SEARCH

SEARCH BY CITATION

Keywords:

  • choice;
  • placebo effect;
  • expectations;
  • decision making;
  • assimilation;
  • positive affect

ABSTRACT

  1. Top of page
  2. ABSTRACT
  3. EXPERIMENT 1
  4. EXPERIMENT 2
  5. GENERAL DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. REFERENCES
  8. Biographies

Every day we use products and treatments with unknown but expected effects, such as using medication to manage pain. In many cases, we have a choice over which products or treatments to use; however, in other cases, people choose for us or choices are unavailable. Does choosing (versus not choosing) have implications for how a product or treatment is experienced? The current experiments examined the role of choice-making in facilitating so-called expectation assimilation effects—or situations in which a person's experiences (e.g., discomfort and pain) are evaluated in a manner consistent with their expectations. In Experiment 1, participants were initially exposed to a baseline set of aversive stimuli (i.e., sounds). Next, some participants were given expectations for two “treatments” (i.e., changes in screen display) that could ostensibly reduce discomfort. Critically, participants were either given a choice or not about which of the two treatments they preferred. Participants in a control condition were not provided with treatment expectations. Results revealed that discomfort experiences assimilated to expectations only when participants were provided with choice. Experiment 2 replicated this finding and provided evidence against the idea that demand characteristics and choice-making unrelated to the core task (i.e., choices without associated expectations) could account for the results. Further, Experiment 2 showed that choosing reduced discomfort because of increased positivity about the treatment. Results are discussed in the context of extant research on choice-making and expectation effects. Copyright © 2013 John Wiley & Sons, Ltd.

Expectations—or future-oriented beliefs that allow individuals to respond to their environments—are critical components of numerous psychological theories and are key determinants of evaluative experiences (Geers & Wellman, 2009; Olson, Roese, & Zanna, 1996). For instance, a voluminous literature has revealed that people's evaluative experiences often assimilate or move toward expectations, such as people reporting pain relief after the application of a placebo treatment with analgesic expectations (Benedetti, 2008; Hodges, Klaaren, & Wheatley, 2000; Lee, Frederick, & Ariely, 2006; Price, Finniss, & Benedetti, 2008; Rief, Hofmann, & Nestoriuc, 2008). However, contemporary research has revealed that, under some conditions, people's expectations do not shape their evaluative experiences at all and, in rare cases, result in contrast effects (Geers & Lassiter, 1999, 2005; Gendolla, Brinkmann, & Scheder, 2008; Hammersley, Finnigan, & Millar, 1998; Hróbjartsson & Gotzsche, 2001; 2004; Shelke et al., 2008; Voudouris, Peck, & Coleman, 1990; Walach, Schmidt, Dirhold, & Nosch, 2002). In the current research, we examined a novel factor that was hypothesized to modulate whether people's evaluative experiences would assimilate or shift toward their expectations. Specifically, we examined whether choosing (versus not choosing) among placebo treatments would be more likely to produce evaluations consistent with expectations—so-called expectation assimilation effects.

Choice-making, expectations, and positive affect

The study of choice-making has enjoyed considerable coverage in psychology over the past six decades. For instance, research has established that choosing can shape preferences (Ariely & Norton, 2008; Brehm, 1956; Harmon-Jones & Mills, 1999; Sharot, Velasquez, & Dolan, 2010), engender positivity (Sharot, De Martino, & Dolan, 2009), direct information processing (Hart et al., 2009; Jonas, Schulz-Hardt, Fischer, & Frey, 2006; Smith, Fabrigar, & Norris, 2008), and increase feelings of personal agency (Leotti, Iyengar, & Ochsner, 2010). Might choice-making also be an important determinant of whether people's experiences assimilate to expectations?

We suggest that the study of choice is logical in the context of expectation effects. Indeed, there are numerous domains in which people make choices between options with unknown but expected effects. For instance, decision makers choose between consumer products, investment opportunities, and medical treatments that vary in their anticipated outcomes. Importantly, the value of choice has become increasingly emphasized over the last several decades, notably in the health domain. For instance, there has been a steadily increasing shift toward patient involvement in healthcare decisions (i.e., “patient-centered medicine” or “shared decision-making”), as evidenced by the proliferation of web-based medical tools and decision aids, self-treatment alternatives, over-the-counter options, and direct-to-consumer medication advertising (Reyes-Ortiz, 1996; Stewart et al., 2003).

Recently, Geers and Rose (2011) provided a framework for understanding the role of treatment choice in the context of the placebo effect, a core example of an expectation assimilation effect (Olson et al., 1996). Placebo effects can be defined as physiological or psychological responses that can be directly attributed to expectations associated with a substance or procedure but not a result of the inherent power of that substance or procedure (Price et al., 2008; Stewart-Williams, 2004). Geers and Rose (2011) outlined several mechanisms to explain why choosing among placebo treatments might increase expectation assimilation effects, compared with conditions where expectations are provided without the availability of choice. Most notably for the current research, they proposed that choice-making should facilitate assimilative responses because of the activation of reward mechanisms in the brain and subsequently experienced positive affect in the context of the placebo treatment. Importantly, there are conceptual reasons that both choice-making and expectations should be connected with a positive affectivity mechanism. First, many placebo models suggest that treatment experiences assimilate to expectations because of the activation of dopaminergic mechanisms in the brain, which are associated with positive affect and reward processing (e.g., da Silva Alves et al., 2011; Diederich & Goetz, 2008; Lin, Wu, Chandra, & Tsay, 1981; Scott et al., 2007, 2008; for reviews, see Atlas, Wager, Dahl, & Smith, 2009; Benedetti, 2008). Interestingly, increasing dopamine also facilitates the construction of positive expectations for the future (Sharot, Shiner, Brown, Fan, & Dolan, 2009). Taken together, these two sets of findings suggest that there is a reciprocal relationship between reward mechanism activation in the brain and positive expectations. Second, the act of choosing has also been associated with increases in positive affect (Sharot et al., 2010; Sharot et al., 2009) and the activation of dopaminergic reward mechanisms (e.g., striatum; Sharot et al., 2009; Sharot et al., 2010; Tricomi, Delgado, & Fiez, 2004; for review, see Leotti et al., 2010).

