Uncertainty Engenders Preference for Status Quo
Hypothesis 1 (H1) posits that increased awareness of uncertainty leads to a greater preference for the status quo While existing empirical evidence supports this idea, it has yet to be directly tested within the theoretical framework presented here, focusing on status quo preference as the key outcome This model suggests that the desire to maintain the status quo is closely linked to deep psychological drives for managing uncertainty, implying that manipulating uncertainty salience will heighten the preference for the status quo.
Multiplicative Dispositional Hypotheses
Hypothesis 2 (H2) suggests that several moderating factors, such as individual differences in uncertainty appraisal, system-justifying attitudes, and sociopolitical indicators, will enhance the effects outlined in Hypothesis 1 Specifically, it is anticipated that higher levels of chronic uncertainty will heighten participants' sensitivity to the experimental manipulation, thereby amplifying the expected impact of the uncertainty manipulation as proposed in H1.
Individuals with high levels of chronic uncertainty may be less influenced by experimental manipulations, as their uncertainty could already be at a maximum This leads to two complementary hypotheses regarding the effects of H1 when considering moderators The primary hypothesis (H2) outlines expected outcomes, while the subordinate hypothesis (H2’s) offers an alternative explanation if evidence supporting H2 is not found.
Hypothesis 2(x) suggests that individual differences in appraisals of uncertainty enhance the effects outlined in Hypothesis 1 Specifically, it posits that individuals who consistently perceive uncertainty as a significant and negative factor will experience a stronger preference for the status quo when faced with situational uncertainty This hypothesis is part of a broader framework that includes various moderators, such as system-justifying attitudes and non-ideological sociopolitical indicators, indicating a structured approach to understanding how different conceptual categories influence behavior.
In this article, two key measures are utilized to evaluate various aspects of uncertainty appraisals: the Intolerance of Uncertainty Scale (IUS) and the Intolerance of Ambiguity Scale (IAS) The IUS, developed by Buhr and Dugas in 2002, assesses a generalized aversion to uncertainty through a four-factor structure that highlights how uncertainty can induce stress, hinder decision-making, promote avoidance of negative situations, and relate to perceptions of unfairness Examples of IUS items include statements like, “Uncertainty makes me uneasy, anxious, or stressed,” and “It’s unfair having no guarantees in life.” Similarly, the IAS focuses on self-reported attitudes towards ambiguous situations, which are defined as those lacking clear outcomes.
“adequately structured or categorized due to of insufficient cues” (Budner, 1962) Sample
An expert who avoids giving a definitive answer may lack depth in their knowledge, while those demanding a simple yes or no fail to grasp the complexity of the issues at hand.
Hypothesis 2(y) posits that system-justifying attitudes and worldviews will amplify the primary effects outlined in Hypothesis 1, forming the theoretical foundation of this dissertation To evaluate this broader construct, I have selected six measures, starting with a fundamental assessment of left-right political ideology, followed by the American System Justification Scale (SJS) developed by Kay and Jost in 2003, which is intended to gauge system justification.
The perceptions of fairness, legitimacy, and justifiability within the American social system are critical, as highlighted by Kay and Jost (2003) For instance, survey items reflect contrasting views, with some respondents believing that the political system functions properly, while others advocate for radical restructuring Additionally, Social Dominance Orientation (SDO) is assessed using the SDO-6 scale, which includes statements such as "It would be good if groups could be equal" and "It’s probably a good thing that certain groups are at the top and other groups are at the bottom." Furthermore, the Belief in a Just World (BJW) scale, developed by Lucas, Zhdanova, and Alexander (2011), measures individuals' beliefs in a fair world where people receive their deserved outcomes, featuring items like "I feel that other people generally earn the rewards and punishments they get in this world" and "Other people usually use fair procedures in dealing with others."
I am accorded.” and “I feel that people generally use methods that are fair in their evaluation of me.”
Hypothesis 2(z) proposes that non-ideological sociopolitical factors will influence the primary effects outlined in Hypothesis 1 This exploratory hypothesis group includes validated measures such as political participation and political attention, each consisting of only one or two items Although the list may seem extensive, it actually contains no more than 13 items, aligning with the typical length of a single scale.
The “interactive” approaches heretofore described under the heading
Multiplicative Dispositional Hypotheses suggest that individual differences in related psychological constructs interact with situational uncertainties in a multiplicative rather than additive way For instance, those with a strong aversion to uncertainty are expected to respond more significantly to uncertain situations.
“reactive” to situational conditions priming uncertainty, thus magnifying their pre- existing attitudes and proclivities
In summary, we propose that the effects outlined in Hypothesis 1 will be most pronounced among individuals who experience significant distress from uncertainty and ambiguity These individuals are likely to have a heightened intrinsic need to engage in cognitive processes that alleviate this distress, leading to a stronger preference for the status quo Additionally, we hypothesize that adherence to system-justifying worldviews will further amplify the impact of primed uncertainty on the preference for maintaining the status quo.
In our study, we focus on the primary variable grouping of interest, denoted as H2(y) We hypothesize that non-ideological socio-political indicators may influence the relationship between the primes and dependent variables, although predictions in this area, labeled H2(z), are exploratory and lack a solid theoretical foundation Our main prediction suggests that individuals who are more politically engaged—characterized by high levels of attention and interest—are likely to experience a magnification or polarization of their attitudes, as they have a greater personal stake and self-identification with the relevant content.
Additive Dispositional Hypotheses
Alternatively, whereas the hypotheses outlined above presume an interaction of the dispositional variables with the situationally manipulated independent variable
Hypotheses 3(x), 3(y), and 3(z) suggest that the influence of dispositional variables will interact additively, rather than multiplicatively, with the effects of uncertainty and mortality salience primes This interaction is expected to result in a preference for the status quo.
The distinction between multiplicative and additive models, though seemingly tedious, is crucial for accurately characterizing the mechanisms that generate the effects of interest The underlying mechanisms are primarily aligned with a multiplicative structure, as discussed earlier, while acknowledging that an additive model could also be a possibility Although additive hypotheses are not explicitly predicted, they provide a useful contrast to the multiplicative hypotheses presented.
The H3 and subsequent nested hypotheses align with the H2 hypotheses, predicting only simple effects from situational and dispositional independent variables (IVs) without significant interaction effects In a hierarchical linear regression analysis, significance is observed at Step 1 for main effects, but not at Step 2 for interactive effects If the additive perspective holds true, each simple effect, linked to the three families of dispositional effects and the uncertainty prime, would operate independently Consequently, the moderator would not amplify attitudinal differences caused by the prime but would instead additively influence the prediction of the dependent variables.
