Creativity at gunpoint : the influence of threat direction, time to respond, and physical aggression on creative defensive responses

(1)

CREATIVITY AT GUNPOINT

Creativity at Gunpoint: The Influence of Threat Direction, Time to Respond, and Physical Aggression on Creative Defensive Responses

Joeke Risseeuw

Studentnummer: 10661247 Supervisor: Yujie Cheng University of Amsterdam

(2)

(3)

CREATIVITY AT GUNPOINT Abstract

Past work on creativity in negative situations revealed that people can be creative in

threatening situations when creativity is functional to reach their goals. Moreover, according to the Dual Pathway to Creativity Model (DPCM), people in negative situations have to cognitively persist in tasks to reach creativity and this requires longer time. Fueled by these two perspectives, I tested the effect of threat direction and response time on creative defensive responses. In this study, participants gave defensive responses to threatening situations

depicted in pictures with a threat facing toward (direct threat) or a threat facing away from the participant (indirect threat). Also, depending on the condition, they were asked to respond immediately (immediate response condition) or were given more time to respond (delayed response condition). Furthermore, aggression was examined to study whether it had an influence on the number of fight tactics and creativity. In general, with more time, people were more creative, regardless of the threat direction. Threat direction only had an influence on specific tactic types such that more original tactics were shown when threat was direct (as opposed to indirect threat) and more original threat exploration tactics were shown when threat was indirect (as opposed to direct). Moreover, physical aggression predicted the number of original fight tactics, especially when threat was direct. Possible explanations, limitations and aims for the future are discussed.

(4)

Creativity at Gunpoint: The Influence of Threat Direction, Time to Respond, and Physical Aggression on Creative Defensive Responses

In their daily lives, people face many threatening situations such as verbal or physical conflict, competitions, or impending deadlines. People also vary widely in their reactions to the same threatening situation: Some people become mind-blank and have no idea what to do, some choose common tactics and some can come up with original ways to deal with the threat. As an illustration, when a victim is being held at gunpoint by an attacker, the victim could negotiate with the attacker instead of freezing. In such cases, creative solutions can be an exceedingly desirable reaction and even life-saving for the victim compared to a common reaction. Therefore, it is of concern to examine factors that contribute to creativity in

threatening situations. Herein, I seek to investigate under what circumstances people can react creatively to a threatening situation and what kind of people can come up with original ways of dealing with threats.

Previous research has investigated typical human reactions in threatening situations (Blanchard, Griebel, Pobbe, & Blanchard, 2011; Blanchard, Hynd, Minke, Minemoto, & Blanchard, 2001) as well as creativity in aversive situations (Dutton, Sandelands, & Staw, 1981; Mikulincer, Paz, & Kedem, 1990). It has been shown that people can still be creative in an aversive situation by being cognitively persistent and by perceiving creativity to be

functional to reach their goals (Baas, De Dreu, & Nijstad, 2011; Roskes, De Dreu, & Nijstad, 2012). However, there is a lack of research on how people respond creatively to the specific threat they face; in most of past research conducted on threatening situations, participants had to pick an option as a reaction to threat instead of generating a reaction completely on their own, thus leaving no space for creative reactions (Blanchard et al., 2001; MacLaren, Best, & Bigney, 2010; Perkins & Corr, 2006). As people have to quickly come up with ideas in threatening situations and a creative reaction can be superior to a common reaction, it is

(5)

important to examine what defensive reactions people generate on their own and to what extend such reactions can be creative. Following the persistent pathway to creativity model and functional creativity perspective (Baas et al., 2011; Nijstad, De Dreu, Rietzschel, & Baas, 2010; Roskes et al., 2012), the current research examines immediate or delayed defensive reactions to either direct or indirect threats. Specifically, I propose that individuals’ creativity in responses to threats increases when they have time to think and react, and this creativity occurs only in domains that are functional and adaptive to regulate the specific threat. Moreover, I seek to explore whether trait aggression of a threatened person might influence creative reactions to coping with threats.

Threats and Defensive Reactions

Threats refer to events and situations that have impending negative or harmful consequences (e.g., sources of potential danger such as a predator), leading humans and non-human mammals to direct their attention and initiate behavioral reactions in response to the threat (Blanchard et al., 2011; Koster, Crombez, Van Damme, Verschuere, & De Houwer, 2004; McNaughton & Corr, 2004; Woody & Szechtman, 2011). Both human and non-human mammals have multiple defensive responses to react to threats, such as risk assessment (e.g., to investigate/explore the situation more), freeze, flight, defensive threat (e.g., screaming or yelling) and defensive attack (e.g., fighting; Blanchard, 1997). The choice of action depends on various features of the threat: The imminence, ambiguity, and chances of escaping the threat, as well as the presence of possible places to hide (Blanchard et al., 2001). For example, if a threat is escapable and/or distant, then people tend to flee. If threats are ambiguous and less dangerous, people will further explore the situation, in an effort to investigate possible risks and choose an adaptive response (Blanchard, Blanchard, & Rodgers, 1991; Blanchard et al., 2011; Perkins & Corr, 2006, Blanchard et al., 2001 ). Moreover, a study on defensive patterns of rats showed that, once physical contact with a predator is unavoidable, rats will

(6)

fight or make noise instead of freezing (Fanselow, 1994). Likewise, once a threat is imminent and there is no apparent escape route, people may revert to defensive attacks, apology,

negotiation, or cooperation instead of freezing (Blanchard et al., 2001; Fernandes et al., 2013). Of these examples, threat imminence serves a crucial role in determining the type of reactions people tend to use to effectively deal with threats. Imminence is strongly influenced by the direction that a threat is facing: Compared to indirect threats (e.g., a gun pointed away from the observer), an observer perceives a direct threat (e.g., a gun pointed toward the observer) as more imminent, as well as more intense, and inescapable (i.e., the observer perceives to less likely successfully hide from the threat; Fernandes et al., 2013). Moreover, direct threats lead to higher levels of arousal than indirect threats (Flykt, Esteves & Öhman, 2007). Furthermore, imminent threats lead to differences in activation of brain regions: Perceiving threats as proximal is associated with higher periaqueductal gray (PAG) activation compared to perceiving threats as distal. The PAG is related to anxiety or fear, as well as fast reflexive behaviors that can help to survive threats (e.g., fight, flight, or freeze; Mobbs, Petrovic, Marchant, Hassabis & Weiskopf, 2007).

In sum, threat imminence influences an individual’s perception of a threat and level of arousal. Therefore, threat imminence stimulates different types of reactions to deal with the threat: Firstly, when a threat is not imminent, it is less arousing, more ambiguous and immediate action is not required, as is the case when a threat is imminent. Thus, people opt for more examination/exploration for non-imminent than imminent threats. Secondly, when threat is imminent it heightens arousal and people tend to fight instead of freezing or fleeing; fighting is most likely to occur when people perceive confrontation with the threat as

(7)

CREATIVITY AT GUNPOINT Creativity and Threats

Creative ideas are novel and useful (Amabile, 1983). Creativity can be measured by fluency (the amount of non-redundant ideas; e.g., Guilford, 1967), flexibility (the amount of ideas in different categories; e.g., De Dreu, Nijstad & Baas, 2011), and originality (the infrequency of useful ideas; e.g., De Dreu & Nijstad, 2008). Threat is usually assumed to undermine creativity: Threat leads to higher arousal (Dutton et al., 1981), elicits negative affect like fear and anxiety (McNaughton & Corr, 2004) and activates an avoidance motivation (Gutnick, Walter, Nijstad & De Dreu, 2012). High arousal leads to narrowed attention and default to primary dominant responses (Dutton et al., 1981). Highly anxious individuals have been shown to use narrower cognitive categories than less anxious

individuals (Mikulincer et al., 1990), and avoidance motivation is usually associated with less risky responses and less flexibility (Gutnick et al., 2012). According to these studies, these threat-induced changes in perception and cognitive processes are negatively related to creative performance.

