Science as a meaning framework : testing belief in science as a way to deal with threats

Graduate School of Psychology

R

M

’

P

T

INTRODUCTION &METHODS

Status: final version Date: 26/08/2014

1. WHO AND WHERE Student

Name :Lisa Baart

Student ID number :5733790

Address :Goudsbloemstraat 139ha

Postal code and residence :1015JM

Telephone number :+31620226625

Email address :lisabaart@gmail.com

Supervisor(s)

Within ResMas (obligatory) :Bastiaan Rutjens

Specialisation :Social Psychology

External supervisor(s), if any :n/a

Second assessor :Joop van der Pligt

Research center / location :Universiteit van Amsterdam

Number of credits : 25

Science as a Meaning Framework: testing belief

in science as a way to deal with threats

By Lisa Baart

ABSTRACT - Perceiving the world as orderly and under control is a fundamental human motivation (Heckhausen & Schulz, 1995) and increases well-being (Fiske & Taylor, 1991). When confronted with experiences of randomness, meaninglessness or lack of control, people compensate by reaffirming a framework or belief system in some other domain. Traditionally, religion has been found to help people to cope with stress and anxiety. The present research investigated whether belief in science can function as an alternative meaning framework that can help people cope with psychological threats. Results show that affirming science in the context of a threat increases self-esteem and decreases feelings of uncertainty. However, no effect of belief in science on threat perceptions related to order and meaning was found. Limitations and suggestions for future research will be discussed.

INTRODUCTION

People are motivated to believe that they live in an orderly world in which things are under control (Heckhausen & Schulz, 1995). Feelings of randomness, meaninglessness or lack of control can be stressful and anxiety provoking (e.g. Maier & Seligman, 1976; Pennebaker & Stone, 2004) and can have detrimental effects on well-being, health and happiness (Rodin & Langer, 1977). Unfortunately, modern society is often complex. We have access to a great deal of information that often does not even concern us directly, but nevertheless can be experienced as threatening. For example, media informs us about wars on the other side of the world or the disastrous effects of climate change that might occur in the future. As a result of the current financial crisis people are unsure of their jobs and if they will be able to pay their mortgage. Recent psychological research tries to explain how people deal with threats like those described above. Research has shown that people can find comfort and meaning in their beliefs (Heine, Proulx & Vohs, 2006) such as religious belief (Inzlicht et al., 2011), national identity (Greenberg et al., 1990) or belief in progress (Rutjens, van der Pligt & van Harreveld, 2010). In this study we will investigate whether belief in science can function as a comprehensive meaning framework that can help people cope with psychological threats.

Building on cognitive dissonance theory, there are several theoretical frameworks that try to explain how people distinct threats such as existential anxiety, uncertainty, uncontrollability and meaninglessness. According to the compensatory control model (CCM; Kay, Gaucher, Napier, Callan & Laurin, 2008) personal control makes the world seem like a less random and unmanageable place (Kay, Whitson, Gaucher & Galinsky, 2011). People respond to losses of personal control by preserving a sense of order by for example adhering to conspiracies, defending their sociopolitical institution or believing in an interventionist God. These processes of compensatory control have been shown to help people cope with the anxiety and discomfort resulting from the lack of personal control. Terror management theory (TMT; Greenberg,

Solomon & Pyszczynski, 1997) asserts that the anxiety caused by mortality is the motivator behind increased support in cultural or religious beliefs. According to this theory, people have a natural desire to live and the realization that death is inevitable produces terror, stress and anxiety. Culture values can serve to manage the terror of death by providing life with meaning. They found that belief in religion can manage the terror of death not only by offering literal immortality (e.g. an afterlife) but also by providing symbolic immortality (e.g. being part of something greater). Other examples of cultural values that help people cope with anxiety caused by mortality salience are national identity (Greenberg et al., 1990) and ethical beliefs (Cuillier, 2009). The meaning maintenance model (MMM; Proulx & Inzlicht, 2012) proposes that people have a need for meaning and continually strive to preserve a functioning meaning framework. When people encounter a threat to meaning, for example through a self-esteem threat, mortality salience or simply by seeing something that does not make sense, they need to regain a sense of meaning. According to the model, people can do so by reaffirming an independent meaning framework such as religion (Van Tongreen & Green, 2010), moralism (Randles, Proulx & Heine, 2010) or nationalism (Heine, Proulx & Vohs, 2006).

These three different approaches reveal a common behavioral pattern: after experiencing a threat, wether it is a threat to control, mortality or meaning, people can compensate for a threat in one domain by reaffirming a belief system in some other domain. Religious belief has shown to be a particularly effective meaning framework to provide a comprehensive worldview that can compensate for lack of control (Kay, Whitson, Gaucher & Galinsky, 2009), alleviate anxiety (Inzlicht, Tullett & Good, 2012) and relieve stress (Ano & Vasconcelles, 2004). For example, Kay et al. (2008) found a causal relationship between lowered perceptions of personal control and the belief in the existence of a controlling God. Similarly, mortality salience (e.g. Burke, Martens & Faucher, 2010, Vail, Rothschild, Weise, Solomon, Pyszczynski & Greenberg, 2010) and experienced meaninglessness (Van Tongreen & Green, 2010) have been shown to result in an increased investment in religion as a cultural worldview. However, in the past decennia Western

society has gone through a process of secularization (Verweij, Ester & Nauta, 1997). There has been a decline in levels of religious practice and affiliation, and religion might not provide the same sense of order, purpose and identity as it used to. It is unlikely that for non-religious individuals religion can function as a compensatory meaning system in the context of threat the same way it does for religious believers (e.g. Arndt, Greenberg & Cook, 2002). For example, mortality salience does not increase supernatural beliefs among atheists (Vail, Arndt & Abdollahi, 2012).

