Sex Differences in Risk Sensitivity Under Positive and Negative Budgets and Predictors of Choice

This article was downloaded by: [Fondren Library, Rice University ]On: 18 November 2014, At: 12:35Publisher: RoutledgeInforma Ltd Registered in England and Wales Registered Number: 1072954 Registeredoffice: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

The Journal of General PsychologyPublication details, including instructions for authors andsubscription information:http://www.tandfonline.com/loi/vgen20

Sex Differences in Risk SensitivityUnder Positive and Negative Budgetsand Predictors of ChoiceHeide K. Deditius Island a , Allen D. Szalda-Petree b &Stephanie C. Kucera ba Pacific Universityb University of Montana-MissoulaPublished online: 07 Aug 2010.

To cite this article: Heide K. Deditius Island , Allen D. Szalda-Petree & Stephanie C. Kucera(2007) Sex Differences in Risk Sensitivity Under Positive and Negative Budgets and Predictors ofChoice, The Journal of General Psychology, 134:4, 435-452, DOI: 10.3200/GENP.134.4.435-452

To link to this article: http://dx.doi.org/10.3200/GENP.134.4.435-452

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the“Content”) contained in the publications on our platform. However, Taylor & Francis,our agents, and our licensors make no representations or warranties whatsoeveras to the accuracy, completeness, or suitability for any purpose of the Content. Anyopinions and views expressed in this publication are the opinions and views of theauthors, and are not the views of or endorsed by Taylor & Francis. The accuracy ofthe Content should not be relied upon and should be independently verified withprimary sources of information. Taylor and Francis shall not be liable for any losses,actions, claims, proceedings, demands, costs, expenses, damages, and other liabilitieswhatsoever or howsoever caused arising directly or indirectly in connection with, inrelation to or arising out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Anysubstantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,systematic supply, or distribution in any form to anyone is expressly forbidden. Terms& Conditions of access and use can be found at http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/loi/vgen20

http://www.tandfonline.com/action/showCitFormats?doi=10.3200/GENP.134.4.435-452

http://dx.doi.org/10.3200/GENP.134.4.435-452

http://www.tandfonline.com/page/terms-and-conditions

http://www.tandfonline.com/page/terms-and-conditions

435435

The Journal of General Psychology, 2007, 134(4), 435–452Copyright © 2007 Heldref Publications

Sex Differences in Risk Sensitivity Under Positive and Negative Budgets

and Predictors of Choice

HEIDE K. DEDITIUS ISLANDPacific University

ALLEN D. SZALDA-PETREE STEPHANIE C. KUCERA

University of Montana–Missoula

ABSTRACT. The authors investigated sex differences in human risk sensitivity by using a computerized choice task with an energy budget analogue. In addition, they explored possible personality predictors of variance sensitivity. The authors modified the traditional energy budget model from those used in risk-sensitive foraging research with nonhu-man animals for appropriate use with a human population. Participants chose between 2 computer icons, 1 yielding a fixed-point reward and the other offering variable points. Men were risk prone in the negative budget and risk averse in the positive budget. Women were risk averse in the negative budget. Personality was not predictive of risk-sensitive bias. Interpreting the results using an evolutionary model, the authors found support for a biological and environmental construct of risk-sensitive behavior.

Keywords: choice, parental investment theory, risk sensitivity, sex differences

OPTIMAL FORAGING THEORY (OFT) states that in a natural environment, it is adaptive for a choosing (foraging) organism to select the food patch or resource option that best suits its needs because the selection increases the forager’s probability of survival (Stephens & Krebs, 1986). Although the term foraging traditionally refers to the acquisition of food by nonhuman animals, it may also include the acquisition of other resources. Consequently, for this study we define foraging as the search, selec-tion, and seizure of any desirable resource by human or nonhuman animals.

OFT provides a comparative framework for understanding choice under varying economic circumstances. Researchers have theorized that the consistent

Address correspondence to Heide K. Deditius Island, Department of Psychology, Pacific Uni-versity, 2043 College Way, Forest Grove, OR 97116, USA; [email protected] (e-mail).

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014

436 The Journal of General Psychology

selection of one resource above another, based on the energetic merits of that choice, has adaptive importance to the organism. Thus, an organism should make choices that will maximize its energetic intake by selecting the resource that yields the greatest net energetic return (Pyke, Pulliam, & Charnov, 1977). OFT indicates that foraging strategies may result from both natural selection processes and current environmental conditions; that is, a specific foraging strategy (e.g., greater sensitivity to the type of reinforcement than to the amount of reinforcement) is the result of heritable variation, the degree to which a trait and behavior are passed on genetically, and immediate environmental conditions. The heritable variation of a particular trait is adaptive if variation in the heritable trait or behavior causes variation in fitness and thus reproductive success (Daly & Wilson, 2002).

Because all organisms require resources for safety, reproduction, nutrition, and health, the forager’s choice behavior should always maximize or minimize resource exploitation in a way that is consistent with the risk–reward relations between three factors: currency (energy gain), time (search and handling), and solving for the option yielding the optimal cost–benefit outcome (Charnov, 1976). The ideal strategy for a forager depends on the individual needs of that organism and the evolutionary pressure exerted through natural selection and reproduc-tive or survival success. Consequently, natural selection will favor individuals within a population who contribute the most to subsequent generations (Tooby & Cosmides, 1990).