However, recent work has offered another view on the link between threat and

creativity: Threat and threat-related status can also motivate and increase creativity when it is functional to deal with the specific problem at hand (i.e. creativity serves to fulfill the goal of avoiding loss or harm; Baas et al., 2011; Roskes et al., 2012). For example, when people are primed with a conflict set, they invest more effort to generate functional conflict tactics and come up with more creative conflict tactics, compared to when they are primed with a cooperation set (De Dreu & Nijstad, 2008). Moreover, studies on the Dual Pathway to Creativity Model (DPCM) have shown that people can achieve creativity not only through cognitive flexibility but also via cognitive persistence. Cognitive flexibility emerges by thinking within broad cognitive categories, or by rapidly switching between different categories. Cognitive persistence is manifested by exploring within one or a few narrow

(8)

categories (De Dreu, Baas & Nijstad, 2008; Nijstad et al., 2010). Several studies examined characteristics related to threats and cognitive persistence: Rodents and humans become more anxious after experiencing threatening events (Kim & Anderson, 2015; Grillon et al., 2004) and people their avoidance behaviors increase after such events (Glotzbach, Ewald,

Andreatta, Puali, & Mühlberger, 2012; Cavallo, Fitzsimons & Holmes, 2010). Such characteristics are related to cognitive persistence: Compared to activating positive moods (e.g., elation), activating negative moods (e.g., fear and anxiety) lead to higher creativity through cognitive persistence (Baas, De Dreu & Nijstad, 2008).Similarly, avoidance-motivated individuals can be as creative as approach-avoidance-motivated individuals by persistently working on the task as long as creativity serves to fulfill a goal (Roskes et al., 2012). Furthermore, time is an important factor in reaching creativity through cognitive persistence. After people have generated several common ideas within limited categories, subsequent ideas become more original and this requires time (Baas et al., 2008).In line with this role of time in creativity in negative situations, a study showed that time pressure weighs stronger on creative performance for high avoidant-motivated individuals than for low motivated individuals. When they are faced with time pressure, high avoidant-motivated individuals perform worse than low avoidant-avoidant-motivated individuals on tasks demanding creative thinking (Roskes, Elliot, Nijstad & De Dreu, 2013).

This work suggests that the influence of threat on creativity is more complex than previously assumed; threat does not necessarily undermine creativity. Instead, it could improve creative thinking when people are aroused by the threat and when creativity helps to solve the specific threat. Moreover, according to the DPCM, threatened people need time to explore more deeply within relatively narrower cognitive categories in order to increase creativity.

(9)

CREATIVITY AT GUNPOINT Aggression

Of course, people exhibit differences in reactions and creativity in reactions to threatening situations. Such differences may be explained by varying degrees of aggression. Trait aggression consists of four characteristics: A tendency to hurt or harm others by verbal (verbal aggression) or physical means (physical aggression), to experience more anger (anger), and have unfriendly, hostile feelings (hostility; (Buss & Perry, 1992). The tendency to display rude, anti-social or verbally aggressive behaviors is related to an orientation to move toward threats when stating responses to human threat situations. In contrast, people who do not tend to engage in such behaviors will avoid threats instead. Likewise, aggression is related to fighting or threatening as a reaction to threats (MacLaren et al., 2010).

Additionally, aggressive traits are associated with higher sensitivity to aggressive stimuli. For example, anger is related to aggressive reactions in hostile situations (Wilkowski & Robinson, 2008), walking forward more (Mayan & Meiran, 2011), and to increased pulling motion (Lewis, Alessandri & Sullivan, 1990). Similarly, in a study by Bushman (1995), participants were asked to compete against an opponent on a reaction time task immediately after

watching a violent or nonviolent videotape. During the reaction time task, the slower

competitor received shocks from the faster competitor. Among individuals who were high in trait aggression, those who watched the violent videotape selected stronger shocks for their competitors to receive than those who watched the nonviolent video tape. When individuals were low in trait aggression, there was no difference between the violent and nonviolent videotape condition. This shows that aggressive-related information can elicit more aggressive responses from people with higher trait aggression.

Several aspects of aggression are also associated with creativity: Hostility is positively related to creativity (Feist, 1998) and trait physical aggression is related to malevolent

(10)

(Cropley, Kaufman & Cropley, 2008). What is benevolent for one person can be harmful for another person. For example, violent self-protection can be benevolent for oneself, but harmful for another. In a study conducted by Lee and Dow (2011), participants were asked to generate ideas on the uses for a pencil and a brick; it was shown that participants high in trait physical aggression generated more malevolent uses for the pencil and the brick than

participants low in trait physical aggression. Present study

The present study examined the influence of threat direction and response time on creative defensive reactions. Also, the moderation effect of aggression in the relationship between threat direction and the amount of (creative) fighting tactics was studied. I manipulated threat imminence by presenting direct versus indirect threats with the use of photographs. Participants thought and typed their most likely reaction to the situations depicted in the photographs either immediately (immediate condition) or a few seconds later (delayed condition).

According to the functional perspective of creativity, threatened people can be creative only in domains that are functional and adaptive to solve the specific type of threat. Compared to indirect threats, direct threats are more imminent (Fernandes et al., 2013) and arousing (Flykt et al., 2007) and the urgency to deal with direct threats is higher. Therefore, compared to indirect threats, people are more aroused and feel more urged to deal with direct threats, leading them to invest effort into coming up with creative responses. Furthermore, I used a threat-related idea generation task in which participants are instructed to indicate their first reaction. Reacting to a threatening situation only allows a few types of response categories to consider (e.g., fighting, fleeing, freezing or exploring), and therefore in threat-relevant domains the topic itself is narrowly defined. This gives a precondition for the persistent pathway: With more time to respond, threatened people can explore deeper within this narrow

(11)

topic and by this exploration within the topic people can eventually heighten their creativity. Therefore, I hypothesize that for overall fluency of ideation, having more time to think will be associated with a greater number of reactions (Hypothesis 1a). I further hypothesize that direct threats will lead to a higher number of reactions than indirect threats (Hypothesis 1b). Furthermore, I expect an interaction effect between threat condition and time to respond condition, such that there is a strong effect of the time to respond condition on the number of reactions when threat is direct, but this effect is moderate for indirect threats (Hypothesis 1c). With more time to respond, people can think about more reactions and thus eventually engage more tactic categories different from an initial idea. Furthermore, compared to indirect threat, direct threat should lead to higher persistence. Therefore, I predict that for flexibility of ideation, direct threats lead to a higher number of tactic categories than indirect threats

(Hypothesis 2a) and in the delayed response condition people will react with a greater amount of tactic categories than in the immediate response condition (Hypothesis 2b). Furthermore, I expect an interaction effect between time to respond condition and threat condition, such that when threat is direct, the effect of time on flexibility is stronger than when threat is indirect (Hypothesis 2c).