In the current study we will investigate whether belief in science can function as an alternative meaning framework that can help people cope with threats. Whereas most Western individuals accept science as a reliable source of knowledge about the world, some perceive it as ‘a superior, even exclusive, guide to reality, and as possessing a unique and central value’ (Farias, Newheiser, Kahane & Toledo, 2013, p.4). This type of attitude can be referred to as ‘belief in science’, or ‘a science based worldview’, and is often accompanied with the rejection of anything supernatural (Dawkins, 2006). Indeed, throughout history, science and religion have repeatedly come into conflict. However, they also share many important characteristics. They both offer ultimate explanations for fundamental questions (Preston & Epley, 2009). Research suggests that the compensatory effects of religion are not driven by the supernatural or transcendent content per se, but by the general qualities of belief such as providing order and meaning. For example, Rutjens, Van der Pligt and van Harreveld (2010) found that after a control-threat people had no preference for intelligent design over the theory of evolution, as long as the latter was presented in an orderly way. This suggests that science, similar to religion, may offer a sense of order (Rutjens, Harreveld & van der Pligt, 2013), meaning, and explanation and satisfies personal goals (Preston, 2012). It is thus plausible that to a certain extent, science can serve the same comforting role that has been associated with religious belief.

As an example of this, Farias et al. (2013) showed that among non-religious individuals, belief in science increases when they are in threatening or stressful situations. They also found

that after mortality salience participants reported a greater belief in science, which was negatively correlated with religiosity and spirituality. Rutjens, Harreveld and Van der Pligt (2013) showed that science can compensate for lack of control in the face of threat by serving order and predictability. However, there has been no research on whether affirming one’s belief in science reduces stress and a sense of order, purpose and identity. Moreover, the exact relationship between pre-existing beliefs about religion and science and belief in science as a way to cope with threats remains unclear. Science and religion have a complex relationship. They are inherently opposed in their explanations and it is argued that belief in one necessarily undermines belief in the other (Dawkins, 2006). Indeed research has shown that increasing the perceived value of one, decreases the automatic evaluation of the other (Preston & Epley, 2008). Kay et al. (2008) showed the same hydraulic effect for belief in God and belief in a government as a defense mechanism. We are therefore also interested in the relationship between attitudes towards science and religion as a way to cope with threats.

In the present study we will investigate whether science, like religion, can function as a comprehensive meaning-making system. We will study the impact of affirmed vs questioned scientific belief on perceptions of threat and negative affect. For seven consecutive days, all participants will receive a newspaper article that contains a threatening message (e.g. “Russia threats to start a third world war). The affirming-science condition will also receive a newspaper article that affirms science as a belief system. The questioning-science condition will also receive a newspaper article that undercuts science, such as articles about scientific fraud or pre-mature scientific breakthroughs that later had to be tempered. Participants in the control condition will receive neutral articles. Every two days we will measure threat perception, that is - how orderly they see the world and how accessible death related cognition is at that point in time, and negative affect in the form of self-esteem and self-uncertainty. Moreover, on the last day we will ask them explicitly about their attitudes towards science and religion. This way we can both look at between-participants condition effects, and the temporal within-participant effects of threats

and questioning vs. affirming science. We hypothesize that belief in science, like religious belief, can serve as a meaning framework. Specifically, we expect that participants who are confronted with information that questions science will show increased threat perceptions and negative affect, such as higher stress and self-uncertainty, than participants who read neutral and science-affirming articles. Moreover, we expect that participants who are confronted with information that affirms science will have lower threat perceptions and lower stress than participants who read neutral or science-questioning texts. Our second hypothesis is that as a meaning-framework science and religion have a hydraulic relationship, meaning that increased belief in one results in a decreased belief in the other.

If science proves to be an effective meaning framework, it will show that we are not restricted to religion in our search for compensatory order and meaning. Moreover, if science can fulfill a fundamental need for order, identity and purpose, it will change the public perception of the role of science in our society, for example science education or the way science is represented in the media.

The study is a 3 condition blind paradigm with condition (affirming science vs. questioning science vs. neutral) as independent variable and threat perception (order, identity and purpose) and attitudes towards science and religion as dependent variables. Participants in all three conditions will receive a daily threat in the form of a newspaper article. Threat perception will be measured on three different moments in time (t=0, t=4 and t=7) and will both serve as a between-subjects variable and a temporal within-subjects variable, meaning that we expect differences between conditions to increase over time.

