Loading...
Please wait, while we are loading the content...
Similar Documents
Confidence and Memory Accuracy : Do we know what we think we know ?
| Content Provider | Semantic Scholar |
|---|---|
| Author | Brocka, Bennett |
| Copyright Year | 2011 |
| Abstract | This paper focuses on the confidence/accuracy (CA) relationship in memory. A seemingly predictable correlation between a memory’s accuracy and a person’s subjective confidence in that memory has been shown to be quite fragile. Leippe (1980) provides evidence showing the potential disassociations between confidence and accuracy, noting that they can be altered independently and that suggests lawyers, judges and police officers to be aware of these effects when interviewing witnesses and evaluating evidence. Deffenbacher (1980) put forth a paper arguing that subjective confidence should not be used as a measure at all in court cases. In fact a multitude of newer research now indicates the weakness of subjective confidence and investigates particular methods of altering confidence or accuracy of a memory independently. This separation proves extremely important for legal cases in the matters of eyewitness testimony and even more everyday memory events. While interesting results have indicated the importance of the encoding process (Chua, Rand-Giovannetti, Schacter, Albert, and Sperling, 2004) and the bilateral parietal lobe (Simons, Peers, Mazuz, Berryhill and Olson, 2010) in the Running Head: CONFIDENCE AND MEMORY ACCURACY 3 creation of confidence, the exact mechanism by which confidence is disassociated from accuracy is not known, however studies have discovered much worthy of deeper examination. Confidence and Memory Accuracy: Do we know what we think we know? Confidence subjectively “feels” like indicator of memory accuracy in our daily lives; the more sure we are of a memory’s correctness, generally the more accurate we seem to turn out to be. But this confidence is purely subjective, so to what extent can it be trusted? Confidence in memory accuracy is not only important when remembering day to day details like a person’s name or phone number, but it is still used as a barometer of accuracy in eyewitness testimony, where the outcome is much more important. Change-blindness related studies by Simons and Running Head: CONFIDENCE AND MEMORY ACCURACY 4 Levin (1998) and Levin et al. (2002) serve as striking examples of poor eyewitness accuracy and serve as a good reason to question eye witness accuracy and perhaps more importantly, why inaccuracies occur even when confidence is high. While confidence has a weak predictive effect of accuracy, the effect is more restricted than one might expect. A correlation of +.41 was found between correct identification and confidence in the Sporer et al. study (1995). This however was only found in those subjects who had correct memories; no significant correlation was found between confidence and incorrect guesses, meaning this correlation is largely unhelpful in real world situations where there is no separate objective measure of accuracy. A further understanding of confidence is important in understanding this correlation though, and a number of theories explain confidence differently. A signal detection approach assumes that the distinction is similar to the old/new stimulus distinction, with Stretch & Wixted’s study (1998) suggesting the criteria for confidence changes as accuracy changes; the clearer the distinction between accurate and inaccurate memories, the higher the confidence level. This model fits well with the apparent link between accurate memories and stronger confidence, but does not explain why they might diverge in some situations. Attribution theory is another perspective where confidence is related to attributing a memory to a past event where the more divergent a memory is from the remembered experience the less confidence the person has in the memory. (Jacoby, Kelley, & Dywan, 1989). Both of these theories are grounded in recognition processes however, where a study by Olsson and Juslin (1999) suggests that encoding is important for the creation of confidence, this is the basis for Chua et al.’s 2004 neural imaging study of encoding. To test their hypothesis Chua et al. gathered 16 subjects and had them perform a series of remembering tasks, associating Running Head: CONFIDENCE AND MEMORY ACCURACY 5 names with faces. Later the subjects were tested on the associations with a forced yes/no response and subjects were asked whether they had high or low confidence in their accuracy. Generally subjects were more correct than not, even when reporting low confidence, but the study was more concerned with the cases where high confidence was attributed to an incorrect response. Subjects underwent fMRI imaging during encoding, and the differences between high confidence and low confidence were examined to discern any differences in neural activation. Certain activation in the left frontal gyrus was found to be associated with high confidence regardless of accuracy, suggesting it is important for subjective confidence. Additional activation of the left frontal gyrus was noticed in high confidence incorrect responses, diverging