In sum, both placebo expectations and choosing have been critically linked to positive affect via dopaminergic reward mechanisms. In typical placebo studies, participants are provided with positive expectations for a “treatment” but are not provided with a choice over the treatment. Although provided (i.e., non-chosen) placebo treatments have been shown to produce changes in dopaminergic activation and positive affect, we suggest that choosing among treatments with positive expectancies should produce even stronger expectation assimilation effects for two, interrelated reasons. First, positive affectivity experienced via choice-making and receiving a placebo expectation for a treatment may lead to more favorable evaluations of the treatment itself and the treatment context (although see the General Discussion section for a description of situations where this may not occur). Feeling more favorable about the treatment context should place people in the best position for their experiences to assimilate to the positive expectation. Second, the act of choosing is often seen as an expression of the self (Leotti et al., 2010), at least in Western, independent cultures (Iyengar & Lepper, 1999; Snibbe & Markus, 2005). If the self, which is viewed positively, becomes committed to a treatment via choice, then the selected option should become infused with positivity as well and facilitate the extent to which people will desire that the selected placebo treatment is effective (for related ideas, see Brehm, 1956; Harmon-Jones & Mills, 1999; Jonas et al., 2006; Sharot et al., 2010). Importantly, an influential placebo model (Price & Barrell, 1984; Price, Barrell, & Barrell, 1985; Price et al., 2008) suggests that the positive affect mechanism for expectation effects is strongest when desires for treatment efficacy co-occur with expectations. That is, placebo effects are strongest when expectations and desires for the treatment to work are both strong—and that the combination of strong desires and expectations operate through affect-based mechanisms. Thus, if choice-making engenders a greater desire for the treatment to be successful, this should amplify the placebo expectation effect via positivity experienced within and about the treatment context.

Current research

On the basis of a recent theoretical framework (Geers & Rose, 2011), it is suggested that choice over treatments should produce greater assimilative effects in the context of placebo expectations. Indeed, one recent study by Rose, Geers, Rasinski, and Fowler (2012) provided preliminary evidence in support of this hypothesis. Participants in experimental conditions were provided with two placebo treatments (i.e., inert hand lotions) that could ostensibly reduce pain from cold water. Some participants were given the opportunity to choose the “treatment” that they felt might work best for them, whereas other participants had the treatment selected for them by the experimenter. Participants permitted to choose among the treatments showed more assimilation to the placebo expectation (i.e., lower pain ratings) than did participants who had the treatment chosen for them. The Rose et al. study was the first to demonstrate that choice plays an important role in shaping placebo expectation effects.

Despite providing a useful preliminary investigation, there are several limitations to the Rose et al. study that necessitate follow-up research. First, the study was conducted in the domain of pain from cold water exposure. Although this finding is expected to have broad theoretical reach, it is nevertheless important to confirm that the results can be generalized to other domains, expectations, and experiences. Second, Rose et al. did not examine whether the impact of choice-making on expectation assimilation effects was mediated by changes in positive affect experienced in the treatment context. Third and finally, Rose et al. included only three conditions in their study: a condition where participants chose among two placebo treatments, a condition where participants had one of two placebo treatments selected for them by the experimenter, and a condition where participants did not have a choice or a placebo treatment. This design, with the absence of a choice/no expectation condition, opens up the possibility that simply having any choice—absent of expectations associated with the choice—might be sufficient to produce this pattern of results on its own. That is, given that making a choice can enhance perceived feelings of control, and given that control is related to reduced stress, discomfort, and pain (Maier, Laudenslager, & Ryan, 1985; Thompson, 1981), it is plausible that choice alone could reduce some discomfort (Rokke & Lall, 1992). Ruling out this alternative explanation is critical for theoretical and practical purposes.

In the current research, two experiments examined the role of choice-making in facilitating expectation assimilation effects using a novel paradigm. To first establish the usefulness of our paradigm, Experiment 1 was conducted as a replication of the three core conditions in the Rose et al. study. In Experiment 2, we investigated the mediating role of positivity regarding the treatment and included the full factorial design where we orthogonally manipulated expectations and choice-making.

EXPERIMENT 1

  1. Top of page
  2. ABSTRACT
  3. EXPERIMENT 1
  4. EXPERIMENT 2
  5. GENERAL DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. REFERENCES
  8. Biographies

In Experiment 1, all participants first listened to and rated their discomfort from a set of aversive sounds. Participants in experimental conditions were then provided with a cover story involving the influence of visual stimuli on auditory perception. Critically, part of this cover story involved providing expectancies for two different “treatments” (i.e., screen colors) that could ostensibly reduce discomfort when listening to additional aversive sounds. Participants in the choice condition were permitted to choose which of the two treatments they preferred, whereas participants in the no choice condition had the treatment chosen for them. Participants in the control condition were not given an expectation or a choice but were still exposed to the treatment in some form (described later). The critical dependent measure was self-reported change in discomfort from the aversive sounds pre-manipulation to post-manipulation.

Participants and design

The sample consisted of 78 undergraduate students from a large, Midwestern university in the United States who participated in the study as part of a psychology course requirement. Participants were randomly assigned to the choice condition, the no choice condition, or the control condition.

Procedures, manipulations, and measures

Participants came to the lab for a study involving “sound evaluation.” Participants were seated at private computers, prompted to put on headphones, and asked to listen to three unique and unpleasant sounds (derived from Bushman & Baumeister, 1998). After listening to each sound for 5 seconds, participants judged how irritated, uncomfortable, and unpleasant they felt on 7-point scales (1 = not at all; 7 = very). The aggregated ratings for each sound were highly related (α = .95) and averaged for analysis as a baseline measure of discomfort.

