c. Discussion

The present experiment investigated the facilitating effect of anger primes on effort-related cardiac response. Based on the IAPE model (Gendolla, 2012), we had predicted that anger primes that were processed during task performance would result in weaker effort-related PEP response than neutral primes and sadness primes. Our findings only partially supported our hypotheses. Preliminary analyses had found that prime effects only occurred during the first minute of the task, presumably because participants disengaged after the first minute. This was not anticipated, because in our previous priming studies (e.g., Gendolla & Silvestrini, 2011), we did not observe such habituation effects. But it is of note that other studies have found that the influence of affect primes can decrease with time and habituation (Murphy, Monahan, & Zajonc, 1995; neutral primes. This supports the IAPE model idea of a facilitating effect of implicit anger during performance. However, also the PEP response in the sadness-prime condition tended to be weaker than in the neutral prime condition. This is not in accordance with our predictions. We interpret this finding in the context of the affect prime effects on participants’ ability and importance of success ratings.

Participants in the sadness-prime condition evaluated their ability as lower than participants in the anger-prime and neutral-prime conditions, suggesting higher task demand in the sadness-prime condition, because demand is inverse to ability (see Wright, 1998). Additionally, the importance of success in the sadness-prime condition was

61 significantly lower than in the other two cells. This suggests that participants in the sadness-prime condition have disengaged, because the high effort that was necessary due to the high subjective demand was not justified, because success importance was relatively low (see Wright, 1998; Stewart, Wright, Hui, & Simmons, 2009; Wright &

Dismukes, 1995)—an effect that is in accordance with the principles of motivational intensity theory (Brehm & Self, 1989). SBP response corresponded to that of PEP, as already found in previous studies (e.g., Gendolla & Silvestrini, 2011; Silvestrini &

Gendolla, 2011c). This can be explained as being caused by cardiac contractility effects, which can systematically influence systolic pressure by their impact on cardiac output (Silvestrini & Gendolla, 2009a, 2009c).

In summary, the present findings bring support to the idea that implicit anger has a facilitating effect during task performance—effort-related cardiac response in the anger-prime condition was lower than in the neutral anger-prime condition. Moreover, no evidence was found that the affect primes induced conscious affective states or that participants were aware of the affect primes’ content. However, given that the present study revealed other unanticipated—though post hoc explicable—results for the effect of implicit sadness, we conducted a second experiment to provide more conclusive results about implicit anger vs. sadness primes’ effect on effort-related cardiac response.

62 II.2.4. Study 2

In this experiment, anger, neutral, and sadness primes were presented during task performance as in Experiment 1. However, we used affect primes from a different database and applied the same priming procedure as in previous studies (e.g., Gendolla &

Silvestrini, 2011). In accordance with the IAPE model (Gendolla, 2012), we expected again a linear affect prime effect on PEP response: PEP reactivity was anticipated to be weaker in the anger-prime condition than in the sadness-prime condition, with the neutral-prime control condition falling in between.

II.2.4.a Method

Participants and Design

Forty-nine healthy undergraduate students from various disciplines of the University of Geneva (28 women, 21 men, mean age 21.08 years, SD = 0.35) were randomly assigned to a 3-cell (Prime: anger vs. sadness vs. neutral) between-persons design³. Data of 9 participants had to be excluded from the statistical analyses because of incomplete data, leaving a final sample of N = 40. Participation was voluntary and remunerated with a small amount of money (10 CHF, approximately 11 USD).

Affect Primes

Low contrast, low resolution, grey scale, front perspective pictures of averaged human facial expressions from the Average Karolinska Directed Emotional Faces (AKDEF), (Lundqvist & Litton, 1998) were used as affective stimuli. Half of the faces were female (AKDEF pictures FNES, FSAS, FANS) and half were male (AKDEF pictures MNES, MSAS, MANS).

3 The distribution of women and men was balanced between the experimental conditions.

63 Apparatus, Physiological Measures

Apparatus and measures were identical with Experiment 1, except for blood pressure assessment. In this study we assessed SBP and DBP with a Dinamap ProCare 300 monitor, (GE Medical Systems Information Technologies Inc., Milwaukee, WI), which uses oscillometry. A blood pressure cuff (DuraCuf) was placed above the brachial artery of participants’ left arm. The cuff automatically inflated in 1-min intervals during habituation and task performance periods. Assessed values were stored on hard drive and later analyzed offline.

