need help to summarize two articles in MLA format atleast one and a half page each

Gender differences in reward sensitivity and information processing during decision-making

Kaileigh A. Byrne & Darrell A. Worthy

Published online: 17 February 2015 # Springer Science+Business Media New York 2015

Abstract Gender differences in reward sensitivity and information processing were examined in two studies using a dynamic decision-making task. In Experiment 1, the optimal strategy involved forgoing an option that provided larger immediate rewards in favor of one yielding larger delayed rewards. In Experiment 2, the optimal strategy was to select the option that provided larger immediate rewards because the delayed reward option never gave larger rewards than the immediate reward option. Foregone reward information was either presented or withheld. In Experiment 1, information regarding foregone rewards biased participants toward the sub-optimal choice, whereas in Ex- periment 2, foregone rewards directed participants toward the optimal option. Males selected the optimal choice more in the delayed rewards task, while females were more biased toward the poorer choice by foregone reward information. In contrast, females outperformed males in the immediate rewards task. The results suggest a gender difference in information processing styles during decision-making.

Keywords Decision-making . Information processing . Risk . Reward . Gender

JEL Classifications D830 . D800

1 Introduction

The conventional wisdom that men never stop to ask for directions while women prefer receiving directions over consulting a map reflects a gender difference in the types of information males and females prefer to utilize when making decisions. People make decisions every day—from choosing brands of foods to purchase at the grocery store, to career choices, to financial investments, decisions are made in a multitude of domains and often can have significant consequences. Consequently, it is important to understand how people make decisions, the information they attend to or ignore, and how individual differences account for variation in decision-making. Much of the prior

J Risk Uncertain (2015) 50:55–71 DOI 10.1007/s11166-015-9206-7

K. A. Byrne: D. A. Worthy (*) Department of Psychology, Texas A&M University, 4235 TAMU, College Station, TX 77843-4235, USA e-mail: [email protected]

research on gender differences in decision-making has focused on risk-taking behavior and attention to gains compared to losses. However, much work remains to be done to develop a full understanding of how men and women differ during decision-making. Specifically, it is not clear if there are gender biases in the use of external information and reward sensitivity and whether such differences are advantageous or detrimental when making decisions.

One consistent finding is that women are less risk-seeking than males (e.g., Croson and Gneezy 2009; Jianakoplos and Bernasek 1998; Powell and Ansic 1997; Wong and Carducci 1991). Risk-taking is a critical aspect of investing, and indeed, women demonstrate significantly greater risk aversive behavior in investing mutual funds compared to men (Dwyer, Gilkeson, and List 2002) and their pensions after retirement (Bajtelsmit and Bernasek 1996). While increased risk-taking in the stock market can result in larger gains, it is still a gamble and can equally result in large losses. Barber and Odean (2001) analyzed a sample of over 35,000 household investments and found that men traded 45% more often than women, but the increased trading also resulted in a larger net loss. It appears that in an effort to maximize the future value of their investments, males take more risks but also pay more for the consequences of this behavior, often leading to counterproductive results.

Reduced risk-taking may be related to other work that suggests that women are more sensitive to losses than men. This is one purported reason for why women have performed more poorly on the Iowa Gambling Task than men (e.g., Bolla et al. 2004; Reavis and Overman 2001; Weller et al. 2010). Females exhibit more loss aversive behavior in the task by avoiding the advantageous decks more after they have yielded a loss compared to males (van den Bos et al. 2013). Other gambling tasks with a gain-loss structure have also shown that females exhibit more risk aversive behavior (e.g., Levin et al. 1988; Brinig 1995; Eckel and Grossman 2008). Even on a riskless task in which participants were asked to rate their willingness to sell their car or purchase a new car, females gave more loss aversive responses (Gaechter et al. 2007).

In addition to risk-taking, females also exhibit reduced risk adjustment. It is important to note that risk adjustment is not necessarily reflective of risk aversive behavior, but rather indicates a strategy to potentially improve a high-risk situation (Wehrung et al. 1989). Gender differences in risk adjustment have been examined in the context of gambling situations. For example, compared to males, females are less likely to increase bets as the probability of a win increases on the Cambridge Gambling Task (Deakin et al. 2004; van den Bos et al. 2012). In this task participants try to earn as many points as possible by choosing which of two boxes an object is hidden in and then betting on their decision. Each object is given a ratio reflecting the likelihood that the object is hidden there. As the likelihood that a given box contained the object increased, males bet higher amounts, but females tended to change their betting amount less even as the likelihood of a win increased. Although no gender differences in risk-taking were observed, the tendency to engage in less risk adjust- ment indicated that females exhibited slightly more cautious betting behavior than males in the task.