Next, participants in the two expectation conditions (the choice condition and the no choice condition) were told that the subsequent part of the study involved an investigation of factors that affect sound perception. A cover story was provided, suggesting that color can have an effect on sensory perception. Participants were told that previous research has shown that green and blue make sounds less unpleasant because people find these particular colors to be relaxing (during funnel debriefing, no participants indicated being suspicious about this claim). Next, the two colors (labeled “Spring Green” and “Sky Blue”) appeared on the screen for 1 minute. Participants in the choice condition were instructed to think about which of these two colors they wanted to use as the background during the next set of sounds. Participants in the no choice condition were told that one of the colors on the screen would be selected for them as their background (participants were randomly assigned to receive either “Spring Green” or “Sky Blue”, although they were not informed about how the selection was made). Participants in the control condition were given no expectations or information about the colors in the context of auditory perception. Instead, participants in this condition had the two colors appear on the screen for 1 minute and were told that the screen would be adjusted to one of these colors during the next set of sounds in order to preserve the life of the computer monitor. It should be noted here that the effect of the actual color presented (“Sky Blue” or “Spring Green”) was analyzed in an analysis of variance (ANOVA) with discomfort as the dependent measure. This analysis did not reveal a main effect of color nor did the color presented interact with choice condition (Fs < 2.3, ps > .10).

Next, using the same procedures described earlier, participants listened to and rated the same three aversive sounds that were used for baseline measures. The aggregated ratings for each sound were highly related (α = .94) and averaged for analysis as a post-manipulation measure of discomfort. Finally, participants answered follow-up questions about the experiment and were debriefed, thanked, and dismissed.

Results and discussion

The critical analysis for our purposes involved how the discomfort ratings differed as a function of measurement timing and condition. The aggregated discomfort ratings (Table 1) were submitted to a 2 (Measurement Timing: baseline or post-manipulation) × 3 (Condition: choice, no choice, or control) mixed-model ANOVA with the first factor as a repeated measure. First, there was a main effect of measurement timing, F(1, 75) = 4.95, p < .05, where discomfort ratings were lower after the manipulation (M = 4.09; SD = 1.54) than at baseline (M = 4.36; SD = 1.51). The main effect for condition was not significant, F(2, 75) = .88, p > .40. However, most important for our purposes was that there was a significant Measurement Timing × Condition interaction, F(1, 75) = 6.45, p < .01. Post-hoc tests showed that discomfort ratings did not differ from baseline to post-manipulation for the control condition or the no choice condition (ps > .50; ds < .19). However, for the choice condition, discomfort ratings were significantly lower post-manipulation (M = 3.85; SD = 1.55) than at baseline (M = 4.98; SD = 1.24; p < .01; d = 1.06). Taken together, these results suggest that participants given a choice showed more assimilation to the placebo expectation (i.e., reduced discomfort) than did participants not given a choice (and participants in a control condition), replicating the core findings of Rose et al. and confirming the usefulness of this paradigm.

Table 1. Discomfort ratings as a function of condition and measurement timing in Experiment 1
Measurement timingChoice conditionNo choice conditionControl condition
MSDMSDMSD
  1. Note. Discomfort ratings were aggregated from ratings of how unpleasant, irritated, and uncomfortable (1 = not at all; 7 = very) participants felt upon listening to three different aversive sounds.

Baseline4.681.243.991.604.401.63
Post-manipulation3.851.553.901.354.541.66

EXPERIMENT 2

  1. Top of page
  2. ABSTRACT
  3. EXPERIMENT 1
  4. EXPERIMENT 2
  5. GENERAL DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. REFERENCES
  8. Biographies

Experiment 2 was similar to Experiment 1, with a few key exceptions. First, we included a condition in which participants had a choice but did not have an expectation associated with the choice. As stated previously, the inclusion of this condition is critical to rule out that simply having any choice—absent of expectations associated with the choice—might be sufficient to reduce discomfort on its own. For instance, given that the act of choosing is associated with positive affect (Leotti et al., 2010), making an unrelated choice could reduce discomfort in and of itself. However, as delineated previously, our core argument is that the combination of choice-making and positive expectations should produce the strongest change in discomfort. To test this notion, Experiment 2 used a full factorial design, crossing Choice (choice or no choice) × Expectation (expectation or no expectation).

Second, Experiment 1 demonstrated that participants having choice showed greater assimilation to the expected impact of the placebo treatment. However, Experiment 1 did not demonstrate why this might occur. Experiment 2 examined a possible mediator: positivity about the treatment context. Recall that placebo expectations and choice-making have been linked to positive affect via dopaminergic reward mechanisms (da Silva Alves et al., 2011; Diederich & Goetz, 2008; Lin et al., 1981; Scott et al., 2007, 2008; Sharot et al., 2010; Sharot et al., 2009; Tricomi et al., 2004; for reviews, see Atlas et al., 2009; Benedetti, 2008; Leotti et al., 2010). As such, we hypothesized that participants making a choice about treatments with expected effects would experience more positive affect during the task, which would impact participants' feelings about the treatment context (i.e., the screen color) and facilitate placebo expectation effects.

Third, we wanted to provide evidence to cast doubt that demand characteristics could account for the results of Experiment 1. In particular, it was possible that participants in the choice condition (versus the no choice condition) felt more responsibility to the experimenter or more personal pressure to appear competent because they had to make a choice. If this is the case, it suggests that participants in the choice condition reported experiencing less discomfort than participants in the no choice condition because of external motivations related to demand characteristics and not their actual discomfort. Experiment 2 included several demand characteristic items to examine this possibility.

Fourth and finally, we sought to include an additional dependent measure that might shed light on participants' experiences beyond self-reported discomfort. In particular, we assessed participant interest in using the same screen color again for future trials. We reasoned that if participants in the choice/expectation condition had a more pleasant experience overall with the treatment, then they would be more willing to use the same screen color again for future trials. In short, this provided a measure of whether our manipulations would shape future decisions.

Participants and design

The sample consisted of 81 undergraduate students (44 female) from a large, Midwestern university in the USA who participated in the study as part of a psychology course requirement (one participant did not follow instructions and was removed from the analysis). Participants were randomly assigned to one cell of a 2 (Choice Condition: choice or no choice) × 2 (Expectation Condition: expectation or no expectation) between-subject design.

