by John E. Grable, Ph.D., CFP®, and Sonya L. Britt, Ph.D.
John E. Grable, Ph.D., CFP®, is a professor of personal financial planning at Kansas State University. He currently serves as the program director of the registered CFP Board undergraduate, MS, and Ph.D. programs at Kansas State.
Sonya L. Britt, Ph.D., is an assistant professor at Kansas State University. She has an expertise in financial therapy and quantitative research methodologies in financial planning.
The purpose of this study was to test for a video narration effect in the context of financial risk-tolerance assessment. A randomized experimental study was designed to compare risk-profile scores of those who completed a pen-and-paper risk-tolerance assessment instrument to risk scores from one of three stimuli treatments:
(a) viewing and listening to the same questions asked by a female narrator,
(b) viewing and listening to the same questions narrated by a male, or
(c) viewing the same questions without narration.
Controlling for financial knowledge:
The use of standardized client data assessment tools, as a component of the financial planning process, is mandated in the United States, Australia, and Europe. In the United States, for example, the Securities and Exchange Commission gives directions that registered investment advisers collect relevant client data prior to making investment recommendations. FINRA also requires brokers to collect similar data. While nearly all financial advisers collect data via a pen-and-paper questionnaire method, the use of written assessments is not required under the law. Rather, documentation, in written form, of the assessment outcome is required. As the use of technology increases, it is important for financial advisers to understand the role technology plays in shaping client attitudes and behaviors.
Of particular importance in the data collection process is the assessment of a client's tolerance for taking financial risk. Financial risk tolerance has been defined in numerous ways. Those working as practitioners sometimes refer to risk tolerance as the level of volatility someone is willing to stomach when investing before they lose sleep at night. Researchers often define risk tolerance as the willingness to engage in an activity in which the outcomes are uncertain. Conceptually, risk tolerance can be thought of as the inverse of risk aversion. Those who are risk tolerant tend to be willing to take higher risks with their assets and cash flow compared to those who are risk averse.
Regardless of how risk tolerance is defined-either practically or academically-there are two common themes that emerge from the definitions used. First, risk tolerance refers to a client's willingness to take risk, not necessarily their capacity or preference for risk taking (Roszkowski & Davey, 2010). Second, regulations that require the assessment of a client's risk tolerance simply require the evaluation; that is, the current regulatory environment does not specify how an adviser should assess a client's tolerance for risk. As long as the assessment offers a defensible level of validity (i.e., accuracy) and reliability (i.e., repeatability) the adviser will be deemed to have met the intent of the rule.
Given the various definitions of risk tolerance, it comes as no surprise that the methods used to assess risk tolerance are quite diverse, as well (Roszkowski, Davey, & Grable, 2005). Some advisers ask current and prospective clients to complete a pen-and-paper survey prior to or during the first client-planner meeting. In these situations the client receives a data gathering form that includes risk-tolerance questions. The client, after completing the document, either mails the form to the adviser or brings it to the first meeting. The adviser, usually working with a paraplanner, inputs the results into a database. Unless the risk-assessment tool is a standardized product, the interpretation of the client's risk profile is determined by the adviser.
Increasingly, advisers are using online questionnaires to automate the assessment process. Alternatively, some advisers believe that they can obtain a better understanding, and a more accurate assessment, of a client's tolerance for risk by asking questions in a face-to-face (i.e., narrated) encounter. This technique provides a spontaneous client response to risk-assessment items. This method of inquiry also adds one or more external stimuli to the assessment process. Key among these is what is generally called the narrator effect. This is the role the adviser's pitch and tone of voice plays in shaping or influencing the client's response to questions.
The purpose of this study was to test for a video narration effect in the context of financial risk-tolerance assessment. Specifically, a randomized experimental study was designed to compare risk-profile scores of those who completed a pen-and-paper risk-tolerance assessment instrument (i.e., the control group) to those who received one of three treatments: (a) they viewed and listened to the same questions being asked by a female narrator, (b) they viewed and listened to the same questions narrated by a male, or (c) they viewed the same questions without narration. It was hypothesized that the possibility of a video narrator effect might be present, and if true, financial advisers (i.e., insurance agents, investment brokers, financial planners, etc.) ought to consider this possibility as a factor that influences responses to risk assessments, especially as they incorporate additional technological evaluation tools into their practice.