Participants

A total of 112 participants participated in this study. 26 participants were excluded from analysis because they filled in less than 4 questionnaires. 9 participants filled in only 6 questionnaires, their data will only be used for the within-participants analysis leaving 21 participants in the questioning condition and 28 participants in both the control- and the affirming condition. The final data set included 77 participants (28 men) between 17 and 41 years of age (M = 20,92, SD = 3,09).

Procedure and materials

Participants were told that they would be participating in a study “about the dynamic aspects of temporal memory”. For 7 consecutive days participants received an online “media package” in Qualtrics. The media package contained three articles: an article containing a threat, an article that, depending on condition, contained either information about scientific fraud and premature scientific breakthroughs (questioning condition), information that presented science as a broad meaning system (affirming condition) or a neutral article (control condition) (see Appendix A for an overview of the articles) and a third, unrelated filler article. After reading the articles, participants are asked to write a small (50 words) summary to make sure they read the articles thoroughly. Next participants will fill in an affect grid to measure their affect. The Affect

Grid (Russell, Weiss & Mendelsohn, 1989) is a single item mood scale that requires participants to place a mark on a 9 x 9 grid to indicate their current mood. The grid has two dimensions: pleasure-displeasure (horizontal) and arousal-sleepiness (vertical). For example, high energy and high frustration (upper right corner) indicate stress. Moreover, we will administer the felt uncertainty scale (McGreggor, Zanna, Holmes & Spencer, 2001), consisting of 19-item that measures uncertainty about the self. Participants indicate to what extent they experience emotions (e.g. uncomfortable, torn) on a 7-point Likert scale (1 = not at all, 7 = extremely). We added the single-item esteem scale (Robins, Hendin, & Trzesniewski, 2001) “I have high self-esteem”.

On day 1 (t = 0), day 4 and day 7, felt uncertainty and threat will be measured (see figure 1 for an overview of the seven days and dependent variables). To measure perceptions threat, participants will complete two questionnaires. The Death Thought Accessibility Task measures the accessibility of death-related thoughts. We will use the Dutch translation of this task as used by Rutjens, van der Pligt and Harreveld (2009). Participants will be presented with 16 Dutch word fragments with one letter missing. They are asked to complete the word. Eleven of the words could be completed as either a death related or a neutral word (e.g. ster., which could be finished as the word sterf, the Dutch translation of to die, or sterk, the Dutch translation of strong), whereas the other five were filler words, and could only be completed in a neutral way (e.g. .open, which can only be completed to neutral words such as lopen, the Dutch translation for walking). Threat to order will be measured with two questions. The questions are ‘the events in my life are largely incidental’ and ‘our lives are ruled by randomness’. These questions are rated on a 7 point scale (1 = completely disagree, 7 = completely agree). Following various kinds of threats, people often turn to abstract conceptions of reality (Jonas et al. 2014). Therefore, action identification is measured next with the behavior identification form (BIF) (Vallacher & Wegner, 1989). This questionnaire distinguishes between low-level construals, that emphasize how to do the action, and high level construals, that emphasize why the action is performed. The BIF

constitutes of 25 items that describes actions (e.g. ‘greeting someone’) and participants can indicate which of the two identifications (e.g. ‘saying hello’ or ‘showing friendliness’) suits them best for that action.

Moreover on day 1 (t=0) we will ask participants to rate their pre-existing beliefs in science and religion. Scientific faith will be measured with a 10-item scale as developed by Farias, Newheiser, Kahane and Tolade (2013). The items describe ideas about science (e.g. ‘Science tells us everything there is to know about what reality consists of’). The scale has been shown to have high internal consisitency (Cronbach’s α .86). We removed the item “In a demon-haunted world, science is a candle in the dark.” because we consider it to be too vague. Religiousness will be measured with a 10-item scale as developed by Plante and Boccaccini (1997). The items describe ideas about religion (e.g. ‘My relationship with God is extremely important to me’). Both questionnaires will be rated on a scale from 1 to 7 (1=strongly disagree to 7=strongly agree).

On the last day attitudes towards science and religion will be measured. We will use a translation and modification of the three questions as formulated by Hayes and Tariq (2000) to measure explicit attitudes toward science and religion. An example of a question is “we believe too often in science, and not enough in feelings and faith” or “Overall, religion does more harm than good”. We will also use two self-made questionnaires in a 7-point matrix table that ask whether the participant finds the idea that science/religion has an influence on society more negative or positive, more bothersome or enjoyable, more unpleasant or pleasant, more meaningless or meaningful and more useless or useful. Moreover we will use a slider on which they can indicate whether they want the financial investment by the government in science to be lower, the same or higher. Two sliders on which they can indicate to what extent their worldview is scientific/religious (0 = not at all, 100 = extremely) and a third 10-point slider without value points on which they can indicate whether their worldview is more scientific or religious. We will use these variables both as a manipulation check, and to look at the presumed hydraulic relationship between science and religion.