Procedures, manipulations, and measures

Participants came to the lab for a study involving “Sound Evaluation.” After listening to and rating baseline aversive sounds following the same procedure as Experiment 1 (α = .95 for the aggregated discomfort measures across the three sounds), participants were randomly assigned to one of four experimental conditions. Participants in the two expectation conditions (the choice–expectation condition and the no choice–expectation condition) were provided with the same cover story used in Experiment 1 about the influence of two different colors (“Spring Green” and “Sky Blue”) on sensory perception. As in Experiment 1, participants in the choice–expectation condition were asked which of the two colors they wanted to use during a new set of sounds, whereas participants in the no choice condition were told that one of the colors would be selected for them. Participants in the two control conditions (the choice–no expectation condition and the no choice–no expectation condition) were not provided with expectations about the screen colors. As in Experiment 1, the screen color for these conditions was either changed to blue or green, but the participants were told that this change was to preserve the life of the computer monitor. Participants in the choice–no expectation condition additionally made a choice that was unrelated to color and the sound task. Immediately before the second set of sounds, these participants made a choice of which two shapes (triangles or squares) they wanted to use during an unrelated spatial rotation task that would ostensibly happen later in the study. Participants in the no choice–no expectation condition were told about the spatial rotation task but did not have the opportunity to make a choice about the shapes that would be used during this task.1 As in Experiment 1, it is also notable that the actual color presented did not have a main effect influence on discomfort ratings nor did it interact with our main manipulations (Fs < 1.2, ps > .20).

After the manipulations, participants again listened to and rated the same three aversive sounds (α = .96 for the aggregated measures across the three sounds). Additionally, there were several supplemental measures. First, to assess demand characteristics, participants rated the extent to which they (i) felt like they owed it to the experimenter to report less irritation than actually experienced, (ii) were concerned about looking competent, (iii) were concerned about what the experimenter thought of them, and (iv) were worried that the experimenter would think they made bad decisions (1 = not at all; 7 = very true of me; α = .80). Second, participants were asked whether they would want the same screen color again for another set of sounds (1 = definitely no; 7 = definitely yes). Third, to provide a measure of positivity about the “treatment” context that would apply to all conditions in the experiment, participants judged their feelings about the presented screen color on the following 9-point semantic differential scales: bad–good; negative–positive; risky–safe; unpleasant–pleasant; and flawed–perfect (α = .85). Finally, participants answered follow-up questions about the experiment and were debriefed, thanked, and dismissed.

Results and discussion

Discomfort analyses

As with Experiment 1, the critical analysis for our purposes involved how the discomfort ratings differed as a function of measurement timing and condition. The aggregated discomfort ratings (Table 2) were submitted to a 2 (Measurement Timing: baseline or post-manipulation) × 2 (Choice Condition: choice or no choice) × (Expectation Condition: expectation or no expectation) mixed-model ANOVA with the first factor as a repeated measure. First, there were no main effects of measurement timing (F = 1.98, p > .15), expectation (F = .002, p > .90), or choice (F = .14, p > .80). Also, all two-way interactions were not significant (Fs < 1.70, ps > .20). However, most important for our purposes was that there was a significant Measurement Timing × Choice Condition × Expectation Condition interaction, F(1, 76) = 5.30, p < . 03. As can be seen in Figure 1, follow-up tests showed that only participants in the critical choice–expectation condition—who chose between treatments (colors) with expected effects—showed significant differences from baseline discomfort (M = 4.42; SD = 1.70) to post-manipulation discomfort (M = 4.00; SD = 1.75) (t = 3.62, p < .01; d = .86). However, for all other conditions—the no choice–expectation condition, the choice–no expectation condition, and the no choice–no expectation condition—baseline discomfort ratings were not significantly different from post-manipulation ratings (ts < 1, ps > .40; ds < .19). Taken together, these results suggest that only participants given a relevant choice about treatments with associated expectations showed the change in discomfort. Participants not given choice and participants given an unrelated choice did not show any differences in baseline and post-manipulation discomfort.

Table 2. Discomfort ratings as a function of choice condition, expectation condition, and measurement timing in Experiment 2
Expectation condition and measurement timingChoiceNo choice
MSDMSD
  1. Note. Discomfort ratings were aggregated from ratings of how unpleasant, irritated, and uncomfortable (1 = not at all; 7 = very) participants felt upon listening to three different aversive sounds.

Expectation    
Baseline4.421.703.961.62
Post-manipulation4.001.754.061.46
No expectation    
Baseline3.881.524.421.60
Post-manipulation3.911.574.311.71
image

Figure 1. Discomfort ratings as a function of choice condition, expectation condition, and measurement timing in Experiment 2

Download figure to PowerPoint

Supplementary analyses

First, in an effort to explore a demand characteristic explanation, recall that participants rated several items assessing the extent to which they were concerned about the experimenter's perceptions of them and about appearing competent. The aggregated demand characteristic ratings were submitted to a 2 (Choice Condition: choice or no choice) × 2 (Expectation Condition: expectation or no expectation) ANOVA. If demand characteristics can account for our results, then participants in the choice–expectation condition should have provided the highest ratings overall. However, there were no significant main effects or interactions on this variable (Fs < 1.2, ps > .30). Also, when including the aggregated demand characteristic index as a covariate in our 2 (Measurement Timing: baseline or post-manipulation) × 2 (Choice Condition: choice or no choice) × 2 (Expectation Condition: expectation or no expectation) ANOVA described earlier for the discomfort measure, the significance level of the critical three-way interaction did not change (F = 5.45, p < .05). Taken together, this suggests that demand characteristics are unlikely to fully account for our results.

Second, as a measure designed to get at future decision making, we asked participants whether they would want to use the same screen color again for another set of sounds. Ratings were submitted to a 2 (Choice Condition: choice or no choice) × 2 (Expectation Condition: expectation or no expectation) ANOVA. As with the discomfort ratings, we anticipated a Choice Condition × Expectation Condition interaction where participants in the choice–expectation condition would show high preference for sticking with the same color. First, there was a significant main effect of the choice condition, where choice participants indicated more interest in having the same screen color, F(1, 76) = 4.71, p < .05. The main effect of the expectation condition was marginally significant, where participants receiving an expectation indicated marginally less interest in having the same screen color, F(1, 76) = 3.19, p = .08. Qualifying these main effects, there was a significant Choice Condition × Expectation Condition interaction, F(1, 76) = 6.04, p < .05. For the two expectation conditions, if participants perceived that the treatment was “working,” they would be more inclined to use the same treatment again. Indeed, participants in the choice–expectation condition provided higher ratings (M = 3.94; SD = 2.53) than did participants in the no choice–expectation condition (M = 1.94; SD = 1.51), t(34) = 2.88, p < .01, indicating relatively more willingness to use the same treatment again for future trials. Participants in the two control conditions did not differ in their interest level across the choice–no expectation condition (M = 3.65; SD = 1.79) and the no choice–no expectation condition (M = 3.78; SD = 1.70), t(42) = .23, p > .80, and were generally ambivalent about using the same screen color again. It is notable, although, that our results do not support that participants in the choice–expectation condition had the strongest preference for sticking with the same color. Indeed, ratings in the choice–expectation condition were not significantly different from the two control conditions (ts < .45, ps > .60) and indicated that, even though their evaluative experiences with the aversive sounds were least distressing, participants in this condition were also quite ambivalent about their decision to stay with their original color.