More time will lead participants to think about more possible reactions and thus eventually leads to a larger number of original tactics. Therefore I predict that longer time to respond will lead to a larger number of original tactics (Hypothesis 3a). As participants’ creativity is more functional in the direct than indirect threat condition, I hypothesize that the number of original ideas will be larger in the direct threat than in the indirect threat condition (Hypothesis 3b). With more time, people will come up with more original ideas, especially when creativity is functional and people are more aroused. Therefore, I expect an interaction effect between the threat condition and the time to respond condition, such that there is a strong effect of time on the number of original tactics when threat is direct, but this effect is

(12)

moderate when threat is indirect (Hypothesis 3c).

Compared to when threat is indirect, people tend to fight more in response to direct threats, such as when they perceive confrontation with a threat as inevitable (e.g., a gun is pointed to a person and there is a small chance of successfully hiding or running). When a threat is indirect (e.g., a predator is looking away), the threat is more ambiguous and a person does not necessarily have to directly act, for example by fighting or fleeing. Instead a person can use functional reactions such as exploring the situation or thinking about options. I therefore hypothesize that the threat condition will influence the number and originality of fight or threat exploration reactions to threat, such that direct threat is expected to lead to a greater number of fight tactics and a greater number of original fight tactics compared to indirect threat (Hypothesis 4a). On the other hand, indirect threat is expected to lead to a greater number of threat exploration tactics and a greater number of original threat exploration tactics compared to direct threat (Hypothesis 4b). As exploratory part of the research, I will also look at the interaction effect between time to respond and threat condition on each tactic type and propose that more time will lead to a larger number of original tactics, yet only the specific tactic type is adaptive to regulate the threat (e.g., fighting when threat is direct or exploring when threat is indirect).

I have argued that there is a positive relation between aggression and fighting as a response to threat (MacLaren et al., 2010). Also, there is a positive relation between reacting with aggression and aggressive stimuli (Wilkowski & Robinson, 2008; Bushman, 1995), and aggression is related to higher creativity in harming others (Lee & Dow, 2011). Therefore, I expect a main effect for aggression, such that aggression increases the number of fight tactics and the number of original fight tactics (Hypothesis 5a). Furthermore, I expect that the effect of threat direction on the number of fight tactics and the number of original fight tactics is moderated by aggression. That is, for individuals higher in aggression, threat direction has a

(13)

greater effect on the number of fight tactics and the number of original fight tactics than for individuals lower in aggression (Hypothesis 5b).

Methods Design and participants

Native Dutch speaking undergraduate students (N = 182, Mage = 22.21, SDage = 3.19, 73.1% female) from the University of Amsterdam participated for research credits or 5 euros. The participants were randomly assigned to one of four conditions in a 2 (threat condition: direct vs. indirect threat) x 2 (time to respond condition: immediate response vs. delayed response) between-subjects design. The dependent variables were overall fluency, within-category fluency, overall flexibility, overall originality, and within-within-category originality. Procedure

Upon arrival to the laboratory, participants signed an informed consent form and were seated in individual cubicles preventing them from seeing and communicating with others. Each cubicle was equipped with a personal computer that displayed all instructions and registered all responses. After giving some personal information (age and gender),

participants completed a questionnaire measuring trait aggression. Then participants started a task in which they were asked to think about and type their response to a specific threatening situation depicted in a photograph. In all conditions, 14 photographs of either a threatening human or animal (Kveraga et al., 2015; see Appendix A) were displayed one at a time in random order. The content of the photographs was related to each condition: In the direct threat condition, the threats were directed to the viewer (the photograph was of a man

pointing a gun at the participant) and in the indirect threat condition, the threats were directed away from the viewer (the photograph was of a man pointing a gun at others). Participants were asked to state their most likely response in the specific situation depicted in each picture (in total 14 responses). For the response time there were two conditions: An immediate

(14)

response condition and a delayed response condition1. In the immediate response condition, participants had 12 seconds in total to type their reaction, and in the delayed response condition, they had 22 seconds in total to type their reaction. After the task, I measured how participants perceived the pictures, as a manipulation check. Lastly, participants were debriefed, rewarded, and dismissed.

Dependent variables

Participants’ defensive reactions were categorized into 7 broad tactics: flight tactics (e.g., “run away”, “walk backwards”), fight tactics (e.g., “strike back”, “find weapons”), freeze tactics (e.g., “stand still”, “make no sound”), threat exploration tactics (e.g., “think about possible solutions”, “assess the situation,” “vigilant monitoring”), cooperative approach tactics (e.g., “cooperate with the offender”, “convince the attacker that it is meaningless to hurt you”, “act friendly”), non-functional avoidance tactics (e.g., “ignore the threat”, “act as if nothing is wrong”), and other unspecified tactics that contained ideas that could not be coded into the former six (e.g., “take precautions”, “seek help”, “scream loudly”, “stay calm”). Fluency, flexibility and originality were three indexes for creativity in this study.

Fluency. The index for overall fluency was measured by the number of non-redundant defensive reactions a participant used. The index for within-category fluency was measured by the number of non-redundant defensive reactions within each broad tactic category.

Flexibility. To measure flexibility, a category diversity index was computed by counting the number of categories a participant used to generate responses. These categories were derived from a coding schema (based on previous study; see appendix B) in which all the responses generated by participants were coded into 30 concrete categories, for example, threaten back, flee, negotiate, cooperate, call for help and so on (and those 30 categories were then grouped into 7 broad categories mentioned above).

1

(15)

Originality. An overall originality index and within category originality index were computed. First, frequency scores from a database of prior research were assigned to participants’ answers. In the case that participants gave multiple ideas in one response, the part before ‘and’, ‘or’ or ‘then’ or a comma would be considered as their tactic. In case a participant said “Stay calm and ….”, the part after the ‘and’ was considered, as staying calm was seen as no action and the part after it was. After assigning a frequency score, the

infrequency was calculated by deducting the frequency score from 1. Tactics were counted as an original tactic in case the infrequency was .95 or higher (Baas, De Dreu & Nijstad, 2011). The sum of the number of original ideas (per category) per participant was the index of (within-category) originality.

Manipulation Checks. The participant’s ratings of how much the pictures induced time pressure (e.g., “I experienced time pressure to come up with a solution to the threats”) and arousal, as well as the degree to which the photographs were found to be negative, unpleasant, threatening (e.g., “I thought the pictures were threatening”), visible, and clear were used to measure the effectiveness of the experiment. Also, the participants rated whether the photographs were directed at and relevant to the participant. All items were measured on a 7-point scale (1 = strongly disagree, to 7 = strongly agree). Furthermore, a task was included for the purposes of assessing arousal and unpleasantness whereby the participant had to select on two axes how aroused and unpleasant they felt. These scores were standardized and used to form averaged scores together with the self-reported items for arousal (3 items, α = .71) and negativity and unpleasantness (4 items, α = .84). Also, the self-reported items measuring time pressure were averaged (2 items, α = .74).

