Psychiatry Research 220 (2014) 226–232

Contents lists available at ScienceDirect

Psychiatry Research journal homepage: www.elsevier.com/locate/psychres

An investigation into reasoning biases, mood and cognitive state, and subclinical delusional ideation Haley Medlin, Debbie Warman n School of Psychological Sciences, University of Indianapolis, 1400 East Hanna Avenue, Indianapolis, IN 46227, United States

art ic l e i nf o

a b s t r a c t

Article history: Received 27 November 2013 Received in revised form 7 July 2014 Accepted 9 July 2014 Available online 16 July 2014

Following research on reasoning and the continuum of delusional ideation, the present study attempted to investigate the impact of different experimentally-induced states (stress, paranoia, and neutral) on the jumping-to-conclusions reasoning bias in individuals with varying levels of subclinical delusional ideation (SDI). Participants (N¼117) completed a measure of subclinical delusional ideation (the Peters et al. Delusions Inventory or PDI; Peters et al., 1999); and were randomly assigned to receive one of two experimental inductions (stress or paranoia), or no experimental induction; their performance on two probabilistic reasoning tasks – one easy and one challenging – was assessed. Although no differences were found between individuals with high vs. low subclinical delusional ideation in the no induction condition or following the paranoia induction, in the stress-induction condition, individuals with high levels of subclinical delusional ideation were significantly less likely to jump to conclusions on the easy reasoning task. No significant effects emerged on the more challenging task. Assessment of post-test paranoid thinking indicated our paranoia induction did not have its intended effect. Importantly, because there was no pre-test of anxiety, paranoid thinking, or reasoning to determine if they shifted after the inductions, results need to be interpreted with caution. & 2014 Elsevier Ireland Ltd. All rights reserved.

Keywords: Probabilistic reasoning Decision-making Stress induction Paranoia induction Psychosis-proneness

1. Introduction Individuals with delusions have considerable biases in their reasoning, as has been demonstrated in decades of studies using probabilistic reasoning tasks (e.g., Huq et al., 1988; Garety et al., 2013). Relative to healthy and psychiatric controls, individuals with delusions jump to conclusions (JTC) on tasks that ask participants to decide how much data they need before making a decision, and they frequently make extremely hasty decisions, often requesting only one or two stimuli before making their decision (e.g., Dudley et al., 1997; Warman et al., 2007; Dudley et al., 2013). In the hopes of increasing understanding of delusional thinking, attention has been paid to individuals lower on the continuum of psychosis – individuals in the general population who do not have psychotic disorders but who endorse a high level of unusual beliefs (e.g., Peters et al., 1999). Results from reasoning studies have revealed some, but far less robust, evidence for the JTC bias for these individuals (e.g., Colbert and Peters, 2002; Warman and Martin, 2006; Zawadzki et al., 2012), indicating that variation in levels of subclinical delusional ideation (SDI) may be a useful way of understanding the process of delusions, though conclusions to date have been inconsistent. Specific factors related to the JTC bias for individuals

n

Corresponding author. Tel.: þ 317 788 2102; fax: þ317 788 2120. E-mail address: [email protected] (D. Warman).

http://dx.doi.org/10.1016/j.psychres.2014.07.014 0165-1781/& 2014 Elsevier Ireland Ltd. All rights reserved.

who have delusions and who are high in subclinical delusional ideation have received considerable attention recently, as examination may shed light on the various processes that exacerbate or reduce the bias (e.g., Ellett et al., 2008; So et al., 2008; Keefe and Warman, 2011; Lee et al., 2011; Freeman et al., 2013; Warman et al., 2013) and, thus, aid in cognitive theories of the acquisition and maintenance of delusions (e.g., Garety and Freeman, 2013). Considering the situations likely to evoke emotion and, potentially, delusional thinking, it may not be surprising that “stressful” situations have received the most attention in reasoning studies that have used experimental inductions to determine their impact on the JTC bias (e.g., Ellett et al., 2008). Although results have not been entirely consistent (e.g., So et al., 2008), there does appear to be evidence that stress exacerbates the already robust relationship between delusions and JTC (e.g., Ellett et al., 2008; Moritz et al., 2009). There appears to be an important relationship between stress and subclinical delusional ideation as well. For example, Keefe and Warman (2011), who induced stress using a speeded subtraction task, found no relationship between subclinical delusional ideation and reasoning under normal (emotionally neutral) conditions, but following the stress induction, high-SDI individuals (individuals high in subclinical delusional ideation) were overconfident in decisions relative to their low-SDI peers. Similarly, White and Mansell (2009) found high-SDI individuals JTC relative to low-SDI individuals, and found these individuals felt rushed; it seems possible they were in a stressed state during the task. Importantly, just