Third and finally, we assessed participants' overall positivity about the treatment context via their feelings about the screen color used during the sound task. The aggregated ratings were submitted to a 2 (Choice Condition: choice or no choice) × 2 (Expectation Condition: expectation or no expectation) ANOVA. As expected, and consistent with the discomfort ratings, there were no main effects (Fs < 1.20, ps > .20), but there was a significant Choice Condition × Expectation Condition interaction, F(1, 76) = 4.74, p < .05. The nature of this interaction was such that participants in the choice–expectation condition had more positive feelings about the color overall (M = 4.80; SD = 1.01) than did participants in the no choice–expectation condition (M = 4.06; SD = 1.22), t(34) = 1.99, p < .05. Participants' ratings in the two no expectation conditions fell somewhere in between and were not significantly different between the choice–no expectation condition (M = 4.40; SD = 0.59) and the no choice–no expectation condition (M = 4.15; SD = 1.08), t(42) = .90, p > .30. Also important was the fact that participants in the choice–expectation condition were more positive overall about the screen color than were participants in the choice–no expectation condition, t(42) = 2.02, p < .05. Given this set of results, the next section more fully examines the role of screen color positivity as a mediator of the effect of choice on discomfort ratings.

Screen color positivity as a mediator

To determine whether the effect of the Choice × Expectation interaction on discomfort ratings was mediated by screen color positivity, a path and bootstrapping analysis were conducted using methods described by Preacher and Hayes (2008). The independent variable for these analyses was the Choice × Expectation interaction term, the dependent variable was the post-manipulation discomfort rating, and the mediator was screen color positivity. Note that these analyses also controlled for the main effects of choice and expectation, as well as baseline discomfort ratings.

  • Step 1: Does the Choice × Expectation interaction term predict discomfort? The first step was to establish that the independent variable (the Choice × Expectation interaction term) predicted the dependent variable (post-manipulation discomfort ratings). Indeed, as already described earlier, the Choice × Expectation interaction was a significant predictor of discomfort (b = −.59, t = −2.07, p < .05).
  • Step 2: Does the Choice × Expectation interaction term predict screen color positivity? The second step was to establish that the independent variable (the Choice × Expectation interaction term) predicted the mediator (positive feelings about the screen color). Indeed, as already described earlier, the Choice × Expectation interaction was a significant predictor of screen color positivity (b = 1.10, t = 2.39, p < .05).
  • Step 3: Does screen color positivity predict discomfort? The third step was to establish that the mediator (positive feelings about the screen color) predicted the dependent variable (post-manipulation discomfort ratings). This analysis showed that more positive feelings were, indeed, associated with less discomfort (b = −.15, t = −2.18, p < .05).
  • Step 4: Was the effect of the Choice × Expectation interaction term on discomfort mediated by screen color positivity? The fourth step tested the mediated path from the Choice × Expectation interaction term (IV) to discomfort ratings (DV) through screen color positivity (mediator). Following Preacher and Hayes (2008), an accelerated-biased-corrected bootstrap analysis was conducted using 5000 resamples. This procedure tests whether a path is different from 0, which is indicated if the confidence interval band does not include 0. Indeed, this bootstrap analysis showed that the mediation path via screen color positivity was significant (95% CI: −.49, −.011). In short, this analysis provides evidence that participants in the expectation–choice condition provided lower discomfort ratings because of their increased positivity about the treatment.
  • Step 5: Was the effect of the Choice × Expectation interaction term on discomfort fully or partially mediated? The final step establishes whether the inclusion of the mediator (screen color positivity) indicates partial or full mediation of the path between the independent variable (the Choice × Expectation interaction term) and the dependent variable (discomfort ratings). Supporting full mediation, when controlling for screen color positivity, there was no longer an association between the Choice × Expectation interaction and discomfort (b = −.42, t = −1.46, p = .15). See Figure 2.
image

Figure 2. Mediation of relationship between the Choice × Expectation interaction (IV) and discomfort (DV) through screen color positivity in Experiment 2. Coefficients were derived from a bootstrap procedure (Preacher & Hayes, 2008). Coefficients in parentheses are for influence of the IV on the DV when controlling for the mediator (**p < .01; *p < .05)

Download figure to PowerPoint

GENERAL DISCUSSION

  1. Top of page
  2. ABSTRACT
  3. EXPERIMENT 1
  4. EXPERIMENT 2
  5. GENERAL DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. REFERENCES
  8. Biographies

The results of the current experiments suggest that choice-making can facilitate expectation assimilation effects. In Experiment 1, participants having choice over which of two treatments they would use in response to an aversive stimulus reported a larger change in discomfort than did participants not having a choice. Participants in a control condition (who did not receive an expectation or a choice) did not show a significant change across the sets of sounds either, suggesting that this result could not be due to habituation to the stimulus. Experiment 2 replicated and extended the findings of Experiment 1 in several ways. First, we experimentally ruled out that simply making any choice would be sufficient to change discomfort; indeed, only when choices were made about treatments with relevant expectations did discomfort ratings decrease. Second, we provided evidence that the results were not due to higher demand characteristics in the choice–expectation condition, such as feeling more pressure to appear competent or appease the experimenter. Third, we showed that participants in the choice–expectation condition not only changed in their feelings of discomfort but were also more inclined to use the same color treatment again in future trials (relative to those in the no choice–expectation condition). Finally, we showed that change in discomfort as a function of the Choice × Expectation interaction was mediated by positivity about the treatment context (i.e., the screen color). This latter result is supportive of a recent theoretical framework by Geers and Rose (2011), suggesting that choice-making and positive expectations should combine to facilitate placebo expectation effects via the activation of positive affectivity and dopaminergic reward mechanisms, which infuse the treatment context.