After the experiment, participants are asked to fill in questions about the purpose of the experiment to check whether they saw through the manipulation. Finally, they will be fully debriefed. RESULTS T0 • Pre-existing beliefs science • Pre-existing beliefs religion • Order • Death thought accessibility • Self-uncertainty • Self-esteem • Grid • Behavior Identification • Articles • Grid T2 • Articles • Grid T3 • Articles • Grid T4 • Articles • Grid • Death thought accessibility • Order • Self-uncertainty • Behavior Identification • Grid T5 • Articles • Grid T6 • Articles • Grid T7 • Articles • Death thought accessibility • Order • Self-uncertainty • Behavior Identification • Grid • Attitudes science and religion • Demographics

In order to assess the attitudes towards science we standardized the scores of the propositions towards science items, the science worldview item and the scientific influence items and added them to make the Zscience variable. We did the same for attitude towards religion (Zreligion). Worldview was measured by subtracting the science worldview item from the religion worldview item. A score above zero indicates a preference towards science, a score below zero indicates a preference towards religion. Successively, we calculated a potential ambivalence score using the formula by Thompson, Zanna, and Griffin (1995): (R + S)/2 - |R – S|. The rationale behind this formula is that strong positive and negative feelings are related to higher levels of ambivalence, in this case related to science and religion. Scores on this measure could range from -50 (not ambivalent at all) to 100 (very ambivalent).

To test the effect of condition (affirming vs questioning vs control) on threat perceptions, well-being and attitudes towards science and religion at t7, we performed an Analysis of Variance (ANOVA) We controlled for pre-existing beliefs in science and religion, belief in god and whether or not a participant completed the questionnaires on time. Additionally, we performed a repeated measures analysis on order, death thought accessibility, pleasure, affect, uncertainty, self-esteem and behavior identification.

Order

There was no significant main effect of condition on order perception at t7, F(2,77) = .439, p = .64712_{. The repeated measures analysis revealed no significant effect of condition on time F(4}

,140) =1.86, p =.121. Planned contrast revealed a significant effect of condition on time between time point 1 (t0) and time point 2 (t4) F(1,70) = 3.15, p<.05. However, this is probably due to the

1 _{We performed an ANCOVA to control for pre-existing beliefs in science and religiousness. As a covariate,} both pre-existing beliefs in science F(1,69) = 5,95, p<.05 and religiousness F(1,69) = 5,51, p<.05 were

significantly related to order perceptions. There was no significant main effect of condition on order perception at t7, F(2,69) = .300, p = .7421 after controlling for the pre-existing beliefs in science and religion. The repeated measures analysis revealed no significant effect of condition on time F(4,70) = 1,78, p = .135 after controlling for the pre-existing beliefs in science and religion.

2

Leaving out participants who did not fully complete the questionnaires in time had no effect on the main effect of order F(2,59) =.471, p =.627 nor on the repeated measures F(2,124) =.2,36, p =.057.

inexplicable large difference at the first time point (t0) between the questioning (M = 8.06, SD = 1.96), the control (M = 8.78, SD = 2.67) and the affirming condition (M = 7.46, SD = 2.86). See also figure 2.

Death thought accessibility

We found no effect of condition on death thought accessibility F(2,75) = 1.62 p = .2043 _{at t7.}

Planned contrasts revealed that participants in the affirming condition scored higher than participants in the control condition, but this effect was not significant t(75) = -1.61, p = .112 (see table 1 for the means and standard deviations). Repeated measures analysis revealed that there was no significant effect of condition on time F(3.31,124.1) = .388, p = .817. See also figure 3.

3 _{Leaving out participants who did not fully complete the questionnaires in time had no effect on the main effect} of DTA F(2,59) =.555, p =.577 nor on the repeated measures F(2,130) =..489, p =.744

7 7,5 8 8,5 9 T0 T4 T7 O rd er Time Questioning Control Affirming 3,5 4 4,5 5 T0 T4 T7 DT A Time Questioning Control Affirming

Figure 2: repeated measures plot of order

Figure 3: repeated measures plot of death thought accessibility

Felt Uncertainty

There was no effect of condition on felt uncertainty F(2,69)=1.405, p=.2524 _{at t7. Planned}

contrasts revealed that participants in the questioning condition (M=61.8, SD=20.8) scored lower than participants in the affirming condition (M=54.1, SD=23.5), meaning that participants in the questioning condition were more uncertain, but this effect was not significant p = .112. Repeated measures analysis revealed no significant effect of condition on time F(4,144) = 1.4, p = .236. However, planned contrasts revealed a significant effect of condition and time between the first (t0) and the third (t7) time point F(2,72) = 3.11, p <.05. The questioning condition (T0: M=66. SD=18.5) showed less decrease in felt uncertainty compared to the control (T0: M=61.1, SD=24.3; T7: M=53.7, SD=22.8) and the affirming condition (T0: M=66.2, SD=20.7) (see figure 4 for the plot).

Behavior Identification Behavior Identification

There was no significant effect of condition on behavior identification F(2,67) = 1.14, p = .326 at t7. Repeated measures analysis revealed no significant effect of condition on time F(3.31, 115.86) = .329, p = .823.

Self-esteem

4 _{Leaving out participants who did not fully complete the questionnaires in time had no effect on the main effect} of felt-uncertainty F(2,59) =.842, p =.436 nor on the repeated measures F(4,126) =.1,15, p =.335.