OFT predicts that organisms will prefer a food source that provides the maximum amount of food per unit of time (Pyke et al., 1977). Thus, organisms should seek to minimize delay to reward and maximize reward amount. If this is true, foragers receiving a mean reward amount should be indifferent to variation in amount of reward per trial. However, many researchers have found that some organisms display a bias toward a variable reward option (Bateson & Kacelnik, 1997; Caraco, 1981, 1982; Caraco, Martindale, & Whitman, 1980). For example, Bateson and Kacelnik found that under a negative energy budget (caloric-intake requirement not yet met), starlings (Sturnus vulgaris) were significantly more likely to choose the variable delay option than the constant delay option.

Risk Sensitivity and Foraging

Risk-sensitive foraging theory (RSF) is a set of models for which the signifi-cance of variation between alternative responses is the main tenet. Because RSF respects the significance of variance associated with differing strategies, research-ers may think of it as variance-sensitive theory. In this article, we define a risky decision as a decision that involves probable variation. The term does not imply an increased danger of predation. A typical RSF procedure affords the organism two or more foraging options (patches) yielding the same average amount of nourishment, but the variability of obtainable food (resource) differs from option

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014

to option (Bateson & Kacelnik, 1995). If the forager always favors the constant or fixed option, the RSF theorist refers to its preference as risk averse. If the forager prefers the option with greater variability, the RSF theorist refers to this bias as risk prone. Consistent with OFT, RSF theorists assume that animals will behave in ways that maximize their inclusive fitness or their direct (immediate relatives) and indirect (extended relatives) reproductive success. According to Smallwood (1996), the key contributions of RSF are the sensitivity of organisms to variability in the reward and the preference of one foraging option over another in relation to the organism’s energetic gain and loss. In addition, the amount is variable about the mean. Thus, the specific foraging choices of an organism may reflect where the animal falls within an energetic continuum or budget. The specific choices may vary depending on whether the organism’s current energy stores are in deficit (negative budget) or surplus (positive budget).

RSF researchers typically use the daily energy budget as an explanatory construct. When the fixed and variable options have the same mean value, the daily energy budget model predicts that risk-prone choices will appear under a negative energy budget, whereas risk-averse choices will appear under a positive energy budget (Smallwood, 1996). RSF researchers have extensively evaluated the daily energy budget model, which has a participant base with an extensive representation of taxonomic classes, including nectarivorous birds (Carter, 1991; Carter & Dill, 1990; Pimm, 1978; Wunderle, Santa-Castro, & Fetcher, 1987), nectarivorous insects (Real, 1980; Real, Ott, & Shvereine, 1982; Waddington, Allen, & Heinrich, 1981), granivorous birds (Bateson & Kacelnik, 1997; Caraco, 1980, 1981, 1983; Caraco et al., 1990; Case, Nichols, & Fantino, 1995; Hamm & Shettleworth, 1987; Tuttle, Wulfson, & Caraco, 1990), migratory birds (Moore & Simm, 1990), cichlid fish (Roche, Dravet, Bolyard, & Rowland, 1998; Young, Clayton, & Barnard, 1990), adult humans (Pietras & Hackenberg, 2001, 2003; Pietras, Locey, & Hackenberg, 2003), and rodents (Barnard & Brown, 1985; Kagel, MacDonald, Green, Battalio, & White, 1986; Kirshenbaum, Szalda-Petree, & Haddad, 2000, 2002; Lawes & Perrin, 1995).

Pietras and colleagues tried to assess risk sensitivity among humans through energy budget simulations by using point requirements and monetary earnings in place of caloric reinforcement (Pietras & Hackenberg, 2001, 2003; Pietras et al., 2003). In one such study, Pietras and Hackenberg (2001) used a small-n within-subject design by which they provided participants with repeat-ed choices between a fixed number and a variable number of points exchange-able for money. Each participant had to meet a point requirement to exchange those points for money at the end of the experimental session. The goal of that study was to develop a human procedure in the area of risk-sensitivity with monetary outcomes analogous to those of the nonhuman animal studies using caloric budgets. The procedure for that study involved 30 forced-choice and 30 free-choice trials, through which the participants learned the point-earn-ing outcomes. They predicted that behavior would match the risk-sensitivity

Island, Szalda-Petree, & Kucera 437

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014


constructs of a daily energy budget despite the conversion of caloric energy into monetary units. In other words, under conditions in which the earning budget conditions were positive (i.e., outcome was certain), the participant would be risk averse, and when the earning budget conditions were negative (i.e., outcome was uncertain), the participant would be risk prone. They found risk-averse behavior under positive budget conditions and risk-prone behavior under negative conditions for 2 of 3 participants, whereas the 3rd participant did not demonstrate strong risk-sensitive biases.

In a second study, Pietras and Hackenberg (2003) again used a small-n design (n = 3) but modified the procedure to maintain the mean number of points as a constant across fixed and variable outcomes while manipulating the probabilities of reward within the variable option to reflect the following distributions: (a) nor-mal, (b) rectangular, (c) exponential, and (d) bivalued (bimodal). Their goal was to investigate human sensitivity to risk as a function of the form of the probability distribution presented. Therefore, the researchers limited verbal instruction. They considered the participants risk averse if their choice behavior was above risk indifference (proportion of choices at 50%) for the fixed option and risk prone if their choice behavior was below risk indifference. The participants demonstrated risk-averse behavior regardless of the probably distribution within the variable-choice option (Pietras & Hackenberg, 2003).

Although Pietras and Hackenberg (2001, 2003) have undertaken the energy budget analogue for risk sensitivity among humans, the comparative litera-ture addressing the potential difference in risk sensitivity between the sexes is incomplete. Because existing research on human risk-sensitive foraging does not address behavioral differences between the sexes with regard to choice (Pietras & Hackenberg, 2001, 2003; Pietras et al., 2003), researchers should consider the differential perceptions of men and women regarding conditions of risk.