Limitations and future directions

As with any study, there are limitations to acknowledge. First, given our focus here on one paradigm (treatment expectations for aversive sounds), extrapolation of these findings to other domains, expectations, and experiences should be made cautiously. Although we suggest our results have broader theoretical implications (Geers & Rose, 2011) and do extend to other contexts (e.g., pain analgesia; Rose et al., 2012), future research in conceptually related paradigms would be useful. For instance, it is important to examine these relationships outside the domain of aversive stimuli and discomfort/pain experiences, such as consumer or investment decisions. Also, it would be useful to explore the role of choice-making in shaping nocebo expectation assimilation effects—situations where expectations for an experience are negative rather than positive (e.g., medication side effects).

Second, our use of self-report measures still leaves open the possibility that response biases may account for a portion of our results. For instance, it is possible that participants' discomfort ratings in choice–expectation conditions reflect lay beliefs, as opposed to actual changes in discomfort. Also, to assess our key mediator, we inquired about positivity for the treatment context rather than assessing underlying physiological evidence of the activation of dopaminergic reward mechanisms or positive affectivity. Although we suggest that our inclusion of demand characteristic items helps assuage these concerns somewhat, it is important for future research to use more subtle designs or measures (e.g., physiological and behavioral) to verify that the present findings involving both our core dependent measure (discomfort experiences) and candidate mechanism (positivity) are not due to response biases.

Third, making choices may not always engender assimilation to expectations. For instance, if making choices is stressful or results in negative affect (e.g., when a decision is very complex or if there are too many choices; Burger, 1989; Iyengar & Lepper, 2000), then people's evaluative experiences may contrast away from the expectations and undercut the benefits of choice (Vohs et al., 2008). Future research that explores the impact of choice-making on expectation effects in these types of situations would be extremely valuable.

Implications for extant research and clinical settings

These results have implications for research literatures on choice-making and expectation effects, as well as for clinical settings. First, numerous studies (and meta-analyses) have revealed that evaluative experiences do not always assimilate toward expectations (Hammersley et al., 1998; Hróbjartsson & Gotzsche, 2001, 2004; Shelke et al., 2008; Voudouris et al., 1990; Walach et al., 2002), and contemporary research has sought to identify conditions that moderate when expectations will and will not produce assimilation effects (versus contrast or null effects). For example, null (or contrast) effects are likely when large amounts of stimulus information are extracted or experienced (Geers & Lassiter, 1999; 2005), cognitive effort is expended (Gendolla et al., 2008; see also Geers & Lassiter, 2003), and information about the stimulus is excluded from one's self-construal (Markman & McMullen, 2003). Our findings offer evidence for a novel moderating factor and suggest that, in the absence of choice, a provided placebo expectation may not always produce assimilation effects. That is, without the addition of the choice variable, we would have concluded that participants were not responsive to the expectation. It appears that, although participants in choice and no choice conditions received similar expectations regarding the color treatment, participants in the choice condition were in a better position to confirm this expectation. In sum, we suggest that it is important for researchers examining expectation effects to incorporate this ecologically valid variable into their research methodologies and conceptual models, and consider this as an explanation for null expectation effects in their studies.

Second, our results bolster and extend prior research on the psychological consequences of choice-making. Our experiments demonstrate that choice availability guides whether people's experiences are assimilated to placebo expectations. This result is particularly consistent with a few key findings in the choice literature. First, choice-making can shape preferences via mechanisms that engender a chosen option as more positive than a non-chosen option (Ariely & Norton, 2008; Brehm, 1956; Harmon-Jones & Mills, 1999; Sharot et al., 2010). Second, and relatedly, other research has shown that the act of choosing increases general positive affectivity via the activation of dopaminergic reward centers in the brain (Leotti et al., 2010; Sharot et al., 2009; Sharot et al., 2010; Tricomi et al., 2004). Our results are compatible with these findings, wherein participants who chose (versus not) among inert treatments may have been particularly prone to desire that their chosen treatment was effective—hence leading to more positive feelings about the treatment and, subsequently, greater assimilation expectation effects.

Third, the current research and other work (Geers & Rose, 2011; Rose et al., 2012) have implications for clinical settings and for randomized-clinical trials involving placebos. For instance, if most treatments work, in part, because of expectations (Brody & Brody, 2000; Harrington, 1997), then providing choice may be an effective and straightforward way to enhance placebo effects in clinical practice. Moreover, the recent movement toward patient-centered medicine has placed active patient involvement and choice as central components of many healthcare encounters and experiences (Reyes-Ortiz, 1996; Stewart et al., 2003). Consequently, an important disconnect now exists between the passive role assigned to individuals in placebo research and the active role individuals regularly assume in modern health care. Addressing this gap—particularly with respect to how active choice-making can shape (placebo) expectation effects—is critical for refining theoretical models and research paradigms, which will ultimately aid in creating more effective therapies, treatments, interventions, and clinical trials.

Conclusions

Expectations can have important consequences for numerous experiences, ranging from consumer satisfaction to healthiness. The current research suggests that choice-making is a critical determinant of whether people's experiences will assimilate to expectations. Our results highlight that it is important to understand the various consequences that everyday choice-making can have on our evaluative experiences.

ACKNOWLEDGEMENTS

  1. Top of page
  2. ABSTRACT
  3. EXPERIMENT 1
  4. EXPERIMENT 2
  5. GENERAL DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. REFERENCES
  8. Biographies

We thank J. Frank Yates and two anonymous reviewers for their helpful comments on a previous version of this manuscript. We also thank Brad Bushman, Kip Williams, and Douglas Geers for their assistance with the audio files.

  1. 1

    Some readers may wonder why we used an unrelated set of stimuli (i.e., shapes) rather than using the same stimuli (i.e., colors) for our two control conditions. Our reasoning was that people have lay beliefs about the impact of color on subjective experiences (in fact, this was one reason why color was chosen as the primary stimuli for our paradigm in the first place). Thus, we were concerned that participants in a choice–no expectation condition using the color paradigm could generate their own expectations about why they were making the choice, potentially contaminating our manipulation and making interpretation of the results difficult.