from high confidence correct situations, indicating that this difference during encoding is related to the separation of accuracy and confidence, as was hypothesized. Additionally, certain medial temporal lobe (MTL) activation was associated with only correct and high confidence associations. These results give us a clue as to how memory and confidence dissociate, but they are not a practical means of investigating whether a particular confident memory is also accurate; this could only be inferred from an fMRI image during the encoding process, which is impossible in any practical situation. However Chua et al.’s 2004 study was, in their estimation, the first study to examine neural images during encoding for the purposes of memory confidence, so it is possible more directly useful evidence could be found. However, if the difference in confidence and accuracy is only represented neutrally during encoding, it seems unlikely if a clear neural difference could be distinguished from a neural image taken only during recall. Discouragingly, Running Head: CONFIDENCE AND MEMORY ACCURACY 6 this could mean there is no way to determine whether an already formed memory is accurate or not. Despite the findings regarding the MTL’s involvement in confidence in Chua et al.’s 2004 study , Simons et al.’s 2010 study found significant evidence of the importance of the parietal lobe’s importance in confidence. Simons et al. tested patients with bilateral and unilateral parietal lobe lesions against healthy controls. The bilateral lesion patients were found to have no significant difference in ability to recall accurately in a recall task, but consistently reported significantly lower confidence in their recall relative to controls or unilateral parietal lobe lesion patients. Interestingly, the patients recall accuracy has been consistently similar to controls’ recall ability despite the fact that the lesion areas of the patients closely match the areas activated in the healthy control’s parietal lobe during recall. A possible explanation of the reduced confidence in the bilateral lesion patients is that they would be less confident in their cognitive abilities overall, but Simons et al. dismiss this as the unilateral lesion group did not have this same confidence deficit, despite being matched for lesion size, and in addition the bilateral lesion patients showed normal levels of confidence in recognition tasks. Instead Simons et al. posit that the bilateral lesion patients may have a deficiency of the vivid episodic detail present in healthy controls (also suggested by studies by Lyle and Johnson 2006, 2007). This hints at an important role of the parietal lobe in episodic memory. Shifting from the neurological basis of confidence and accuracy to the functional basis, one of the most significant causes of reduced memory accuracy is misleading post-event information (MPI), as described by Koriat, Goldsmith, and Pansky (2000). In a laboratory MPI paradigm, subjects will be exposed to an event or bit of information, later misled about a Running Head: CONFIDENCE AND MEMORY ACCURACY 7 particular detail (often by asking “leading questions”) and subsequently tested, subjects memories typically show impairment about the detail they were misled about. Interestingly the MPI effect seems strongest when the misleading information is presented in a way that suggests a subject to recall the original event, possibly meaning the misleading detail is integrated with the original memory. MPI is perhaps most dangerous as subjects often associate high confidence with the false memories created due to MPI (Zaragoza and Mitchell 1996). MPI is also interesting in that it appears to have developmental and age related components, suggested both by Koriat et al. (2000) and a study on developmental effects on the CA relationship by Howie and Roebers (2007). Howie and Roebers found many interesting results regarding subjective confidence in children. Children (test groups of 8 and 10 years of age) were generally found to have a poorer internal CA calibration when presented with misleading questions in trials, where adults remained more accurate in their confidence calibrations in these situations. This suggests that the questioning of children must be handled especially carefully and that their confidence ratings might be considered less accurate. Children performed as well as adults in these tests when given only unbiased, not misleading questions, but in real world applications the control over misleading and non-misleading questions cannot be assumed. While the effect of misleading information on confidence appears stronger in children, a misinformation effect in the elderly appears present as well. A study by Jacoy et al... (2005) showed that older adults (age 64 to 87) were ten times more likely to produce a false memory than younger adults (18-27) after being presented with misleading information (this is also discussed in Koriat et al. 2000). Running Head: CONFIDENCE AND MEMORY A |
| File Format | PDF HTM / HTML |
| Alternate Webpage(s) | http://professional.education.uiowa.edu/bbrocka/MemoryEssay.pdf |
| Language | English |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