Review of Literature
The importance of narration as a factor that influences consumer attitudes and behaviors has been understood for quite a long time in the marketing field. However, a review of the financial services, financial planning, personal and consumer finance, and household finance literature, related to client assessments and risk tolerance, shows a paucity of research devoted to this topic. It might be that firms conduct non-published research that is designed to benefit their sales force, and as such, the information is generally known but not publicly available. It is more likely, however, that the financial planning profession has not generally considered the role narration plays in the data gathering process.
Voice narration refers specifically to a voice of authority that is used to sell, promote, or describe a product or idea in an effort to solicit a consumer response (Johnson & Young, 2002; Whipple & McManamon, 2002). For the purposes of this study, narration is described as a media or video source, rather than a direct one-on-one sales presentation. Matching gender narration with a product, based on the anticipated financial user, has become an important marketing tool. It is known, for example, that "female spokespersons and announcers for a female gender-imaged product will produce better attitudes toward the presenters in the ad" (Whipple & McManamon, 2002, p. 90). In effect, this means that for products such as household supplies and personal care products advertisers are better served by using female narration. Only in consumer purchase situations where there is a gift-buying scenario or in advertisements of male-orientated products does it appear that a male's narration is more effective than that of a female. With neutral gender-oriented products, both male and female narration generates similar consumer responses.
The point of sharing information from the Whipple and McManamon (2002) study is to indicate that narration may have an impact on client responses to financial planning intake form items as well. Consider research reported by Edworthy, Hellier, and Rivers (2003). They noted that female voices have an advantage over men's narration in being able to portray urgency. Arrabito (2009) noted that this is because a woman's voice has a higher pitch and pitch range, while men tend to remain constant in their pitch while talking. Male narration is typically associated with qualitative factors such as strength, dependability, and authority, whereas women's voices tend to be associated with childhood and nurturing (McMinn, Brooks, Triplett, Hoffman, & Huizinga, 1993). In other words, the evidence suggests that a person's perception or judgment of a scenario can be influenced, in part, by the sex of the narrator.
According to Jones, Feinberg, DeBruine, Little, and Vukovic (2010), there is a biological preference toward narration where women prefer masculine men's voices and men prefer feminine women's voices. These preferences appear to be a response to "reflect adaptations that identify high-quality (e.g., healthy) mates" (Jones et al., p. 57). Jones et al. classified feminized female voices as those with raised pitch, while masculine male voices were those with lowered pitch. Jones and associates termed these voice preferences as an opposite-sex bias. It appears that, at least subconsciously, women and men respond differently to voice narration as a way to send a signal that they find the voice representative of a fertile (in terms of biological reproduction) possible mate. The preference for a low pitched male voice among women, but not men, may be indicative of the desire among women to identify a strong mate. For men, a high pitched female voice may suggest a reproductive female that has potential as a mate. The literature related to this hypothesis does not suggest that women and men necessarily act on these impulses in an objective sexual manner, but rather that voice pitch influences the way people respond to commands and inquiries. It is possible that if a narration effect exists in relation to risk-tolerance assessment, some of the bias exhibited by women and men for one voice over another might be related to reproductive adaptive preferences.
The possibility of a gender effect extends beyond video narration. Among financial risk-tolerance research scientists it is generally acknowledged as true that women exhibit a lower risk tolerance than men (Arano, Parker, & Terry, 2010; Grable, 2008; Jianakoplos & Bernasek, 1998; Kohler, 1996; Neelakantan, 2010; Schubert, Brown, Gysler, & Brachinger, 1999; West, Moskal, Dziuban, & Rumbough, 1996). Consider a risk-tolerance investigation conducted with one of the largest independent risk databases. Hallahan, Faff, and McKenzie (2004) used the Finametrica dataset and found that female investors scored significantly lower than male investors in terms of tolerance for financial risk. This finding held true even when controlling for other relevant demographic factors (e.g., marital status, income, etc.). Weber, Siebenmorgen, and Weber (2005) noted a similar gender response pattern in the domain of risk taking. They attributed lower female scores to women respondents reporting feelings of low competence when making financially risky choices. In a noteworthy study, Neelakantan (2010) also documented a gender bias in relation to risk tolerance where women were again found to be less risk tolerant than men. Rather than looking only at a subjective measure of risk tolerance, Neelakantan was able to objectively calibrate holdings in individual retirement accounts to determine that differences in risk tolerance can have a major impact on wealth accumulation. According to Neelakantan, women were likely to amass 10 percent less wealth than similar men as a result of risk choices.
