaculation (Haensel, Rowland, Kallan, & Slob, 1997; Strassberg, Slob, Rowland, de Gouveia Brazo, & Tan, submitted for publication) have compared data on ejaculatory response obtained from the laboratory or from daily diaries of sexual activities at home. In both studies, at the group level, data from these two sources showed a high degree of consistency, suggesting that laboratory-generated data do indeed relate to sexual functioning outside the lab. We have just now begun the task of ascertaining whether the ejaculatory latency of individual men in the laboratory correlates well with their latency at home with their partner. Obviously, there is a need for further efforts in this regard, particularly where responses are highly dependent on contextual factors that are less easily simulated in the laboratory.

CONCLUSION

The preceding discussion was intended to pinpoint significant issues in the laboratory study of sexual response. This listing was not meant to be exhaustive (see Rosen & Beck, 1988, required reading for anyone wanting greater depth in this field), but rather to inform an audience of sexologists about the value, limitations, and caveats of this type of research. One of the more promising aspects of sexual psychophysiological research is that many of the problems discussed above can be addressed. Problems arising from the lab session itself are, for the most part, under the control of the experimenter and simply require more painstaking efforts to ensure high quality studies. A number of relevant extraneous variables that may impact outcome measures have already been identified, and these should routinely be considered for evaluation in any sexual psychophysiological study. Such extraneous variables might be handled in a variety of ways. For example, as part of a preliminary analysis of the data, comparisons across groups or treatment conditions on these variables could be undertaken. Where differences are significant, these variables could be entered as control variables (or covariates) in the regression equation or analysis of variance. Alternatively, all factors could be entered as control variables in the statistical analysis; however, this strategy would require much larger groups of subjects.

The laboratory study of sexual response will continue to be limited by its inability to address broader relationship factors and sociocultural variables, both of which are critical to understanding healthy sexual response. Such issues simply cannot be examined by psychophysiological research as it is currently conceptualized and designed. Furthermore, despite progress in recent years, issues of external validity and the application of experimental research procedures to understanding individual rather than group behaviors will continue to challenge sexual psychophysiological methodology. Nevertheless, the contribution made by this field to the understanding of sexual arousal and response is unique. It has provided a good testing ground for many ideas presumably related to sexual disorders and deviancies and, more importantly, it has uncovered enough new knowledge and raised enough new questions about sexual response to have established legitimacy as an alternate way of learning about sexual response.

REFERENCES

Abel, G. G., Barlow, D. H., Blanchard, E. B., & Guild, D. (1977). The components of rapists' sexual arousal. Archives of General Psychiatry, 34, 895-903.

Abel, G. G., Osborn, C., Anthony, D., & Gardos, P. (1992). Current treatments of paraphiliacs. Annual Review of Sex Research, 3, 255-290.

Abrahamson, D. J., Barlow, D. H., & Abrahamson, L. S. (1989). Differential effects of performance demand and distraction on sexually functional and dysfunctional males. Journal of Abnormal Psychology, 98, 241-247.

Alzate, H., & Londono, M. L. (1984). Vaginal erotic sensitivity. Journal of Sex & Marital Therapy, 10, 49-56.

Amberson, J., & Hoon, P. (1981). Hemodynamics of sequential orgasm. Archives of Sexual Behavior, 14, 351-360.

Barbaree, H. E., Marshall, W. L., & Lanthier, R. (1979). Deviant sexual arousal in rapists. Behaviour Research & Therapy, 17, 215-222.

Barlow, D. (1986). Causes of sexual dysfunction: The role of anxiety and cognitive interference. Journal of Consulting and Clinical Psychology, 54, 140-157.

Beck, J. G., & Baldwin, L. E. (1994). Instructional control of female sexual responding. Archives of Sexual Behavior, 23, 665--684.

Beck, J. G., Barlow, D. H., Sakhiem, D. K., & Abrahamson, D. J. (1987). Shock threat and sexual arousal: The role of selective attention, thought content, and affective states. Psychophysiology, 24, 165-172.