H. Medlin, D. Warman / Psychiatry Research 220 (2014) 226–232

as the results for delusions and stress have been inconsistent in terms of the impact on JTC, the same is true for studies of subclinical delusional ideation (e.g., Lincoln et al., 2010b), indicating the relationship between mood or cognitive state and delusions is complex and in need of further investigation. As noted, the relationship between JTC and delusions is quite clear, but the impact of induced mood or cognitive state is less so. One possible explanation for the inconsistency in studies on delusions/ subclinical delusional ideation and reasoning is that JTC is simply not as strongly related to mood or cognitive state as might be expected. Another possibility is that how mood or cognitive state is induced has been so varied across studies that conclusions are hard to draw. Indeed, experimental manipulations or inductions have been conducted in a number of ways, including a stressful test (a speeded subtraction task; Keefe and Warman, 2011), buying a newspaper in a busy shopping area (Ellett et al., 2008), asking participants to describe an anxiety-provoking situation they experienced (Lincoln et al., 2010a), a loud noise (Lincoln et al., 2010b), and anxiety-evoking music (Moritz et al., 2009). This variability has, perhaps, made conclusions challenging to draw. In addition, direct comparisons of various induced unpleasant mood or cognitive states have not, to date, been investigated (see Lee et al., 2011), though such comparisons may help illuminate relationships between them and the JTC bias. Two states hypothesized to be particularly relevant to delusions are anxiety or stress (Moritz et al., 2009) and also paranoia, which can range from social evaluative concerns to concerns of severe threat (see Green et al., 2011; Garety and Freeman, 2013). Their theoretical relationship with the continuum of delusional ideation suggests that these states would provide a useful comparison in terms of reasoning evoked. The present study was designed to test a number of questions left unanswered in extant research—to investigate the relative impact of a stress induction, a paranoia induction, and no induction on the reasoning of individuals with varying levels of subclinical delusional ideation. It was expected, based on previous research (e.g., Colbert and Peters, 2002), that high-SDI individuals would JTC relative to low-SDI individuals. Further, following recent findings (e.g., Lincoln et al., 2010a; Keefe and Warman, 2011), it was expected that individuals who were given a stress induction would JTC relative to individuals in the no-induction condition and that this would be particularly prominent in high SDI individuals. Finally, due to the relationship between paranoia and delusional thought, and the possibility that induced paranoia influences how individuals with unusual beliefs process information and make decisions, it was expected that high-SDI individuals in the paranoia-induction condition would have the most profound JTC bias, relative to low-SDI individuals and also to high-SDI individuals in the other conditions. To decrease the transparency of the study (i.e., so participants would not find it obvious we were using manipulations to impact reasoning), no pre-test of anxiety or paranoia was conducted; instead, participants were randomly assigned to an experimental condition and their anxiety and paranoia were evaluated only after the induction was completed. Although this procedure likely made the participants less aware of the aims of the study, it does limit conclusions that can be drawn about whether findings are due to the actual inductions themselves.

2. Method 2.1. Participants Individuals (n¼ 117) were recruited from the community and from a university setting. Approximately 50% of the sample consisted of individuals in the general community (n¼ 59), and the remainder were from an undergraduate student population (n¼ 58). Although examining only undergraduates is a primary recruitment strategy in studies of subclinical delusional ideation (e.g., McKay et al., 2006; Lincoln et al., 2010a; Balzan et al., 2012), we recruited from the general population as well (through advertisements) in an effort to increase the generalizability of our

227

findings. Participants were required to be at least 18 years of age, able to speak and read English, and able to provide written informed consent. Participants could not have a current or previous diagnosis of a schizophrenia-spectrum disorder or Bipolar Disorder. Participants from the community received $25.00, while college students received course credit for their participation in the study. 2.2. Materials 2.2.1. Peters et al. Delusions Inventory Study participants completed the Peters et al. Delusions Inventory (PDI; Peters et al., 1999), a 40-item self-report instrument assessing for multiple dimensions of delusional ideation. The PDI has been used frequently in research on subclinical delusional ideation and reasoning (e.g., Warman et al., 2007; LaRocco and Warman, 2009; Lincoln et al., 2010c). The PDI has high internal consistency (α ¼0.88) and test–retest reliability (r ¼0.82; Peters et al., 1999). 2.2.2. State-Trait Anxiety Inventory (STAI) The State Anxiety scale of the STAI (Spielberger, 1983), a 20-item measure of state anxiety for adults, was used as a manipulation check to assess the degree of anxiety after the experimental inductions. The STAI has good psychometric properties, including test–retest reliability coefficients that range from 0.34 to 0.62. 2.2.3. Modified Paranoia Checklist A modified version of the 18-item Paranoia Checklist (Freeman et al., 2005) was used to assess the degree of paranoia after the inductions. The original self-report measure was developed to measure paranoid ideation and has good internal consistency (αZ 0.90) and convergent validity. Following Lincoln et al., 2010a, this study employed a modified version of the scale that was designed to assess state paranoia. It was further modified to make the questions more applicable to nonclinical populations. Modifications from the original scale are italicized in the following example: “At the moment, I believe that people deliberately try to irritate me.” Participants were asked to rate only conviction and distress for each item. 2.2.4. Filler task A filler task in which participants sorted a stack of names (written one per index card) into alphabetical order was conducted simultaneously with the paranoia induction, in order to have the participants engaged in a relatively mindless activity. The filler task was administered to individuals in the noinduction condition as well, in order to maintain consistency in time requirements across conditions. The names used in this untimed task were selected through the use of a random (first) name generator (http://www.kleimo.com/random/name. cfm) that draws from a list provided by the U. S. Census Bureau. 2.3. Procedures Within each sample type (community vs. college), participants were randomly assigned to one of the three experimental groups, to ensure equal representation of community and college student participants across the conditions. Following provision of informed consent, participants completed the PDI, followed by the experimental manipulations that are described in the following section. The filler task was administered concurrently with the paranoia induction. As indicated, participants in the “no induction” condition were also administered the filler task. Following the completion of the experimental manipulation (or no manipulation), all participants completed the reasoning tasks, followed by one assessment of anxiety state and one assessment of paranoid thinking. 2.3.1. Experimental manipulations Participants within each group (community or college student) were randomly assigned to receive a stress induction, a paranoia induction, or no induction. 2.3.2. Stress induction Following the design of Keefe and Warman (2011), participants randomly assigned to the stress induction completed a speeded subtraction task designed to be somewhat stressful (see Sgoutas-Emch et al., 1994; Tohill and Holyoak, 2000). During this task, a stopwatch was used to mark the elapsed time, and it was also expected to increase stress levels. Furthermore, mistakes were verbally corrected by the examiner as they occurred, and after 25 s elapsed, the participant was informed of the remaining time and told that he/she is “too slow.” After 45 s elapsed, the participant was asked to stop and was told that the task would be repeated at the end of the experiment, although it was not actually repeated. 2.3.3. Paranoia induction Individuals assigned to the paranoia induction completed the study in a room containing a two-way mirror. In this condition, the presence of the two-way mirror in the room was emphasized, as the use of a two-way mirror to heighten paranoia and self-focused attention has been successful in previous studies (e.g., Fenigstein and Vanable, 1992; Smari et al., 1994). Individuals were told before beginning the