Procedure

The study had been approved by the local ethical committee. The experiment was run in individual sessions and the procedure was very similar to that of Experiment 1. After having signed consent, participants were seated in front of a computer. The experimenter attached the blood pressure cuff and the electrodes and left the participant alone. The procedure started with biographical questions. Then participants indicated their affective state with 2 items of the positive (happy, joyful) and 2 items of the negative (sad, depressed) scales of the UWIST mood check list (Matthews, Jones & Chamberlain, 1990) on 7- point scales (1 - “not at all”, 7 - “very much”)⁴.

After these ratings, we recorded cardiovascular baseline activity as in Experiment 1 (8 min). After this, participants received the instructions (“Please respond as quickly and accurately as possible”) for the adapted version of the d2 mental concentration test (Brickenkamp & Zillmer, 1998). Task trials started with a fixation cross (1000 ms), followed by pictures of facial expression (26 ms—the presentation time was longer, because the administered averaged facial expression had much lower contrast than the affect primes in the first experiment). In each condition 1/3 of the faces showed an emotional expression (anger vs. sadness) while the other 2/3 showed emotionally neutral faces. This procedure has been found to be the most effective in the present priming paradigm (Silvestrini & Gendolla, 2011b). The facial expression pictures were immediately backward masked with a random dot pattern (133 ms), followed by the

4 Due to an error, we did not assess participants’ anger ratings as in the first study. However, as Experiment 1, also our previous studies administering anger primes and assessing explicit anger have not revealed any evidence that the anger primes made participants explicitly angry (Freydefont et al., 2012; Freydefont &

Gendolla, 2012).

64 task stimulus. From here on, the trial structure was identical with that of Experiment 1.

Participants worked on 36 trials after a training session with 10 trials including correctness feedback and only neutral facial expressions.

After the task, participants made the same task-related ratings as in the previous study and the same affect ratings as at the beginning of the procedure. This was followed by the same funnel debriefing procedure as in Experiment 1.

II.2.4.b. Results

Cardiovascular Baselines

Averages of values of the last two minutes of the baseline measurement period constituted the cardiovascular baseline values (Cronbach’s αs > .93)⁵. Cell means and standard errors appear in Table 3. Three-cell one-way ANOVAs found no significant baseline differences (ps > .08). Preliminary ANOVAs found also no gender effects on baseline values of DBP (p > .88) and HR (p > .21). However, there were significant gender effects on SBP and PEP baseline values. Men had higher SBP (M = 110.05, SE = 2.85 vs. M = 98.74, SE = 2.29), F(1, 38) = 7.94, p < .01, η² = .17, and longer PEP (M = 106.32, SE = 2.06 vs. M = 100.53, SE = 1.68) than women, F(1, 38) = 4.75, p < 04, η² = .11.

5 We calculated the cardiovascular baseline values from the 2 last min of the habituation period, because there was a decline in assessed values over the first 6 min. For the last 2 min of the habituation period, the values were stable and did not differ significantly from one another for all cardiovascular indexes (ps

65 Table 3: Means and Standard Errors (in Parentheses) of the Cardiovascular Baselines Values for the Experiment 2.

Anger primes Neutral primes Sadness primes

PEP 104.00

Note: PEP = pre-ejection period (in ms), SBP = systolic blood pressure (in mmHg), DBP

= diastolic blood pressure (in mmHg), HR= heart rate (beats/min).

Cardiovascular Reactivity

Cardiovascular reactivity scores were determined as in Experiment 1. Preliminary analyses found no significant gender effects on any indices (ps > .17). Gender was therefore not considered in the further analyses. Moreover, ANCOVAs found no significant associations between baseline values and reactivity scores (all ps > .23). The 1-minute scores showed high internal consistency across the entire task performance period (Cronbach’s Alphas > .97). Moreover, we observed no significant time effect on PEP reactivity (ps > .49). Consequently, we analyzed cardiovascular responses for the entire task performance period by averaging the 1-min reactivity scores.