RSF predicts that during extreme negative conditions when foragers are less likely to come out ahead (positive budget) in exploiting a fixed resource path, they use a variable resource source. Although the variable resource options offer less certainty of outcomes, they may provide the forager an opportunity to survive or succeed in mating, feeding, or garnering necessary resources for other biologi-cal needs. The foraging strategy switch from risk averse to risk prone reflects the forager’s attempt to survive by seeking the variable option. The flexibility of risk preference based on environmental context may be different for men and women. If Trivers’ (1985) evolutionary explanation (i.e., parental investment theory) generalizes to conditions of choice, women should behave conserva-tively. Natural selection would favor women who demonstrate conservation, as maternal responsibility and high parental investment necessitate. Men could switch their foraging strategies more readily because the bulk of their investment lies with their own well-being, as the woman assumes the majority of parental investment. For example, if male mammals generally adhere to a polygamous mating strategy, they do not have the burden of safeguarding or provisioning their

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014

offspring and therefore may not demonstrate risk aversion across all contexts. A female mammal may also choose to engage in a variety of mating strategies: monogamy, polyandry, or polygamy. However, under most circumstances, she will be responsible for providing for and rearing her offspring. Therefore, female choice behavior may reflect a more conservative orientation because of the much higher parental investment.

The Present Study

In this study, we experimentally tested the OFT and RSF models in a human adult population. Our goal was to assess the credibility of these models as explanatory constructs for the sensitivity of the choosing organism, specifically male and female human adults, to variability. Therefore, we predicted that the sensitivity of the foraging organism to variability would dictate the risk strategy that the individual uses in making his or her selection of a choice.

We posed two primary hypotheses. Hypothesis 1 predicted that men would select a greater-than-chance number of constant option choices (risk-averse bias) under the positive budget (winning condition) and a greater-than-chance number of variable option choices (risk-prone bias) under the negative budget (losing condition). How-ever, we predicted that women would select a greater-than-chance number of constant option choices (risk-averse bias) under both the positive (winning) condition and the negative budget (losing) conditions. Researchers have not addressed sex differences in choice. However, using the parental investment model as an evolutionary frame-work for female choice behavior, we anticipated that women would demonstrate the greatest risk aversion in the positive budget condition.

Hypothesis 2 was exploratory and evaluated behavioral correlates of risk-sensitive foraging behavior within the domains of the five-factor personality measure, Revised NEO Personality Inventory (NEO PI-R; Costa & McCrae, 1995). There are individual differences between foraging conspecifics, organisms of the same species, regardless of whether they are human or nonhuman. In light of this circumstance, researchers can infer that it is possible that individual dif-ferences in character or personality could mediate an organism’s choice behavior. Because our sample included male and female human adults, the inclusion of a measure sensitive to individual differences in personality and specifically with a risk-related component was relevant. Extraversion and excitement seeking are positively correlated with high-risk sexual behavior (Miller et al., 2004), alco-hol use (Cookson, 1994), and openness to experience (Aluja, Garcia, & Garcia, 2003). Similarly, openness to experience is positively associated with sensation seeking (Eisenman, Grossman, & Goldstein, 1981) and with socially acceptable risk taking and sensation seeking (Diehm & Armatas, 2004). Although risk sensi-tivity does not refer to danger directly, sensation seeking may reflect sensitivity to variability and therefore a proclivity for more risk-prone behavior. Consequently, we predicted that individuals who demonstrate higher risk-prone responding also


Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014


would score higher on the domains of extraversion and openness to experience in the five-factor model of personality as measured by the NEO PI-R (Costa & McCrae, 1992).

Method

Participants

Participants were male (n = 90) and female (n = 145) psychology students from an undergraduate, introductory psychology course. Men’s mean age was 20 years (SD = 3.75 years), whereas women’s mean age was 21 years (SD = 5.43 years). We excluded from analysis participants who showed no variation in choice across all free-choice trials (e.g., they exclusively chose one option across all 75 trials) because of their possible lack of understanding or engagement in the task (n = 5).

Apparatus

We used two Dell computers operating under Windows 95 to display the stimuli and record responses. Each computer had a serial mouse, and the task was programmed in Visual Basic 6.0 by the investigators. Both computers used SVGA monitors that were set at a resolution of 640 × 480 pixels.

Measures

The NEO PI-R Form S is a self-administered measure with 240 items appro-priate for men and women of all ages (Piedmont, 1998). The NEO PI-R Form S may provide a valid and reliable means for the investigation of correlates between personality factors and choice behavior and how those personality dimensions may influence risk preferences in a choice task. We selected this measure on the basis of the broad applicability and flexibility of the scale and a reliability generalization by Caruso (2000). Furthermore, researchers have widely cited the NEO-PI-R in an array of risk-related behavioral research, including sensation seeking, impulsivity, and risk taking (Costa & McCrae, 1992).

Procedure

We divided the experimental task into two phases. Phase 1 consisted of 10 forced-choice trials used to familiarize participants with the two types of point outcomes: fixed and variable. Phase 2 consisted of 75 free-choice trials used to access choice bias.

Phase 1. We delivered 10 forced-choice trials. For each trial, the computer pre-sented only one choice box option. We required participants to mouse-click on

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014

the choice box three times to select the box. Because of the ease of selection in using a mouse, we had three clicks designate a commitment to the choice and also allow the participant an opportunity to change the choice selection during the free-choice trials. After the three mouse clicks, the task’s program displayed on the screen the point value for that box for 3 s, and the next forced-choice trial commenced. The program randomly assigned the order of choice-box pre-sentations with the provision that it presented each choice box five times. On the completion of the forced-choice trials, the program presented an instruction screen indicating the start of Phase 2.