228

H. Medlin, D. Warman / Psychiatry Research 220 (2014) 226–232

filler task that someone “may or may not” be observing and evaluating their efforts from the other side of the two-way mirror. The phrasing was hypothesized to induce a degree of ambiguity that is reflective of the types of situations that may exacerbate paranoid ideation in individuals who are actively-deluded. Furthermore, the examiner noted that she “may” receive text messages updating her of the participant's progress at various intervals throughout the session. The examiner's phone then buzzed twice at intervals of approximately 1 min while participants completed the sorting task.

2.3.4. No induction The third group completed the filler task but did not receive an induction and served as the control condition.

2.3.5. Decision-making tasks All participants completed one trial each of an easy (85:15) and a difficult (60:40) version of the draws-to-decision beads task (originally developed by Phillips and Edwards, 1966). The script for explanation of both tasks was adapted from Keefe and Warman (2011). Participants were told that two jars (A and B) had 100 beads in complementary ratios of two different colors (e.g., 85 pink/15 blue and 85 blue/15 pink). After the jars were hidden, beads were drawn one at a time (with replacement) in a predetermined order from the selected jar, at the request of the participant. Individuals were told to request as many stimuli as they would like (up to 20 draws), and were asked to make their decision only when they were certain about their answer. Once the participant was certain of which jar the beads were coming from, no more beads were drawn and the participant stated his/her decision. The order in which the beads were drawn for the 85:15 version was as follows: P B B B B P B B B P P B B B B B B B B B, where “P” indicates a pink bead was drawn and “B” indicates a blue bead was drawn. For the 60:40 version, the order was: P P P B P B P P B P P B B B B P P P B P. The order of decision-making tasks was counterbalanced within each condition.

2.3.6. Posttest assessment of anxiety and paranoid thinking The state version of the STAI and the modified Paranoia Checklist were administered to all participants after reasoning tasks were completed. A debriefing form was reviewed with participants at the conclusion of the study.

2.4. Statistical analyses For preliminary analyses, independent sample t-tests (sample type – college vs. community; sex) and Pearson correlations (age) were used to analyze group differences in or association with data-gathering (jumping to conclusions). For the 85:15 reasoning task (primary analysis), a 3 (induction group: no induction, stress induction, or paranoia induction)  2 (degree of subclinical delusional ideation – high vs. low) Univariate Analysis of Variance was conducted with draws-to decision as the dependent variable. For the 60:40 reasoning task (primary analysis), a 3  2 Univariate Analysis of Covariance was conducted with draws-todecision as the dependent variable and age as a covariate. To assess differences in anxiety state and/or paranoid thinking across the experimental groups following the experimental manipulation/ induction, one Univariate Analysis of Variance was conducted for paranoia, and another for anxiety state. Please see Section 3 for posthoc procedures.

3. Results 3.1. Group classification Following procedures from previous research (e.g., Warman, 2008; Balzan et al., 2012), individuals were classified as low or high in subclinical delusional ideation (low-SDI vs. high-SDI) according to whether their total score on the PDI fell above or below the median for the entire sample. Fifty-eight individuals were classified as high-SDI, with scores falling above the median of 91, and 58 individuals were classified as low-SDI. One participant had a score that fell exactly at the median; consistent with prior research (e.g., Laroi and Van der Linden, 2005), this participant was excluded from analyses. The median in the present study was somewhat lower than has been found in previous research using the PDI-40 as a measure of subclinical delusional ideation (see Warman et al., 2007, median¼97.5; Warman, 2008, median ¼100; Keefe and Warman, 2011, median ¼106.5), despite this study recruiting from the general population as opposed to only undergraduate students. Demographics of the sample can be

Table 1 Participant characteristics by level of subclinical delusional ideation (SDI). High-SDI

Low-SDI

n

%a

n

%

Gender Male Female Race Caucasian African–American Other Missing

18 40

31.0 69.0

15 43

25.9 74.1

44 8 5 1

75.9 13.8 8.6 1.7

74.1 10.3 6.9 0

Age

22.19 (5.96)

43 6 4 0 28.48 (11.70)

Note: Mean age of participants according to their level of SDI presented, along with standard deviations [M (S.D.)]. a

Percentages are in reference to classification as high- vs. low-SDI.

Table 2 PDI scores and number of participants for individuals classified as having high or low levels of subclinical delusional ideation across the three conditions of the study.

Stress induction Paranoia induction No induction

High-SDI

Low-SDI

n

n

19 22 17

PDI score M

S.D.

159.05 169.50 151.35

40.66 61.91 52.30

20 17 21

PDI score M

S.D.