Hypothesis 3(x) suggests that individual differences in how people generally perceive uncertainty are directly linked to their preference for the status quo However, these differences do not influence the strength of the effects generated by the uncertainty prime.
Hypothesis 3(y) suggests that individual variations in system-justifying attitudes and worldviews are directly linked to preferences for the status quo However, these individual differences do not influence the strength of the effects generated by the uncertainty prime.
Hypothesis 3(z) suggests that individual differences in sociopolitical indicators are directly linked to preferences for the status quo, while these differences do not influence the strength of effects generated by the uncertainty prime.
The initial predictions will be evaluated for their raw effects, without the inclusion of control variables or mediation analysis Subsequently, these hypotheses will undergo testing with the four "Thoughts and Feelings Measures" incorporated into the regression model It is important to distinguish that "controls" pertain to these specific measures, while "control condition" refers to participants who have been randomly assigned to the control prime The four measures assess the extent to which participants considered or experienced specific thoughts and feelings.
“death and dying,” “anxious,” “insecure,” and “threatened.” Lastly, each of the moderational regression models was tested for mediated moderation criteria with the
In the analysis of "thoughts and feelings measures," it was determined that testing for mediated moderation is not feasible when a single variable functions both as a control and a mediator Consequently, mediational tests of moderation were conducted without the inclusion of control variables.
The inclusion of "thoughts and feelings measures" was initially intended to serve as control variables, with mediational tests being a secondary consideration We anticipate that analyses including these controls will reveal that the impact of uncertainty on status quo preference remains significant Additionally, in line with previous studies, the Positive Affect Negative Affect Scale (PANAS) will be utilized to assess any unintended effects of the salience manipulation on both positive and negative emotions (Van den Bos et al 2005, p 96) Furthermore, administering this scale after the prime can function as a time delay and a potential filler task.
H1 Uncertainty/Mortality Salience (situationally manipulated) →Status Quo
Preference H2 Multiplicative Dispositional Hypotheses Moderating factors expected to magnify the effect specified by H1
H2x Uncertainty Salience * Generalized appraisals of uncertainty → Status
Quo Preference H2y Uncertainty Salience * System justifying attitudes and worldviews
→Status Quo Preference H2z Uncertainty Salience * (non- ideological) sociopolitical indicators
H3x Uncertainty Salience + Generalized appraisals of uncertainty → Status
Quo Preference H3y Uncertainty Salience + System justifying attitudes and worldviews →
Status Quo Preference H3z Uncertainty Salience + (non- ideological) sociopolitical indicators→
Status Quo Preference Following from these hypotheses, I shall perform two experiments Experiment 1 will employ a manipulation with three conditions: uncertainty salience, mortality
The study will replicate the setup used by Van den Bos et al (2005) with a focus on salience and control group salience through the Life Event Inventory (LEI) method It will incorporate previously discussed moderating variables, alongside controls and mediation tests Following Experiment 1, a pilot test will explore alternative methods for priming the manipulated mental states, diverging from the LEI approach Experiment 2 will then implement this new method of inducing uncertainty, mortality salience, and control, aiming to test the same hypotheses as Experiment 1 while varying the independent variable's priming method.
Experiment 1 aimed to validate the core hypothesis (H1) and its variations, proposing that manipulated uncertainty leads to increased preference for the status quo This hypothesis is yet to be fully developed and tested.
The moderational and interactive hypotheses will explore how individual differences in uncertainty appraisals, system-justifying attitudes, and non-ideological social and political variables influence the relationship between our prime and various aspects of status quo preference This research builds on previous studies while introducing novel methodologies, paving the way for deeper insights into established theories like Uncertainty Management and System Justification.
This experiment involved a manipulated categorical independent variable with three levels: uncertainty salience, mortality salience, and a control group Additionally, it utilized thirteen continuous moderators as a second independent variable in subsequent models A hybrid measure was employed to assess participants' thoughts and feelings related to perceived threats.
The study utilized three distinct assessments to measure status quo preference, employing the Status Quo Preference Scale, Attitudes Towards Reformers, and Support for Regime Change Higher scores on these scales indicated a stronger preference for the status quo, more negative attitudes towards reformers, and greater support for regime change, respectively Insecurity, death, and anxiety were included as control variables in the analysis, as detailed in the Appendix.
In the first experiment, I manipulated uncertainty using a method similar to the uncertainty salience priming procedure established by Van den Bos (2001), which is grounded in previous uncertainty management research (Van den Bos, 2005, 2009) Participants were prompted to describe the emotions elicited by uncertainty and to detail their physical sensations associated with feeling uncertain This approach has been widely validated in the uncertainty management literature as an effective means of inducing state uncertainty Additionally, I incorporated mortality salience manipulation, following the methodology of Van den Bos et al (2005), who first compared the effects of uncertainty salience and mortality salience manipulations directly.
Two-Way Interaction Models (Moderator by Condition)
Interaction terms were generated by multiplying predictor variables and incorporating these products into a hierarchical linear regression model Each moderator was analyzed individually across two models, with the second model accounting for four control variables: death thoughts, anxious thoughts, insecure thoughts, and threat thoughts To ensure clarity, only regression models yielding significant or marginally significant omnibus results were further examined.
In the regression analysis, the main effects of the condition and the moderator were evaluated in step 1, while the interaction between condition and the moderator (Condition X Moderator) was assessed in step 2 Control variables were included alongside the main effects in the initial step for the "with controls" models Prior to analysis, all continuous variables underwent linear transformations to achieve a mean of 0 and a standard deviation of 0.5, rescaling them from -0.5 (representing one standard deviation below the mean) to 0.5 (one standard deviation above the mean) Consequently, the variables were transformed into Z-scores and divided by 2, ensuring that subsequent references to "low" and "high" levels of these variables are consistently coded.
Two, two-way interactions were found to be significant at the p < 05 level: (a)
MS X Political Ideology, and (b) US X Attention to Politics The abbreviations “US” and
In this article, "MS" denotes the Uncertainty Salience and Mortality Salience dummy coded variables Study 1 revealed significant regression results solely for the dependent variable of attitudes towards reformers, while the status quo preference scale and support for regime change did not produce any significant findings in experiment 1.
In this study, we employed dummy coding for all regression models to analyze the interaction of political ideology across different experimental conditions, specifically uncertainty salience and mortality salience, compared to a control condition The uncertainty salience variable (US) was assigned a value of 1 for participants exposed to the uncertainty condition, while those in the other two conditions received a score of 0 Similarly, the mortality salience variable (MS) was scored as 1 for participants in the mortality salience condition, with the other conditions marked as 0 The control condition was represented with a value of 0 for both the US and MS variables, allowing for clear comparisons across the experimental groups.