Phase 2. Phase 2 consisted of five blocks of 15 free-choice trials totaling 75 total free-choice trials. The free-choice trials were the same as the forced-choice trials with the exception that both choice options were presented concurrently. In addi-tion, the program presented the current trial number in the upper right-hand cor-ner of the screen so the participant could track the number of trials remaining.

We randomly assigned participants to a budget condition—winning and ahead (positive budget) or losing and behind (negative budget)—yielding four independent sex-and-budget groups. We used a sham competitive component to manipulate the budget conditions. The free-choice instructions indicated that participants would compete for accumulated points against an introductory psychology student at another university, although there was no competitor. The budget status during the task depended only on random assignment to either the positive or negative budget.

Clicking on a Start box at the bottom of the instructions initiated an Internet- connecting screen with a 10-s countdown to log-on with a linking designation akin to those used in Internet gaming websites. We modeled the connecting screen after Internet computer games in which participants must link with others to play interactive games.

Once the connecting screen indicated that both participants had logged on, a status screen with a status bar displaying Even on a winning–losing continuum displayed for 10 s and preceded the 75 free-choice trials. The computer provided feedback during the free-choice trials on relative performance to the supposed competitor after each trial block (15 trials) with a 10-s presentation of the status screen. Thus, after the onset of the task, there were four presentations of the status screen throughout the session. The location of the status bar along the winning–losing continuum indicated whether the participant’s point total was behind or ahead of the supposed competitor. This status was not contingent on the participant’s choices but was purely based on budget condition. Participants assigned to the winning condition always received status-screen feedback indi-cating that they were ahead of the competitor (i.e., winning), and those assigned to the losing condition always received status-screen feedback indicating that they were behind the competitor (i.e., losing). The consequent trial blocks did not reset the counter to zero. The participant merely received feedback from the


Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014


presentation of the status screen of his or her current status relative to the com-petitor after each block.

To ensure believability of the manipulation, small movements of the status bar within the assigned losing–winning gradient contributed to the perception of a contingency between point status and choice strategy. However, we programmed the movement of the status bar to be random, and it was not contingent on par-ticipants’ choices during the particular trial block and did not change the value of the feedback in terms of budget condition (see Figure 1).

Participant selection of the variable option yielded 1 or 5 points—p (1 point) = 0.5, p (5 points) = 0.5—whereas selection of the fixed option exclusively yielded

FIGURE 1. Status screen presented in all 15 trials.

{

Status bar

Status barometer

Very ahead

Ahead

Slightly ahead

Even

Slightly behind

Behind

Very behind

The red bar to your right indicates how far ahead, “winning,” or behind, “losing,” you are with your competitor.

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014

3 points, p (3 points) = 1.0. We pseudorandomly assigned the choice boxes (blue or yellow) to the left or right screen position so that neither choice box presented in the same screen position for more than two consecutive trials. Additionally, we counterbalanced choice-box value (fixed or variable) across box color.

Once the experimental task was complete, we asked the participants to com-plete the NEO PI-R Form S using the Psychological Assessment Resource test booklet and response sheet. After participants completed the personality question-naire, we verbally debriefed or provided a debriefing handout to all participants.

Results

We conducted all statistical analyses of choice data using the proportion of risk-prone choices. We calculated the risk-prone proportion by dividing the number of variable option choices by the total number of trials, excluding the first block of 15 trials (trials = 60).

We excluded the first block of 15 trials from all calculations because we used it solely to establish the believability of the competition manipulation, and thus the actual budget manipulation did not begin until Trial 16, after the first status screen presentation. Additionally, the first block of 15 trials is unsuitable as a baseline condition for comparison with subsequent conditions because of the limited resolution in scaling associated with proportions from 15 trials (e.g., a 1-unit change in choice results in a 6.67% change in proportion). Moreover, from an analytical perspective, the distributions associated with proportions that are drawn from widely different event sizes, in this case 15 versus 60 trials, are not appropriate for comparison.

We conducted one-sample t tests for each sex-and-budget group using the proportion of risk-prone choices to determine whether risk-prone choices devi-ated from chance performance (proportion = .50). The one-sample t tests for men demonstrated risk-sensitive behavior for both budget groups (see Figure 2). Under the positive budget, men demonstrated a significant risk-averse bias (M = 0.43, SE = 0.04), t(41) = –2.42, p < .05, d = 0.36, whereas under the negative bud-get, they demonstrated a significant risk-prone bias (M = 0.55, SE = 0.02), t(46) = 2.97, p < .05, d = 0.44. The one-sample t tests for women failed to show risk sensitivity under the positive budget (M = 0.48, SE = 0.02, t(79) = 0.02, p > .05, d = 0.001); but, the tests for women succeeded in showing risk-averse bias under the negative budget (M = 0.47, SE = 0.02), t(64) = –2.09, p < .05, d = 0.26.

We conducted a 2 (budget) × 2 (sex) analysis of variance (ANOVA) by using the proportion of risk-prone choices to determine whether sex and budget interacted. Results from the ANOVA revealed a main effect for budget, F(1, 230) = 5.50, p < .05, f = 0.15; no main effect for sex, F(1, 230) = 0.61, p > .05, f = 0.05; and a significant Budget × Sex interaction, F(1, 230) = 11.23, p < .01, f = 0.22. Post hoc analysis of the interaction using Tukey’s Honestly Significant Difference (HSD; .05) revealed that the male-and-negative budget group was


Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014


significantly more risk-prone than were either the male-and-positive budget group or the female-and-negative budget group. All other pairwise comparisons were not significantly different, all ps > .05. Although the analysis indicated a significant Sex × Budget interaction, the post hoc analysis showed that the inter-action was based solely on the effectiveness of the budget manipulation on male choice and the ineffectiveness of the budget manipulation on female choice.