Aggression

Aggression was measured with The Aggression Questionnaire (Buss & Perry, 1992) which aims to measure an individual’s tendency to hurt or harm others by verbal or physical

(16)

means and to experience anger and hostility. The questionnaire included 29 items measuring four facets of aggression: physical aggression (PA; e.g., “If somebody hits me, I hit back”, 9 items, α = .84), verbal aggression (VA; e.g., “When people annoy me, I may tell them what I think of them”; 5 items, α = .61), anger (AG; e.g., “When frustrated, I let my irritation show”; 7 items, α = .81) and hostility (HS; e.g., “I wonder why sometimes I feel so bitter about things”; 8 items, α = .8). Participants answered on a 7-point scale (1 “Extremely

uncharacteristic” to 7 “Extremely characteristic”). Results Data treatment

Participants were asked to report 14 reactions to the photographs, in case a participant typed less than half (less than 7) of the reactions, the participant did not qualify to have completed the task and was therefore considered as an outlier. One participant was considered as an outlier and excluded from further analyses. The resulting sample size was 181 (Mage = 22.22, SDage = 3.2, 73% female, Nimmediate+direct = 46, Ndelayed+direct = 45, Nimmediate+indirect = 45, Ndelayed+indirect = 45). Greenhouse-Geisser corrections were given in case the assumption of sphericity was violated during repeated measures analyses.

Manipulation Check

To verify the effectiveness of the manipulations, I entered participants’ ratings of time pressure and arousal, as well as negativity and unpleasantness, threat direction, self-relevance, visibility, and clarity of the pictures in separate ANOVAs with threat condition and time to respond condition as between-subjects variables. Firstly, there was a significant main effect of time to respond condition on the experience of time pressure, F(1, 177) = 33.2, p < .001, ηp2 = .16. Participants in the immediate response condition experienced more time pressure (M = 5.32, SD = 1.41) than participants in the delayed response condition (M = 3.92, SD = 1.82). Secondly, there was a significant main effect of threat condition on participants’ experience of

(17)

the direction of the threat, F(1, 177) = 205.6, p < .001, ηp2 = .54. Participants in the direct threat condition reported that the animals or people in the pictures looked more in the participant’s direction (M = 6.02, SD = 1.14) than in the indirect threat condition (M = 3.13, SD = 1.54). In addition, threat condition had a marginally significant effect on how personally relevant participants reported the pictures to be, F(1, 177) = 3.52, p = .062, ηp2 = .02.

Participants in the direct threat condition found the pictures to be more relevant to themselves (M = 3.78, SD = 1.77) than participants in the indirect threat condition (M = 3.31, SD = 1.58). Lastly, there was a marginally significant effect of threat condition on how visible participants reported the pictures to be, F(1, 177) = 3.24, p = .074, ηp2 = .02. Participants in the direct threat condition found the pictures to be more visible (M = 6.04, SD = .88) than participants in the indirect threat condition (M = 5.78, SD = 1.1). All the other effects of time to respond condition, threat condition or their interaction on time pressure, arousal, negativity, threat direction, self-relevance, visibility and clarity were not significant, Fs > 1.5, ps > .20.

In short, negativity and unpleasantness and the clarity remained equal across groups and thus did not influence the experiment. Moreover, the time to respond manipulation was successful: Participants with less time to respond reported more time pressure than

participants in the delayed response condition. Furthermore, the threat direction manipulation was successful: Participants in the direct threat condition reported the pictures to be more directed and relevant to themselves than participants in the indirect threat condition. Arousal, however, remained equal across groups. In the direct threat condition participants rated the pictures to be more visible than in the indirect threat condition, however, the ratings indicated that the visibility was high enough that both groups could successfully conduct the tasks.

(18)

Fluency

To test the effects of threat condition, time to respond condition, and their interaction on the number of responses (overall fluency), I entered the total number of ideas generated per participant in a 2 (threat condition: direct vs. indirect threat) x 2 (time to respond condition: immediate vs. delayed response) ANOVA with threat condition and time to respond condition as between-subjects factors. The main effect of time to respond condition was significant, F(1, 177) = 11.77, p = .001, ηp2 = .06. Participants with more time to respond generated more answers (M = 13.59, SD = .70) than participants who had to respond

immediately (M = 13.08, SD = 1.22). The main effect of threat condition (F < 1, p = .653) and the interaction effect between threat condition and time to respond condition on the number of participant’s responses to the pictures (F < 1, p = .997) were not significant.

Flexibility

To test the effects of threat condition, time to respond condition, and their interaction on the number of narrow categories of defensive reactions a participant used (flexibility), I entered the total number of narrow categories in a 2 (threat condition: direct vs. indirect threat) x 2 (time to respond condition: immediate vs. delayed response) ANOVA, with threat condition and time to respond condition as between-subjects factors.

Whereas the main effect of threat condition was not significant, F(1, 177) < .01 p = .949, ηp2 < .01, the main effect of time to respond condition was significant, F(1, 177) = 5.86, p = .016, ηp2 = .03. Participants with more time to respond generated responses in more different tactic categories (M = 6.05, SD = 1.58) than participants who had to respond

immediately (M = 5.54, SD = 1.33). Moreover, the interaction effect between threat condition and time to respond condition was significant, F(1, 177) = 4.13, p = .044, ηp2 = .02 (see figure 1). Pairwise comparisons showed that when threat was facing away, the participants with more time to respond reported responses in more categories (M = 6.23, SD = 1.6) than

(19)

participants who had less time to respond (M = 5.31, SD = 1.24), F(1, 177) = 9.86, p = .002, ηp2

= .05. However, when threat is facing toward the participant, there was no such difference between the time to respond conditions on the number of categories of participants’ defensive tactics, F(1, 177) = .076, p = .783, ηp2 < .001 (Mdirect+immediate = 5.76, SDdirect+immediate = 1.35; Mdirect+delayed = 5.84, SDdirect+delayed = 1.55).

(20)

Figure 1. Differences in flexibility between the two threat (direct and indirect) and time to respond (immediate and delayed response) conditions.

0 1 2 3 4 5 6 7

Direct Threat Indirect Threat

Fle xib ilit y _Immediate response Delayed response

(21)

CREATIVITY AT GUNPOINT Originality

To test the effects of threat, time to respond, and their interaction on the originality of the responses, I submitted the number of original ideas per participant to a 2 (threat condition: direct vs. indirect threat) x 2 (time to respond condition: immediate vs. delayed response) ANOVA, with threat condition and time to respond condition as between-subjects variables. Figure 2 shows that the main effect of time to respond condition on the number of original ideas was significant, F (1,177) = 10.27, p = .002, ηp2 = .06. In the delayed response condition participants had a greater number of original tactics (M = 3.70, SD = 2.35) than in the

immediate response condition (M = 2.70, SD = 1.96). Furthermore, the main effect of threat condition was significant, F(1, 177) = 4.87, p = .029, ηp2 = .03. Participants in the direct threat condition reported a larger number of original tactics (M = 3.54, SD = 2.44) than participants in the indirect threat condition (M = 2.84, SD = 1.92). The interaction between threat and time to respond condition was marginally significant, F(1, 177) = 3.42, p = .066, ηp2 = .02.