In the initial step of the analysis, both the US and MS dummy codes were included alongside all main effects and relevant controls, as well as the moderator of political ideology (PIDEO) The linear regression results indicated that there was no significant effect of the US on attitudes toward reformers compared to the control dummy code, with a coefficient of B = 0.048.
SE = 0.097, t(173) = 0.636, p = 526), with all controls); nor for MS (B = 0.007, β 0.007, SE = 0.088, t(177) = 0.082, p = 935), without controls (see Table 6); (B = -0.096, β = -0.089, SE = 0.123, t(173) = -0.783, p = 435), with all controls (see Table 7)
Table 6 Hierarchical Multiple Linear Regression Analysis for Attitude Towards
Reformers Regressed Onto Condition and the Moderator of Political Ideology, Without Controls, Study 1
Uncertainty salience (US) 048 087 045 042 087 039 Mortality salience (MS) 007 088 007 -.004 088 -.004 Political ideology (PIDEO) -.327*** 072 -.324 -.186 118 -.184
Note Referent group for Condition = Control
The analysis demonstrated a significant main effect of political ideology on attitudes toward reformers, with leftists and liberals showing more positive attitudes compared to conservatives, as indicated by a regression coefficient of B = -0.327 (β = -0.324, SE = 0.072, t(177) = -4.527, p < 001) This effect persisted even after controlling for additional variables, with a coefficient of B = -0.326 (β = -0.322, SE = 0.071, t(173) = -4.591, p < 001), confirming established findings in existing literature.
In Step 2 of the analysis, dummy code x moderator interaction terms were included, revealing a marginally significant interaction between MS and political ideology in predicting attitudes toward reformers (B = -0.308, β = -0.174, SE = 0.173, t(175) = -1.787, p = 076) This interaction became significant when all controls were added to the model (B = -0.360, β = -0.203, SE = 0.169, t(171) = -2.131, p = 035) Conversely, the interaction between US and political ideology did not produce significant results, both without controls (B = -0.138, β = -0.075, SE = 0.177, t(175) = -0.782, p = 436) and with controls (B = -0.156, β = -0.084, SE = 0.175, t(171) = -0.888, p = 376) For detailed results, refer to Tables 6 and 7.
Table 7 Hierarchical Multiple Linear Regression Analysis for Attitude Towards
Reformers Regressed Onto Condition, the Moderator of Political Ideology, With
Uncertainty salience (US) 062 097 058 055 097 052 Mortality salience (MS) -.096 123 -.089 -.134 123 -.124 Political ideology (PIDEO) -.326*** 071 -.322 -.164 114 -.163 Death thoughts (DT) 147 120 145 185 121 183 Anxious thoughts (AT) -.275* 111 -.272 -.282* 111 -.279 Insecure thoughts (IT) -.006 127 -.006 -.006 127 -.006 Threat thoughts (TT) 272* 121 266 262* 121 255
Note Referent group for Condition = Control
Simple slopes analyses Without controls, simple slopes analyses revealed that for
Conservatives, those primed with mortality salience exhibited more negative attitudes towards reformers, though these effects did not reach the threshold of statistical
75 significance (B = 0.150, β = 0.139, SE = 0.119, t(175) = 1.262, p = 209) For Liberals, those primed with mortality salience exhibited more positive attitudes towards reformers, likewise falling short of statistical significance (B = -0.159, β = -0.147, SE = 0.128, t(175)
With controls in the model, however, significant effects did emerge For
Conservatives, the mortality salience prime had no effect (B = 0.046, β = 0.043 SE 0.139, t(171) = 0.333, p = 739) For Liberals, however, the mortality salience prime caused an increase in favorable attitudes towards reformers (B = -0.314, β = -0.290, SE 0.159, t (171) = -1.973, p = 05 exactly) See Figure 1
The estimated marginal means for attitudes towards reformers were analyzed, revealing significant variations influenced by political ideology, as shown in Figure 1 The results include error bars that represent ± 1 standard error from the mean, highlighting the precision of the findings from Study 1.
Mediated moderation analyses were conducted to explore whether thoughts and feelings measures mediate the significant two-way interaction identified These analyses are exploratory and not linked to specific theories or hypotheses central to the main theoretical claims The absence of support for mediated moderation does not impact the evaluation of previous hypotheses Throughout the analyses, there was a complete lack of evidence for mediated moderation models Since thoughts and feelings measures cannot function as both control and mediator simultaneously, the mediated moderation analyses were performed without controls To suggest mediated moderation through multiple regression, the potential mediator must be correlated with the dependent variable, as outlined by Bucy & Tao (2007), Muller et al (2005), Hayes (2009), and Edwards & Lambert.
In a study conducted in 2007, it was found that among four measures of thoughts and feelings, only "threat" thoughts showed a significant correlation with attitudes towards reformers (r(193) = 184, p = 01), indicating that threat will be tested as a mediator The other three potential mediators—anxiety (r(193) = -.03, p = 68) and death—did not demonstrate any correlation with the dependent variable in the regression models.
The analysis indicates that the criteria for mediated moderation are not met, as evidenced by the correlation results: secure attachment (r(193) = 111, p = 13) and insecure attachment (r(193) = 070, p = 33) To achieve the "initial criteria" outlined by Muller et al (2005) and Bucy & Tao (2007), it is essential that three specific p values from the relevant equations reach statistical significance.
Y = β 10 + βUS 11 + βMS 12 + βB 13 + βUSxB 14 + βMSxB 15 + ε 1 p MSxB = 076 (1)
M = β 20 + βUS 21 + βMS 22 + βB 23 + βUSxB 24 + βMSxB 25 + ε 2 p MSxB = 957 (2)
Y = β 30 + βUS 31 + βMS 32 + βB 33 + βUSxB 34 + βMSxB 35 + βM + βMB + ε 3 p MsxB = 007 (3)
The study examines the interaction between mortality salience (MS) and political ideology (B) as it relates to attitudes towards reformers (Y) The analysis reveals that the interaction of MS and B does not achieve statistical significance, with a coefficient of t(175) = 0.05 and p = 957 Consequently, the initial criteria for mediated moderation, which rely on three significance values, are not satisfied in this instance.
In the context of mediated moderation, if the initial criteria are satisfied with all three p-values showing statistical significance, we then assess the secondary criteria This requires that the coefficient of βMSxB from the first equation must decrease in magnitude when additional factors from the third equation are included Specifically, βMSxB35 should be less than βMSxB15 However, in this case, the coefficient for the interaction between MS and Political Ideology actually increased, with βMSxB15 at -0.308 and βMSxB35 at -0.484, thereby ruling out the possibility of mediated moderation.