We conducted correlations (r) to assess the predictive validity of personality score on the NEO PI-R and the proportion of risk-prone choices. Among women, there were no significant correlations between any NEO PI-R domain or facet scores and proportion of risk-prone choices (see Tables 1 and 2). Among the men, there were no significant correlations between the NEO PI-R domain scores and proportion of risk-prone choices, but there were two significant facet-scale correlations. Activity, r(230) = –.229, p < .05, and tender-mindedness, r(230) = –.217, p < .05, were both negatively correlated with proportion of risk-prone choices. Although the results did not reveal a relation between the domains of extraversion and openness to experience, activity is a facet scale within the domain of extraversion. Additionally, both of the facets activity and tender-mindedness were directionally consistent with RSF.

Those who score low on tender-mindedness are less moved by others’ needs and appeals to pity. They consider themselves pragmatic realists and rational decision makers (Costa & McCrae, 1992). Low scores on activity indicate less energy expenditure as related to a relaxed and leisurely manner. Both of these

FIGURE 2. Proportion of risk-prone (RP) choices by men and women under positive and negative energy budgets. Error bars represent standard deviations.

Men Women

0.7

0.6

0.5

0.4

0.3

Prop

ortio

n of

RP

Cho

ices

Negative Budget Positive Budget

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014

facets’ scores fall within the theoretical conception of RSF because the correla-tions are negative, indicating that risk-prone responses were elevated. However, other anticipated relationships, such as excitement seeking, were not correlated


TABLE 1. Correlations Between NEO Personality Inventory-Revised Form S (P. T. Costa, Jr. & R. R. McCrae, 1992) Domain Scores and Proportion of Risk-Prone (RP) Choices

Pearson’s r

Domain Mena Womenb Combinedc

Neuroticism –.030 .062 .016Extraversion –.065 –.034 –.049Openness to experience –.116 –.087 –.100Agreeableness –.069 .103 .238Conscientiousness –.172 –.072 –.109

an = 90. bn = 145. cN = 235.

TABLE 2. Correlations Between NEO Personality Inventory-Revised Form S (P. T. Costa, Jr. & R. R. McCrae, 1992) Facet Scales for Extraversion and Agreeableness and the Proportion of Risk-Prone (RP) Choices

Pearson’s r

Facet scale Mena Womenb Combinedc

Extraversion E1: Warmth –.132 .008 –.056 E2: Gregariousness .092 –.026 .018 E3: Assertiveness –.010 –.139 –.085 E4: Activity –.229* –.012 –.099 E5: Excitement seeking .048 .017 .035 E6: Positive emotions –.083 .037 –.019Agreeableness A1: Trust –.039 .042 .007 A2: Straightforwardness –.034 .130 .056 A3: Altruism –.109 .047 –.018 A4: Compliance .07 .044 .048 A5: Modesty .056 .067 .047 A6: Tender-mindedness –.217* .096 –.042

an = 90. bn = 145. cN = 235. *p < .05.

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014


with risk-prone behavior, and therefore researchers should interpret the two facet correlations cautiously.

A measure’s reliability is actually a product of the sample’s scores and not a product of the measure. Therefore, the reliability of a measure changes with the change in its sample. We used Cronbach’s alpha to determine consistency in responding among the participants. The combined reliability coefficients ranged from .81 to .88, indicating reasonable domain score reliability for the NEO PI-R (see Table 3). The score reliability for the domain scales was in the high-reliability range for this measure.

Discussion

As we predicted in Hypothesis 1, men and women demonstrated risk sensitivity in their choice patterns. Men showed a significant risk-prone bias under the negative budget and a significant risk-averse bias under the positive budget. Furthermore, women were risk averse under both conditions, although significantly so under the negative budget only. Therefore, the results support Hypothesis 1 fully for men and partially for women. This finding is consistent with Pietras and Hackenberg’s (2001) results: Of their three participants, 1 man and 1 woman demonstrated risk sensitivity, whereas the other participant (female) was risk indifferent.

Although small-n designs are typical in comparative research, one of our goals was to apply a group design to a computerized choice task by using an energy bud-get model. Pietras and Hackenberg (2001, 2003) applied a minimum point require-ment in their manipulation, whereas our model used an energy budget analog of points exchangeable for a monetary reward. Instead of requiring an arbitrary point value as the minimum requirement, we required participants to win the computer game to receive their monetary reward. Therefore, the minimum requirement was to win the game against the competitor. This expectation simulated a more extreme

TABLE 3. Cronbach’s Alpha Reliabilities for Combined NEO Personality Inventory-Revised Form S (P. T. Costa, Jr. & R. R. McCrae, 1992)

Domain Cronbach’s α M SD

Neuroticism .87 92.74 18.10Extraversion .83 116.80 17.10Openness to experience .81 118.71 19.38Agreeableness .86 115.38 17.44Conscientiousness .88 111.18 18.84

Note. N = 235.

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014

energy state, an all-or-nothing requirement. Sternberg, Bokat, Kass, Alboyadjian, and Gracely (2001) demonstrated that the inclusion of competition in (a) an athletic activity, track, and (b) a passive activity, a computer game, increased interest and investment of human participants in the prescribed activity.