Pairwise comparisons revealed that among the immediate response condition, participants in the direct threat condition reported more original tactics (M = 3.33, SD = 2.14) than

participants in the indirect threat condition (M = 2.04, SD = 1.54), F(1, 177) = 8.24, p = .005, ηp2

= .04. There was no such difference among participants in the delayed response condition (F < 1, p > .01). Furthermore, when threat was indirect, participants in the delayed response condition reported a larger number of original tactics (M = 3.64, SD = 1.94) than participants in the immediate response condition (M = 2,04, SD = 1,54), F(1, 177) = 12.70, p < .001, ηp2 = .07, whereas there was no difference among participants in the direct threat condition (F < 1, p > .01; Mdelayed+direct = 3.76, SDdelayed+direct = 2.72).

(22)

Figure 2. Differences in originality between the two threat (direct and indirect) and time to respond (immediate and delayed response) conditions. 0 0,5 1 1,5 2 2,5 3 3,5 4 4,5

Direct Threat Indirect Threat

N umb er o f o rig in al id ea s Immediate response Delayed response

(23)

CREATIVITY AT GUNPOINT Within-category fluency

To test the effects of threat condition and time to respond condition on the number of tactics in different tactic categories (within-category fluency), I submitted the number of tactics generated within each defense category to a 2 (threat condition: direct vs. indirect threat) x 2 (time to respond condition: immediate vs. delayed response) x 7 (tactic type: flight, fight, freeze, exploration, cooperation, nonfunctional avoidance, other unspecified) repeated measures ANOVA, with threat and time to respond as between-subjects factors and tactic type as within-subjects factor. The multivariate main effect of tactic type was significant, F(6, 172) = 473.05, p < .001, ηp2 = .94 (Greenhouser-Geisser correction: p < 0.001, ηp2 = 0.52).

Threat condition. The interaction between tactic type and threat condition was significant, F(6, 172) = 3.32, p = .004, ηp2 = .10 (Greenhouse-Geisser correction: p = 0.001, ηp2

= 0.03; see figure 3). Pairwise comparisons revealed that the effect of threat condition was significant for the number of flight tactics, F(1, 177) = 7.53, p = .007, ηp2 = .04. Participants in the direct threat condition generated more flight tactics (M = 5.22, SD = 2.58) than

participants in the indirect threat condition (M = 4.28, SD = 1.98). Moreover, threat condition had a significant effect on the number of cooperation tactics, F(1, 177) = 13.18, p < .001, ηp2 = .07. Participants in the indirect threat condition reported more cooperation tactics (M = 3.93, SD = 1.55) than participants in the direct threat condition (M = 2.99, SD = 1.94).

However, threat condition did not have a significant effect on the number of fight tactics (F < 1, p = .007), freeze tactics (F < 1, p = .371), threat exploration tactics (F < 1 p = .528), nonfunctional avoidance tactics (F < 1, p = .677) and other unspecified tactics (F < 1, p = .389).

Time to respond condition. There was a significant interaction effect between tactic type and time to respond condition, F(6, 172) = 2.07, p = .059, ηp2 = .07 (see figure 4; Greenhouse-Geisser correction: p = .052, ηp2 = .01). Pairwise comparisons showed several

(24)

effects of the time to respond condition on the number of each tactic type: Firstly, participants in the delayed response condition reported to fight more often (M = 2.23, SD = 1.9) than participants in the immediate response condition (M = 1.49, SD = 1.52), F(1, 177) = 8.3, p = .004, ηp2 = .05. Similarly, participants in the delayed response condition, reported a greater number of threat exploration tactics (M = .38, SD = .76) than participants in the immediate response condition (M = .16, SD = .43), F(1, 177) = 5.45, p = .021, ηp2 = .03. However, the main effect of time to respond condition was not significant on the number of flight tactics (F(1, 177) = 2.09, p = .15, ηp2 = .01), freeze tactics (F < 1, p = .361), cooperation tactics (F < 1, p = .76), non-functional avoidance tactics (F < 1, p = .677) and other unspecified tactics (F < 1, p = .956).

(25)

Figure 3. The differences between the two threat (direct and indirect threat) conditions of the number of tactic types used per participant.

Figure 4. The differences between the two time to respond (immediate and delayed) conditions of the number of tactic types used per participant.

0 1 2 3 4 5 6 N um be r of ta ct ics Direct Threat Indirect Threat 0 1 2 3 4 5 6 N um be r of l t act ics Immediate Delayed

(26)

Within-category originality

To test the effects of threat condition and time to respond condition on the number of original ideas within each tactic type, I entered the number of original ideas within a category in a 2 (threat condition: direct vs. indirect threat) x 2 (time to respond condition: immediate vs. delayed response) x 7 (number of original ideas within category: flight, fight, freeze, exploration, cooperation, nonfunctional avoidance, other unspecified) repeated measures ANOVA, with threat and time to respond as between-subjects factors and the number of original ideas within a category as within-subjects factor.

The multivariate main effect of tactic type was significant, F(6, 172) = 179.16, p < .001, ηp2 = .48 (Greenhouse-Geisser correction: p < .001). The interaction between time to respond condition and tactic type was not significant, F(6, 172) = 1.51, p = .179, ηp2 = .05 (Greenhouse-Geisser correction: p = .552).

Threat condition. The interaction between threat direction condition and tactic type was significant, F(6, 172) = 7.92, p < .001, ηp2 = .217 (Greenhouse-Geisser correction: p < .001; see figure 5). Pairwise comparisons showed that the effect of threat condition was significant on the number of original fight tactics, F(1, 177) = 4.33, p = .039, ηp2 = .02. Participants that reacted to direct threats generated a higher number of original fight tactics (M = .45, SD = .83) than participants that reacted to indirect threats (M = .23, SD = .56). Furthermore, the effect of threat condition was significant on cooperation tactics, F(1, 177) = 40.24, p < .001, ηp2 = .19. Participants in the indirect threat condition generated a higher number of original cooperation tactics (M = 3.60, SD = 1.56) than participants in the direct threat condition (M = 2.01, SD = 1.80). Lastly, the effect of threat condition was significant on other unspecified tactics, F(1, 177) = 4.57, p = .034, ηp2 = .02. Participants facing direct threats had a higher number of original other unspecified tactics (M = 1.21, SD = 1.11) than participants facing indirect threats (M = .86, SD = 1.11). Threat condition had no effect on the

(27)

number of original flight, freeze, threat exploration and nonfunctional avoidance tactics, Fs < 1, ps > .01.