In the analysis of political attention, the initial step involved entering the US and MS dummy codes along with all main effects and relevant controls, alongside the moderator of political attention (ATTN) The linear regression results indicated no significant effect of the US on attitudes toward reformers, with values showing B = -0.002 (p = 983) without controls and B = 0.003 (p = 974) with all controls Similarly, for MS, no significant effect was found, as evidenced by B = 0.020 (p = 824) without controls and B = -0.058 (p = 641) with all controls.
Table 8 Hierarchical Multiple Linear Regression Analysis for Attitude Towards Reform Regressed Onto Condition and the Moderator of Attention to Politics, Without Controls, Study 1
Uncertainty salience (US) -.002 088 -.002 -.002 088 -.002 Mortality salience (MS) 020 089 019 026 089 025 Political attention (ATTN) 079 073 079 -.090 134 -.090 +
Note Referent group for Condition = Control
The analysis in Step 1 indicated that political attention did not have a significant impact on attitudes toward reformers, both in the model without controls (B = 0.079, β = 0.079, SE = 0.073, t(189) = 1.094, p = 276) and in the model with controls (B = 0.059, β = 0.059, SE = 0.071, t(185) = 0.828, p = 408).
Table 9 Hierarchical Multiple Linear Regression Analysis for Attitude Towards Reform Regressed Onto Condition, the Moderator of Attention to Politics, With Controls, Study 1
Uncertainty salience (US) 003 098 003 -.009 098 -.008 Mortality salience (MS) -.058 123 -.054 -.066 123 -.062 Political attention (ATTN) 059 071 059 -.130 132 -.130 Death thoughts (DT) 083 118 083 094 119 094 Anxious thoughts (AT) -.276* 114 -.276 -.274* 113 -.274 Insecure thoughts (IT) 003 131 003 043 135 043 Threat thoughts (TT) 328** 120 328 298* 127 298
Note Referent group for Condition = Control
Step 2 Each of the dummy code x moderator interaction terms were entered in step 2 Regression analyses revealed a significant interaction between US and political attention predicting attitudes toward reformers, without controls (B = 0.373, β = 0.206,
SE = 0.187, t(187) = 1.991, p = 048), which became marginally significant when all controls were entered into the model(B = 0.339, β = 0.188, SE = 0.187, t(183) = 1.808, p
= 072) The interaction between MS and political attention predicting attitudes towards reformers did not yield significant results without controls (B = 0.135, β = 0.086, SE 0.176, t(187) = 0.768, p = 444), or with controls (B = 0.208, β = 0.132, SE = 0.175, t(183) = 1.190, p = 236) See Tables 8 and 9
Analysis Summary
The Manipulation Check and Thoughts/Feelings measures demonstrate that the primes were effective in eliciting the desired affective response in each condition
The anticipated primary effects of the prime were not observed; however, two significant regression models emerged The first model, which examined the interaction between mortality salience (MS) and political ideology, aligned with the original hypothesis H2(y), revealing that under mortality salience, conservatives exhibited a stronger aversion to reformers while liberals showed an increased preference for them Conversely, the interaction between US attention and attitudes towards reformers in uncertainty conditions did not support any of the initial hypotheses Regression analyses were conducted both with and without controls, as detailed in the respective tables throughout the study.
To mitigate mono-operation bias and enhance the robustness of theoretical claims, Experiment 2 aimed to replicate Experiment 1 using an alternative priming method A pilot study was conducted to evaluate two new priming techniques: list recall and word search The pilot assessed various uncertainty priming materials, including list recall uncertainty, word search uncertainty, life event inventory uncertainty (used in Study 1), and LEI control (also from Study 1), based on three manipulation check item ratings The most effective priming method would be chosen for Experiment 2, ensuring that participants in the uncertainty condition experience greater uncertainty salience compared to mortality salience and control topics, as well as the other uncertainty priming methods tested The list recall method involved participants memorizing and recalling five words associated with uncertainty: uncertain, shaky, gamble, dicey, and wavering.
84 consisted of asking the participant to locate these same words hidden in a word search puzzle
A study involving 82 American participants was conducted via Amazon Mechanical Turk, utilizing three one-way analyses of variance (ANOVA) to assess differences across four conditions in relation to three dependent variables Significant results were found for uncertainty salience, with F(3, 78) = 14.173, p < 001, η² = 353, warranting further pairwise comparisons Post-hoc analysis using Tukey’s HSD revealed that the word search uncertainty condition (M = 5.273, SD = 1.549) significantly exceeded the control condition (M = 2.952, SD = 2.334, p < 001) Additionally, the LEI uncertainty condition (M = 6.400, SD = 0.940) also showed a significantly higher mean than the control condition (M = 2.952).
The study revealed a significant difference in mean ratings between the LEI uncertainty condition and the word search uncertainty condition, with a p-value of less than 001 The standard deviation was calculated at 2.334, indicating a notable variation in responses No other significant group differences were observed.
Table 10 ANOVA of Manipulation Check by Condition, Pilot Study
List recall Word search Control Life event inventory Dependent variable M SD M SD M SD M SD F p η 2 Tukey’s HSD
Note Referent group for Condition = Control
Table 11 Post Hoc Comparisons from Tukey’s Honestly Significant Difference Test, Pilot Study
List recall Life event inventory -2.189** 0.568 001
Word search Life event inventory -1.127 0.548 176
Life event inventory List recall 2.189** 0.568 001
Life event inventory Word search 1.127 0.548 176
List recall Life event inventory 0.916 0.684 541
Word search Life event inventory -0.355 0.659 950
Life event inventory List recall -0.916 0.684 541
Life event inventory Word search 0.355 0.659 950
List recall Life event inventory -0.192 0.497 980
Word search Life event inventory -0.395 0.479 843
Life event inventory List recall 0.192 0.497 980
Life event inventory Word search 0.395 0.479 843
The findings indicate that both the LEI and Word Search manipulations significantly increased uncertainty compared to the control condition, with the LEI showing a stronger effect While both methods primed uncertainty thoughts more effectively than the control, the Word Search was less effective than the LEI in eliciting uncertainty salience Consequently, for experiment 2, only the Word Search prime remained viable, as the pilot study revealed that none of the uncertainty primes influenced participants’ mortality salience or control topic salience ratings.
In experiment 2, the theoretical rationale and hypotheses remained consistent with experiment 1, with the only change being the transition from LEI to a Word Search format The methods, procedures, and analysis protocols were identical across both studies Participants engaged with a unique puzzle for each of the three conditions: the uncertainty salience condition featured words such as uncertain, shaky, gamble, dicey, and wavering; the mortality salience condition included death, mortal, grave, tombs, and demise; while the control condition presented words like television, dust, goggles, afternoon, and spot After being randomly assigned to one of the three conditions, participants were tasked with locating all hidden words within a two-minute time limit before proceeding to a follow-up questionnaire.