Beggs, V. E., Calhoun, K. S., & Wolchik, S. A. (1987). Sexual anxiety and female sexual arousal: A comparison of arousal during sexual anxiety and sexual pleasure stimuli. Archives of Sexual Behavior, 16, 311-319.

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.

Bozman, A. W., & Beck, J. G. (1991). Covariation of sexual desire and arousal: The effects of anger and anxiety. Archives of Sexual Behavior 20, 47-60.

Carmichael, M., Humbert, R., Dixen, J., Palmisano, G., Greenleaf, W., & Davidson, J. (1987). Oxytocin increase in human sexual response. Journal of Clinical Endocrinology and Metabolism, 64, 27-31.

Cranston-Cuebas, M., & Barlow, D. H. (1990). Cognitive and affective contributions to sexual dysfunction: An integrative and interdisciplinary review. Annual Review of Sex Research, 1, 119-161.

Davis, R. C., & Buchwald, A. M. (1957). An exploration of somatic response patterns: Stimulus and sex differences, Journal of Comparative and Physiological Psychology, 50, 44-52.

Elliott, A. N., & O'Donohue, W. T (1997). The effects of anxiety and distraction on sexual arousal in a nonclinical sample of heterosexual women. Archives of Sexual Behavior, 26, 607-624.

Freund, K. (1963). A laboratory method for diagnosing predominance of homo- or hetero-erotic interest in the male. Behaviour Research & Therapy, 1, 85-93.

Freund, K., Chan, S., & Coulthart, R. (1979). Phallometric diagnosis with "nonadmitters." Behaviour Research & Therapy, 17, 451-457.

Geer, J. H., & Fuhr, R. (1976). Cognitive factors in sexual arousal: The role of distraction. Journal of Consulting and Clinical Psychology, 44, 238-243.

Geer, J. H., & McGlone, M. (1990). Sex differences in memory for erotica. Cognitions and Emotions, 4, 71-78.

Haensel, S., Rowland, D. L., Kallan, K., & Slob, A. K. (1997). The effects of clomipramine in men with premature ejaculation and in controls. Journal of Urology, 156, 1310-1315.

Heiman, J. R. (1980). Female sexual response patterns: Interactions of physiological, affective, and contextual cues. Archives of General Psychiatry, 37, 1311-1316.

Heiman, J. R., & Rowland, D. L. (1983). Effects of instructions on affective and physiological response in sexually dysfunctional and normal men. Journal of Psychosomatic Research, 27, 105-116.

Heiman, J. R., Rowland, D. L., Hatch, J., & Gladue, B. (1991). Genital and psychological response following sexual arousal in women. Archives of Sexual Behavior, 7, 171-186.

Holden, C. (1976). House chops sex-pot probe. Science, 192, 450. Incrocci, L., Hop, W. C. J., & Slob, A. K. (1996). Visual erotic and vibrotactile stimulation and intracavernous injection in screening men with erectile dysfunction: A 3 year experience with 406 cases. International Journal of Impotence Research, 8, 227-232.

Incrocci, L., & Slob, A. K. (1994). Visual sexual stimulation and penile vibration in screening men with erectile dysfunction. International Journal of Impotence Research, 6, 227-229.

Janssen, E., Everaerd, W., van Lunsen, R. H., & Oerlemans, S. (1994). Validation of a psychophysiological waking erectile assessment (WEA) for the diagnosis of male erectile disorder. Urology, 43, 686-695.

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

Kaplan, H. S. (1974). The new sex therapy. New York: Brunner/Mazel.

Koukounas, E., & McCabe, M. (1997). Sexual and emotional variables influencing sexual response to erotica. Behaviour Research & Therapy, 35, 221-231.

Laan, E., & Everaerd, W. (1995). Determinants of female sexual arousal: Psychophysiological theory and data. Annual Review of Sex Research, 6, 32-76.