To increase participant interest and investment in earning high points and consequently in winning the computer task, we included indirect competition with an Internet competitor. Because it is likely that most foraging organisms encounter a competitor in any given resource area, the inclusion of competition in the computer task seemed both relevant and appropriate. Competitors impose positive and negative energy budgets in both natural and artificial environments. In accord with the daily energy budget model, we introduced positive and nega-tive budget conditions as winning and losing, respectively, against an assumed Internet competitor on a computerized game of choice.

In our task, the participant was unable to seek alternatives. Assuming that paren-tal investment theory (Trivers, 1985) applies to conditions of choice, only women should behave conservatively. Natural selection likely favors women who dem-onstrate conservation, as necessitated by maternal responsibility and high parental investment. Applying these ideas to risk, women should be risk averse under modest-risk and high-cost conditions. The female participants in this study demonstrated risk indifference in the positive budget. Therefore, either the daily-energy-budget model is incorrect in explaining choice behavior for women in a positive-budget condition or a potential confound systematically varied with our procedure.

If the theoretical model of parental investment is applicable to male choice behavior, then it is reasonable to conclude that men may use different foraging strategies than do women on the basis of their own budget constraints and those that the environment imposes on the forager. Because all decisions may affect fit-ness, there may be a biological tendency for a particular risk preference based on the strategies that most benefit the individual. We assert that even simple decisions may reflect these biases and that the sex of the forager will reflect differing risk strategies. For example, it may be expected that men will use a risk-prone strategy in the negative budget, because men may interpret high stakes with less aversion than may women. However, for women, the perception of those same high stakes may be even more extreme because the consequences of that choice will affect both their own well-being and the health and welfare of their offspring.

Future investigations of risk-sensitive behavior may warrant the inclusion of the participant’s motivational and competitive assessment of the task. Men may have experienced a slightly greater investment in optimizing their point cache in comparison with women in this task. In addition, if men were more motivated in the task in comparison with women, this difference could influence the risk strategy. Last, if the competitor induced greater distress in one sex than another, this difference could also confound the data. The levels of discomfort that the competitor induced in the male and female participants of our study are unknown. Because researchers have suggested that women are more likely to experience


Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014


discomfort as a consequence of direct competition (Benenson et al., 2002; Fülöp, 2002; Gneezy, Niederle, & Rustichini, 2003), it is difficult to know whether dis-comfort played a role in the choice behavior of our female participants.

The current investigation of risk sensitivity did show risk-sensitive behav-ior among men and women (although only in the negative budget for women). Further refinement of our methodology is essential for cogent identification of environmental influences of risk preference. One way to approach this need is to manipulate delay rather than points. Manipulating delay rather than points may provide a perceived opportunity to leave the study early, if researchers assume that participants are only engaging in the study to fulfill a credit requirement. This manipulation is a reasonable alteration to the method because theories such as scalar expectancy theory predict that differences vary with the unit of energy researchers manipulate. When variability occurs in food amount (or, applied to humans, money or points for money), researchers expect risk aversion. When variability occurs in delay, researchers expect risk-prone behavior (Bateson & Kacelnik, 1996).

Pietras et al. (2003) investigated risk sensitivity by using a small-n (n = 4) within-subject energy budget model that they designed to account for temporal constraints. In that study, the energy analogue was time. The variable and fixed choices represented a fixed delay of 10 s, and the variable choice was either 2 or 18 s, with a time constraint for each trial block. If the participant’s delay counter did not exceed the delay threshold for the given block, then the researchers added 10 points to the participant’s point counter. If the points exceeded the threshold, then the participant received no points. The results of that study revealed risk aversion under the positive budget and risk proneness under the negative budget. Although, consistent with our findings, the researchers uncovered a higher-than-predicted frequency of choices for the fixed option under the negative budget condition, Pietras et al. did not explore sex differences in participants’ variability of choice preference.

Regarding our exploratory hypothesis, Hypothesis 2, the NEO PI-R revealed no predictors of risk sensitivity. Therefore, Hypothesis 2, that domain scores for extraversion and openness to experience would be predictive of choice prefer-ence, was unsupported. However, there were two significant negative correla-tions between the risk-prone bias and the facet scale for activity in the domain of extraversion and the facet scale of tender-mindedness in the domain of agreeable-ness. Both of these facets’ scores fall within the theoretical conception of RSF because the correlations are negative, indicating that risk-prone responses are elevated. There are several explanations regarding the inability of the personal-ity measure to predict risk-sensitive behavior, the first being the evolutionary premise of risk sensitivity. If risk-sensitive behavior is truly an evolved behavior as a consequence of natural selection, then risk behavior should be unrelated to individual differences in personality and thus the environmental context should predominantly constrain or influence it.

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014

A second explanation for the poor predictive ability of the NEO PI-R is that personality measures that researchers claim predict risk taking and sensation seek-ing refer to risk differently than does the foraging literature. Within the context of personality, researchers generally have defined risk as a sensitivity to peril or danger. Individuals who are or are not risky fall toward the ends of the sensation-seeking gradient, indicating they are less sensitive to danger or that they need greater environmental stimuli to be aroused (Zuckerman, 1994). In the context of foraging or risk-sensitive behavior, risk has little to do with sensitivity to danger but instead refers to one’s sensitivity to the unpredictable variability in an outcome. The theory of risk sensitivity asserts that switching between a constant option and variable options is fundamentally a part of survival and that the environment—not the personality—draws on this behavioral strategy. Thus, a personality instrument that measures conventional risk is not calibrated to risk in the context of risk sensi-tivity because they are two different things, linked only by semantics. Risk and risk sensitivity are not analogous, and therefore it may be disadvantageous to assume that measures that are predictive of one would be predictive of the other.