50 55 60 65 70 T0 T4 T7 Fel t U nc er ta in ty Time Questioning Control Affirming

Figure 4: repeated measures plot of felt uncertainty

We found no significant main effect of condition on self-esteem Welch’s F(2,44) = 1.87, p = .1665 _{at t7. However, post hoc tests revealed a significant difference between the affirming}

condition (M=5.19, SD=0.89) and the questioning condition (M=4.4, SD=1.5) (p<.05) and a trend between the difference of the control condition (M=5.04, SD=1.09) and the questioning condition (p=.103). The repeated measures analysis revealed a significant effect of condition on time F(4,140) = 3.25, p<.05. Planned contrasts revealed a significant effect of condition and time between time point 1 (t0) and time point 3 (t7), F(2,70) = 3.76, p<.05. Participants in the questioning condition (T0: M=4.75, SD=1.25) experienced decreased self-esteem over time while the effect for the control condition (T0: M=4.7, SD=1.17) and the affirming condition was opposite (T0: M=5.35, SD=0.79) (see figure 5 for the plot).

5

Leaving out participants who did not fully complete the questionnaires in time had no effect on the main effect of self-esteem F(2,59) =1.333, p =.272 and removed the effect on the repeated measures F(4,124) =.2,33, p =.06.

4 4,5 5 5,5 T0 T4 T7 Se lf-es teem Time Questioning Control Affirming

DTA FUS BIF Order Self-esteem

N M SD N M SD N M SD N M SD N M SD

Questioning 22 4.73 2.18 22 61.23 4.92 22 11.57 4.59 22 8.42 2.23 22 4.5 1.47 Control 28 4.75 1.75 28 58.44 4.65 26 9.77 4.93 28 8.35 2.93 28 5.04 1.07

Affirming 28 3.93 1.82 28 50.62 4.45 28 11.83 5.13 28 7.93 2.22 26 5.19 0.85

Figure 5: repeated measures plot of self esteem

Grid

We found no effect from condition on pleasantness. Welch’s F(2.42)= 2.147. p =.129 at t7. However. post hoc tests revealed a significant difference between the questioning and affirming condition (p<.05) and a trend between the questioning and control condition (p = .107) (see table 2 for the means and standard deviation). We found no significant effect of condition on stress F(2.74)= .212. p= .809. Repeated measures analysis revealed no significant effect of condition on time for pleasure F(10.12.328.97)= .771. p = .658. or stress. F(11.56. 375.71) = 1.04. p =.414.

Ambivalence

There was no significant effect of condition on ambivalence F(2.66) = .323. p = .725 at t7. Attitudes towards science and religion

There was a significant correlation between the standardized scores of attitudes towards science and religion. r =-.35. p (one tailed) <.05. We found no significant effect of condition on attitudes towards science F(2.65) = .834. p = .43967 _{or religion F(2.66) = .648. p = .526 at t7.}

We used a MANOVA to check the attitude items for science and religion individually. We found no effect of condition on investment in science F(2.59) = .511. p = .602. worldview F(2.66) = 1.51. p = .229. propositions science F(2.66) = .541. p = .585. propositions religion F(2.66) = .234. p = .792. influence religion F(2.66) = .136. p = .873 or influence science F(2.66) = 1.42. p = .251. Post hoc analysis revealed that the questioning condition (M = 72.63. SD = 16.12) had a less scientific worldview than the affirming condition (M=76.18. SD=16.4) but that the effect was non-significant (p = .105). Also. the affirming condition (M=29.45. SD=3.66) wanted greater

6

We performed an ANCOVA to control for pre-existing beliefs in science. As a covariate, pre-existing beliefs in science F(1,65) = 26,35, p< .05 was significantly related to order perceptions. We found no effect of condition on attitudes towards science F(2,65) = .461, p = .633 after controlling for pre-existing beliefs towards science as a covariate.

7

Leaving out participants who did not fully complete the questionnaires in time had no effect on the main effect of science F(2,59) =.986, p =.379 religion F(2,59) =1,134, p =.329

Table 1: means and standard deviations at t7

influence of science than the questioning condition (M= 27.53. SD=2.93). but again the effect was non-significant (p = .101).

DISCUSSION

Traditionally. religious belief has been a comprehensive worldview that can compensate for lack of control (Kay et al.. 2009). alleviate anxiety (Inzlicht. Tullet & God. 2012) and relieve

Zscience Zreligion Pleasant Stress Ambivalence

N M SD N M SD N M SD N M SD N M SD

Questioning 21 -.17 .64 22 .17 .79 22 186.99 61.87 22 150.66 59.17 22 -4 34.77

Control 27 .03 .8 27 0 .966 27 209.59 48.67 27 139.33 57.35 27 -6.67 33.20

Affirming 26 .11 .77 26 -.1 .75 28 217.39 33.31 28 144.63 64.40 26 -14.1 35.39

stress (Ano & Vasconcelles. 2004) by providing order. purpose and identity in the face of psychological threat. We hypothesizedthat belief in science. like belief in religion. can serve as a

comprehensive meaning framework. Specifically. we expected that participants confronted with information that affirms science will have lower threat perceptions and lower negative affect than participants who read information that questions science. Contrary to our expectations. we found no significant effect of condition on order perception. death thought accessibility. behavior identification. pleasantness. arousal or ambivalence. We did find a trend in the effect of condition on felt uncertainty and self-esteem; participants in the questioning condition were more uncertain and had less self-esteem than participants in the affirming condition. Moreover. our within-subjects analysis showed that the affirming condition became less uncertain and gained self-esteem over the course of the experiment. as opposed to the questioning condition.