In addition to the narration effect, financial knowledge is telling of individuals' financial perceptions and behaviors and consequently has gained importance as a research topic over the past decade. Hilgert and Hogarth (2003) noted that better informed consumers typically make fewer problematic decisions in the consumer marketplace, which they found to be associated with economic security and enhanced levels of well-being. Of particular importance in the Hilgert and Hogarth study was their finding that consumers at the household level prefer to learn about money management through media sources, including television and video. Interestingly enough, many financial advisers are also turning to video and computer media as a way to communicate with clients. When viewed holistically, video, as an information source preference, may inadvertently be introducing the narration effect into the development, perception, and reporting of household finance information, including the assessment of financial risk tolerance. Hilgert and Hogarth reported that the use of videos, as a source of information, is effective because the media source reflects a visual learning preference among consumers. They also hypothesized that video may be a time-preference tool that reduces learning efforts while increasing knowledge. Controlling for financial knowledge then is important when testing for the main effects of video narration and gender on risk tolerance.
Four research questions were of interest in this study. First, is there a gender (sex) difference in risk-tolerance scores between women and men? Second, is there a main narrator effect in relation to risk-tolerance assessment? Third, is subjective financial knowledge associated with risk-tolerance scores? Finally, is there an interaction between gender and narration on risk-tolerance scores? The following hypotheses were tested:
H1: There is no difference between women and men in terms of financial risk-tolerance scores.
H2: There is no difference in risk-tolerance scores based on the voice narration method used to solicit responses to risk questions.
H3: There is no difference in risk-tolerance scores based on financial knowledge.
H4: There is no interaction between gender and narration on risk-tolerance scores.
Seventy-seven men and women took part in this study1. Participants were recruited from the undergraduate and graduate student population at a Midwestern U.S. university. Participants were split equally between women and men. The mean age of those in the study was 22.82 years (SD = 5.99), with a range of 18 to 56 years. The majority of respondents were single, non-Hispanic White, with incomes ranging from near $0 to over $15,000. Financial knowledge was measured by asking participants, as part of the pre-test to "Rate yourself on your level of knowledge about personal finance issues and investing." This question was used as a proxy for subjective financial knowledge. A 10-step scale was used to record answers, with 1 indicating the lowest level of knowledge and 10 being the highest level. The mean score for the sample was 5.99 (SD = 1.88). Men reported a mean knowledge score of 6.30 (SD = 1.65), whereas women reported a mean score of 5.68 (SD = 2.06). A t test indicated that the difference in knowledge between women and men was not statistically significant. Any missing data were replaced using the mean score of the same gender for the variable.
Participants were recruited from classes using a combination of faculty announcements and handouts advertising the study. Those who participated in the experiment received a $10 gift card for use at Wal-Mart. Ten women and 10 men were randomly assigned to one of 4 groups: (a) pen-and-pencil, (b) video with female voice narration, (c) video with male voice narration, and (d) video with no narration. That is, each group was comprised of an equal number of women and men. Prior to the stimuli each participant completed a pre-test. The pre-test consisted of attitudinal and behavioral assessment items, including the item that assessed subjective financial knowledge.
Stimuli. The stimuli in the experiment consisted of female, male, and no narration of a risk-tolerance assessment questionnaire presented via video. The questionnaire consisted of 13 multiple-choice risk items. The measure was originally designed by Grable and Lytton (1999). The questionnaire has historically shown a reasonable level of validity and reliability, with Cronbach's alpha estimates ranging from .70 to .85. In this study, Cronbach's alpha for the participant group was .74.
A control group was established by randomly assigning 10 women and 10 men to complete the questionnaire using a traditional pen-and-paper method. Each question was presented on a separate 8½ x 11-inch white piece of paper. Every participant was asked to sit in a quiet assessment room apart from the researcher when completing the questionnaire. This approach was taken to mirror the setting of a financial advisory office. Participants circled their answer preference for each item. Scores were then summed, with higher scores representing an elevated tolerance for financial risk-i.e., a higher risk profile.