59.45 49.71 54.19

23.97 28.13 27.53

found in Table 1; specifics about PDI scores and group classification can be found in Table 2. 3.2. Preliminary analyses A number of factors were examined to determine their relationship to the dependent variables (number of beads requested on each task). There were no differences between college and community samples in number of stimuli drawn on either the 85:15 beads task, t(114) ¼ 1.60, P ¼0.11 (M undergraduates ¼ 5.29, M community ¼4.62) or the 60:40 beads task, t(114) ¼ 1.74, P¼ 0.084 (M undergraduates ¼6.29, M community ¼5.14). As a result, no comparisons between these two groups will be conducted or considered further. The demographic variables of age and sex were also explored. There was a trend toward a relationship between age and draws-todecision on the 85:15 task, r(116)¼  0.17, P¼ 0.07, but this result did not achieve statistical significance and the magnitude of the correlation was small, so this relationship will not be considered further. Age, however, was significantly correlated with draws-to-decision on the 60:40 task, r(116)¼ 0.25, P¼ 0.006, indicating the younger the participant, the more stimuli he/she requested on this task. Although this is a weak correlation (ro0.30), following a conservative approach, age will be entered as a covariate in the primary 60:40 beads task analysis. In addition, sex was examined. No effect for sex emerged for either the 85:15 task, t(114)¼1.69, P¼0.10, or the 60:40 task, t(114)¼ 0.04, P¼0.97, thus sex will not be considered further. 3.3. Primary analyses 3.3.1. The 85:15 reasoning task The main hypotheses of the study were that high-SDI individuals would request fewer stimuli than low-SDI individuals, that stress would exacerbate the bias, and that high-SDI individuals in

H. Medlin, D. Warman / Psychiatry Research 220 (2014) 226–232

Fig. 1. Average number of stimuli requested in 85:15 task across the three different experimental conditions, by level of subclinical delusional ideation.

the paranoia induction would have the most profound JTC bias. No support was found for subclinical delusional ideation being related to data gathering, F(1, 110) ¼0.30, P¼ 0.58, and no support was found for a main effect of induction condition, F(2, 110) ¼0.07, P ¼0.94. A significant interaction between level of subclinical delusional ideation and induction condition, however, did emerge, F(2, 110) ¼ 3.18, P ¼0.046 (see Fig. 1). Post-hoc tests revealed that regardless of level of subclinical delusional ideation, no differences emerged in how many stimuli participants requested across the three conditions, (low-SDI: F(2, 55) ¼1.26, P ¼0.29, high-SDI: F(2, 55) ¼2.03, P ¼0.14). Additional follow-up tests were conducted to determine the differences between low-SDI and high-SDI individuals for each of the three conditions. For individuals in the stress-induction group, high-SDI participants (M¼5.63; S.D. ¼2.27) requested more beads than did low-SDI individuals (M ¼4.25; S.D. ¼ 1.68), t(37) ¼  2.17, P¼ 0.04. For individuals in the paranoia-induction group, high-SDI (M ¼5.27; S.D. ¼2.57) and low-SDI (M ¼4.77; S.D. ¼ 1.99) participants requested the same number of beads before making a decision, t(37) ¼  0.67, P ¼0.50. Finally, in the no-induction group, high-SDI (M ¼4.24; S.D. ¼1.20) and low-SDI (M ¼5.43; S.D. ¼3.12) participants requested the same number of beads before making a decision, t(36) ¼ 1.49, P¼ 0.15.

3.3.2. The 60:40 reasoning task The same hypotheses listed for the 85:15 task were examined for the 60:40 task. As with the 85:15 task, there was no support for a relationship between level of subclinical delusional ideation and data-gathering, F(1, 109) ¼0.002, P ¼0.97 and no support for a main effect for induction, F(2, 109) ¼ 0.36, P ¼0.70. Further, no level of subclinical delusional ideation  induction interaction emerged, F(2, 109) ¼0.83, P ¼0.44. In sum, no significant effects emerged on this task, as can be seen in Fig. 2.1 1

Some studies on the continuum of delusional ideation and reasoning have used “extreme-responding” (i.e., whether any participants decide on the basis of one or two stimuli) as an outcome variable (e.g., Freeman et al., 2008; Lincoln et al., 2010b; see Fine et al., 2007 for review). As such, it was considered worth exploring differences in extreme-responding in the present study, and whether extremeresponding was related to level of subclinical delusional ideation or induction condition. However, the low rates of extreme-responding in the present study violated assumptions of the statistical test and resulted in considerable loss of power. Due to the loss of power, the finding of no difference between groups was not considered interpretable, and as a result, these analyses/findings are not reported or considered further. Of note to the reader and to aid comparison across studies, however, extreme responding (ER; requested two or fewer stimuli) results are reported here: for high-SDI individuals: in the no induction group, 5.9% (1/17) ER’d on the 85:15 and 60:40 tasks; in the stress induction group, 5.3% (1/19) on the 85:15 task, 10.5% (2/19) 60:40 task; in the paranoia induction group 0% (0/22) on 85:15 task, 4.5% (1/22) 60:40 task. For low-SDI individuals: in the no induction group: 4.8% (1/21) on the 85:15 task, 9.5% (2/21) 60:40 task; in the stress induction

229

Fig. 2. Average number of stimuli requested in 60:40 task across the three different experimental conditions, by level of subclinical delusional ideation.

3.4. Posttest assessment of anxiety and paranoid thinking Following the experimental manipulation, analyses were conducted to determine if state anxiety, as measured by the STAI, and paranoia, as measured by the Paranoia Checklist, were heightened in the stress induction and paranoia induction groups, respectively. In examination of reported level of state anxiety, a significant effect emerged for experimental induction group, F(2, 114) ¼4.49, P ¼0.01. The Games–Howell post-hoc procedures (run due to a violation of the assumption of homogeneity of variance) revealed that the stress induction group reported a higher level of anxiety on the STAI (M¼ 37.15; S.D. ¼11.11) compared to the no-induction group (M¼30.62; S.D. ¼6.64). No differences in anxiety levels, however, were found between paranoia-induction group (M ¼35.21; S.D. ¼ 11.24) and the other two groups. In contrast, no statistically-significant differences were found in the levels of paranoia reported across experimental induction conditions, F(2, 114) ¼2.28, P ¼0.11. The participants in the paranoia induction group (M¼ 23.82, S.D. ¼5.63) did not endorse conviction in paranoid beliefs at a level that significantly differed from the average endorsement of individuals in the stress induction group (M ¼24.85, S.D. ¼ 8.36) or the no induction group (M ¼21.87, S.D. ¼ 3.91). Our paranoia induction, then, did not have the intended impact as there was no evidence for increased paranoid thinking following this manipulation. In sum, none of the study hypotheses were supported. Although a delusion-prone effect emerged in the stress induction condition for the easy reasoning task, results were opposite of those expected. Further, although we intended to heighten paranoid thinking in the paranoia condition, there was no evidence our manipulation was effective.