66 Table 4: Means and Standard Errors (in Parentheses) of the Cardiovascular Reactivity Values for the Experiment 2.

Anger primes Neutral primes Sadness primes

SBP 4.91

Note: PEP = pre-ejection period (in ms), SBP = systolic blood pressure (in mmHg), DBP

= diastolic blood pressure (in mmHg), HR= heart rate (beats/min).

PEP Reactivity

Cell means are depicted in Figure 3. The linear contrast for PEP reactivity was significant, F(1, 37) = 6.20, p < .02, η² = .17, while the orthogonal quadratic contrast was not (p > .35). Moreover, focused comparisons revealed that PEP reactivity in the anger-prime condition (M = -1.04, SE = 0.77) was significantly weaker than in the sadness-prime condition (M = -4.08, SE = 0.13), t(37) = 2.49, p < .02, η² =.14. PEP reactivity in the neutral-prime condition fell between both cells (M = -1.56, SE = 0.53) and differed significantly from the sadness-prime condition, t(37) = 2.06, p < .05, η² =.10, while the difference between the anger-prime and the neutral-prime conditions was not significant (p > .67).

67 Figure 5: Means and standard errors of the pre-ejection period reactivity (in ms) during cognitive task for the Experiment 2.

SBP, DBP, and HR Reactivity

Cell means and standard errors appear in Table 4. Neither the linear (ps > .14), nor the quadratic (ps > .40) contrasts were significant for SBP, DBP, or HR responses.

Task Performance

Three-cell one-way ANOVAs of reaction times (in milliseconds) for correct responses (M = 735.45, SE = 27.74) and the percentage of correct responses (M = 95.55

%, SE = 0.02) revealed no significant effects between experimental conditions (ps > .39).

68 Task Rating

Three-cell one-way ANOVAs of participants’ ratings of task difficulty (M = 1.82, SE = 0.23), mobilized effort (M = 2.45, SE = 0.19), capability (M = 5.70, SE = 0.25), the subjective value of success (M = 4.88, SE = 0.22), and the importance of success (M = 5.25, SE = 0.21) found no significant effects (ps > .15).

Affect Rating

We calculated average mood scores for the positive and negative affect items of the UWIST scale taken before and after the task (Cronbach’s as > .67). A 3 (prime) x 2 (time) mixed model ANOVA of these scores revealed no significant effects (ps > .38;

before task: M = 5.20, SE = 0.18; after the task: M = 5.23, SE = 0.15). An ANCOVA assessing possible significant associations between the post-task affect scores and the reactivity score of PEP revealed no significant association (ps >.16)

Funnel Debriefing

Fifty-two percent of the participants spontaneously reported to have seen faces when the experimenter asked them to describe a trial of the d2 task. However, only 24%

of these individuals were able to differentiate the faces’ gender, and only one participant reported to have seen a neutral emotional expression. Moreover, nobody guessed a link between the emotional faces and effort mobilization.

69 II.2.5. General Discussion

The present studies investigated the facilitating effect of implicit anger during task performance in terms of its effect on effort-related cardiovascular response. Based on the IAPE model (Gendolla, 2012), implicit anger (as implicit happiness) activated in achievement situations should make knowledge about performance ease accessible, resulting in experiences of lower subjective demand during task performance. Given that effort is proportional to task demand as long as success is possible and justified (Brehm &

Self, 1989; see Wright & Kirby, 2001; Gendolla et al., 2012 for reviews), anger primes should result in relatively low effort if people perform under “do-your-best” instructions.

By contrast, implicit sadness should activate knowledge about performance difficulty, resulting in experiences of higher subjective demand and leading to higher effort intensity than implicit anger stimuli under this condition. Gendolla and Silvestrini (2011) have reported initial evidence for these predictions. The present research extended those previous studies by considering a neutral-prime control condition, which was expected to fall in between the implicit anger and sadness conditions, and using affect primes from a different database.