The remaining participants were randomly assigned to the video assessment groups with female, male, or no narration, with an equal gender distribution in each group. Proxies for masculinized and feminized voices were chosen by the researchers based on pitch variation analysis. A number of narrator volunteers were evaluated. The selected female was chosen because of her feminized pitch variation when speaking. It would be nearly impossible to confuse the voice with that of a male. On the other hand, a male narrator was chosen as a result of pitch variation analysis that showed a more narrow range of pitch and a lower overall tone. Participants heard only the voice narration for their particular group.
Participants in the narration groups were asked to sit in comfortable arm chairs while watching the risk-assessment questions appear via video on a big-screen television. The viewing environment was set up to resemble a video conferencing facility at a financial advisory firm. The progression of the questions was timed, so that the slides progressed at the same rate for the three stimuli groups. Participants could read the questions as they were being narrated, but they could not see the narrator. Participants used an i>clicker to choose their response to each question. An i>clicker looks and works similar to a television remote control. About one-half of participants had used an i>clicker in a classroom situation prior to the study; upon completion of the experiment participants were asked about the difficulty in using the tool. No one indicated that they had trouble with the assessment format. The tool has five buttons, with each button corresponding to a letter in the alphabet-A, B, C, D, or E. If a participant wanted to choose answer C, for example, they would click the button and their response would be recorded electronically. Scores, representing risk profiles, were then calculated by the researchers by summing responses for each participant.
Method. An analysis of covariance (ANCOVA) was used to test the research questions and hypotheses. Specifically, financial risk-tolerance scores from the 13-item risk questionnaire were used as the dependent variable. Two between-subjects factors were tested: (a) gender, with men coded 0 and women coded 1 and (b) narration effect, with male voice coded 1, female voice coded 2, no voice code 3, and the control group coded 4. The covariate was a participant's subjective financial knowledge. Financial knowledge was controlled in this study for two reasons. First, the research team was unable to pre-determine the level and extent of personal finance and investing knowledge participants knew prior to the study. The risk literature shows, quite clearly, that financial knowledge and risk tolerance tend to be highly correlated (see Grable, 2008). If knowledge, as a known covariate was not controlled in this study, true variations in responses to questions might have been overlooked. Second, the prediction of risk scores, based on any narration effect, was an important anticipated outcome for this study. Financial knowledge was chosen as the covariate to reduce the possibly of explaining financial knowledge versus a narration effect in the analyses.
The ANCOVA model was found to be statistically signficant, F8,71 = 4.97, p < .001. Results indicated that both gender and the type of narration had a main effect on risk-tolerance scores. It was noted that men did report, on average, a higher mean risk score (M = 27.20) compared to women (M = 24.43), and the difference was statistically significant (F1,78 = 5.63, p < .01). The first null hypothesis was not supported.
There was also a main effect for narration on risk tolerance. This test compared risk scores for those responding to a female narration (M = 26.10), male narration (M = 26.52), no narration (M = 27.18), and pen-and paper (M = 23.45). Participants in the narrated groups scored higher than the control group, and the differences were significant (F3,76 = 3.20, p < .05). Simple contrast comparisons indicated significant differences between the control group and the three stimuli groups, with significance ranging from p = .01 to p = .04. The second null hypothesis was not supported.
It was also noted that the covariate, financial knowledge, was significantly positively associated with risk-tolerance scores (F1,78 = 12.95, p < .001). That is, those who reported a higher subjective knowledge level exhibted greater risk-tolerance scores. As such, the third null hypothesis was not supported.
Most importantly, there was a significant interaction effect between the the type of narration (i.e., female, male, none, and control group) and the gender of the participant, on calculated risk-tolerance scores (F3,76 = 2.74, p < .05). The interaction is represented in Figure 1. As shown in the figure, females and males were affected differently by the type of narration used in the experiment. Specifically, when controlling for financial knowledge, women were predicted to exhibit higher risk scores (M = 26.86 v M = 26.17) when responding to male narrated questionning. The opposite was true when questions were narrated by a female. Men exhibited higher estimated scores than women (M = 28.30 v M = 23.91). There was a significantly wider difference in risk scores between women and men when the risk questions were viewed but not narrated. Men scored highest with this stimulus (M = 29.80 v M = 24.57). Women's scores, when compared to those with male narration, were estimated to be lower. Finally, men in the control group displayed the lowest estimated tolerances for risk. Women in the control group were also predicted to have a relatively low risk tolerance (M = 23.50), but slightly higher than that for men (M = 23.40). Overall, the model explained 28.70 percent (Adjusted R-squared) of variance in risk-tolerance scores. Financial knowledge alone explained 15.40 percent of the variability in tolerance for risk.