The exploration of the interaction between threat condition and time to respond condition on within-category fluency and originality

The three-way interaction between threat condition, time to respond condition, and tactic type was examined as exploratory part of this study. One analysis was for the within-category fluency and one for within-within-category originality. For within-within-category fluency, the three-way interaction was not significant after the Greenhouse-Geisser correction, F(6, 172) = 2.12, p = .053, ηp2 = .07 (Greenhouse-Geisser correction: p = .122). Pairwise comparisons showed several interaction effects between the time to respond condition and threat condition on the number of a specific tactic type: Firstly, when threat was direct, participants in the immediate response condition had a greater number of fight tactics (M = 2.24 SD = 2.07) than participants in the delayed response condition (M = 1.46 SD = 1.64), F(1, 177) = 4.75, p = .031, ηp2 = .03. This difference was marginally significant when threat was indirect, F = 3.59, p = .06, ηp2 = .06. Participants in the delayed response condition reported a larger number of fight tactics (M = 2.22, SD = 1.73) than participants in the immediate response condition (M = 1.53, SD = 1.39). Secondly, when threat was indirect, participants in the delayed response condition reported a larger number of threat exploration tactics (M = .47, SD = .87) than participants in the immediate response condition (M = .13, SD = .34), F = 6.62, p = .011, ηp2 = .04. There was no such difference when threat was direct, F = .52, p = .49, ηp2 < .01. No significant effects of the interaction between threat condition and time to respond condition were found on the number flight, freeze, nonfunctional avoidance and other unspecified tactics, Fs < 2, ps > .1.

Regarding within-category originality, the three-way interaction between time to respond, threat direction, and tactic type emerged as marginally significant (see Figure 6a &

(28)

6b), F(6, 172) = 2.45, p = .107, ηp2 = .058 (Greenhouse-Geisser correction: .065). Several interaction effects between threat condition and time to respond condition on the number of original tactics within a tactic type were observed: Firstly, when threat was direct, the delayed response condition showed a larger number of original fight tactics (M = .60, SD = .94) than the immediate response condition,(M = .30, SD = .70), F = 3.97, p = .048, ηp2 = .02. When threat was indirect, there was no substantial difference between the time to respond

conditions, F(1, 177) = .02, p = .882, ηp2 < .01.Secondly, when threat was indirect,

participants in the delayed response condition (M = .47, SD = .87) reported a greater number of original threat exploration tactics than participants in the immediate response condition (M = .13, SD = .34), F(1, 177) = 6.62, p = .011, ηp2 = .04. However, when threat was direct there was no difference between the time to respond groups, F(1, 177) = .52, p = .47, ηp2 > .01. Thirdly, a marginally significant effect on the number of original other unspecified tactics was observed such that when threat was indirect, participants in the immediate response condition reported a larger number of original other unspecified tactics (M = 1.41, SD = 1.19) than participants in the delayed response condition (M = 1, SD = 1), F(1, 177) = 3.18, p = .076, ηp2

= .02.Yet there was no such difference when threat was direct, F = .45, p = .505, ηp2 < .01. There was no significant interaction between time to respond condition and threat condition on the number of original flight, freeze, cooperation or nonfunctional avoidance tactics, Fs < 2, ps > .15.

(29)

Figure 5. The differences between the direct and indirect threat condition of the number of original responses of each tactic type.

0 0,5 1 1,5 2 2,5 3 3,5 4 N umb er o f o rig in al ta ct ic s Direct Indirect

(30)

Figure 6a. The differences between the time to respond (immediate and delayed response) conditions of the number of original responses of each tactic type when threat was direct.

0 0,5 1 1,5 2 2,5 N um bdr o f o rig ina l i de as Direct Immediate Direct Delayed

(31)

Figure 6b. The differences between the time to respond (immediate and delayed response) conditions of the number of original responses of each tactic type when threat was indirect.

0 0,5 1 1,5 2 2,5 3 3,5 4 Indirect Immediate Indirect Delayed

(32)

Aggression

Descriptive statistics and intercorrelations are shown in table 1. Physical aggression (PA) positively correlated with the number of fight tactics (r = .24, p = .001): With higher PA, participants responded more often with fight tactics to the pictures. Gender negatively correlated with PA (r = -.34, p < .001), verbal aggression (VA; r = -.23, p = .002), and the number of fight tactics (r = -.26, p = .001). Women had lower PA and VA and a lower number of fight tactics than men. The originality of the fight tactics correlated significantly with threat condition (r = -.15, p = .041), such that there were more original fight tactics when threat was direct. Furthermore, the number of fight tactics was related to the number of

original fight tactics (r = .38, p < .001): With more fight tactics, the number of original fight tactics increased.

To further test whether aggression influences the number and originality of fight tactics and whether aggression moderates the link between threat condition and the number and originality of fight tactics, I conducted a hierarchical regression analysis in three steps. As the current data showed significant correlations and previous studies showed that males tend to fight more frequently (MacLaren, Best, & Bigney, 2010) and are in general more

aggressive than females (Struber, Luck & Roth, 2008; Archer, 2004), I dummy-coded gender (1 – male, 0 – female) as a control variable in the first step. For the second step, the four components of aggression (PA, VA, AG and HS) were centered around the means and threat direction (TD) was dummy coded (-1 – indirect threat, 1 – direct threat). For the third step, I created the interaction terms by multiplying each separate centered component of aggression with the dummy-coded TD variable.

As can be seen in table 2, gender predicts the number of fight tactics in the first step, R2 = .07, F(1, 179) = 13.18, p < .001. Men have a higher number of fight tactics compared to women (β = .26, t = 10.74, p > .001). In the second step I added PA, VA, AG, HS and TD to

(33)

the regression. The model remained significant, R2 = .11, ΔR2 = .04, F(6, 174) = 3.43, p = .003. PA (β = .21, t = 2.3, p = .023) positively predicted the number of fight tactics, whereas VA (p = .962), AG (p = .14), HS (p = .301) and TD (p = .997) did not. In the third step including the interaction terms, the model was still significant, R2 = .07, ΔR2 = .01, F(10, 170) = 2.13, p = .025. However, the interactions between PA and TD (p = .339), VA and TD (p = .769), AG and TD (p = .805), HS and TD (p = .725) did not predict the number of fight tactics.

To test the effects of gender, PA, VA, AG, HS and TD on the number of original fight tactics and whether aggression moderates the link between TD and the number of original fight tactics, I conducted a regression analysis in three steps. The first step included dummy-coded gender (1 – male, 0 – female), the second step included centered PA, VA, AG, HS and dummy-coded TD (-1 – indirect, 1 – direct). For the third step, I created the interaction terms by multiplying each separate centered component of aggression with the dummy-coded TD variable.

As shown in table 3, the first step with gender does not predict the number of original fight tactics, R2 = .01, F(1, 179) = 1.48, p = .225. In the second step, R2 = .06 ΔR2 = .05, F(6, 174) = 1.77, p = .108, TD emerged as a significant predictor (β = .15, t = 2.04, p = .043), meaning that with direct threat, the number of original fight tactics increased. PA (p = .155), VA (β = -.14, t = -1.62, p = .107), AG (p = .956) and HS (p = .539) did not predict the number of original fight tactics. In the final step, R2 = 1, ΔR2 = .04, F(6, 174) = 1.77, p = .108, the interaction between PA and TD emerged as a marginally significant predictor, β = .17, t = 1.89, p = .060. Whereas the interactions between VA and TD (p = .796), AG and TD (p = .338) and HS and TD (p = .095) did not predict the number of original fight tactics. Simple slopes analyses revealed that direct threat increased the number of original fight tactics compared to indirect threat for participants high in PA (+1 SD, b = .23, SE = .08, β = .32, t =

(34)

2.79, p = .006), yet threat direction had no effect for participants low in PA (-1 SD, b = -.01, SE = .08, β = -.02, t = -.14, p = .887; see figure 7).

