magazine issue 2 2012 / Issue 14

It’s your choice! – Or is it really?

• written by Anna Sagana, Melanie Sauerland & Harald Merckelbach
• edited by Ashley Waggoner Denton

From Change Blindness to Choice Blindness

Given the difficulty people have in noticing changes to visual stimuli, one may wonder what would happen if these changes concerned the decisions people make. For example, imagine that you are sitting outdoors on a sunny spring day and hear that the person sitting next to you orders black coffee. The waiter comes back with a cappuccino. Do you think you would notice this? Drawing from the findings on change blindness, there is reason to believe that you would not. What, however, if it were you who ordered black coffee and ended up with a cappuccino? Obviously, there is a difference between passively observing the scene and being actively involved in decision making and ordering. In the latter case, it is your intentions and actions that are targeted. This should make a difference – shouldn’t it?

To examine whether choice blindness also occurs in other situations, Hall and colleagues (2010) invited supermarket customers to sample two different kinds of jams and teas. After participants had tasted or smelled both samples, they indicated which one they preferred. Subsequently, they were purportedly given another sample of their preferred choice. On half of the trials however, these were samples of the non-chosen jam or tea (see http://www.youtube.com/watch?v=_VPcl04Adh8). Again, only about one-third of the participants detected this manipulation. Based on these findings, Hall and colleagues (2010) proposed that choice blindness is a phenomenon that occurs not only for choices involving visual material, but also for choices involvinggustatory and olfactory information. Recently, the phenomenon has also been replicated for choices involving auditory stimuli (Sauerland, Sagana, & Otgaar, 2012). Specifically, participants had to listen to three pairs of voices and decide for each pair which voice they found more sympathetic or more criminal. The voice was then presented again, however, the outcome was manipulated for the second voice pair and participants were presented with the non-chosen voice. Replicating the findings by Hall and colleagues, only 29% of the participants detected this change.

Merckelbach, Jelicic, and Pieters (2011b, Experiment 2) investigated choice blindness for intensity ratings of one’s own psychological symptoms. Their participants had to rate the frequency with which they experienced 90 common symptoms (e.g., anxiety, lack of concentration, stress, headaches etc.) on a 5-point Likert scale. Prior to a follow-up interview, the researchers inflated ratings for two symptoms by two points. For example, when participants had rated their feelings of shyness, as 2 (i.e.,occasionally), it was changed to 4 (i.e., all the time). This time, more than half (57%) of the 28 participants were blind to the symptom rating escalation and accepted it as their own symptom intensity rating. This demonstrates that blindness is not limited to recent preference selections (Hall et al., 2010; Johansson et al. 2005), but can also occur for intensity and frequency ratings that rely on long-term episodic memory.

Inspired by these findings, Sauerland, Schell, Collaris, Reimer, Schneider, and Merckelbach (2012) tested blindness for one’s own history of norm-violating behaviors. Specifically, they explored whether choice blindness might be relevant to suspects’ false confessions. Participants were presented with an 18-item questionnaire about their history of norm violation which asked them to rate how often they had displayed each behavior on a 4-point Likert scale. Participants were later interviewed about four of those items, two of which had covertly been manipulated. For example, the response to the item “I have never committed small-scale shop lifting” was changed from never to sometimes. As expected, a substantial proportion (25%) of participants failed to notice changes of their own ratings. Importantly, detection varied as a function of the interval between completing the questionnaire and being interviewed about the manipulated outcome. Participants in Experiment 1 who were presented with the manipulated outcome after 10 minutes were more likely to notice the change in their response compared to Experiment 2 participants who were presented with the manipulation after a week (75% and 55%, respectively). Thus, the longer interval resulted in a decrease of the detection rate.