Gender differences during decision-making have also been attributed to differences in information processing (van den Bos et al. 2013). The way in which a decision- making situation is contextualized influences the information processing styles indi- viduals utilize (Slovic and Lichtenstein 1983). For example, in a bidding context, dollar

56 J Risk Uncertain (2015) 50:55–71

values might be the most salient feature that individuals use to make decisions. However, in an investment scenario, opportunities that are both immediately cost effective and also maximize future returns may be most salient. Thus, the situations that involve single compared to multiple salient features entail different information processing styles (Schoemaker 1998). Males attend more to global information by focusing on a single aspect of an overall task while women attend more to detailed information by combining multiple aspects of a task (Williams and Meck 1991; Andreano and Cahill 2009; Cahill 2006). For example, females attend to both the frequency of wins and losses of each deck as well as the long-term pay-off structure on the Iowa Gambling Task (van den Bos et al. 2013). In contrast, males make choices on the Iowa Gambling Task (IGT) based more on the long-term pay-off of each deck alone, which may explain why females underperform, relative to males on the IGT. Females tend to select the disadvantageous deck (Deck B) that leads to larger imme- diate reward and infrequent but very large losses more than males, while males are better able to globally process both the frequencies of wins with the magnitude of their values, thus attending to the long-term pay-offs of each deck (van den Bos et al. 2013). Furthermore, the selectivity model of information processing proposes that males process information selectively, relying more on the overall objective individual cues and heuristics, while females are more likely to comprehend and integrate all available details, including both subjective and objective information (Meyers-Levy 1989). Because the IGT only utilizes objective cues, the frequency and magnitude of wins and losses, it is unclear how males and females would perform on a decision-making task if external information were also provided. For example, proficiency at focusing on the long-term rewards on the IGT relies on objective cues alone, but whether males also process information globally when subjective external information is involved remains ambiguous.

The present study seeks to resolve this issue by directly examining how gender affects attention to immediate versus long-term rewards and external information. To do this, we utilize a dynamic, frequency-driven decision-making task with a choice-history dependent reward structure in which the reward offered on each trial depends on the individual’s selection of choices on prior trials (Byrne and Worthy 2013; Gureckis and Love 2009; Worthy et al. 2011; Worthy et al. 2012). This type of reward structure mirrors real-world decision-making situations in which the long-term consequences are contingent on choices made previously, such as choosing to invest a sum of money rather than spend it immediately.

Figure 1 shows the reward structure used in the Increasing-Optimal task in Exper- iment 1. The Increasing option always provides a smaller immediate reward on each trial, compared to the alternative option (the Decreasing option). However, rewards provided by both options are a function of how often the Increasing option has been chosen over the previous ten trials, as indicated on the x-axis. For example, the reward options for the 11th trial depend on choices made on the first 10 trials of the task. Similarly, the reward options for the 31st trial depend on the choices made on trials 21– 30. Rewards increase as the Increasing option is selected more often, but decrease as the Decreasing option is selected more often. In this task the optimal action is to select the Increasing option on every trial, even though it provides smaller immediate rewards. Consequently, the task is frequency-driven as performance is determined by the number of times the Increasing option is selected.

J Risk Uncertain (2015) 50:55–71 57

We manipulate the information that is presented on each trial by showing half of participants the foregone reward for the non-chosen option (Fig. 2). The foregone reward information should bias participants toward the sub-optimal alternative by making it salient that the Decreasing option always provides a larger immediate reward than the Increasing option. This manipulation allows us to examine whether females generally prefer options that provide larger immediate reward more than males, or whether females are more likely to use all available information to guide decisions. If females have a general preference for immediate reward then they should select the Increasing option less, and perform more poorly throughout the task regardless of whether foregone reward information is presented or withheld. However, if females are less able to process information selectively, compared to males, then they may be biased toward the sub-optimal decreasing option only when presented with foregone reward information. This would lead to equivalent performance for males and females when foregone reward information is absent, but poorer performance for females when foregone reward information is present.

In Experiment 2 we further examine how gender affects dynamic decision-making by using a Decreasing-Optimal task in which the reward structure is altered so that the Decreasing option is now the optimal choice (Fig. 3). In this task the gain in future reward from selecting the Increasing option is too small to match the larger immediate

0

20

40

60

80

100

0 1 2 3 4 5 6 7 8 9 10

P oi

n ts

E ar

n ed

Selection of the Increasing Option Over the Previous 10 Trials

Decreasing Option

Increasing Option

Maximum of Increasing Option

Minimum of Decreasing Option

Fig. 1 Reward structure for the Increasing-optimal task in Experiment 1

Fig. 2 Sample screenshot from the foregone present (right) and foregone absent (left) condition of the Increasing-optimal decision-making task

58 J Risk Uncertain (2015) 50:55–71

rewards consistently provided by the Decreasing option. Selecting the Increasing option is ultimately futile because the maximum reward that can be obtained from the Increasing option is smaller than the minimum reward that can be obtained from selecting the Decreasing option. If males outperform females on the Increasing-Optimal task in Experiment 1, then one possibility is that males have a general bias toward improving reward on future trials, while females are biased toward immediate reward. This bias would lead to better performance in Experiment 1, but poorer performance in the Decreasing-Optimal task in Experiment 2. Alternatively, males could be biased toward selectively filtering out the foregone reward information which would lead to poorer performance, relative to females, in Experiment 2 when foregone reward information is present, but equivalent performance when foregone reward information is absent.

2 Experiment 1

2.1 Method

Participants One hundred and fourteen (57 female, 57 male) undergraduate students at Texas A&M University participated in the experiment for course credit. In our 2 (Foregone Rewards: Present vs. Absent) x 2 (Gender: Female vs. Male) between participants design, participants were randomly assigned to one of the two reward conditions. There were 56 participants (27 females) in the foregone rewards present condition and 58 participants (30 females) in the foregone rewards absent condition.