Our second hypothesis concerned the hydraulic relationship between science and religion. We expected that increased belief in science would result in a decreased belief in religion. and vice versa. We found indeed that belief in science is negatively correlated with religious belief. However. there was no effect of our manipulation on attitudes towards science and religion. so we cannot draw any conclusion about their hydraulic relationship. The lack of effect on attitudes towards science is surprising because we actively manipulated these variables. and the lack of a significant effect means that our manipulation failed or was not strong enough. This could explain the lack of effect of condition on threat perceptions and negative affect. If for example the participants in the affirming condition did not actually have an increased belief in science. they would not be able to use it as a meaning framework to lower their threat perception. This lack of a manipulation effect raises questions to our findings that participants in the questioning condition felt more uncertain and had less self-esteem than participants in the affirming condition. Since we found no difference in attitude towards science between these two conditions it would be premature to assign the effect to belief in science. It could also be possible

that the articles about scientific malfunction constituted a threat in itself. explaining the lowered self-esteem and self-certainty of the questioning condition.

Another explanation for the failed manipulation might be the type of articles we used. Science is a broad concept. and even though we strived to a wide variety of science-related topics. they might have failed to influence belief in science in general. For example. it is possible that reading an article about a scientific breakthrough in the medical domain would only result in a temporary affirmation of medical science. and not of science in general. The long-term aspect of this study could in this case be a disadvantage because the effect of affirmation in one domain might fade over time making it impossible to affect belief in science in general. In our results. we found little difference between threat perceptions and negative affect at day 4 compared to day 7. Therefore. future research could adapt a simpler. shorter design. where participants read a wider variety of science-affirming articles in a shorter time period. which might also lower the participant drop-out.

Another possible reason why the manipulation. and therefore our hypothesis. failed is the number of participants in this study. The power analysis we did before conducting this study recommended at least 30 participants per condition. but due to a high participant drop-out rate during our study we did not reach this number. For example for the attitudes towards religion variable we had only 22 participants in the questioning condition and 28 participants in the affirming- and control condition. This gives us a power of 0.34. or a 66% chance at a type II error (Cohen. 1988). The lack of statistical power is a big limitation of this study and in order to conclude no effect of belief in science on threat perception and negative affect this study should be replicated with enough participants. Our inability to show any within-subjects effect of threat on negative affect underlines this. Previous research has shown that in the face of threat people have increased stress (Ano & Vasconcelles. 2004) and anxiety (Inzlicht. Tullett & Good. 2012). Even our control condition did not show increased anxiety or stress after receiving a threat for

seven consecutive days which implies either a lack of power or a too weak manipulation of threat.

There is a great deal of research that showed that religious belief increases in the face of threats such as lack of control (Kay. Gaucher. Napier. Callan & Laurin. 2008). existential anxiety (Greenberg. Solomon & Pyszczynski. 1997) or lack of meaning (Van Tongreen & Green. 2010). We argued that since science and religion share many important characteristics. e.g. they both offer ultimate explanations for fundamental questions (Preston & Epley. 2009) and both provide order and meaning (Rutjes. Van der Pligt & Van Harreveld. 2010). the same effect might apply to science. However. we found no increase in belief in science in the face of threats. These results contradict the findings by Farias et al. (2013) that secular individuals increase their belief in science to help them deal with stressful and anxiety provoking situations. Research has shown that an increase of religious belief in the face of threat alleviates anxiety (Inzlicht. Tullett & Good. 2012) and relieves stress (Ano & Vasconcelles. 2004). We expected a similar result for belief in science but since we did not have an increased belief in science this effect was not demonstrable.

Kay et al. (2008) have argued that different meaning making systems have a hydraulic relationship. An analogy of this is a glass of water that only when full offers a satisfying sense of order and meaning. Much of the time. personal control is insufficient to completely fill the glass. Therefore. in order to reach a satisfying level of control. people need to complement external belief systems. In this analogy. when one external belief system rises. the other goes down. Kay et al. (2008) showed this hydraulic effect for belief in God and belief in a government. In this study we failed to show the same effect for religious belief and belief in science but this is probably due to the aforementioned lack of threat. Future research could look more precisely at the relationship between religious belief and belief in science as meaning systems and could also include other meaning systems such as belief in a government or nationalism.

In conclusion. this study failed to show whether science. like religion. can function as a comprehensive meaning making system. Nevertheless. it seems too early to completely reject the hypotheses made in this study and an altered replication of this study could help obtain final answers.