4. Discussion The present study was an investigation of the jumping-toconclusions (JTC) bias in individuals with varying levels of subclinical delusional ideation under conditions intended to induce either stress or paranoia. There was no relationship between level of subclinical delusional ideation and reasoning and/or induction on the 60:40 (difficult) version of the reasoning task used. On the 85:15 (easy) version of the reasoning task, however, individuals who varied in their levels of subclinical delusional ideation displayed data-gathering differences in the stress induction condition, though findings were counter to hypothesis and prior research (e.g., Lincoln et al., 2010a; Keefe and Warman, 2011) in that (footnote continued) group: 15% (3/20) 85:15 task, 15% (3/20) 60:40 task; in the paranoia induction group: 11.8% (2/17) 85:15 task, and 17.6% (3/17) for 60:40 task.

230

H. Medlin, D. Warman / Psychiatry Research 220 (2014) 226–232

high-SDI participants requested more stimuli before making a decision than did low-SDI participants in the stress induction condition. Counter to our hypotheses, no effect for the paranoia induction emerged, and, in fact, there was no evidence our paranoia induction had its intended effect on participants’ cognitive state. Although the primary aim of the present study was to investigate how a stress or paranoia induction influences data gathering for high- and low-SDI individuals, the primary question that can be raised from the present findings, given our methodology, is whether our design actually allows for such determinations to be made. After all, we did not assess mood or cognitive state (or reasoning) before the induction was presented, making it impossible for us to determine whether or not the stress or paranoia inductions had any influence on the participant relative to before the induction. Our study was intended to be a replication and extension of Keefe and Warman’s (2011) study; like Keefe and Warman, we used a between-groups design to evaluate reasoning differences – and we added a posttest assessment of mood and cognitive state to provide further information about one potential confounding variable. The potential limitations of this approach are, however, essential to note. We cannot guarantee participants were similar in their state anxiety or paranoia or reasoning at the start of the study, before the induction began, which is reason to call the current conclusions into question. Our posttest-only assessment of mood and cognitive state as a “manipulation check” has been used numerous times in the past, such as by Oaksford et al. (1996), yet such a methodological approach has significant limitations. It is likely valuable to consider that like Oaksford and colleagues, we relied upon random assignment to be the “great equalizer” (Heiman, 2001; Baumeister and Bushman, 2011) across our three experimental conditions, such that any baseline differences would be theoretically equally represented across experimental groups – and thus any differences between groups on the “manipulation checks” could be expected to be due to the effects of the induction. We took this approach due to concerns that facevalid questions on paranoia and anxiety at the outset of the study would prime the participants to respond differently to the experimental conditions and make our findings less meaningful. Still, without a pre–post assessment, it is impossible to know that groups were indeed no different in their baseline levels of anxiety and paranoia. As a result, our conclusions regarding the effectiveness of our inductions – that the stress induction did lead to increased anxiety but the paranoia induction did not result in any increase in paranoia – may not, given the limitations of our design, even be valid or reasonable conclusions. A pre–post assessment of anxiety and paranoid thinking would eliminate concern that baseline differences may have contributed to post-induction differences between groups, and as such, would be ideal for future research in this field. In sum, although our intent was to investigate the relative impact on reasoning of stress and paranoia inductions relative to no induction, our posttest-only assessment of anxiety and paranoid thinking and reasoning, despite random assignment to induction groups, limits or precludes our ability to make any claims about these influences. The problem of experimentally investigating the impact of induced stress and paranoid thinking has been experienced by other researchers in this area (e.g., So et al., 2008). Music has been used to affect mood (Moritz et al., 2009) as have noises (Lincoln et al., 2009), and even instructions to worry (Freeman et al., 2013), and results have been quite varied. Specific paranoia inductions have not been widely investigated. Although one recent study determined that paranoid thoughts could be induced by having the study interrupted by a confederate and/or having someone outside a testing room laugh loudly (Green et al., 2011), rates of paranoid attributions for these events were very low. Fenigstein and Vanable (1992) used a two-way mirror in an effort to determine its effects on self-consciousness