Materials and procedure Participants completed the decision-making task on PC computers using Psychtoolbox for Matlab (version 2.5). Participants completed 250 trials and were given a goal of earning 16,000 points by the end of the task. The rewards given on each trial were based on the number of times participants had selected the Increasing option over the past ten trials (Fig. 1). On the first trial it was assumed

0

20

40

60

80

100

120

140

0 1 2 3 4 5 6 7 8 9 10

P oi

n ts

E ar

n ed

Selection of the Increasing Option Over the Previous 10 Trials

Decreasing Option

Increasing Option

Minimum of Decreasing Option

Maximum of Increasing Option

Fig. 3 Reward structure for the Decreasing-optimal decision-making task in Experiment 2

J Risk Uncertain (2015) 50:55–71 59

that participants had alternated selecting the Increasing and Decreasing option over the previous ten trials. Thus, they began the task with a vector of [0, 1, 0, 1, 0, 1, 0, 1, 0, 1] where a ‘1’ indicated an Increasing option selection and a ‘0’ indicated a Decreasing option selection. The sum of the numbers in this vector determined the place they were on the x-axis of Fig. 1. On the first trial of the task a ‘1’ or ‘0’ was added, depending on which option had been selected, and the last number in the vector was removed. This continued throughout the task, and after ten trials the rewards participants received were completely determined by their past behavior.

Point values increased or decreased by five-point increments. Participants started the task with 55 points for the Increasing option, which increased to 60 points if selected again, then 65 points and so forth until the deck value reached 80 points (Table 1). Once 80 points were reached, the Increasing option repeatedly gave a value of 80 points until the Decreasing option was selected. Similarly, participants started the task with 65 points for the Decreasing option, which decreased to 60 points if selected again and continued to decrease to a minimum value of 40 points. Once 40 points were reached, participants repeatedly earned 40 points if they continued to select the immediately rewarding Decreasing option. Thus, selecting the Decreasing option repeatedly provid- ed 40 points after it was selected ten or more times in a row. In both the foregone rewards present and absent conditions, participants were shown the points earned upon selecting each option and their point total was updated. Participants in the foregone rewards present condition received additional feedback about the value of the foregone reward for the non-chosen option on each trial (Fig. 2). For example, if they earned 40 points by selecting the Decreasing option, they were informed that they would have

Table 1 Pattern of points earned in the Increasing-Optimal task for the Increasing and Decreasing options

Increasing-option selections Decreasing-option selections

Increasing option Decreasing option Increasing option Decreasing option

55 65 55 65

60 70 55 65

60 70 50 60

65 75 50 60

65 75 45 55

70 80 45 55

70 80 40 50

75 85 40 50

75 85 35 45

80 90 35 45

80 90 30 40

80 90 30 40

80 90 30 40

Participants begin with 55 points for the Increasing option and 65 points for the Decreasing option. If the Increasing option is repeatedly selected, individuals will earn 80 points on each trial after the first ten trials. If the Decreasing option is repeatedly selected, individuals will earn 40 points on each trial after the initial ten trials. Thus repeatedly selecting the Increasing option leads to a 40 point advantage compared to the Decreasing option. Switching between decks follows the same pattern.

60 J Risk Uncertain (2015) 50:55–71

received only 30 points had they selected the Increasing option. However, selecting the Increasing option repeatedly in this example would lead to an increase in points earned on each selection, from 35 points to 40 points and so forth until it reached the maximum value of 80 points.

2.2 Results

2.2.1 Analysis of points earned on the Increasing-optimal task

Because participants were given the goal of earning 16,000 points by the end of the task, overall task success was assessed by computing the total points earned on the task. A 2 (Foregone Rewards: Present vs. Absent) X 2 (Gender: Male vs. Female) ANOVA revealed a main effect of foregone reward information, F(1, 113)=9.51, p<.01, partial η2=.08 in which individuals who were not presented with the misleading foregone reward information (M=16,163, SD=1987) earned more points than those presented with foregone reward information (M=14,838, SD=2844). These results suggest that, on average, viewing misleading information about foregone rewards prevented individuals from reaching the goal in the task, whereas on average, individuals who were not shown foregone rewards information reached the goal. Furthermore, a main effect of gender showed that males (M=16,034, SD=2008) earned more points on the task overall compared to females (M=14,991, SD=2876), F(1, 113)=6.25, p<.01, partial η2=.05. The foregone rewards X gender interaction was significant, F(1,110)= 4.10, p<.05, partial η2=.04, and follow-up pairwise comparisons revealed that males earned significantly more points (M=15,804, SD=2058) than females (M=13,801, SD=3223) on the task when foregone rewards were presented, t(54)=−2.79, p<.01, but there was no difference in points earned between males’ (M=16,272, SD=1963) and females’ (M=16,062, SD=2038) performance when foregone reward information was absent, t(56)=−.40, p=.69.