REFERENCES

meta-analysis. Journal of Clinical Psychology .61. 461-480.

Arndt. J.. Greenberg. J.. & Cook. A. (2002). Mortality salience and the spreading activation of worldview-relevant constructs: Exploring the cognitive architecture of terror management. Journal of Experimental Psychology: General. 131. 307-324.

Burke. B. L.. Martens. A.. & Faucher. E. H. (2010). Two decades of terror management theory: A meta-analysis of mortality salience research. Personality and Social Psychology

Review. 14. 155–195

Cohen. J. (1988). Statistical power analysis for the behavioral sciences. Lawrence Erlbaum Associates publishers. New Jersey.

Cuillier. D. (2009). Mortality Morality: effect of Death Thoughts on Journalism Students’ Attitudes Toward Relativism. Idealism and Ethics. Journal of Mass Media Ethics. 24. 40-58.

Dawkins. R. (2006). The God delusion. Boston/New York: Houghton Mifflin.

Farias. M.. Newheiser. A.-K.. Kahane. G. & de Toledo. Z.. (2013). Scientific faith: Belief in science increases in the face of stress and existential anxiety. Journal of Experimental Social Psychology. in

press.

Greenberg. J.. Solomon. S.. & Arndt. J. (2008). A basic but uniquely human motivation: Terror management. In J. Y. Shah & W. L. Gardner (Eds.). Handbook of motivation science (pp. 114–

134). New York. NY: Guilford Press.

Greenberg. J.. Pyszczynski. T.. Solomon. S.. Rosenblatt. A.. Veeder. M.. Kirkland. S. & Lyon. D. (1990). "Evidence for terror management II: The effects of mortality salience on reactions to those who threaten or bolster the cultural worldview". Journal of Personality and Social Psychology. 58: 308-

318.

Fiske . ST.. & Taylor . SE . (1991). Social cognition . New York: McGraw - Hill.

Gray. J. (2004). Heresies: Against progress and other illusions. London. UK: Granta Books.

Greenberg. J.. Solomon. S.. & Pyszczynski. T. (1997). Terror management theory of self-esteem and social behavior: empirical assessments and conceptual refinements. In M. P. Zanna (Ed.). Advances in experimental social psychology. Vol. 29. (pp. 61−139) New York: Academic Press.

Hayes. B. C.. & Tariq. V. N. (2000). Gender differences in scientific knowledge and attitude towards \ science: a comparative study of four Anglo-American nations. Public Understanding of Science. 9.

433-447.

Heckhausen. J.. & Schulz. R. (1995). A life-span theory of control. Psychological Review. 102. 284–304.

Heine. S. J.. Proulx. T. & Vohs. K. D. (2006). The meaning maintenance model: On the coherence of human motivations. Personality and Social Psychology Review. 10. 88-110.

Inzlicht. M.. Tullett. A. M.. & Good. M. (2011). The need to belief: A neuroscience account of religion as a motivated process. Religion. Brain and Behaviour. 1. 192-212.

Kay. A. C.. Gaucher. D.. McGregor. I.. & Nash. K. (2010). Religious belief as compensatory control.

Kay. A. C.. Gaucher. D.. Napier. J. L.. Callan. M. J.. & Laurin.K. (2008). God and the government: Testing a compensatory control mechanism for the support of external systems. Journal of Personality and Social Psychology. 95. 18–35.

Kay. A. C.. Whitson. J. A.. Gaucher. D.. & Galinsky. A. D. (2009). Compensatory control: Achieving order through the mind. our institutions. and the heavens. Current Directions in Psychological Science. 18. 264–268.

Koole. S.L.. Van den Berg. A.E. (2005). Lost in de wilderness: Terror Management. Action orientation and Nature Evaluation. Journal of Personality and Social Psychology. 6. 1014-1028.

Maier. S. F.; Seligman. M.E. (1976). Learned helplessness: Theory and evidence. Journal of Experimental Psychology: General. 105. 3-46

McGregor. I.. Zanna. M.P.. Holmes. G.J.. Spencer. S.J. (2001). Compensatory conviction in the face of

personal uncertainty: Going to extremes and being oneself. Journal of Personality and Social Psychology. 80. 3. 472-488.

Nash. K.. McGregor. I.. & Inzlicht. M. (2010). Line bisection as a neural marker of approachmotivation. Psychophysiology. 47(5). 979–983.

Pennebaker. J. W.. & Stone. L. D. (2004). Translating traumatic experiences into language:

Implications for child abuse and long-term health. In L. J. Koenig. L. S. Doll. A. O’Leary. & W. Pequegnat (Eds.). From child sexual abuse to adult sexual risk: Trauma. revictimization. and intervention.

(pp. 201–216). Washington. DC: American Psychological Association.

Preston. J. L. (2011). Religion is the opiate of the masses (but science is the methadone).

Religion. Brain and Behaviour. 1. 231-233.

Preston. J.. & Epley. N. (2009). Science and God: An automatic opposition between ultimate explanations. Journal of Experimental SocialPsychology. 45. 238-241.