and paranoid thoughts and found that for people high in public self-consciousness, or those who endorsed high awareness of how they are viewed by others, the two-way mirror had quite an impact on how “watched” individuals felt, even when the mirror was not directly discussed with participants. It was our hope that by using the two-way mirror and having a direct conversation about it with participants and noting they were possibly being observed and evaluated, we would enhance the effects found in prior studies of two-way mirrors and induce paranoia. This did not happen. It is possible another variable, such as public selfconsciousness, would be important to assess to determine its impact in paranoia-induction conditions such as the one used in this study, and this variable itself may be more relevant to the current research than paranoia per se. Another potential way of understanding the limitation of our paranoia induction is to consider that the distinctions between anxiety and paranoia may not particularly useful. Garety and Freeman (2013), for example, note the similarities between anxious and paranoid thinking, noting “persecutory thoughts are considered an extension of anxious and depressive concerns about the person's own vulnerability and lack of worth” (p. 330). In addition, Lincoln et al. (2009) and Lincoln et al. (2010b) found that stress impacted individuals' levels of paranoia, even when paranoia was not experimentally manipulated on its own. Interestingly, however, they found that while stress increased the level of paranoia for individuals with schizophrenia, subclinical paranoid symptoms in healthy participants actually reduced under stress. Further investigations into how to uniquely influence paranoia rather than general stress or anxiety would be particularly useful for understanding the relationship between paranoid thinking and data-gathering; it appears we have not yet achieved a proper method to investigate this relationship. While the JTC bias has been found quite consistently for individuals with delusions regardless of various experimental conditions and stimuli used (e.g., Fine et al., 2007), the findings are far less consistent for individuals with higher levels of subclinical delusional ideation (e.g., Warman et al., 2007) and the reason for these discrepancies continues to be unclear. In the present study, the experimental inductions, when considered independent of level of subclinical delusional ideation, did not negatively affect participants' data-gathering. Based on prior research that has demonstrated that higher anxiety is related to decreased data-gathering (e.g., Bensi et al., 2010), it was expected that individuals, regardless of level of subclinical delusional ideation, would JTC more when stress or paranoia was induced than when it was not. Unexpectedly, the findings of Bensi and colleagues were not replicated in the present study, as participants requested the same number of stimuli in the no-induction, stress-induction, and paranoia-induction conditions. However, a recent study noted that the responses of highly paranoia-prone individuals to stressful situations can help explain the emotional reaction they experience (Westermann et al., 2012). It is possible, considering the individuals in the present study were functioning well (i.e., they were either college students or managing in the community), that these were individuals with good coping strategies even under stressful or high self-focus conditions, thus making the impact of the inductions less powerful. The present findings for level of subclinical delusional ideation and reasoning were unexpected despite some inconsistency in reports of the relationship between the two in the extant literature. In the stress-induction condition, high-SDI individuals requested more stimuli than low-SDI individuals on the easy reasoning task. This is in direct contradiction to the study that the present study is based on (Keefe and Warman, 2011), which demonstrated greater JTC by high-SDI individuals than low-SDI individuals. The way that JTC was understood in the Keefe and Warman study may help illuminate the differences in results

H. Medlin, D. Warman / Psychiatry Research 220 (2014) 226–232

between the two studies; they examined draws-to-decision but did not request participants wait until certainty. Instead, they had participants decide when they felt ready and then assessed confidence in decisions. Under those conditions, individuals with varying levels of subclinical delusional ideation requested the same number of stimuli, but high-SDI individuals were more confident in their decisions, which Keefe and Warman equated with JTC. The present study requested individuals be certain before deciding and did not include an assessment of participant confidence. It is also important to note that Keefe and Warman's study did not include the 85:15 task that revealed such unexpected results in the present study. Finally, Keefe and Warman had more participants in each condition than we had in the present study – they had approximately 30 in each cell as opposed to the present study which had approximately 20. It is possible had our sample size per cell been higher, our results would have more closely resembled those of Keefe and Warman's. Given the inconsistencies in the literature related to subclinical delusional ideation, it seems reasonable that any effect in the population is quite small. As a result, the small sample size of the present study is quite limiting in terms of drawing conclusions. Thus despite the many similarities between the two studies, including identical stress tests for the stress induction, enough methodological differences may exist to partially explain the divergent findings. Other explanations for the unique findings of this study may also arise from similarly-surprising results in the literature. In conjunction with the present study, findings by Lincoln et al. (2010b) and Warman (2008) may call into question the expectation that stress or emotionally-salient stimuli (respectively) will exacerbate jumping to conclusions in nonclinical populations, and particularly in individuals with higher levels of subclinical delusional ideation. In the present study, more cautious reasoning was exhibited by high-SDI individuals under stress relative to low-SDI peers on the easy reasoning task. The Lincoln et al. (2010b) and Warman (2008) studies both found that stress or emotional stimuli resulted in more cautious reasoning across their nonclinical populations, regardless of level of subclinical delusional ideation. This seems contradictory to the current understanding of the continuum of delusional ideation and associated reasoning biases, where high-SDI (relative to low-SDI) individuals are expected to exhibit reasoning styles more similar to deluded populations. However, when clinical and nonclinical populations are directly compared, there is some suggestion that non-deluded individuals as a group respond differently to anxiety or stress inductions than their deluded peers; while deluded individuals exhibit greater JTC than peers following an anxiety induction, this was not the case for the other groups (Moritz et al., 2009). It is possible that while stress (and perhaps paranoia) is experienced differently by clinical and nonclinical populations and related to differences in reasoning styles, there are more similarities than differences in the way that high-SDI and low-SDI individuals experience stressful situations. The present study had a number of strengths. First, the present study tested a more representative sample than many studies that relied only on undergraduates to draw conclusions (e.g., Lincoln et al., 2010a; Balzan et al., 2012). Half of the present sample was recruited from the general population, which allowed for significantly more heterogeneity in the sample, particularly in terms of age and sex (see Warman, 2008). Importantly, we found no differences in datagathering for our community relative to our university sample, indicating studies that draw from university samples exclusively may be quite useful for understanding reasoning processes in the general population. Second, the present study used a no-induction control condition, as opposed to numerous studies that used comparisons groups that may, due to the procedures used (e.g., inducing relaxation), have affected decision-making in their own right (e.g., Ellett et al., 2008; So et al., 2008). A number of limitations