The new manipulation method, known as word search prime, effectively addresses specific validity threats associated with thought generation in the LEI uncertainty-prime context Experiment 1 revealed that recalling uncertain moments can indirectly trigger mortality salience This poses a potential issue, as participants in the uncertainty condition may evoke memories related to death, linking uncertainty with heightened mortality awareness.
The priming method used in experiment 2 addresses the limitations of the life event inventory (LEI) by eliminating the confounding factors associated with generated memories While experiment 1 focused on recalling autobiographical memories to evoke specific emotions, experiment 2 utilized implicit concept activation through a word search prime, highlighting the differences in psychological effects This distinction may arise from the self-referencing effect linked to autobiographical information generation, which can subtly impact the relationship between the independent variable construct of uncertainty and its operationalization Therefore, replicating the design with the word search manipulation mitigates the risks of mono-operation bias The study involved 197 American participants recruited via Amazon Mechanical Turk.
A significant difference was found in the self-report ratings of mortality salience among three experimental conditions, with an F-value of 9.196 (p < 001, η² = 087) Post-hoc analysis using the Tukey HSD test revealed that participants in the mortality salience condition (M = 2.861, SD = 1.878) reported significantly higher ratings compared to those in the uncertainty condition (M = 1.776, SD = 1.346).
< 001) The mortality salience condition was also significantly greater than the control topic salience condition (M = 1.877, SD = 1.536; p = 002)
The manipulation check revealed no significant differences in uncertainty salience and control topic salience across conditions For group comparisons, refer to the “Uncertainty salience check,” “Control topic salience check,” and “Mortality salience check” in Tables 12 and 13 Unlike the pilot study, the uncertainty puzzle prime did not lead to increased self-reported thoughts about uncertainty when compared to the control group.
In comparing the manipulation check item analyses of Experiment 1 and Experiment 2, significant differences emerge Experiment 1 revealed that both LEI primed uncertainty and LEI primed mortality salience resulted in higher ratings for their respective manipulation check items compared to the control condition and each other Notably, LEI primed uncertainty led to an increase in mortality salience ratings, although not as pronounced as those in the mortality salience condition Conversely, LEI primed mortality salience also elevated uncertainty salience ratings, yet again, not to the same extent as observed in the uncertainty salience condition.
Table 12 ANOVA of Manipulation Check and Thoughts/Feelings Variables by Condition, Study 2
Between groups effect M SD M SD M SD F p η 2 Tukey’s
Table 13 Post Hoc Comparisons from Tukey’s Honestly Significant Difference Test, Study 2
In experiment 2, the word search primed mortality salience resulted in significantly higher ratings on mortality salience manipulation check items compared to the other two conditions However, it did not lead to increased uncertainty salience, unlike experiment 1 with LEI Additionally, the word search primed uncertainty condition did not enhance mortality salience levels This suggests that the potential confound observed in experiment 1, where uncertainty priming inadvertently influenced mortality salience, was not present in experiment 2 Interestingly, the uncertainty priming condition in experiment 2 did not yield increased self-reported uncertainty salience, as indicated by the manipulation check item ratings.
Despite the expectation that manipulation check results from the pilot test would replicate in experiment 2, they did not This discrepancy may stem from the timing of the PANAS and dependent variable measures, which were administered after the prime in experiments 1 and 2, unlike the pilot where participants answered manipulation checks immediately after exposure Consequently, the delay and potential conceptual interference in experiment 2 could have diminished the effects observed in the pilot test regarding the uncertainty crossword prime.
In Experiment 2, the prime may not have activated the necessary cognitive-emotional constructs to influence the dependent variables Alternatively, participants engaged in the word search might have been less aware of the prime's impact on their thoughts and feelings, suggesting that the prime could still affect them subconsciously This latter scenario appears plausible, as several significant interactions in the regression models of Experiment 2 occurred despite weak outcomes on the uncertainty salience manipulation check, echoing findings from Schwarz & Clore's studies.
Research indicates that participants reported a more positive mood on sunny days, but this effect diminished when they became aware of the weather's potential influence on their feelings Similarly, in a study, the introduction of the PANAS scale between the prime and manipulation check may have obscured participants' awareness of the prime's effects on their mental state, allowing it to impact dependent variables without showing differences in manipulation check items This suggests that overt primes are often consciously disregarded, while subtler primes can influence participants without detection Filler tasks are frequently used in survey research to divert participants' attention from the prime's effects.
2011) for a discussion of this and similar effects in priming and automaticity research
Two-Way Interaction Models (Moderator by Condition)
Interaction terms were generated by multiplying variables and incorporating the products as predictor variables in a hierarchical linear regression model Consistent with experiment 1, each moderator was individually tested for each dependent variable across two models, with the second model incorporating four control variables: (a) death thoughts, (b) anxious thoughts, (c) insecure thoughts, and (d) threat thoughts The coding procedures followed in this experiment were identical to those used in experiment 1.
Significant two-way interactions influencing the status quo preference scale were identified, specifically between the United States and political ideology, as well as between the United States and party identification Additionally, notable two-way interactions affecting attitudes towards reformers were observed, particularly between Mississippi and political ideology.
MS X Social Dominance Orientation, (c) MS X System Justification, (d) US X
Intolerance of Uncertainty, (e) US X Political Self-Efficacy, and (f) US X Trust in
Government Two-way interactions significant for the dependent variable of regime change included: (a) US X Belief in a Just World, and (b) US X Trust in Government
Table 15 ANOVA of Three Dependent Variables by Condition, Study 2
Between groups effect M SD M SD M SD F p η 2 Tukey’s
Interactions: Predicting Status Quo Preference
In the analysis of political ideology, the initial step involved entering the US and MS dummy codes, along with all main effects and relevant controls, alongside the moderator of political ideology (PIDEO) The linear regression results indicated no significant effect of the US on status quo preference, with coefficients showing B = 0.085 (β = 0.081, SE = 0.088, t(180) = 0.967, p = 335) without controls and B = 0.095 (β = 0.091, SE = 0.079, t(176) = 1.201, p = 231) with controls Similarly, no significant effect was found for MS, with B = 0.048 (β = 0.046, SE = 0.088, t(180) = 0.552, p = 581) without controls and B = 0.001 (β = 0.001, SE = 0.080, t(176) = 0.012, p = 991) with all controls.