2.2.2 Analysis of Increasing-optimal selections

The total proportion of Increasing option selections was computed and used as the dependent variable (Fig. 4). Points earned on the task are a direct function of selections of the Increasing-optimal option, r=1.00, p<.01, and as a result, the following analysis reflects the proportion of optimal Increasing option deck selections needed to reach the point goal. A 2 (Foregone Rewards: Present vs. Absent) X 2 (Gender: Male vs. Female) ANOVA revealed a main effect of foregone reward information, F(1, 113)=9.47, p<.01, partial η2=.08, and a main effect of gender, F(1, 113)=6.51, p<.01, partial η2=.06. Participants in the foregone reward absent condition (M=.62, SD=.20) selected the Increasing-optimal option significantly more than participants in the foregone rewards present condition (M=.49, SD=.29), and males (M=.61, SD=.21) chose the Increasing- optimal option significantly more often than females (M=.50, SD=.30). Since this task entails binary choices, random deck selections would be indicated by a deck selection rate of 0.5. Values greater than 0.5 reflect a tendency to choose the Increasing-optimal deck, whereas values less than 0.5 indicate a preference for choosing the sub-optimal Decreasing option. Thus, these results reflect a tendency for males to learn to choose the

J Risk Uncertain (2015) 50:55–71 61

Increasing-optimal option, but females choose the Increasing-optimal option at chance and do not learn the optimal decision strategy. There was also significant foregone rewards X gender interaction, F(1, 110)=4.37, p<.05, partial η2=.04. Pairwise comparisons within each foregone reward information condition showed that males (M=.59, SD=.21) selected the optimal Increasing option significantly more than females (M=.38, SD=.33) in the foregone rewards present condition, t(54)=−2.87, p<.01. However, in the foregone rewards absent condition, males (M=.63, SD=.20) and females (M=.61, SD=.21) performed equivalently, t(56)=−.39, p=.70.

To assess differences in learning between genders, the proportion of optimal Increas- ing deck selections was computed in 50-trial blocks and analyzed using a 5 (block) X 2 (gender) X 2 (foregone reward condition) repeated measures ANOVA. Figure 5 shows selection of the Increasing-optimal decks in each of five 50-trial blocks in the foregone reward present (a) and absent (b) conditions. The results indicated a significant effect of block, F(4, 440)=90.39, p<.001, partial η2=.45, a significant block X foregone reward condition interaction, F(4, 440)=5.36, p<.01, partial η2=.05, a significant block X gender interaction, F(4, 440)=4.68, p<.01, partial η2=.04, and a significant block X foregone reward condition X gender interaction, F(4, 440)=5.61, p<.01, partial η2=.05.

To decompose the three-way interaction we performed a 5 (block) X 2 (gender) repeated measures ANOVAwithin each foregone reward condition. Within the foregone rewards present condition there was a significant effect of block, F(4, 216)=59.03, p<.001, partial η2=.52, and a significant block X gender interaction, F(4, 216)=9.14, p<.001, partial η2=.15. For males in the foregone reward condition a repeated measures ANOVA revealed a very strong effect of block, F(4, 112)=56.93, p<.001, partial η2=.67. For females in the same condition a repeated measures ANOVA revealed a significant effect of block, F(4, 104)=11.14, p<.001, partial η2=.30, although the effect size was less than half the size of the effect for males. This suggests that both males and females showed improved performance over the course of the task, but males showed much stronger improvement compared to females. Within the foregone rewards absent

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

AbsentPresent

P ro

p or

ti on

o f

In cr

ea si

n g

O p

ti on

S

el ec

ti on

s

Foregone Reward Condition

Female

Male

Fig. 4 Proportion of Increasing-optimal deck selections in Experiment 1 by gender and foregone reward information condition. Note. Because values lower than 0.5 (shown by the dotted line in the figure) indicate a tendency to choose the sub-optimal option, females chose the Decreasing sub-optimal deck more when foregone rewards are present, but chose the Increasing-optimal option more in the absence of foregone rewards information. Error bars represent standard error of the mean

62 J Risk Uncertain (2015) 50:55–71

condition there was a significant effect of block, F(4, 224)=33.60, p<.001, partial η2=.38, but the block X gender interaction was not significant, F(4, 224)<1, p=.97.

2.3 Discussion

The results of Experiment 1 suggest that females gave greater weight to foregone reward information than males, resulting in a greater proportion of Decreasing deck selections and, consequently, sub-optimal performance. Because attention to foregone rewards information in this task results in earning more points on each immediate trial, but fewer points over the course of the entire task, both males and females earned fewer points on the task when foregone rewards were presented and, on average, failed to reach the goal when foregone reward information was available. In contrast, on average both males and females reached the goal in the absence of foregone rewards. Thus, relying on the foregone reward cues in this task leads to task failure. However, in the

a

b

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

P ro

p or

ti on

o f

In cr

ea si

n g

O p

ti on

S

el ec

ti on

s

Trial Block

Female

Male

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

50 100 150 200 250

50 100 150 200 250

P ro

p or

ti on

o f

In cr

ea si

n g

O p

ti on

S

el ec

ti on

s

Trial Block

Female

Male

Fig. 5 Proportion of Increasing deck selections in the Increasing-optimal task in Experiment 1 by 50-trial block for the foregone rewards present (a) and foregone rewards absent (b) conditions. Note. The dotted line indicates chance (M=.50) performance. Error bars represent standard error of the mean

J Risk Uncertain (2015) 50:55–71 63

absence of foregone reward information, females and males selected the optimal Increasing option equivalently. In the presence of foregone reward information, males on average earned 468 points fewer than in the foregone absent task, but females earned 2261 fewer points, indicating that the presence of foregone reward information led females to lose almost five times as many points than males on the task. These results suggest that females are more likely to rely on all available information resources, both the foregone reward information and the reward given by a selected option, when making decisions and that females are not simply biased toward maximizing immediate reward. This type of decision-making behavior indicates a detailed processing style, in this case evidenced by females considering the objective reward cues as well as the subjective external information about foregone rewards.