Figure 7. Interaction between physical aggression (PA) and threat direction (indirect and direct threat). 0 0,1 0,2 0,3 0,4 0,5 0,6 0,7

Indirect Threat Direct Threat

N u m b er of or igi n al f igh t t ac ti cs Low PA High PA

(35)

CREATIVITY AT GUNPOINT Table 1.

Descriptive Statistics and Intercorrelations of Variables

Variable 1 2 3 4 5 6 7 8 M (SD) 1. Physical Aggression - 2.74 (1.11) 2. Verbal aggression .42** - 3.53 (.84) 3. Anger .48** .43** - 2.9 (1.07) 4. Hostility .35** .32** .56** - 3.15 (1.11) 5. Threat condition -.02 .04 -.02 -.02 - 1.5 (.50) 6. Gender -.34** -.23** -.07 -.04 -.07 - 1.73 (.45)

7. Number of fight tactics .24** .10 .02 .09 .01 -.26** - 1.86 (1.75) 8. Number of original fight tactics .09 -.09 -.03 -.05 -.15* -.09 .38** - .34 (.72) Note. Gender (1 = Men; 2 = Women), Threat (1 = direct, 2 = indirect).*p < .05. **p < .001.

(36)

Table 2.

Regression coefficients of the hierarchical regression analysis for the number of fight tactics

Step 1 Step 2 Step 3

b SE β b SE β b SE β Gender 1.03 .28 .26** .78 .31 .2* .79 .31 .20* PA .33 .14 .21* .34 .15 .21* VA -.01 .18 0 -.03 .18 -.02 AG -.23 .16 -.14 -.26 .17 -.16 HS .14 .14 .09 .16 .14 .10 TD 0 .13 0 0 .13 0 PA * TD .13 .14 .08 VA * TD -.05 .18 -.03 AG * TD -.04 .17 -.03 HS * TD -.05 .14 -.03 R² .07 .11 .11 Adj R² .06 .08 .06 ∆R² .07 .04 0 F-value 13.18** 3.43* 2.13* ∆F 13.18** 1.45 .26

PA= physical aggression; VA= verbal aggression; AG = anger; HS = hostility; TD = threat direction. †p < .1. *p < .05. **p < .001

(37)

Table 3.

Regression coefficients of the hierarchical multiple regression analyses for originality

Step 1 Step 2 Step 3

b SE β b SE β b SE Β Gender .15 .12 .09 .15 .13 .09 .16 .13 .10 PA .09 .06 .13 .08 .06 .12 VA -.12 .07 -.14 -.11 .08 -.12 AG 0 .07 -.01 -.04 .07 -.06 HS -.04 .06 -.06 -.01 .06 -.01 TD .11 .05 .15* .11 .05 .15* PA* TD .11 .06 .17† VA* TD -.02 .08 -.02 AG * TD .07 .07 .1 HS * TD -.1 .06 -.15 R² .01 .06 .1 Adj R² 0 .03 .04 ∆R² .01 .05 .04 F-value 1.48 1.77 1.79† ∆F 1.48 1.82 1.77

PA = physical aggression; VA = verbal aggression; AG = anger; HS = hostility; TD = threat direction. †p < .1. *p < .05. **p < .001

(38)

Discussion

Based on current theories about creativity, the present study tested the notion that threatened people can be more creative in dealing with threatening situations when they are given sufficient time. Originality was the key-index for creativity in the present study, as both (within-category) fluency and flexibility are precursors for originality, yet originality is not necessarily a precursor for (within-category) fluency and flexibility (De Dreu et al., 2008). The present study showed partial support for the functional creativity perspective and the persistence pathway to creativity: Firstly, people had a higher number of original tactics with more time. Secondly, when threat was direct people were more original than when threat was indirect. Thirdly, with more time, people their originality significantly increased when

reacting to indirect threat. However, contrary to expectations, this increase was not significant for reactions to direct threats. Moreover, when responses were delayed (as opposed to

immediate responses), people enhanced their creativity in functional domains: People had higher originality in fighting when threat was direct and exploring threats when threat was indirect. For aggression, PA predicted the number of fight tactics, TD predicted the number of original fight tactics and PA moderated the link between TD and the number of original fight tactics.

The role of time and threat direction

In line with the hypotheses regarding the role of time, participants whose responses were delayed reported a larger number of tactics (fluency), categories (flexibility), and original tactics (originality) compared to people who had to respond immediately. This supports the persistence pathway: More time enables people to enhance creativity in an aversive situation (Baas et al., 2008).

For the threat conditions, it was expected that direct threat would be related to higher fluency, flexibility and originality than indirect threat. Only for originality it showed an effect

(39)

of threat condition: The direct threat condition showed higher originality than the indirect threat condition. Moreover, there was an interaction effect between time to respond condition and threat condition on flexibility: Only when threats were indirect, there was higher

flexibility in the delayed response condition (as opposed to the immediate response

condition). To explain these differences in flexibility, humans enter a defensive mode called circa-strike when a threat is very imminent and contact with a threat is very likely (Fanselow, 1994). In this defensive mode humans tend to fight or flee. As these are only two types of categories, this could mean that direct threat would only evoke reactions in very few

categories compared to the indirect threat condition, where people reported more categories. Thus in the indirect threat condition people were enabled to explore more categories than in the direct threat condition and time may have been beneficial to explore these categories. This may explain the larger flexibility of delayed responses in the indirect threat condition.

There also was an interaction effect between threat and time to respond condition on the number of original tactics: Only when people had to respond immediately, the direct threat group showed significantly higher originality than the indirect threat group. Moreover, when looking at the effect of time, there was an increase in originality for delayed reactions

(compared to immediate reactions) in both threat conditions. However, only for the indirect threat group this increase with delayed reactions was significant (see figure 2). To speculate about this pattern of the interaction between the threat and time to respond condition, the lack of increase in originality with the added time for direct threats, may have been due to that (1) direct threat was urgent enough to solve and there was already sufficient time for people to think and come up with creative responses in the immediate response condition or (2) people in the direct threat condition required even more time in the delayed response condition to enhance creativity above what was shown in the immediate response condition. Therefore, added time may currently not have increased the creativity of reactions to direct threats. To

(40)

clarify the mechanism of time when people responded to indirect threats, indirect threats may have been sufficiently urgent for people to invest energy and thus enhance creativity when responses were delayed. Yet the urgency may not have been great enough to enhance creativity by quickly investing effort in the short time span of the immediate response

condition. Therefore, the indirect threat group may also have been significantly lower than the direct threat group when responses were immediate. Also, it should be noted that in the

current research, the threat conditions were not related to differences in participants’ reported arousal. This could have undermined the expected influence of the threat condition on the amount of cognitive resources people invested to enhance creativity. Prior studies have reported ambiguous differences in arousal in response to direct or indirect threats: On one hand, people did not differ in how aroused they became from either a direct or indirect threat, yet threatening stimuli were more arousing than neutral stimuli (Fernandes et al., 2013). However, Flykt et al. (2007) reported that compared to indirect threats, direct threats are perceived as more imminent and therefore lead to more arousal. Whereas Fernandes et al. (2013) used self-reports of arousal, Flykt et al. (2007) used skin conductance responses (SCR) as measurements for arousal. As this research used participants’ ratings of their arousal

instead of SCR measures, the current findings regarding arousal in the threat conditions are in line with Fernandes et al. (2013).