Laan, E., Everaerd, W., van Bellen, G., & Hanewald, G. (1994). Women's sexual and emotional responses to male and female produced erotica. Archives of Sexual Behavior, 23, 153-169.

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

Lohr, B. A., Adams, H. E., & Davis, J. M. (1997). Sexual arousal to erotic and aggressive stimuli in sexually coercive and noncoercive men. Journal of Abnormal Psychology, 106, 230-242.

Magoun, H. W. (1981). John B. Watson and the study of human sexual behavior. The Journal of Sex Research, 17, 368-378.

Malamuth, N. M. (1986). Predictors of naturalistic aggression. Journal of Personality and Social Psychology, 50, 953-962.

Malamuth, N. M., & Check, J. V. (1980). Penile tumescence and perceptual responses to rape as function of victims' perceived reactions. Journal of Applied Social Psychology, 10, 528-547.

Masters, W. H., & Johnson, V. E. (1966). Human sexual response. Boston: Little, Brown.

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

McAnulty, R. D., & Adams, H. E. (1992). Validity and ethics of penile circumference measures of sexual arousal: A reply to McConaghy. Archives of Sexual Behavior, 21, 177-186.

McConaghy, N. (1989). Validity and ethics of penile circumference measures of sexual arousal: A critical review. Archives of Sexual Behavior, 18, 357-369.

McConnell, J. V. (1974). Understanding human behavior. New York: Holt, Rinehart & Winston. 1974.

Meston, C. M., & Gorzalka, B. B. (1996). Differential effects of sympathetic activation on sexual arousal in sexually dysfunctional and functional women. Journal of Abnormal Psychology, 105, 582-591.

Mitchell, W. B., DiBartolo, P. M., Brown, T. A., & Barlow, D. H. (1998). Effects of positive and negative mood on sexual arousal in sexually functional males. The Archives of Sexual Behavior, 27, 197-208.

Morokoff, P. (1986). Volunteer bias in the psychophysiological study of female sexuality. The Journal of Sex Research, 22, 35-51.

Morokoff, P., & Heiman, J. (1980). Effects of erotic stimuli on sexually functional and dysfunctional women: Multiple measures before and after sex therapy. Behaviour Research & Therapy, 18, 127-137.

Mosovich, A., & Tallefero, A. (1954). Studies of EEG and sex function at orgasm. Diseases of the Nervous System, 15, 218-220.

Moss-Morris, R., & Petrie, K. J. (1997). Cognitive distortions of somatic experiences: Revision and validation of a measure. Journal of Psychosomatic Research, 43, 293-306.

Over, R., & Koukounas, E. (1995). Habituation of sexual arousal: Product and process. Annual Review of Sex Research, 6, 187-223.

Palace, E. M., & Gorzalka, B. B. (1990). The enhancing effects of anxiety on arousal in sexually functional and dysfunctional women. Journal of Abnormal Psychology, 99, 403-411.

Palace, E. M., & Gorzalka, B. B. (1992). Differential patterns of arousal in sexually functional and dysfunctional women: Physiological and subjective components of sexual response. Archives of Sexual Behavior, 21, 135-160.

Quinsey, V. L. (1973). Methodological issues in evaluating the effectiveness of aversive therapies for institutionalized child molesters. Canadian Psychological Review, 14, 350-361.

Quinsey, V. L., & Chaplin, T. C. (1984). Stimulus control of rapists and non-sex offenders' sexual arousal. Behavioral Assessment, 6, 169-176.

Rosen, R. C., & Beck, J. G. (1988). Patterns of sexual arousal: Psychophysiological processes and clinical applications. New York: Guilford.

Rowland, D. (1996). The psychobiology of sexual arousal and behavior. In L.Diamant & R. McAnulty (Eds.), Psychobiology of sexual orientation, behavior, and gender identity (pp. 19-42). Westport, CT: Greenwood.