231

remained, however. Most importantly, perhaps, is that our paranoia induction did not have the desired impact – there was no evidence our paranoia induction resulted in increased paranoid thinking in our study. Although investigations such as these are in their infancy, the failure of the expected increases in paranoia to emerge was unexpected. A further limitation of this study is the possibility of baseline differences in decision-making between groups, as we used a between-group, not a pre–post design. As a result, it is impossible to know that groups were indeed no different in their baseline decision-making, anxiety, or paranoid thinking. We chose an approach we hypothesized would reduce problems with practice effects that may have influenced findings in previous studies (e.g., Peters and Garety, 2006), and consistent with other studies examining reasoning under multiple experimental conditions (e.g., Warman et al., 2013), we relied on random assignment to be the “great equalizer” instead of a pre–post design (Heiman, 2001; Baumeister and Bushman, 2011). Still, however, no guarantee of these equal distributions can be made. Finally, no determination was made of whether participants understood the reasoning tasks. Recent research has indicated the beads task is perhaps more complicated to participants than had been considered (e.g., Balzan et al., 2012), and results of these popular studies of reasoning should be interpreted in light of this possibility. That said, there was no indication in our study that miscomprehension was related to level of subclinical delusional ideation or the experimental manipulations. Instead, it is possible that miscomprehension was wide-spread in our study and had instructions been clearer, perhaps more results would have emerged. In sum, the present study aimed to investigate subclinical delusional ideation under various experimental conditions, with a focus on stress and paranoia. No differences emerged between the reasoning of individuals with high vs. low levels of subclinical delusional ideation, except under stress, when high-SDI participants unexpectedly exhibited more cautious reasoning than their peers on an easy version of the reasoning task. The finding that individuals in the paranoia induction group did not display reasoning differences relative to their peers may be related to the fact that their reported levels of paranoid thinking were also no different relative to their peers – our paranoia induction appeared to not have its intended impact. Importantly, due to the lack of a pre–post design, we cannot make any determination that reasoning changed as a result of our inductions since we did not do any pre-test of their reasoning performance. It appears that the relationship between subclinical delusional ideation and reasoning may not be all that robust, may be limited to conditions that were not present in the current study, or may actually be counter to expectations in some circumstances. Future investigations would benefit from continued exploration of reasoning under different experimental conditions, in combination with and in addition to reasoning with different stimuli, such as alternative neutral stimuli (e.g., Lincoln et al., 2010a), emotional stimuli (e.g., Warman et al., 2007), or even delusional/improbable stimuli (e.g., LaRocco and Warman, 2009). Given the potential impact of learning more about how individuals along the continuum of delusional ideation process information, such refinement and synthesis are expected to be invaluable for developing theories of psychosis (e.g., Garety and Freeman, 2013) and also interventions to help individuals who may be at risk for developing psychotic disorders learn how to process information in healthier ways.

References Balzan, R., Delfabbro, P., Galletly, C., 2012. Delusion-proneness or miscomprehension? A re-examination of the jumping to conclusions bias. Australian Journal of Psychology 64, 100–107. Baumeister, R.F., Bushman, B.J., 2011. Social Psychology and Human Nature, 2nd edition Cengage, Sand Francisco, CA.

232

H. Medlin, D. Warman / Psychiatry Research 220 (2014) 226–232

Bensi, L., Giusberti, F., Nori, R., Gambetti, E., 2010. Individual differences and reasoning: a study on personality traits. British Journal of Psychology 101, 545–562. Colbert, S.M., Peters, E.R., 2002. Need for closure and jumping-to-conclusions in delusion-prone individuals. Journal of Nervous and Mental Disease 190, 27–31. Dudley, R., Daley, K., Nicholson, M., Shaftoe, D., Spencer, H., Cavanagh, K., Freeston, M., 2013. ‘Jumping to conclusions’ in first-episode psychosis: a longitudinal study. British Journal of Clinical Psychology 52, 380–393. Dudley, R.E.J., John, C.H., Young, A.W., Over, D.E., 1997. Normal and abnormal reasoning in people with delusions. British Journal of Clinical Psychology 36, 243–258. Ellett, L., Freeman, D., Garety, P.A., 2008. The psychological effect of an urban environment on individuals with persecutory delusions: the Camberwell walk study. Schizophrenia Research 99, 77–84. Fenigstein, A., Vanable, P.A., 1992. Paranoia and self-consciousness. Journal of Personality and Social Psychology 62, 129–138. Fine, C., Gardner, M., Craigie, J., Gold, I., 2007. Hopping, skipping or jumping to conclusions? Clarifying the role of the JTC bias in delusions. Cognitive Neuropsychiatry 12, 46–77. Freeman, D., Garety, P.A., Bebbington, P.E., Smith, B., Rollinson, R., Fowler, D., Kuipers, E., Ray, K., Dunn, G., 2005. Psychological investigation of the structure of paranoia in a non-clinical population. British Journal of Psychiatry 186, 427–435. Freeman, D., Pugh, K., Garety, P., 2008. Jumping to conclusions and paranoid ideation in the general population. Schizophrenia Research 102, 254–260. Freeman, D., Startup, H., Dunn, G., Cernis, E., Wingham, G., Pugh, K., Cordwell, J., Kingdon, D., 2013. The interaction of affective with psychotic processes: a test of the effects of worrying on working memory, jumping to conclusions, and anomalies of experience in patients with persecutory delusions. Journal of Psychiatric Research 47, 1837–1842. Garety, P.A., Freeman, D., 2013. The past and future of delusions research: from the inexplicable to the treatable. British Journal of Clinical Psychiatry 203, 327–333. Garety, P., Joyce, E., Jolley, S., Emsley, R., Waller, H., Kuipers, E., Bebbington, P., Fowler, D., Dunn, G., Freeman, D., 2013. Neuropsychologcial functioning and jumping to conclusions in delusions. Schizophrenia Research 150, 570–574. Green, C.E.L., Freeman, D., Kuipers, E., Bebbington, P., Fowler, D., Dunn, G., Garety, P. A., 2011. Paranoid explanation of experience: a novel experimental study. Behavioural and Cognitive Psychotherapy 39, 21–34. Heiman, G.W., 2001. Understanding Research Methods and Statistics, 2nd ed. Houghton Mifflin, Boston. Huq, S.F., Garety, P.A., Hemsley, D.R., 1988. Probabilistic judgements in deluded and non-deluded subjects. The Quarterly Journal of Experimental Psychology A 40, 801–812. Keefe, K.M., Warman, D.M., 2011. Reasoning, delusion-proneness, and stress: an experimental investigation. Clinical Psychology & Psychotherapy 18, 138–147. LaRocco, V.A., Warman, D.M., 2009. Probability estimations and delusionproneness. Personality and Individual Differences 47, 197–202. Laroi, F., Van der Linden, M., 2005. Metacognitions in proneness towards hallucinations and delusions. Behavior Research and Therapy 43, 1425–1441. Lee, G., Barrowclough, C., Lobban, F., 2011. The influence of positive affect on jumping to conclusions in delusional thinking. Personality and Individual Differences 50, 717–722. Lincoln, T.M., Lange, J., Burau, J., Exner, C., Moritz, S., 2010a. The effect of state anxiety on paranoid ideation and jumping to conclusions: an experimental investigation. Schizophrenia Bulletin 36, 1140–1148. Lincoln, T.M., Peter, N., Schafer, M., Moritz, S., 2009. Impact of stress on paranoia: an experimental investigation of moderators and mediators. Psychological Medicine 39, 1129–1139.