Because we are assuming an evolutionary model to explain risk sensitivity, an investigation of the issue of choice preference on the basis of parenting status may be a more applicable issue than that of personality. The energetic cost of parenting and the way participants change their perceptions of valued resources could influ-ence how they respond to variation in resource rewards. It may be of value for future researchers to explore the relation between menstrual phases and risk preference.

Insight into the mechanism and proclivity of male and female choice behav-ior would facilitate greater understanding in the contexts of gambling, econom-ics, addiction, and cognition. However, because applications of RSF and OFT to research are just beginning, further examination of choice behavior must be addressed before conclusive statements can be made concerning human risk sen-sitivity, a traditionally nonhuman animal mode of choice behavior.

AUTHOR NOTES

Heide K. Deditius Island is an assistant professor of psychology specializing in comparative behav-ioral neuroscience at Pacific University in Forest Grove, Oregon. Her primary research interests are the examination of comparative behavior using an integration of psychological, ecological, and evolutionary models. Allen D. Szalda-Petree is a professor of psychology specializing in comparative animal behavior at The University of Montana–Missoula. His primary research interests involve the examination of animal behavior using an integration of psychological, ecological, and evolutionary bases. Stephanie C. Kucera is the development director of the Brain Injury Association of Montana, a statewide nonprofit organiza-tion working for brain injury prevention, research, education, and advocacy. Her research interests include implementation of early intervention and long-term support programs for survivors of traumatic brain injury. She is also an adjunct faculty member of the Montana Technology Institute.

REFERENCES

Aluja, A., Garcia, O., & Garcia, L. F. (2003). Relationships among extraversion, openness to experience, and sensation-seeking. Personality and Individual Differences, 35, 671–680.


Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014


Barnard, C. J., & Brown, C. A. J. (1985). Risk-sensitive foraging in common shrews (Sorex araneus L.). Behavioral Ecology and Sociobiology, 16, 161–164.

Bateson, M., & Kacelnik, A. (1995). Preferences for fixed and variable food sources: Variability in amount and delay. Journal of the Experimental Analysis of Behavior, 63, 313–329.

Bateson, M., & Kacelnik, A. (1996). Risky theories: The effects of variance on foraging decisions. American Zoologist, 36, 402–434.

Bateson, M., & Kacelnik, A. (1997). Starlings’ preferences for predictable and unpredict-able delays to food. Animal Behaviour, 53, 1129–1142.

Benenson, J. F., Roy, R., Waite, A., Goldbaum, S., Linders, L., & Simpson, A. (2002). Greater discomfort as a proximate cause of sex differences in competition. Merrill-Palmer Quarterly, 48, 225–247.

Caraco, T. (1980). On foraging time allocations in a stochastic environment. Ecology, 6, 119–128.

Caraco, T. (1981). Energy budgets, risk and foraging preferences in dark-eyed juncos (Junco hyemalis). Behavioral Ecology and Sociobiology, 8, 213–217.

Caraco, T. (1982). Aspects of risk-aversion in foraging white-crowned sparrows. Animal Behavior, 30, 719–727.

Caraco, T. (1983). White-crowned sparrows (Zonotrichia leucophyrs): Foraging prefer-ences in a risky environment. Behavioral Ecology and Sociobiology, 12, 63–69.

Caraco, T., Blanckenhorn, W. U., Gregory, G. M., Newman, J. A., Recer, G. M., & Zwicker, S. M. (1990). Risk-sensitivity, ambient temperature affects foraging choice. Animal Behaviour, 39, 338–345.

Caraco, T., Martindale, S., & Whitam, T. S. (1980). An empirical demonstration of risk-sensitive foraging preferences. Animal Behavior, 28, 820–830.

Carter, R. V. (1991). A test of risk-sensitive foraging in wild bumblebees. Ecology, 72, 888–895.

Carter, R. V., & Dill, L. (1990). Why are bumblebees risk-sensitive foragers? Behavioral Ecology and Sociobiology, 2, 121–127.

Caruso, J. C. (2000). Reliability generalization of the NEO Personality Scales. Educa-tional and Psychological Measurement, 60, 236–254.

Case, D. A., Nichols, P., & Fantino, E. (1995). Pigeons’ preference for variable-interval water reinforcement under widely varied water budgets. Journal of Experimental Analy-sis of Behavior, 64, 299–311.

Charnov, E. L. (1976). Optimal foraging: The marginal value theorem. Theoretical Popula-tion Biology, 9, 129–136.

Cookson, H. (1994). Personality variables associated with alcohol use in young offenders. Personality and Individual Differences, 16, 179–182

Costa, P. T., Jr., & McCrae, R. R. (1992). The Revised NEO Personality Inventory manual. Odessa, FL: Psychological Assessment Resources.

Costa, P. T., Jr., & McCrae, R. R. (1995). Domains and facets: Hierarchical personal-ity assessment using the Revised NEO Personality Inventory. Journal of Personality Assessment, 64, 21–50.

Daly, M., & Wilson, M. (2002). Editorial introduction: Two special issues on risk. Evolu-tion and Human Behavior, 23, 1–2.

Diehm, R., & Armatas, C. (2004). Surfing: An avenue for socially acceptable risk-taking, satisfying needs for sensation seeking and experience seeking. Personality and Indi-vidual Differences, 36, 663–677.

Eisenman, R., Grossman, J. C., & Goldstein, R. (1981). Undergraduate marijuana use as related to internal sensation novelty seeking and openness to experience. Journal of Clinical Psychology, 36, 1013–1019.

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014

Fülöp, M. (2002, December). Competition in educational settings. Paper presented at a meeting of the Faculty of Education, University of Ljubljana, Slovenia.