However, it is unclear whether males’ performance is a product of a focus on maximizing future reward or from selectively filtering out the foregone reward infor- mation. In order to address this issue, in Experiment 2 participants performed a similar task, but the reward structure is altered so that the immediately rewarding Decreasing option is now the optimal choice. If females outperform males when foregone reward information is both presented and withheld, then we could conclude that males are biased toward maximizing future reward, even when it is counterproductive to do so. However, if males are biased toward selective information processing, then they should perform worse than females only when foregone reward information is presented.

3 Experiment 2

3.1 Method

Participants Eighty (41 female, 39 male) undergraduate students at Texas A&M University participated in the experiment for course credit. In our 2 (Foregone Re- wards: Present vs. Absent) x 2 (Gender: Female vs. Male) between participants design, participants were randomly assigned to one of the two reward conditions. There were 40 participants (21 female) in the foregone rewards present condition and 40 partici- pants (20 female) in the foregone rewards absent condition.

Materials and procedure The general procedure for the two-choice dynamic decision- making task was the same as Experiment 1, except that the Decreasing option now provided substantially larger immediate rewards than the Increasing option (Fig. 3). Choosing the Decreasing option resulted in a decrease in rewards (optimal), while selecting the Increasing option led to an increase in rewards (sub-optimal). Although selecting the Increasing option still led to larger rewards for both options on future trials, it was now the sub-optimal choice because the gain in future reward from selecting it would never make up for the greater immediate reward that could be obtained from selecting the Decreasing option. At the beginning of the task participants received 30 points for selecting the Increasing option and 90 points for selecting the Decreasing option (Table 2). Repeatedly selecting the Increasing option in this task resulted in a maximum point value of 55 points on each trial. In contrast, repeatedly selecting the Decreasing-optimal option led to a gain of 65 points on each trial, and thus provided ten more points on each trial if repeatedly selected. All other aspects of the task were identical to those in Experiment 1.

64 J Risk Uncertain (2015) 50:55–71

3.2 Results

3.2.1 Analysis of points earned on the Decreasing-optimal task

As in Experiment 1, we first examined the total points earned by each participant in the task. Points earned on the task directly reflected the proportion of times the Decreasing option was selected, r=.99, p<.001. A 2 (Foregone Rewards: Present vs. Absent) X 2 (Gender: Male vs. Female) ANOVA revealed a main effect of foregone reward information (F(1, 79)=6.36, p<.01, partial η2=.07) and gender (F(1, 79)=3.96, p<.05, partial η2=.05). Individuals in the foregone reward present condition earned more points (M=15,686, SD=629) than those in the foregone reward absent condition (M=15,356, SD=527), and females (M=15,648, SD=601) selected the optimal Decreasing option more than males (M=15,387, SD=576). The foregone reward X gender interaction was not significant, F(1, 76)=1.14, p=.29.

3.2.2 Analysis of Decreasing-optimal selections

The total proportion of optimal Decreasing option selections are plotted in Fig. 6. A 2 (Foregone Rewards: Present vs. Absent) X 2 (Gender: Male vs. Female) ANOVA revealed a significant main effect of both foregone reward information, F(1, 76)=6.05, p<.05, partial η2=.07) and gender, F(1, 76)=3.96, p=.05, partial η2=.05, however the gender X foregone reward information interaction was not significant, F(1, 76)=1.62,

Table 2 Pattern of points earned in the Decreasing-optimal task for the Increasing and Decreasing option selections

Decreasing-option selections Increasing-option selections

Increasing option Decreasing option Increasing option Decreasing option

30 90 30 90

30 90 35 95

25 85 35 95

25 85 40 100

20 80 40 100

20 80 45 105

15 75 45 105

15 75 50 110

10 70 50 110

10 70 55 115

5 65 55 115

5 65 55 115

5 65 55 115

Participants begin with 30 points for the Increasing option and 90 points for the Decreasing option. If the Decreasing option is repeatedly selected (left), individuals will earn 65 points on each trial after the first ten trials. If the Increasing option is repeatedly selected (right), individuals will earn 55 points on each trial after the initial ten trials. Thus selecting the Decreasing option results in an overall ten-point advantage over the Increasing option if repeatedly selected. Switching between decks follows the same pattern

J Risk Uncertain (2015) 50:55–71 65

p=.21). Participants in the foregone reward present condition (M=.75, SD=.22) selected the optimal Decreasing option significantly more than participants in the foregone reward absent condition (M=.63, SD=.19), and females (M=.73, SD=.21) selected the optimal Decreasing option significantly more often than males (M=.64, SD=.21).