Figure 1. Estimated Risk-Tolerance Mean Scores Based on Interaction Between Gender and Voice Narration
Discussion and Conclusion
The results from this study can be used to inform the practice of financial planning. As explained by Most, Sorber, and Cunningham (2007), "Gender provides a powerful social heuristic for structuring incoming information" (p. 287). In this study there were statistically significant differences between women and men in terms of risk tolerance as well as differences among the groups used to assess risk tolerance. Further, there was also an important interaction between gender and narration that is noteworthy. When controlling for financial knowledge-a predisposing personal factor-women were more likely to have a predicted tolerance for risk higher than that for men when responding to questions narrated by a male voice. When the same questions were narrated by a female voice, men's estimated scores on the risk assessment were significantly higher than for women. Women in the no narration group scored between women who responded with a male and female voice, but lower than men in the no narration group. Overall, men in the no voice group were predicted to have the highest risk-tolerance scores. The pen-and-paper control group exhibited the lowest estimated scores, controlling for financial knowledge.
The interaction results noted in this study suggest that the way in which clients receive and respond to data gathering questionnaires can have a significant impact on their subjective judgments. These results tend to lend support to Hilgert and Hogarth's (2003) assertion that consumers have a preference for video interaction in the domain of money and household finance. Regardless of the narration category, respondents who viewed the risk questions scored higher than those who answered the same questions with a traditional pen-and-paper process. This does not mean that the video approach is more accurate; the results simply indicate that the use of video (such as through online risk tolerance assessments that clients may be asked to complete prior to meeting with their financial adviser), as a way to assess risk tolerance, has an influence on the way questions are perceived and answered. Clearly, the traditional pen-and-paper method of risk-tolerance assessment provides the lowest risk scores for women and men. Practically speaking, when controlling for financial knowledge, there were no real differences between women and men on the pen-and-paper test.
As predicted in the literature, women responded differently than men when listening to male narrated questions. It is possible that what is occurring is similar to what Jones et al. (2010) hypothesized, in that women are subconsciously sending a message of increased tolerance for risk when responding to a man's voice. It was also noted that men's scores on risk tolerance were higher when responding to questions narrated by a female. Again, this might be a biological reaction prompting a signal of enhanced willingness to take risks as a trigger of male strength and vitality.
What is most intriguing is the finding that men's scores, controlling for financial knowledge, were highest when they read the questions via video but did not hear a voice. It may be that men have a preference for video assessment. That is, men may be predisposed to responding to direct video instructions, whereas women prefer greater interaction through narration. The male preference for video may be, as Hilgert and Hogarth (2003) noted, a method that men use to reduce the time needed to respond to questions. That is, video may be perceived by men, more so than for women, as a time saving device that matches their learning style.
Regardless of the underlying cause of these differences, women and men in this study reacted very differently when exposed to the video assessment stimuli. The implications for financial planners are multifaceted. First, the use of technology as a data gathering tool appears to alter the way in which participants evaluate questions. Financial advisers ought to expect that scores will be higher for those who respond to a video-based assessment compared to a traditional pen-and-paper assessment. Second, the gender of the narrator does matter. Women tend to respond more positively to risk tolerance questions when hearing a male's voice. Men are apt to respond more positively with a female's voice. Third, it is possible that advisers who meet with clients to discuss financial risk-tolerance issues, in terms of establishing an assessment of attitudes, run the risk of influencing client responses. Although there is no direct evidence for this hypothesis, it is possible that male advisers who work with female clients may find that their client's responses are skewed to the high side, in comparison to a benchmark pen-and-paper assessment. Women advisers who ask questions of male clients will find the same pattern of response, but to an even greater extent.
There is insufficient data from this study or literature on the topic to know if the pen-and-paper method or a narrated video approach provides a more accurate description of a client's tolerance for risk. What is known, however, is narrated assessment techniques do skew responses, as does the simple choice to use a video assessment technique. A female adviser working with a woman client should expect lower risk answers. If the same client worked with a male adviser, accounting for the client's subjective level of financial knowledge, the client's risk responses will be much higher. The safest approach is to prescribe a standard risk-assessment technique within a firm. That is, firms should choose one consistent method of client assessment. For example, if a pen-and-paper approach is to be used, the tactic should be employed in all client situations. If, on the other hand, a firm prefers a narrated method then all client assessments might be conducted using a male-to-male or female-to-female style. Great variation from pen-and-paper assessment will occur in mixed-gender assessments.