In support of our predictions, both experiments revealed the weakest cardiac PEP response–our primary dependent variable referring to effort (see Kelsey, 2012; Wright, 1996)–in the anger-prime conditions. Moreover, in Experiment 1 SBP reactivity corresponded to the pattern of PEP, which is in line with other studies that have successfully operationalized effort in terms of SBP response (see Gendolla & Richter, 2010; Wright & Kirby, 2001). However, the absence of significant effects on SBP reactivity in Experiment 2 is not problematic. SBP responses in previous studies were explained by the systematic impact of cardiac contractility on cardiac output and thus on SBP (Wright, 1996). But, as already pointed out in the introduction, SBP also relies on vascular resistance and can be masked by it. Thus, finding predicted effects on PEP reactivity without finding effects on blood pressure is not surprising, because PEP is the much purer index of beta-adrenergic impact on the heart. Most relevant, the present studies did not find that shortened PEP was accompanied by decreased blood pressure or HR. Thus, the PEP effects are hardly explicable by pre-load (ventricular filling) or after-load (arterial pressure) effects (see Sherwood et al., 1990).

Unexpectedly, our first study found only weak instead of the anticipated strong PEP reactivity in the sadness-prime condition. Above we have already discussed plausible

70 reasons for this finding. Participants’ ability and importance of success ratings suggest that task demand in the sadness-prime condition was high while success importance was rather low. That is, the subjectively high necessary effort was hardly justified, resulting in disengagement (see Wright, 1998; Stewart et al., 2009; Wright & Dismukes, 1995) according to the principles of motivational intensity theory. Moreover, it is possible that the high contrast affect primes used in Experiment 1 elicited automatic strategies for affect regulation in the implicit sadness condition (cf. Taylor, 1991). Experiment 2, in which we used low contrast affect primes, found support for the expected linear increase in effort-related cardiac response from the anger-prime to the sadness-prime condition, with the neutral prime condition falling in between. However, according to our interpretation, both of the present experiments found support for the IAPE model (Gendolla, 2012) idea of a facilitating effect of implicit anger in effort-mobilization.

Although PEP response in the sadness-prime condition of Experiment 1 was—for reasons discussed above—not stronger than in the anger-prime condition, it is of note that PEP reactivity in the anger-prime cell was significantly weaker than in the neutral prime control condition. This indicates the facilitating effect of implicit anger. Experiment 2 revealed the expected linear increase in PEP response from the anger-prime to the sadness-prime conditions with the neutral prime condition falling in between.

As already observed in our previous studies, the funnel debriefing procedures of both present experiments did not find any evidence for the possibility that participants were aware of the affect primes’ content or that they suspected a link between the

“flickers” and effort mobilization. This suggests that the affect priming procedure was implicit (see Gawronski, Hofmann, & Wilbur, 2006). Moreover, the affect primes had significant effects on PEP response without having any effects on consciously experienced mood. We are aware that zero effects, like the present ones on conscious affect, have to be interpreted with great caution. Not significant is not identical with “no effect”. However, as it stands, we note that we have at least not found any evidence that the primes have induced conscious affective states. As outlined in more details in the IAPE model (Gendolla, 2012), we explain affect prime effects on effort-related cardiovascular response with the implicit activation of mental representations of affective states (see Niedenthal, 2008; Zemack-Rugar et al., 2007). Being activated in the context of task performance, this affect knowledge (anger is associated with ease and sadness is associated with difficulty) influences the experience of task demand “online” during

71 performance. Consequently, we observed effects on effort intensity in accordance with the principles of motivational intensity theory (Brehm & Self, 1989).

Unlike Experiment 1 and some of our previous studies (Gendolla & Silvestrini, 2011; Lasauskaite, Gendolla, & Silvestrini, in press; Silvestrini & Gendolla, 2011c), Experiment 2 did not reveal significant affect prime effects on demand ratings. We attribute this to measurement error, which is explicable by the fact that the demand measures after performance must have suffered from a number of biases, which are typical for retrospective ratings (see Robinson & Clore, 2002). The IAPE model predicts implicit affect effects on experienced demand during performance. But this is hardly assessable without interrupting performance and thus influencing the process of interest.

Consequently, future studies should test for implicit associations between implicit affect and the difficulty and ease concepts, using appropriate methods (see De Houwer, Teige-Mocigemba, Spruyt, & Moors, 2009).