Table 16 Hierarchical Multiple Linear Regression Analysis for Status Quo Preference Regressed Onto Condition and the Moderator of Political Ideology, Without Controls, Study 2
Uncertainty salience (US) 085 088 081 082 087 079 Mortality salience (MS) 048 088 046 057 087 055 Political ideology
Note Referent group for Condition = Control
Table 17 Hierarchical Multiple Linear Regression Analysis for Status Quo Preference Regressed Onto Condition, the Moderator of Political Ideology, With Controls, Study 2
Uncertainty salience (US) 095 079 091 091 077 087 Mortality salience (MS) 001 080 001 013 079 012 Political ideology (PIDEO) -.200** 065 -.204 021 117 021 Death thoughts (DT) 017 098 018 -.009 096 -.009 Anxious thoughts (AT) 208* 094 210 204* 093 206 Insecure thoughts (IT) -.179 117 -.180 -.196+ 115 -.198 Threat thoughts (TT) 435*** 121 445 467*** 120 477
Note Referent group for Condition = Control
Step 1 revealed a marginally significant main effect of political ideology predicting status quo preference in the model without controls (B = -0.142, β = -0.145, SE
The study revealed that leftist/liberals showed a slightly greater preference for the status quo than anticipated, as indicated by the findings (B = -0.200, β = -0.204, SE = 0.065, t(176) = -3.067, p = 003) after controlling for variables Notably, the control variable of anxious thoughts was also statistically significant in this analysis.
In the control model, a significant relationship was found between anxious thoughts and a preference for the status quo (B = 0.208, β = 0.210, SE = 0.094, t(176) = 2.203, p = 029), indicating that higher levels of anxiety are linked to a stronger inclination to maintain existing conditions Additionally, the control variable of threat thoughts also demonstrated statistical significance in the initial step of the analysis.
103 model (B = 0.435, β = 0.445, SE = 0.121, t(176) = 3.588, p < 001), such that increases in threat thoughts were associated with a greater status quo preference
The estimated marginal means for the dependent variable of status quo preference are presented in Figure 3, highlighting the influence of political ideology while controlling for other factors The error bars indicate a range of ± 1 standard error from the mean, based on the findings from Study 2.
Regression analyses indicated a significant interaction between US and political ideology in predicting status quo preference, both without controls (B = -0.469, p = 007) and with controls (B = -0.435, p = 006) Conversely, the interaction between MS and political ideology did not produce significant results, either without controls (B = -0.243, p = 186) or with controls (B = -0.154, p = 347) Additionally, threat thoughts emerged as a statistically significant control variable in the model (B = 0.467, p < 001), as did anxious thoughts.
Simple slopes analyses indicated that among Conservatives, individuals primed with uncertainty salience demonstrated a stronger preference for the status quo (B = 0.317, β = 0.303), highlighting the significant impact of political ideology on decision-making in uncertain contexts.
SE = 0.122, t(178) = 2.595, p = 010) With controls in the model, the remained statistically significant (B = 0.308 β = 0.295, SE = 0.109, t(174) = 2.825, p = 005)
Statistical significance was not achieved for Liberals primed with uncertainty salience without controls (B = -0.152, β = -0.146, SE = 0.123, t(178) = -1.234, p = 219), or with controls (B = -0.126, β = -0.121, SE = 0.110, t(174) = -1.150, p = 252) See Figure 3
Mediated moderation procedures are consistent with the methodologies outlined by Bucy & Tao (2007) and Muller et al (2005), applied only when basic criteria among variables are satisfied Although empirical support for mediated moderation models is currently absent, this does not impact the assessment of the proposed hypotheses The importance of the equations presented by Bucy & Tao (2007) and Muller et al (2005) remains significant in this context.
Y = β 10 + βUS 11 + βMS 12 + βB 13 + βUSxB 14 + βMSxB 15 + ε 1 p USxB = 007** (1)
M = β 20 + βUS 21 + βMS 22 + βB 23 + βUSxB 24 + βMSxB 25 + ε 2 p USxB = 723 (2)
Y = β 30 + βUS 31 + βMS 32 + βB 33 + βUSxB 34 + βMSxB 35 + βM + βMB + ε 3 p USxB = 0071** (3)
In this study, "US" represents the independent variable of uncertainty salience, while "MS" signifies the independent variable of mortality salience The variable "B" acts as a moderator, reflecting political ideology, and "M" serves as the mediator of threat Finally, "Y" denotes the dependent variable, which is the preference for the status quo The focus is on the interaction between uncertainty salience and political ideology.
In our analysis of the USxB interactions, we found that the coefficient of the USxB interaction in the second equation was not statistically significant (t(178) = -0.36, p β USXB = 723) This indicates that the initial criteria for mediated moderation are not met, preventing us from exploring this avenue further Additionally, the relevant coefficient did not decrease from equation 1 to equation 3, which further disqualifies the possibility of mediated moderation.
In the analysis of party identification, the initial step involved entering the US and MS dummy codes, along with all main effects and applicable controls, alongside the moderator of party identification (PPARTY) The linear regression results indicated that there was no significant effect of the US on status quo preference, with values showing B = 0.064 (β = 0.065, SE = 0.083, t(164) = 0.766, p = 445) without controls, and B = 0.078 (β = 0.079, SE = 0.078, t(160) = 0.995, p = 321) with all controls Similarly, the analysis for MS showed no significant effect, with B = -0.001 (β = -0.001, SE = 0.085, t(164) = -0.012, p = 991) without controls, and B = -0.021 (β = -0.021, SE = 0.080, t(160) = -0.257, p = 797) with all controls, as detailed in Tables 18 and 19.
Table 18 Hierarchical Multiple Linear Regression Analysis for Status Quo Preference Regressed Onto Condition and the Moderator of Party Identification, Without Controls, Study 2
Uncertainty salience (US) 064 083 065 068 083 070 Mortality salience (MS) -.001 085 -.001 018 085 018 Party identification
Note Referent group for Condition = Control
Table 19 Hierarchical Multiple Linear Regression Analysis for Status Quo Preference Regressed Onto Condition, the Moderator of Party Identification, With Controls, Study 2
Uncertainty salience (US) 078 078 079 082 078 084 Mortality salience (MS) -.021 080 -.021 -.002 080 -.002 Party identification (PPARTY) -.155* 068 -.169 -.010 117 -.011 Death thoughts (DT) -.011 105 -.011 -018 104 -.018 Anxious thoughts (AT) 214* 094 227 216* 093 229 Insecure thoughts (IT) -.187 117 -.196 -.196+ 117 -.205 Threat thoughts (TT) 353** 122 382 356** 122 385
Note Referent group for Condition = Control
107 Step 1 revealed a significant main effect of party identification predicting status quo preference in the model without controls (B = -0.244, β = -0.265, SE = 0.070, t(164)
A study found that Democrats displayed less negative attitudes towards the status quo compared to Republicans, with significant statistical results (B = -3.495, p = 001) Even after controlling for various factors, this difference remained significant (B = -0.155, β = -0.169, SE = 0.068, t(160) = -2.275, p = 024) Additionally, anxious thoughts were positively correlated with a preference for the status quo (B = 0.214, β = 0.227, SE = 0.094, t(160) = 2.285, p = 024), indicating that higher levels of anxiety were associated with a stronger inclination to maintain existing conditions Similarly, threat-related thoughts also showed significant effects in the initial control model.