As in Experiment 1, learning rate was assessed by examining each 50-trial block (Fig. 7). A 5 (block) X 2 (gender) X 2 (foregone reward condition) repeated measures ANOVA revealed a main effect of block, F(4, 304)=22.39, p<.001, partial η2=.22. The interaction between block and foregone reward condition, (F(4, 304)=.75, p=.56), block and gender, (F(4, 304)=.38, p=.82), and the block X gender X foregone reward condition interaction were all non-significant F(4, 304)=1.20, p=.31). Males and females both selected the Decreasing option less often as they progressed through the task, but females consistently chose the optimal Decreasing option more than males on each trial block.

4 General discussion

In Experiment 2 females selected the optimal Decreasing choice, which maximized immediate reward, more often than males both in the presence and absence of foregone reward information. Males showed a tendency to ignore the foregone reward informa- tion and select the Increasing option, which led to larger future rewards, even though it was counterproductive. When foregone reward information was present, both females and males chose the Decreasing-optimal option more overall compared to the Increas- ing option, but males showed a decline in choosing the Decreasing-optimal option over time, indicating less of a reliance on the subjective foregone reward information and a subsequent shift toward a global processing style for the remainder of the task. Because females are not biased toward maximizing future reward, they may more easily recognize the salient feature in the task—the difference between points earned and

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Foregone Present Foregone Absent

P ro

p or

ti on

o f

D ec

re as

in g

O p

ti on

S el

ec ti

on s

Foregone Reward Condition

Female

Male

Fig. 6 Proportion of Decreasing-optimal deck selections in Experiment 2 by gender and foregone reward information condition. Note. The dotted line indicates chance (M=.50) performance. Error bars represent standard error of the mean

66 J Risk Uncertain (2015) 50:55–71

foregone rewards provided by the foregone information cues. In the absence of additional information, males begin the task by choosing the sub-optimal Increasing option more than females, indicating an initial bias to maximize future rewards, even if it is sub-optimal. In the second half of the task, males and females both chose the Decreasing-optimal deck less. While females began the task successfully, they may not have fully learned the decision reward structure completely by the end of the 250 trials. Thus, while males’ performance remained relatively unchanged in Experiment 1 when foregone reward information was presented or withheld, females showed a greater bias toward the sub-optimal Decreasing option when foregone information was presented. The combined results suggest that males favor maximizing future rewards and are able to ignore foregone reward information that biases them toward the option that provides

a

b

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1 P

ro p

or ti

on o

f D

ec re

as in

g O

p ti

on S

el ec

ti on

s

Trial Block

Female

Male

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

50 100 150 200 250

50 100 150 200 250

P ro

p or

ti on

o f

D ec

re as

in g

O p

ti on

S el

ec ti

on s

Trial Block

Female

Male

Fig. 7 Proportion of Decreasing deck selections in the Decreasing-optimal task in Experiment 2 by 50-trial block for the foregone rewards present (a) and foregone rewards absent (b) conditions. Note. The dotted line indicates chance (M=.50) performance. Error bars represent standard error of the mean

J Risk Uncertain (2015) 50:55–71 67

larger immediate reward. In contrast, females utilize external information to make decisions, even when such information may lead them toward the poorer choice, as in Experiment 1. However, in the absence of such biasing information, females are able to adopt the strategy that is optimal for the task, regardless of whether it entails selecting the option that leads to larger immediate or delayed reward.

It is important to note that unlike the Increasing-Optimal task in Experiment 1, when the Decreasing option is the optimal solution in Experiment 2, performance decreased rather than increased over trials. The discrepancy between the enhanced learning of the optimal strategy as the task progressed in Experiment 1 and the decline in performance in Experiment 2 can be attributed to the saliency of the optimal option presentation between the two tasks over the course of the experiment. In Experiment 1, the difference in immediate reward between the optimal Increasing and sub-optimal Decreasing options begins at ten points with the Increasing option being ten points less than the Decreasing option. This makes the Decreasing option initially appear optimal as it provides larger immediate rewards, but the saliency of the larger rewards provided by the Decreasing option may decrease as the task progresses. In contrast, in Experiment 2 the difference in immediate reward between the sub-optimal Increasing and optimal Decreasing options is 60 points, and the Decreasing-optimal option is immediately made salient, but should decline as the task progresses, which may cause participants to select the Decreasing option less often over time. Thus, in the Decreasing-Optimal task we observe a decline in choosing the optimal option over time as the saliency of the larger immediate rewards for the Decreasing option also declines. Consequently, in the absence of additional information, males initially choose the option with a higher immediate reward, the more salient task feature. When additional information is present and the sixty point difference in the rewards offered by each option is made more apparent, both males and females benefit. However, only males need to overcome a bias related to the association between frequency and increase in reward, and thus their bias toward a global processing style needs to be overridden in order to succeed on the task.

Furthermore, because females shifted their behavior in the presence of external information, it appears that females view external information as more salient than reward value frequency. Females assess the difference in the foregone rewards and rewards they earned on each selection and integrate this information with the frequency of reward, reflecting a detailed processing style. In contrast, males are unaffected by the additional external information and base their decision off of the salient feature of the task, the frequency of Increasing reward values, which follows a global information processing style. These findings are consistent with both the selectivity theory of information processing and prior work that examined gender differences in decision- making information processing styles on the Iowa Gambling Task (Meyers-Levy 1989; van den Bos et al. 2013). Both styles have advantages and disadvantages depending on the context. These processing style benefits may reflect differences in reliance on decision-making heuristics. While heuristics are often used to adapt to common situations, they may also result in a bias that is unsuccessful in decision-making under uncertainty (Cubitt and Sugden 2001). When subjective external information is bene- ficial, females do not have to overcome a bias towards maximizing future reward and tend to have an advantage, but when this information is misleading, a detailed processing style leads to sub-optimal decisions because attention to only the overall goal of the task resulted in the highest long-term rewards.