Gneezy, U., Niederle, M., & Rustichini, A. (2003). Performance in competitive environ-ments: Gender differences. Quarterly Journal of Economics, 118, 1049–1074.

Hamm, S. L., & Shettleworth, S. J. (1987). Risk aversion in pigeons. Journal of Experi-mental Psychology: Animal Behavior Processes, 13, 376–383.

Kagel, J. H., MacDonald, D. N., Green, L., Battalio, R. C., & White, S. (1986). Risk aver-sion in rats (Rattus norvegicus) under varying levels of resource availability. Journal of Comparative Psychology, 100, 95–100.

Kirshenbaum, A. P., Szalda-Petree, A. D., & Haddad, N. F. (2000). Risk-sensitive forag-ing in rats: The effects of response-effort and reward-amount manipulations on choice behavior. Behavioral Processes, 50, 9–17.

Kirshenbaum, A. P., Szalda-Petree, A. D., & Haddad, N. F. (2002). Increased effort requirements and risk-sensitivity: A comparison of delay and magnitude manipulations. Behavioral Processes, 61, 109–121.

Lawes, M. J., & Perrin, M. R. (1995). Risk-sensitive foraging behaviour of the round-eared elephant shrew (Macroscelides proboscideus). Behavioral Ecology and Sociobiology, 37, 31–37.

Miller, J. D., Lynam, D., Zimmerman, R. S., Logan, T. K., Leukefeld, C., & Clayton, R. (2004). Personality and Individual Differences, 36, 1611–1626.

Moore, F. R., & Simm, P. A. (1990). Risk-sensitive foraging by a migratory bird (Den-droica coronata). Experimentia, 42, 1054–1056.

Piedmont, R. L. (1998). The revised NEO Personality Inventory: Clinical and research applications. New York: Plenum Press.

Pietras, C. J., & Hackenberg, T. D. (2001). Risky choice in humans under positive energy budgets. Journal of Experimental Analysis of Behavior, 76, 1–19.

Pietras, C. J., & Hackenberg, T. D. (2003). Human risky choice across four probability distributions. Psychological Record, 53, 443–457.

Pietras, C. J., Locey, M. L., & Hackenberg, T. D. (2003). Human risky choice under tem-poral constraints: Tests of an energy-budget model. Journal of Experimental Analysis of Behavior, 80, 59–79.

Pimm, S. L. (1978). An experimental approach to the effects of predictability on com-munity structure. American Zoology, 18, 797–808.

Pyke, G. H., Pulliam, H. R., & Charnov, E. L. (1977). Optimal foraging: A selective review of theory and tests. Quarterly Review of Biology, 52, 137–154.

Real, L. A. (1980). Fitness, uncertainty and the role of diversification in evolution and behavior. American Naturalist, 155, 623–638.

Real, L. A., Ott, J., & Shvereine, E. (1982). On the tradeoff between the mean and the variance in foraging, effect of spatial distribution and color preference. Ecology, 63, 1617–1623.

Roche, J. P., Dravet, S. M., Bolyard, K., & Rowland, W. (1998). Risk-sensitivity in forag-ing Jack Dempsey cichlids (Cichlasoma octofasciatum). Ethology, 104, 593–602.

Smallwood, P. D. (1996). An introduction to risk-sensitivity: The use of Jensen’s inequal-ity to clarify evolutionary arguments of adaptation and constraint. American Zoology, 36, 392–401.

Stephens, D. W., & Krebs, J. R. (1986). Foraging theory. Princeton, NJ: Princeton Uni-versity Press.

Sternberg, W. F., Bokat, C., Kass, L., Alboyadjian, A., & Gracely, R. H. (2001). Sex-dependent components of the analgesia produced by athletic competition. Journal of Pain, 2, 65–74.

Tooby, J., & Cosmides, L. (1990). On the universality of human nature: The role of genet-ics and adaptation. Journal of Personality, 58, 17–67.


Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014


Trivers, R. (1985). Social evolution. Santa Cruz, CA: Cummings.Tuttle, E. M., Wulfson, L., & Caraco, T. (1990). Risk-aversion, relative abundance

of resources and foraging preferences. Behavioral Ecology and Sociobiology, 26, 165–171.

Waddington, K. D., Allen, T., & Heinrich, B. (1981). Floral preferences of bumblebees (Bombus edwardsii) in relation to intermittent versus continuous rewards. Animal Behaviour, 29, 779–784.

Wunderle, J. M., Santa-Castro, S., & Fetcher, N. (1987). Risk-averse foraging in banana-quits on negative energy budgets. Behavior Ecology and Sociobiology, 21, 249–255.

Young, R. L., Clayton, H., & Barnard, C. J. (1990). Risk-sensitive foraging in bitterlings, Rhodeus sericus: Effects of food requirement and breeding site quality. Animal Behav-iour, 40, 288–297.

Zuckerman, M. (1994). Behavioral expressions and biosocial bases of sensation-seeking. Cambridge, England: Cambridge University Press.

Manuscript submitted March 9, 2005Revision accepted for publication December 19, 2005

��

��

��

��

��

��

��

��

��

�

�

��

��

� ��

� ��

� ��

� ��

� ��

� ��

� ��

� ��

� ��

� ��

� ��

� ��

� ��

� ��

� ��

��

Dow

nloa

ded

by [

Fond

ren

Lib

rary

, Ric

e U

nive

rsity

] a

t 12:

35 1

8 N

ovem

ber

2014

Sex Differences in Risk Sensitivity Under Positive and Negative Budgets and Predictors of Choice

Documents

Transcript of Sex Differences in Risk Sensitivity Under Positive and Negative Budgets and Predictors of Choice