= 0.353, β = 0.382, SE = 0.122, t(160) = 2.887, p = 004), such that increases in threat thoughts were associated with a greater status quo preference
Regression analyses indicated a significant interaction between US and party identification in predicting status quo preference, both without controls (B = -0.378, p = 033) and with controls (B = -0.340, p = 039) Conversely, the interaction between MS and party identification did not show significant results, either without (B = -0.131, p = 438) or with controls (B = -0.109, p = 489) Additionally, the control variable of threat thoughts was statistically significant in the model (B = 0.356, p = 004), as was anxious thoughts (B = 0.216, p = 022) For detailed results, refer to Tables 18 and 19.
Simple slopes analyses indicated that Republicans primed with uncertainty salience showed a significantly higher preference for the status quo, with a coefficient of B = 0.257 (β = 0.263, SE = 0.124, t(162) = 2.082, p = 039) when no controls were applied This effect remained statistically significant in the model with controls (B = 0.252, β = 0.257, SE = 0.115, t(158) = 2.186, p = 030) In contrast, Democrats did not show significant results for uncertainty salience, both without controls (B = -0.120, β = -0.123, SE = 0.118, t(162) = -1.019, p = 310) and with controls (B = -0.088, β = -0.090, SE = 0.110, t(158) = -0.801, p = 424).
Hypothesis 1 (H1) suggested that increased salience of uncertainty would lead participants to favor the status quo and similar concepts However, the anticipated main effects of uncertainty and mortality salience on preferences for the status quo, attitudes towards reformers, and support for regime change were not observed in either study.
In the following discussion, the terms "higher" and "more" refer to the significant effects observed in the primed condition compared to the control Hypothesis 2 (H2) is divided into three parts, with the core premise suggesting that increased affective experiences, triggered by US and MS primes (primarily anticipated for US), interact with selected moderating factors This interaction is expected to lead participants to show a greater preference for the status quo and related constructs, resulting in multiplicative interaction effects The three subsets of Hypothesis 2 (x, y, z) illustrate the distinct effects anticipated for various conceptual categories of moderators employed in this study, with Hypothesis 2(x) focusing on individual differences.
143 uncertainty sensitivity, H2(y) denoting differences in system justifying attitudes and worldviews, and H2(z) denoting differences in non-ideological sociopolitical factors
In relation to Hypothesis 2(x), only Study 2 revealed a statistically significant effect, showing a notable interaction between the US and the Intolerance of Uncertainty Scale (IUS), although this effect became marginally significant when controls were applied Specifically, participants with high intolerance of uncertainty displayed increased positive attitudes towards reformers in uncertain conditions, a finding not observed in Study 1 and not initially predicted, thus leading to post-hoc speculation Our theoretical framework suggests that those intolerant of uncertainty should typically show negative attitudes towards reformers Further investigation is necessary to assess the replicability of this effect and determine if it is a statistical anomaly Ultimately, Hypothesis 2(x) lacked supporting evidence.
Hypothesis 2(y) suggests that attitudes and worldviews related to system justification will influence the effects of Hypothesis 1 Specifically, individuals with strong system-justifying worldviews are expected to experience a magnification or polarization of their baseline attitudes when primed, known as the “extremity effect.” This differs from a “conservative shift,” which implies a uniform push of attitudinal preferences in a single direction.
The analysis reveals significant effects related to Hypothesis 2(y), particularly highlighting the role of political ideology (PIDEO) as a moderator influencing attitudes towards reformers Across multiple studies, a consistent interaction was observed, with liberals demonstrating more positive attitudes towards reformers under mortality salience conditions Although the anticipated effect was expected to be stronger in the uncertainty condition, it was still present, albeit not statistically significant in both experiments The significant interaction between mortality salience and political ideology was primarily driven by leftist/liberals, while a non-significant trend was noted for right-wing/conservatives In the second experiment, while none of the simple slopes achieved statistical significance, the pattern mirrored that of the first experiment, reinforcing the reliability of the findings across both studies.
In both studies, similar patterns were observed in variables related to attitudes and worldviews, although they did not replicate consistently Specifically, in Experiment 2, social dominance orientation (SDO) demonstrated a comparable polarization effect under conditions of mortality salience.
The study revealed that the baseline effect was intensified for individuals with both high and low Social Dominance Orientation (SDO) When examining the American System Justification Scale (SJS) as a moderator in the regression analysis, the anticipated pattern from Hypothesis H2(y) was confirmed Specifically, individuals with high SJS exhibited a greater aversion to reformers compared to those with low SJS, both in general and within the control condition.
The study revealed polarized effects related to MS, primarily influenced by individuals high in SJS Furthermore, in study 2, a significant relationship was found between the Belief in a Just World (BJW) scale and support for regime change When BJW was introduced as a moderator in the uncertainty condition, participants with low BJW showed increased favorable attitudes towards regime change, while those with high BJW exhibited the opposite response.
In experiment 2, a notable regression concerning the Status Quo Preference Scale was observed, with political ideology acting as a moderator This regression revealed a US x Moderator interaction, indicating that under conditions of uncertainty, pre-existing attitudes polarized as predicted, with right-wing conservatives displaying heightened negativity towards reformers, while left-wing conservatives showed reduced negativity Further analysis of this interaction confirmed that the significant effects were predominantly influenced by conservative participants Additionally, the variable of Party Identification mirrored this pattern, reinforcing the observed effects.
Participants identifying as Republicans demonstrate clear polarization, reflecting significant effects that align with the initial predictions outlined in hypothesis H2(y).
Hypothesis 2(z) suggested that non-ideological sociopolitical indicators would moderate the effects of H1, with individuals who feel more politically involved and efficacious experiencing heightened polarization of their baseline responses This hypothesis was more exploratory and less central compared to H2(x, y) The following discussion will focus on the relevant effects observed within this hypothesis grouping.
In Study 1, a marginally significant interaction between U.S context and political attention was observed, indicating that individuals with high political attention displayed more negative attitudes toward reformers when faced with uncertainty However, this effect was not anticipated by Hypothesis 2(z), did not achieve statistical significance at the p < 05 level during the analysis of simple slopes, and was not replicated in subsequent studies Therefore, we regard this finding as potentially spurious or outside the theoretical framework of this dissertation.