68 J Risk Uncertain (2015) 50:55–71

Current research on gender differences in decision-making is limited to risk-taking and responsiveness to gains and losses. Our findings indicate additional mechanisms to account for gender differences in decision-making, namely, reward sensitivity and information processing. This study reveals that gender differences in decision-making are more complex than prior research has demonstrated. In accordance with the selectivity model of information processing, males initially reject additional informa- tion, the foregone rewards information, and focus more on improving rewards on future trials than maximizing immediate reward. As selective processors, males tend to use individual cues to base their decisions off of and perform worse when integrating multiple sources of information, like the actual rewards provided with the information regarding foregone rewards. Consequently, they performed worse than females in Experiment 2 even when the external information was helpful. On the other hand, females, as comprehensive processors, depend on external information to guide their decisions, even if it results in a sub-optimal choice that hampers future opportunities. When the external information is detrimental like in the Increasing-Optimal task in Experiment 1, females prefer to maximize immediate reward, rather than improve long- term outcomes.

Gender differences in preferences for immediate and delayed rewards have been examined in the context of delay discounting. The results have been mixed, with some studies showing higher rates of male delay discounting (Kirby and Marakovic 1996; Silverman 2003), and others showing that females prefer immediate rewards more (Logue and Anderson 2001; Reynolds et al. 2006). Females’ superior ability in exploiting the Decreasing option in the Decreasing-Optimal task in Experiment 2 demonstrates a preference for immediate reward compared to males when it is optimal to maximize immediate reward. However, females in Experiment 1 who did not receive foregone reward information performed comparably to males, suggesting that they are able to maximize delayed reward when the immediate reward strategy is futile. In contrast, males selected the Increasing option more than females even when it was sub- optimal to do so, as in the Decreasing-Optimal task.

4.1 Limitations

Some limitations of the current study should be considered when generalizing these findings to others. Firstly, the study population was limited to undergraduate students. It is possible that level of education or age plays a role in information-processing and preference for reward in decision-making. Secondly, the decision-making experiment was contextualized in the domain of gambling using a card game. It is important to consider the context of these results when generalizing to other domains, such as financial, marketing or professional areas. Future research should be conducted to determine whether educational level, age, or context influences gender differences in decision-making.

4.2 Implications and Future Directions

Overall, the results of our study suggest distinctive gender decision-making styles, where males are more global, selective processors of information and show a cognitive bias towards maximizing long-term benefits, and females are more detailed,

J Risk Uncertain (2015) 50:55–71 69

comprehensive processors of information and can maximize either immediate or long- term benefits in different situations. Both styles have benefits and drawbacks depending on the situation and context. If external information results in long-term benefits, then females have an advantage. On the other hand, if such information leads to a bias toward a poorer strategy, then males are less susceptible to such a bias. These results may provide a mechanism for prior research showing females as less risk-seeking than males. Gender differences in information processing may account for some differences in risk-taking. For example, Barber and Odean’s (2001) finding that men trade more on the stock market but receive larger losses is consistent with our results showing that males become fixated on maximizing long-term gains.

Beyond financial decision-making, these findings may also be extended to market- ing and career choice. For example, females may be more attentive to information in advertising campaigns and utilize such information when purchasing products. In the workplace, males may focus on long-term benefits of their career and opportunity for growth, while females may place more weight on immediate benefits of a job offer. Future research should be aimed at considering the applicability of these findings to such areas as marketing and professional fields. While the results of this investigation demonstrate clear differences and biases in decision-making, often the best way to make a decision is to adapt the strategy to a given task. Both males and females should be aware of these biases and consider more flexible strategies when making a decision, such as being open to external resources, but not leaning too much on them to make decisions, and considering both immediate and long-term effects of each decision.

References

Andreano, J. M., & Cahill, L. (2009). Sex influences on the neurobiology of learning and memory. Learning & Memory, 16(4), 248–266.

Bajtelsmit, V. L., & Bernasek, A. (1996). Why do women invest differently than men? Financial Counseling and Planning, 7, 1–10.

Barber, B. M., & Odean, T. (2001). Boys will be boys: gender, overconfidence, and common stock investments. The Quarterly Journal of Economics, 116(1), 261–292.

Bolla, K. I., Eldreth, D. A., Matochik, J. A., & Cadet, J. L. (2004). Sex-related differences in a gambling task and its neurological correlates. Cerebral Cortex, 14(11), 1226–1232.

Brinig, M. F. (1995). Does mediation systematically disadvantage women? William and Mary Journal of Women and the Law, 2, 1–34.

Byrne, K. A., & Worthy, D. A. (2013). Do narcissists make better decisions? An investigation of narcissism and dynamic decision-making performance. Personality and Individual Differences, 55, 112–117.