To inspect the roles of threat direction and time to respond on each separate tactic type, I looked at the number (within-category fluency) and originality (within-category originality) of tactics in the different conditions. Regarding within-category fluency, it showed an effect of time: When responses were delayed, people reported more fight or threat exploration tactics compared to when responses were immediate. However, threat condition did not influence the number of such tactics. This could mean that people only report an adaptive response such as fighting or exploring to solve the threat after they have had enough

(41)

time to think of their options and possibly discard other options such as obeying or running away from the threat. A reason for this necessity of time to respond with such adaptive tactics, could be that in the current research participants typed their reactions as their own ideas instead of choosing their reaction from a list of options as done in previous research

(Blanchard et al., 2001; MacLaren et al., 2010; Perkins & Corr, 2006). To come up with the most adaptive tactic for oneself, the self-generation of ideas might have required more time.

Findings regarding the interaction between threat and time to respond condition showed a non-significant pattern that in line with the notion that only with time, people come up with expected tactics in each threat condition. Participants reported opting to explore the situation or think of a plan more often when they were granted more time to think. Also, when threat was direct, people reported more fight tactics when responses were delayed, yet not when threat was indirect. However, the present study has not provided clear results on this.

Data also showed an unpredicted effect of threat condition on the number of cooperation and flight tactics: People fled more when threat was facing towards them than when threat faced away and the reverse was true for the number of cooperation tactics. This contradicts prior studies that showed that people mostly cooperate when a threat is imminent and mostly flee when they see chances of escaping the threat (Blanchard 2001, Fernandes et al. 2013). Thus (1) direct threats may not have been as imminent that people perceived no chances of successfully escaping from the threat or (2) indirect threats were not perceived as equally threatening that people reported to flee. For the cooperation responses it may have been that direct threats indeed lead to people act in line with circa-strike behavior as fighting or fleeing (Fanselow, 1994) and people perceived more space to (think of ways to) cooperate when threats were indirect.

Regarding within-category originality, threat condition was influential for the number of original fight tactics, in line with expectations. When threat was direct, participants came

(42)

up with more original fight tactics than when participants reacted to indirect threats. As an additional finding, there was also an effect of threat condition on cooperation: In the indirect threat condition, people reported more original cooperation tactics than in the direct threat condition. It may have been that cooperation was found to be more adaptive when reacting to indirect than direct threats and thus people invested effort into coming up with creative cooperation tactics under indirect threatening circumstances. Also, there was higher within-category fluency for cooperation in the indirect threat condition, which supports the notion that within-category fluency is related to increased originality (Baas et al., 2008).

There were no effects of time on within-category originality. However, equal to expectations, there were interaction effects between threat direction and the time people had to respond: When responses were delayed, participants had more original fight tactics in the direct threat condition, but not in the indirect threat condition. Moreover, when responses were delayed, participants had higher originality in threat exploration tactics in the indirect threat condition, yet not in the direct threat condition. For fighting, it can be explained that if a direct threat evokes a fight response as a necessity to deal with the situation (Fanselow, 1994; Blanchard et al. 2011; Fernandes et al., 2013), then people can be more creative by focusing their cognitive resources solely on fight tactics, similarly to when people are more creative when dealing with conflict and they are in a conflict set (De Dreu & Nijstad, 2008). A similar mechanism may account for the threat exploration tactics: As people explore a threat(ening situation) more when threats are more ambiguous (Blanchard, Blanchard, & Rodgers, 1991; Blanchard et al., 2011; Perkins & Corr, 2006) and indirect threats were expected to be more ambiguous than direct threats in terms of the danger they posed, it may have been most beneficial to assert resources to thinking of such threat exploration tactics when threats were indirect. In sum, when people are thinking of information related to fighting or threat

(43)

creativity in such tactics. Time is beneficial for this: With more time people can explore more within a category and thus can come up with more original ideas (Baas et al., 2008).

Therefore, when fighting or exploring the threat(ening situation) helps to solve the specific threatening situation, added time can help people to come up with more original fight or threat exploration tactics.

The role of aggression

The four facets of aggression were examined to see whether they influenced people’s tendency to fight in threatening situations. As previous and current findings showed gender is related to PA (Lee & Dow, 2011; Archer, 2004) and a tendency to fight (MacLaren et al., 2010), I controlled for gender. It was expected that higher aggression would lead to more fight tactics and more original fight tactics, especially when threat was direct as compared to

indirect. Of the four components of aggression, it showed that only PA predicted the number of fight tactics: People higher in PA more often responded with fight tactics than people lower in PA. There was no interaction between TD and PA on the number of fight tactics, which is at odds with the notion that aggressive behavior tends to be elicited by more aggressive situations and even more so for persons higher in aggression (Bushman, 1995; Wilkowski & Robinson, 2008). However, when threat was direct, people had more original fight tactics in comparison to when threat was indirect. Moreover, PA moderated this link between TD and the number of original fight tactics, such that the comparison between indirect and direct threats showed that only for people higher in PA was there an effect of threat direction on the number of original fight tactics; for people lower in PA this effect did not occur. People higher in PA would respond with more original fight tactics in response to direct threats (compared to indirect threats), whereas for people lower in PA this difference did not occur. In sum, this supports the notion that people higher in PA are more creative in fighting (Lee & Dow, 2011) and this effect only showed in this study when participants were reacting to more

(44)

personally relevant aggressive-related situations (direct threats). As for now, it seems that the theory that people will invest more cognitive energy to be creative within a category if doing so helps to fulfill people’s current goals (Baas et al., 2011, Roskes et al., 2012), is in line with the current results: If a situation requires aggression to solve it, people may invest more energy in coming up with fight-related tactics (e.g., looking for an object to use as a weapon). People higher in aggression could have had more aggressive ideas and thus eventually reach more creative ideas. Therefore, people higher in aggression would benefit from being in an aggression provoking situation and come up with more creative tactics.

Among all results, several have implications: Firstly, in line with the persistence pathway of the DPCM (De Dreu, Baas & Nijstad, 2008; Nijstad et al., 2010), it shows that time can indeed enhance creativity in threatening situations. Therefore, when expecting an individual to generate a creative outcome, the valence and time limit of a task or situation are important to be acknowledged. For example, a too tightly imposed deadline may not lead to optimal creative outcomes, as there is not sufficient time to explore options and generate a creative idea in a pressuring situation. Instead, for creativity to thrive, a pressuring task should preferably have sufficient time, so people can explore beyond unoriginal ideas. Secondly, the exploratory analyses show that when looking at specific tactic types, added time exclusively seems beneficial when a tactic is adaptive to solve the current situation (e.g. fighting when threats are direct). This implies that time only aids in enhancing creativity if efforts are focused to the goal of generating a creative solution. For individuals to direct their efforts to reach a goal by being creative, tasks may be framed as serious and performance oriented (Baas et al., 2008). Moreover, people may be instructed to identify problems, gather information, and generate alternatives and ideas (Zhang & Bartol, 2010), so they will persistently direct their efforts towards exploring options and enhancing creativity. In such cases, people may benefit from extra time to enhance creativity. Thirdly, PA increases