In Study 2, an interaction between political self-efficacy (PSE) and uncertainty was observed among US participants Specifically, individuals with high political self-efficacy demonstrated significantly more positive attitudes towards reformers when faced with uncertainty.
In Study 2, an unexpected effect was observed regarding Trust in Government (TG) in the U.S.; participants with high levels of government trust showed increased support for reformers when primed with uncertainty This finding contradicts the initial hypothesis, which suggested that lower trust would correlate with more favorable attitudes towards reformers.
A significant finding emerged regarding the relationship between Trust in Government (TG) and support for regime change, indicating that individuals with high trust in government showed reduced support for regime change when faced with uncertainty This aligns with hypothesis 2(z) However, it's important to note that Trust in Government was unevenly distributed across conditions, leading to a correlation with the independent variable and violating assumptions of interactive regression models Consequently, any conclusions drawn from this moderator should be approached with caution, as empirical support for the hypothesis 2(z) grouping remains limited.
Hypothesis 3 consists of three parts (x, y, z) related to specific moderator variables It suggests that increased uncertainty salience, along with thirteen individual difference moderators, will lead participants to favor the status quo in an additive manner However, the studies revealed no main effects from the manipulated independent variable, and the anticipated multiplicative effects were not supported, indicating that Hypothesis 3 lacks empirical backing.
The absence of main effects in both studies indicates a rejection of hypothesis 1, even in the absence of moderator interactions This finding is theoretically significant for several reasons, particularly within the context of uncertainty management literature (Hogg, 2000).
Research from 2004 to 2005, alongside terror management literature (Greenberg et al., 2008), utilized similar independent variable manipulations and consistently demonstrated significant effects on dependent variables related to cultural worldviews, norms, and values, supporting Hypothesis 1 presented in this study.
The unexpected lack of replication of established effects raises questions, especially since an identical manipulation (the LEI in experiment 1) did not yield significant results This situation may be viewed as a "fail to replicate" scenario However, key differences between the current studies and previous research could account for this non-replication, particularly the absence of status quo preference measures in earlier studies Additionally, the shift from traditional university psychology laboratories to online surveys may have influenced outcomes, highlighting the need for further research on the implications of survey methods Recent reviews suggest that understanding differences in participant pools is a valuable avenue for future exploration.
Research indicates that in-person college samples and Mturk participants exhibit similar behaviors, suggesting that differences in participant characteristics are not the main reason for the absence of significant effects in these experiments While the use of Mturk introduces a novel method for participant recruitment, there may still be undetected systematic differences between these populations that warrant further investigation.
Our study focused on dependent variables that reflect preference for the status quo and its manifestations, such as attitudes towards reformers and support for regime change, rather than cultural worldviews or norms While we anticipated that findings related to cultural perspectives would influence status quo preferences due to their conceptual overlap, this was not guaranteed Cultural norms and worldviews embody the status quo, suggesting potential similarities in effects However, it's important to note that a worldview or norm differs from an attitude, as evidenced by participants, particularly those identifying as right-leaning or Republican during the Obama administration, who may perceive the political status quo as conflicting with their personal worldviews.
Recent discussions within the social science community present a strong case for requiring multiple replications of an effect prior to making bold claims of causality
Recent replication attempts have revealed that previously well-established priming effects may not produce consistent results, indicating that subtle procedural differences could significantly impact outcomes This inconsistency raises concerns about the reliability of past findings and calls into question long-standing theoretical conclusions in the field (Pashler & Wagenmakers, 2012; Stroebe & Strack).
To address the issue, it is essential to conduct both repeated exact procedural replications and varied conceptual replications The consistency of effects related to political ideology across two studies alleviates concerns, as the predicted theoretical effect was observed in the same manner in different implementations of the independent variable manipulation Furthermore, similar interaction effects were noted among various variables within the attitudes and worldviews category, indicating that the effects relevant to hypothesis H2(y) were not mere coincidences, especially concerning the significant variable of political ideology and its related factors.
The experimentwise type I error is a significant concern, as the analysis involves 39 comparisons due to 13 Moderator Variables and 3 Dependent Variables, considering all controls This high number of statistical comparisons increases the risk of type I errors, with a calculated likelihood of approximately 86% for at least one false positive in each study, as per Shaffer's (1995) formula: 1 - (1 - 05)^(39) Consequently, the substantial volume of statistical models utilized raises the probability of erroneous findings in the results.
The attention effect observed in the study was unpredictable and inconsistent with established theories, leading to the conclusion that it is likely spurious In study 2, while some interaction effects related to H2(z) were noted, they were limited, and the significant US x TG effect on attitudes toward reformers contradicted H2(z) Although the US x TG interaction did support H2(z) regarding support for regime change, this was the only evidence favoring the hypothesis Due to the uneven distribution and random assignment of trust in government across conditions in experiment 2, any conclusions drawn should be approached with caution.
H2(y) has received support from various studies regarding political ideology as a moderator, particularly in Experiment 2, which examined conceptually related variables within system-justifying attitudes and worldviews Caution is advised when drawing conclusions from Study 2, as its results do not replicate those of Study 1 However, the consistent pattern observed among related moderator variables suggests the robustness of these effects If these effects were merely due to experimentwise type I error, it would be improbable for such a similar pattern to occur repeatedly in alignment with the a-priori predictions of H2(y) Therefore, individuals with strong system-justifying attitudes and worldviews exhibit polarized pre-existing attitudes when exposed to mortality salience or uncertainty salience This polarization is most pronounced regarding political ideology, particularly in attitudes towards reformers, as consistently replicated across experiments.
Terror Managements Versus Uncertainty Management
The current project aimed to validate findings by Van den Bos et al (2005), highlighting the superior strength and applicability of uncertainty salience (US) manipulations compared to mortality salience (MS) manipulations Initially, the hypothesized effects included interactions between MS and moderators, but the primary focus was on the US manipulation Notably, four out of seven significant findings in the H2(y) family demonstrated a US x Moderator interaction, while three out of seven involved other interactions.
MS x Moderator interaction In the strict sense of having predicted effects primarily for
The author suggests that support for the a-priori formal hypotheses was somewhat lacking in the context of the US, rather than MS However, it was anticipated that all hypotheses would be evident for both US and MS manipulations, as prior research has established this consistent pattern.
The analysis of MS reveals that it does not align with the status quo preference measures, such as the status quo preference scale, attitudes towards reformers, and support for regime change, which are considered larger dependent variable constructs However, the consistent interactions between MS/US and the moderator regarding H2(y) indicate a recurring pattern, reinforcing the validity of this set of hypotheses.
Research indicates that the interaction between prime and political ideology significantly influences attitudes towards reformers, demonstrating a predicted pattern of extremity, magnification, or polarization effects.