Moreover, the affect primes had not revealed effects on task performance in term of the number of correct responses or reaction times during the mental concentration task.

However, it has to be considered that the link between effort and performance is complex.

Effort and performance are not interchangeable constructs. Effort refers to the mobilization of resources for instrumental behavior, while performance refers to its outcome and performance depends on effort, ability, and strategy use, rather than effort alone (Locke & Latham, 1990). That is, the relation between effort and performance is complex. For the present research, it is most relevant that the present experiments provide evidence for the facilitating effect of implicit anger in terms of effort-related cardiac response.

73 II.3. Study 3

: Beyond Valence: The Differential Effect of Masked Anger and Sadness Stimuli on Effort-Related Cardiac Response.

Reprint of: Freydefont, L., Gendolla, G. H. E., & Silvestrini, N. (2012). Beyond valence:

The differential effect of masked anger and sadness stimuli on effort-related cardiac response. Psychophysiology 49, 665-671. doi: 10.1111/j.1469-8986.2011.01340.x

II.3.1.Abstract

This experiment investigated the moderating effect of masked anger versus sadness primes on objective task difficulty’s impact on effort-related cardiovascular response. Cardiovascular measures (ICG and blood pressure) were assessed during a habituation period and an easy versus difficult short-term memory task during which participants were exposed to masked emotional facial expressions. As expected, sadness primes led to stronger cardiac pre-ejection period (PEP) responses than anger primes when the task was easy. When the task was difficult, we observed the reversed pattern.

Here, anger primes led to stronger PEP reactivity than sadness primes. Heart rate responses described the corresponding pattern. The results demonstrate that masked anger and sadness primes have different effects on cardiac response in easy and difficult tasks.

The effect of anger primes resembles the facilitating effect of happiness primes observed in previous studies.

Descriptors: Implicit affect priming, Anger, Effort, Cardiovascular reactivity, Active coping

75 II.3.2. Introduction

Numerous studies have found that exposure to masked affective stimuli systematically influences evaluative judgments and human behavior (e.g., Murphy &

Zajonc, 1993; Öhman, Flykt, & Lundquist, 2000; Winkielman, Berridge, & Wilbarger, 2005). Apparently, implicit affect priming has these effects through the activation of mental representations of emotions (Zemack-Rugar, Bettman, & Fitzsimons, 2007). As masked words or objects can activate semantic knowledge in long-term memory (see Förster & Liberman, 2007 for a review), masked affective stimuli can influence evaluative judgments and behavior by activating emotion concepts (see Niedenthal, 2008).

Recent studies from our laboratory found that implicit affect priming also influences subjective evaluations of task difficulty and effort-related cardiovascular response (Gendolla & Silvestrini, 2011a). Participants who processed masked sadness stimuli during the performance of attention or short-term memory tasks showedstronger responses of both cardiac pre-ejection period (PEP) and systolic blood pressure (SBP) and evaluated task difficulty to be higher than participants primed with masked happiness or anger stimuli. These findings are in accordance with the principles of motivational intensity theory (Brehm & Self, 1989; Wright & Kirby, 2001), which posits that effort is mobilized proportionally to subjective task difficulty as long as success is possible and justified. The Gendolla and Silvestrini (2011) experiments demonstrated for the first time that anger primes, in comparison with sadness stimuli, were associated with subjective ease and corresponding low cardiovascular response. However, participants in these

Recent studies from our laboratory found that implicit affect priming also influences subjective evaluations of task difficulty and effort-related cardiovascular response (Gendolla & Silvestrini, 2011a). Participants who processed masked sadness stimuli during the performance of attention or short-term memory tasks showedstronger responses of both cardiac pre-ejection period (PEP) and systolic blood pressure (SBP) and evaluated task difficulty to be higher than participants primed with masked happiness or anger stimuli. These findings are in accordance with the principles of motivational intensity theory (Brehm & Self, 1989; Wright & Kirby, 2001), which posits that effort is mobilized proportionally to subjective task difficulty as long as success is possible and justified. The Gendolla and Silvestrini (2011) experiments demonstrated for the first time that anger primes, in comparison with sadness stimuli, were associated with subjective ease and corresponding low cardiovascular response. However, participants in these