Lincoln, T.M., Peter, N., Schafer, M., Moritz, S., 2010b. From stress to paranoia: an experimental investigation of the moderating and mediating role of reasoning biases [Research letter]. Psychological Medicine 40, 169–171. Lincoln, T.M., Ziegler, M., Mehl, S., Rief, W., 2010c. The jumping to conclusions bias in delusions: specificity and changeability. Journal of Abnormal Psychology 119, 40–49. McKay, R., Langdon, R., Coltheart, M., 2006. Need for closure, jumping to conclusions, and decisiveness in delusion-prone individuals. Journal of Nervous and Mental Disease 194, 422–426. Moritz, S., Veckenstedt, R., Randjbar, S., Hottenrott, B., Woodward, T.S., von Eckstaedt, F.V., Schmidt, C., Jelinek, L., Lincoln, T.M., 2009. Decision making under uncertainty and mood induction: further evidence for liberal acceptance in schizophrenia. Psychological Medicine (39), 1821–1829. Oaksford, M., Morris, F., Grainger, B., Williams, J.M.G., 1996. Mood, reasoning, and central executive processes. Journal of Experimental Psychology: Learning, Memory & Cognition 22, 476–492. Peters, E., Garety, P., 2006. Cognitive functioning in delusions: a longitudinal analysis. Behaviour Research and Therapy 44, 481–514. Peters, E.R., Joseph, S.A., Garety, P.A., 1999. The measurement of delusional ideation in the normal population: introducing the PDI (Peters et al. Delusions Inventory). Schizophrenia Bulletin 25, 553–576. Phillips, L.D., Edwards, W., 1966. Conservatism in a simple probabilistic inference task. Journal of Experimental Psychology 72, 346–354. Sgoutas-Emch, S.A., Cacioppo, J.T., Uchino, B.N., Malarkey, W., Pearl, D., KiecoltGlaser, J.K., Glaser, R., 1994. The effects of an acute psychological stressor on cardiovascular, endocrine, and cellular immune response: a prospective study of individuals high and low in heart rate reactivity. Psychophysiology 31, 264–271. Smari, J., Stefansson, S., Thorgilsson, H., 1994. Paranoia, self-consciousness, and social cognition in schizophrenics. Cognitive Therapy and Research 18, 387–399. So, S.H., Freeman, D., Garety, P., 2008. Impact of state anxiety on the jumping to conclusions delusion bias. Australian and New Zealand Journal of Psychiatry 42, 879–886. Spielberger, C.D., 1983. State-Trait Anxiety Inventory for Adults. Mind Garden, Palo Alto, CA. Tohill, J.M., Holyoak, K.J., 2000. The impact of anxiety on analogical reasoning. Thinking and Reasoning 6, 27–40. Warman, D.M., 2008. Reasoning and delusion proneness: confidence in decisions. The Journal of Nervous and Mental Disease 196, 9–15. Warman, D.M., Martin, J.M., Lysaker, P.H., 2013. Jumping to conclusions and delusions: the impact of discussion of the bias on the bias. Schizophrenia Research 150, 575–579. Warman, D.M., Lysaker, P.H., Martin, J.M., Davis, L., Haudenschield, S.L., 2007. Jumping to conclusions and the continuum of delusional beliefs. Behaviour Research and Therapy 45, 1255–1269. Warman, D.M., Martin, J.M., 2006. Jumping to conclusions and delusion-proneness: the impact of emotionally salient stimuli. Journal of Nervous and Mental Disease 194, 760–765. Westermann, S., Kesting, M., Lincoln, T.M., 2012. Being deluded after being excluded? How emotion regulation deficits in paranoia-prone individuals affect state paranoia during experimentally induced social stress. Behavior Therapy 43, 329–340. White, L.O., Mansell, W., 2009. Failing to ponder? Delusion-prone individuals rush to conclusions. Clinical Psychology and Psychotherapy 16, 111–124. Zawadzki, J.A., Woodward, T.S., Sokolowski, H.M., Boon, H.S., Wong, A.H.C., Menon, M., 2012. Cognitive factors associated subclinical delusional ideation in the general population. Psychiatry Research 197, 345–349.

An investigation into reasoning biases, mood and cognitive state, and subclinical delusional ideation.

Following research on reasoning and the continuum of delusional ideation, the present study attempted to investigate the impact of different experimen...
502KB Sizes 0 Downloads 5 Views