REFERENCES

  1. Top of page
  2. ABSTRACT
  3. EXPERIMENT 1
  4. EXPERIMENT 2
  5. GENERAL DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. REFERENCES
  8. Biographies
  • Ariely, D., & Norton, M. I. (2008). How actions create—not just reveal—preferences. Trends in Cognitive Sciences, 12, 1316.
  • Atlas, L. Y., Wager, T. D., Dahl, K., & Smith, E. E. (2009). Placebo effects. In J. T. Cacioppo, & G. G. Berntson (Eds.), Handbook of neuroscience for behavioral psychologists (pp. 12361259). Hoboken, NJ: John Wiley & Sons, Inc.
  • Benedetti, F. (2008). Mechanisms of placebo and placebo-related effects across diseases and treatments. Annual Review of Pharmacology and Toxicology, 48, 3360.
  • Brehm, J. W. (1956). Post-decision changes in the desirability of choice alternatives. Journal of Abnormal and Social Psychology, 52, 384389.
  • Brody, H., & Brody, D. (2000). The placebo response. New York, MY: Harper Collins Publishers.
  • Burger, J. M. (1989). Negative reactions to increases in perceived personal control. Journal of Personality and Social Psychology, 35, 351363.
  • Bushman, B. J., & Baumeister, R. F. (1998). Threatened egotism, narcissism, self-esteem, and direct and displaced aggression: Does self-love or self-hate lead to violence? Journal of Personality and Social Psychology, 75, 219229.
  • da Silva Alves, F., Scmitz, N., Figee, M., Abeling, N., Hasler, G., van der Meer, J., … van Amelsvoort, T. (2011). Dopaminergic modulation of the human reward system: A placebo-controlled dopamine depletion fMRI study. Journal of Psychopharmacology, 25, 538549.
  • Diederich, N. J., & Goetz, C. G. (2008). The placebo treatments in neurosciences: New insights from clinical and neuroimaging studies. Neurology, 71, 677684.
  • Geers, A. L., & Lassiter, G. D. (1999). Affective expectations and information gain: Evidence for assimilation and contrast effects in affective experience. Journal of Experimental Social Psychology, 35, 394413.
  • Geers, A. L., & Lassiter, G. D. (2003). Need for cognition and expectations as determinants of affective experiences. Basic and Applied Social Psychology, 25, 313325.
  • Geers, A. L., & Lassiter, G. D. (2005). Affective assimilation and contrast: Effects of expectations and prior stimulus exposure. Basic and Applied Social Psychology, 27, 143154.
  • Geers, A. L., & Rose, J. P. (2011). Treatment choice and placebo expectation effects. Social and Personality Psychology Compass, 5, 734750.
  • Geers, A. L., & Wellman, J. A. (2009). Expectations. In D. Sanders, & K. R. Scherer (Eds.), Oxford companion to emotion and the affective sciences. (pp. 164165). New York: Oxford University Press.
  • Gendolla, G. H., Brinkman, K., & Scheder, D. (2008). Ego involvement moderates the assimilation effect of affective expectations. Motivation and Emotion, 32, 213220.
  • Hammersley, R., Finnigan, F., & Millar, K. (1998). Verbal expectancies and performance after alcohol. Addictive Behaviors, 23, 489496.
  • Harmon-Jones, E., & Mills, J. (Eds.) (1999). Cognitive dissonance: Progress on a pivotal theory in social psychology. Washington, D.C.: American Psychological Association.
  • Harrington, A. (1997). The placebo effect: An interdisciplinary exploration. Cambridge, MA: Harvard University Press.
  • Hart, W., Albarracín, D., Eagly, A. H., Brechan, I., Lindberg, M. J., & Merrill, L. (2009). Feeling validated versus being correct: A meta-analysis of selective exposure to information. Psychological Bulletin, 135, 555588.
  • Hodges, S. D., Klaaren, K. J., & Wheatley, T. (2000). Talking about safe sex: The role of expectations and experience. Journal of Applied Social Psychology, 30, 330349.
  • Hróbjartsson, A., & Gotzsche, P. C. (2001). Is the placebo powerless? An analysis of clinical trials comparing placebo with no treatment. The New England Journal of Medicine, 344, 15941601.
  • Hróbjartsson, A., & Gotzsche, P. C. (2004). Is the placebo powerless? Update of a systematic review with 52 new randomized trials comparing placebo with no-treatment. Journal of International Medicine, 256, 91100.
  • Iyengar, S., & Lepper, M. (2000). When choice is demotivating: Can one desire too much of a good thing? Journal of Personality and Social Psychology, 79, 9951006.
  • Iyengar, S. S., & Lepper, M. (1999). Rethinking the value of choice: A cultural perspective on intrinsic motivation. Journal of Personality and Social Psychology, 76, 349366.
  • Jonas, E., Schulz-Hardt, S., Fischer, P., & Frey, D. (2006). Biased information seeking after decisions: Cognitive and motivational biases. In D. Frey, H. Mandl, & L. von Rosenstiel (Eds.), Knowledge and action (pp. 2351). Ashland, OH, US: Hogrefe & Huber Publishers.
  • Lee, L., Frederick, S., & Ariely, D. (2006). Try it, you'll like it: The influence of expectation, consumption, and revelation on preferences for beer. Psychological Science, 17, 10541059.
    Direct Link:
  • Leotti, L. A., Iyengar, S. S., & Ochsner, K. N. (2010). Born to choose: The origins and value of the need for control. Trends in Cognitive Sciences, 14, 457463.
  • Lin, M. T., Wu, J. J., Chandra, A., & Tsay, B. L. (1981). Activation of striatal dopamine receptors induces pain inhibition in rats. Journal of Neural Transmission, 51, 21322.
  • Maier, S. F., Laudenslager, M. L., Ryan, S. M. (1985). Stressor controllability, immune function, and endogenous opiates. In F. R. Brush, & J. B. Overmier (Eds.), Affect, conditioning, and cognition: Essays on the determinants of behavior (pp. 183201). Hillsdale, NJ: Lawrence Erlbaum Associates.
  • Markman, K. D., & McMullen, M. N. (2003). A reflection and evaluation model of comparative thinking. Personality and Social Psychology Review, 7, 244267.