Rowland, D. L., Cooper, S. E., & Heiman, J. R. (1995). A preliminary investigation of affective and cognitive response in men before and after treatment in a sex therapy program. Journal of Sex & Marital Therapy, 21, 3-20.

Rowland, D. L., Cooper, S. E., & Slob, A. K. (1996). Genital and psychoaffective response to erotic stimulation in sexually functional and dysfunctional men. Journal of Abnormal Psychology, 105, 194-203.

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

Rowland, D. L., den Ouden, A. H., & Slob, A. K. (1994). The use of vibrotactile stimulation for determining sexual potency in the laboratory in men with erectile problems: Methodological considerations. International Journal of Impotence Research, 6, 153-161.

Rowland, D. L., & Heiman, J. R. (1991). Self-reported and genital arousal changes in sexually dysfunctional men following a sex therapy program. Journal of Psychosomatic Research, 35, 1-11.

Rowland, D. L., Heiman, J. R., Hatch, J., Gladue, B., Doering, C., & Weiler, S. (1987). Endocrine, psychological, and genital response to sexual arousal in men. Psychoneuroendocrinology, 12, 149-158.

Rowland, D. L., & Slob, A. K. (1995). Understanding and diagnosing sexual dysfunction: Recent progress through psychophysiological and psychophysical methods. Neuroscience & Biobehavioral Reviews, 19, 201-209.

Sakhiem, D. K., Barlow, D. H., Beck, J. G., & Abrahamson, D. J. (1984). The effect of an increased awareness of erectile cues on sexual arousal. Behaviour Research & Therapy, 23, 151-158.

Saunders, D. M., Fisher, W. A., Hewitt, E. C., & Clayton, J. P. (1985). A method for empirically assessing volunteer selection effects: Recruitment procedures and responses to erotica. Journal of Personality and Social Psychology, 49, 1703-1712.

Simon, W. T., & Schouten, P. G. (1991). Plethysmography in the assessment and treatment of sexual deviance: An overview. Archives of Sexual Behavior, 20, 75-92.

Slob, A. K., Bax, C., Hop, W., Rowland, D. L., & van der Werff ten Bosch, J. J. (1996). Sexual arousability and the menstrual cycle. Psychoneuroendocrinology, 20, 1996.

Slob, A. K., Blom, J., & van der Werff ten Bosch, J. J. (1990). Erection problems in medical practice: Differential diagnosis with a relatively simple method. Journal of Urology, 143, 46-50.

Slob, A. K., Rowland, D., Blom, J., & van der Werff ten Bosch. J. J. (1991). Psychological factors affect erectile response to papaverine. Urology, 38, 294-295.

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, 251-257.

Strassberg, D. S., Slob, A. K., Rowland, D. L., de Gouveia Brazo, C., & Tan, P. (1998). Clomipramine in the treatment of rapid (premature) ejaculation. Manuscript submitted for publication.

Tiefer, L. (1996). The medicalization of sexuality: Conceptual, normative, and professional issues. Annual Review of Sex Research, 7, 252-282.

Trivedi, N., & Sabini, J. (1998). Volunteer bias, sexuality, and personality. Archives of Sexual Behavior, 27, 181-196.

Wincze, J. P., Venditti, E., Barlow, D., & Mavissakalian, M. (1980). The effects of a subjective monitoring task in the physiological measure of genital response to erotic stimulation. Archives of Sexual Behavior, 9, 533-545.

Wolchik, S. A., Braver, S. L., & Jensen, K. (1985). Volunteer bias in erotic research: Effects of intrusiveness of measure and sexual background. Archives of Sexual Behavior, 14, 93-107.

Worminth, J. S. (1986). Assessing deviant sexual arousal: Physiological and cognitive aspects. Advances in Behavior Research & Therapy, 8, 101-137.

Manuscript accepted July 23, 1998

Address correspondence to David L. Rowland, Department of Psychology, Valparaiso University, Valparaiso, IN 46383; e-mail: David. [e-mail]