Cahill, L. (2006). Why sex matters for neuroscience. Nature Neuroscience Reviews, 7, 477–484. Croson, R., & Gneezy, U. (2009). Gender differences in preferences. Journal of Economic Literature, 47, 448–598. Cubitt, R. P., & Sugden, R. (2001). Dynamic decision-making under uncertainty: an experimental investiga-

tion of choices between accumulator gambles. Journal of Risk and Uncertainty, 22, 103–128. Deakin, J., Aitken, M., Robbins, T., & Sahakian, B. J. (2004). Risk taking during decision-making in normal

volunteers changes with age. Journal of the International Neuropsychological Society, 10, 590–598. Dwyer, P. D., Gilkeson, J. H., & List, J. A. (2002). Gender differences in revealed risk taking: evidence from

mutual fund investors. Economics Letters, 76(2), 151–158. Eckel, C. C., & Grossman, P. J. (2008). Forecasting risk attitudes: an experimental study using actual and

forecast gamble choices. Journal of Economic Behavior & Organization, 68(1), 1–17. Gaechter, S., Johnson, E. J., & Hermann, A. (2007). Individual-level loss aversion in riskless and risky

choices. CeDEx Discussion Paper 2007-02. University of Nottingham. Gureckis, T. M., & Love, B. C. (2009). Learning in noise: dynamic decision-making in a variable environ-

ment. Journal of Mathematical Psychology, 53, 180–193.

70 J Risk Uncertain (2015) 50:55–71

Jianakoplos, N. A., & Bernasek, A. (1998). Are women more risk averse? Economic Inquiry, 36, 620–630. Kirby, K. N., & Marakovic, N. N. (1996). Delay-discounting probabilistic rewards: rates decrease as amounts

increase. Psychonomic Bulletin & Review, 3(1), 100–104. Levin, I. P., Snyder, M. A., & Chapman, D. P. (1988). The interaction of experiential and situational factors

and gender in a simulated risky decision-making task. Journal of Psychology, 122, 173–181. Logue, A. W., & Anderson, Y. D. (2001). Higher education administrators: when the future does not make a

difference. Psychological Science, 12, 276–281. Meyers-Levy, J. (1989). Gender differences in information processing: a selectivity interpretation. In P.

Cafferata & A. Tybout (Eds.), Cognitive and affective responses to advertising (pp. 219–260). MA: Lexington Books.

Powell, M., & Ansic, D. (1997). Gender differences in risk behaviour in financial decision-making: an experimental analysis. Journal of Economic Psychology, 18(6), 605–628.

Reavis, R., & Overman, W. H. (2001). Adult sex differences on a decision-making task previously shown to depend on the orbital prefrontal cortex. Behavioral Neuroscience, 115(1), 196–206.

Reynolds, B., Ortengren, A., Richards, J. B., & De Wit, H. (2006). Dimensions of impulsive behavior: personality and behavioral measures. Personality and Individual Differences, 40, 305–315.

Schoemaker, P. H. (1998). Determinants of risk-taking: behavior and economic views. Journal of Risk and Uncertainty, 6, 49–73.

Silverman, I. W. (2003). Gender differences in delay of gratification: a meta-analysis. Sex Roles, 49, 451–463. Slovic, P., & Lichtenstein, S. (1983). Preference reversals: a broader perspective. American Economic Review,

73, 590–605. van den Bos, R., de Visser, L., Van de Loo, A. J. A. E., Mets, M. A. J., Van Willigenburg, G. M., & Homberg,

J. R. (2012). Sex differences in decision-making in adult normal volunteers are related to differences in the interaction of emotion and cognitive control. In K. O. Moore & N. P. Gonzalez (Eds.), Handbook on psychology of decision-making: new research (pp. 179–198). Hauppage: Nova Science Publisher Inc.

van den Bos, R., Homberg, J., & de Visser, L. (2013). A critical review of sex differences in decision-making tasks: focus on the Iowa gambling task. Behavioural Brain Research, 238(1), 95–108.

Wehrung, D. A., Kam-Hon, L., Tse, D. K., & Vertinsky, I. B. (1989). Adjusting risky situations: a theoretical framework and empirical test. Journal of Risk and Uncertainty, 2, 189–212.

Weller, J. A., Levin, I. P., & Bechara, A. (2010). Do individual differences in Iowa gambling task performance predict adaptive decision making for risky gains and losses? Journal of Clinical and Experimental Neuropsychology, 32(2), 141–50.

Williams, C. L., & Meck, W. H. (1991). The organizational effects of gonadal steroids on sexually dimorphic spatial ability. Psychoneuroendocrinology, 16, 1–3.

Wong, A., & Carducci, B. J. (1991). Sensation seeking and financial risk taking in everyday money matters. Journal of Business and Psychology, 5(4), 525–530.

Worthy, D. A., Gorlick, M. A., Pacheco, J. L., Schnyer, D. M., & Maddox, W. T. (2011). With age comes wisdom: decision-making in younger and older adults. Psychological Science, 22, 1375–1380.

Worthy, D. A., Otto, A. R., & Maddox, W. T. (2012). Working-memory load and temporal myopia in dynamic decision-making. Journal of Experimental Psychology. Learning, Memory, and Cognition, 38(6), 1640– 1658.

J Risk Uncertain (2015) 50:55–71 71

Copyright of Journal of Risk & Uncertainty is the property of Springer Science & Business Media B.V. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.