Robins. R. W.. Hendin. H. M.. & Trzesniewski. K. H. (2001). Measuring Global Self-Esteem: Construct Validation of a Single-Item Measure and the Rosenberg Self-Esteem Scale.

Personality and Social Psychology Bulletin. 27. 151-161.

Rodin. J.. & Langer. E. (1977). Long-term effects of a control-relevant intervention. Journal of Personality and Social Psychology.35.897-902

Rutjens. B. T.. van der Pligt. J.. & van Harreveld. F. (2010). Deus or Darwin: Randomness and belief in theories about the origin of life. Journal of Experimental Social Psychology. 46. 1078–1080.

Rutjens. B. T.. van Harreveld. F.. & van der Pligt. J. (2010). Yes we can: Belief in progress as compensatory control. Social Psychological & Personality Science. 1. 246–252.

Rutjens. B. T.. van Harreveld. F.. & van der Pligt. J. (2013). Step by step: finding compensatory order in science. Current directions in psychological science. 22. 250-255.

Rutjens. B. T.. van Harreveld. F.. van der Pligt. J.. Kreemers. L. M.. & Noordewier. M. K. (2013). Steps. stages. and structure: Finding compensatory order in scientific theories. Journal of Experimental Psychology: General.

Plante. T.G.. Boccaccini. M.T. (1977) The Santa Clara Strength of Religious Faith Questionnaire Pastoral Psychology. 45. 375–386

Preston. J.L.. Epley. N. (2008). Science and God: an automatic opposition between ultimate explanations.

Journal of experimental social psychology. 45. 238-241.

Proulx. T.. Heine. S. J.. & Vohs. K. D. (2010). When is the unfamiliar the uncanny? Meaning affirmative after exposure to absurdist literature. humor. and art. Personality and Social Psychology Bulletin. 36.

817–829.

Proulx. T.. & Inzlicht. M. (2012). The five “A"s of meaning maintenance: Finding meaning in the theories of sense-making. Psychological Inquiry. 23(4). 317–335.

Pyszczynski. T.. Greenberg. J.. & Solomon. S. (1999). A dual-process model of defense against

conscious and unconscious death-related thoughts: An extension of terror management theory.

Psychological Review. 106. 835-845.

Randles. D.. Proulx . T.. & Heine. S. J. (2010). Turn-frogs and careful-sweaters: Non-conscious perception of incongruous word pairings provokes fluid compensation. Journal of Experimental Social Psychology. 47. 246-249

Russell. J. A.. Weiss. A.. & Mendelsohn. G. A. (1989). Affect grid: A single-item scale of pleasure and arousal. Journal of Personality and Social Psychology. 57. 493–502.

Shariff. A. F.. & Norenzayan. A. (2007). God is watching you: Priming God concepts increases prosocial behavior in an anonymous economic game. Psychological Science. 18. 803-809.

Vail. K. E.. Rothschild. Z. K.. Weise. D.. Solomon. S.. Pyszczynski. T.. & Greenberg. J. (2010). A terror management analysis of the psychological functions of religion. Personality and Social Psychology Review. 14. 84-94.

Vallacher. R. R.. & Wegner. D. M. (1989). Levels of personal agency: Individual variation in action identification. Journal of Personality & Social Psychology. 57. 660-671.

Van Tongreen. D.. & Green. J. D. (2010). Combating meaninglessness: On the automatic defense of meaning. Personality and Social Psychology Bulletin. 36. 1372-1384.

Verweij. J.. Ester. P.. & Nauta. R. (1997). Secularization as an economic and cultural phenomenon: A cross-national analysis. Journal for the Scientific Study of Religion. 36. 309–324.

Weary. G.. & Edwards. J. A. (1994). A measure of dispositional causal uncertainty. Journal of Personality and Social Psychology. 67. 308- 318.

APPENDIX A

Neutral Articles

- Fashion Millionaire interrogated in art theft case - Brits come up with new literature prize

- Dutch divided over flexible working - Eat in the heart-attack and die - Children's book new king

- Spicy soup burns hole in stomach man

- Train stationary for one-and-a-half hours in Roosendaal - Letter of the day

- Huge demand for goat milk

- The Netherlands not ready for flu pandemic - Nearly 500 pounds of cheese stolen from factory - One in four women suffers from obesity

- North sea jazz festival stays in Rotterdam - Parents not so much time for child

Threat articles

- Massive bee mortality major threat to food - Climate change threatens humanity

- Big rise in number of murders in Amsterdam - U.S. takes threat North Korea seriously - Dutch dikes unstable

- Russia wants to start World War III - Nearly a quarter Dutch youth unemployed

Affirming science articles

- New compelling evidence for Big Bang Theory - Electrical hat improves learning ability

- Cause obesity explained

- Exciting science: artificial blood ready for testing on humans - Life elixir for mice discovered

- Stem cells from teeth suitable for treating brain - Toddler with HIV cured by rapid approach

Questioning science articles

- Psychologist do not want snoopers

- Immunologist Singapore falsified and invented - New bacterium from Antartica appears a mistake - No major discovery curiosity after all

- Science fails when it comes to millions of diseased - VU professor invented more than Diederik Stapel - Can we blame the economic crisis on science