The results from this experiment are, by nature, exploratory. This is the first experimental study to look at the role of video narration as a factor influencing risk-tolerance attitudes. The findings are intriguing and warrant further study. Specifically, a future experimental test should introduce another stimulus, namely, a visual representation of the narrator. Obviously, this brings into play other issues related to the relative attractiveness of the narrator, as well as environmental factors such as clothing, color, and background effects. However, the introduction of a visual narrator may help answer questions related to not only the narrator effect, but also perceived client biases based on the gender of the planner.
1 Three men were unable to fully complete the video with no narration group test, so the mean score of the seven men who did complete the video with no narration were substituted for the three missing scores. All future references to the men's scores for video with no narration group assumes the 10 scores that were obtained rather than purely 10 unique men.
Arano, K., Parker, C., & Terry, R. (2010). Gender-based risk aversion and retirement asset allocation. Economic Inquiry, 48, 147-155.
Arrabito, G. R. (2009). Effects of talker sex and voice style of verbal cockpit warnings on performance. Human Factors, 51, 3-20.
Edworthy, J., Hellier, E., & Rivers, J. (2003). The use of male or female voices in warning systems: A question of acoustics. Noise and Health, 6, 39-50.
Grable, J. E. (2008). Risk Tolerance (pp. 1-20). In J. J. Xiao (Ed.), Advances in Consumer Financial Behavior Research. New York: Springer.
Grable, J. E., & Lytton, R. H. (1999). Financial risk tolerance revisited: The development of a risk assessment instrument. Financial Services Review, 8, 163-181.
Hallahan, T. A., Faff, R. W., & McKenzie, M. D. (2004). An empirical investigation of personal financial risk tolerance. Financial Services Review, 13, 57-78.
Hilgert, M. A., & Hogarth, J. M. (2003). Household financial management: The connection between knowledge and behavior. Federal Reserve Bulletin, 89, 309-322.
Jianakoplos, N. A., & Bernasek, A. (1998). Are women more risk averse? Economic Inquiry, 36, 620-630.
Jones, B. C., Feinberg, D. R., DeBruine, L. M., Little, A. C., & Vukovic, J. (2010). A domain-specific opposite-sex bias in human preferences for manipulated voice pitch. Animal Behaviour, 79, 57-62.
Kohler, M. P. (1996). Risk-taking behavior: A cognitive approach. Psychological Reports, 78, 489-490.
McMinn, M. R., Brooks, S. D., Triplett, M. A., Hoffman, W. E., & Huizinga, P. G. (1993). The effects of God language on perceived attributes of God. Journal of Psychology and Theology, 21, 309-321.
Most, S. B., Sorber, A. V., & Cunningham, J. G. (2007). Auditory stroop reveals implicit gender associations in adults and children. Journal of Experimental Social Psychology, 43, 287-294.
Neelakantan, U. (2010). Estimation and impact of gender differences in risk tolerance. Economic Inquiry, 48, 228-233.
Roszkowski, M. J., & Davey, G. (2010). Risk perception and risk tolerance changes attributable to the 2008 economic crisis: A subtle but critical difference. Journal of Financial Services Professionals, 64(4), 42-53.
Roszkowski, M. J., Davey, G., & Grable, J. E. (2005). Insights on measuring risk tolerance from psychology and psychometrics. Journal of Financial Planning, 18 (4), 68-76.
Schubert, R., Brown, M., Gysler, M., & Brachinger, H. W. (1999). Financial decision-making: Are women really more risk-averse? The American Economic Review, 89, 381-385.
Weber, E. U., Siebenmorgen, N., & Weber, M. (2005). Communicating asset risk: How name recognition and the format of historic volatility information affect risk perception and investment decisions. Risk Analysis, 25, 597-609.
West, G. B., Moskal, P. D., Dziuban, C. D., & Rumbough, L. P. (1996). Gender and marital differences for risk taking among undergraduates. Psychological Reports, 78, 315-320.
Whipple, T. W., & McManamon, M. K. (2002). Implications of using male and female voices in commercials: An exploratory study. Journal of Advertising, 31, 79-91.