  • Olson, J. M., Roese, N. J., & Zanna, M. P. (1996). Expectancies. In E. T. Higgins, & A. W. Kruglanski (Eds.), Social psychology: Handbook of basic principles (pp. 211238). New York: The Guilford Press.
  • Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods, 40, 879891.
  • Price, D. D., & Barrell, J. J. (1984). Some general laws of human emotion: Interrelationships between intensities of desire, expectation, and emotional feeling. Journal of Personality, 52, 389409.
  • Price, D. D., Barrell, J. E., & Barrell, J. J. (1985). A quantitative-experiential analysis of human emotions. Motivation and Emotion, 9, 1938.
  • Price, D. D., Finniss, D. G., & Benedetti, F. (2008). A comprehensive review of the placebo effect: Recent advances and current thought. Annual Review of Psychology, 59, 565590.
  • Reyes-Ortiz, C. A. (1996). Patient-centered medicine. Journal of the American Medical Association, 275, 11561166.
  • Rief, W., Hofmann, S. G., & Nestoriuc, Y. (2008). The power of expectation—Understanding the placebo and nocebo phenomenon. Social and Personality Psychology Compass, 2, 16241637.
  • Rokke, P. D., & Lall, R. (1992). The role of choice in enhancing tolerance to acute pain. Cognitive Therapy and Research, 16, 5365.
  • Rose, J. P., Geers, A. L., Rasinski, H. M., & Fowler, S. L. (2012). Choice and placebo expectation effects in the context of pain analgesia. Journal of Behavioral Medicine, 35, 462470.
  • Scott, D. J., Stohler Egnatuk, C. M., Wang, H., Koeppe, R. A., & Zubieta, J. K. (2007). Individual differences in reward responding explain placebo-induced expectations and effects. Neuron, 55, 325336.
  • Scott, D. J., Stohler, C. S., Egnatuk, C. M., Wang, H., Koeppe, R. A., & Zubieta, J. K. (2008). Placebo and nocebo effects are defined by opposite opioid and dopaminergic responses. Archives of General Psychiatry, 65, 220231.
  • Sharot, T., De Martino, B., & Dolan, R. J. (2009). How choice reveals and shapes expected hedonic reaction. Journal of Neuroscience, 29, 37603765.
  • Sharot, T., Shiner, T., Brown, A. C., Fan, J., & Dolan, R. J. (2009). Dopamine enhances expectation of pleasure in humans. Current Biology, 19, 20772080.
  • Sharot, T., Velasquez, C. M., & Dolan, R. J. (2010). Do decisions shape preference? Evidence from blind choice. Psychological Science, 21, 12311235.
  • Shelke, A. R., Roscoe, J. A., Morrow, G. R., Colman, L. K., Banerjee, T. K., & Kirshner, J. J. (2008). Effect of a nausea expectancy manipulation on chemotherapy-induced nausea: A University of Rochester cancer center community clinical oncology program study. Journal of Pain and Symptom Management, 35, 381387.
  • Smith, S. M., Fabrigar, L. R., & Norris, M. E. (2008). Reflecting on six decades of selective exposure research: Progress, challenges, and opportunities. Social and Personality Psychology Compass, 2, 464493.
  • Snibbe, A. C., & Markus, H. R. (2005). You can't always get what you want: Educational attainment, agency, and choice. Journal of Personality and Social Psychology, 88, 703720.
  • Stewart, M., Brown, J. B., Weston, W. W., McWhinney, I. R., McWilliam, C. L., & Freeman, T. R. (2003). Patient-centered medicine: Transforming the clinical method (2nd ed). Abingdon, UK: Radcliffe Medical Press Ltd.
  • Stewart-Williams, S. (2004). The placebo puzzle: Putting together the pieces. Health Psychology, 23, 198206.
  • Thompson, S. C. (1981). Will it hurt less if I can control it? A complex answer to a simple question. Psychological Bulletin, 90, 89101.
  • Tricomi, E. M., Delgado, M. R., & Fiez, J. A. (2004). Modulation of caudate activity by action contingency. Neuron, 41, 281292.
  • Vohs, K. D., Baumeister, R. F., Schmeichel, B. J., Twenge, J. M., Nelson, N. M., & Tice, D. M. (2008). Making choices impairs subsequent self-control: A limited-resource account of decision making, self-regulation, and active initiative. Journal of Personality and Social Psychology, 94, 883898.
  • Voudouris, N. J., Peck, C. L., & Coleman, G. (1990). The role of conditioning and expectancy in the placebo response. Pain, 43, 121128.
  • Walach, H., Schmidt, S., Dirhold, T., & Nosch, S. (2002). The effects of a caffeine placebo and suggestion on blood pressure, heart rate, well-being, and cognitive performance. International Journal of Psychophysiology, 43, 247260.

Biographies

  1. Top of page
  2. ABSTRACT
  3. EXPERIMENT 1
  4. EXPERIMENT 2
  5. GENERAL DISCUSSION
  6. ACKNOWLEDGEMENTS
  7. REFERENCES
  8. Biographies
  • Jason Rose is an Assistant Professor of Psychology at the University of Toledo. His research interests include social comparison, risk perception, norm communication, and the consequences of choice. He holds a PhD from the University of Iowa and a BA from Purdue University at Fort Wayne.

  • Andrew Geers is a Professor of Psychology at the University of Toledo. His research focuses on health behavior, optimism, assimilation and contrast effects, and the psychological components of analgesia and medical treatments. He received his BA from the University of Cincinnati and his PhD from Ohio University.

  • Stephanie Fowler is a doctoral candidate in the Psychology program at the University of Toledo. Her research is centered on the social psychological processes in responses to health threats and behavior change. She holds a BS from the University of North Florida and an MA from the University of Toledo.

  • Heather Rasinski is a doctoral candidate in the Psychology Department at the University of Toledo. Her research interests include prejudice, gender, choice, and health behavior. She earned her BA at the University of Akron and her MA at the University of Toledo.