Eyewitness Reliability involving Fires
Eyewitness testimony is often one of the few pieces of information available to a fire investigator. This information may be useful in determining the area of origin, which is a crucial step in determining the cause of the fire. Opinions developed from eyewitness testimony can be presented as expert testimony, which is often a controlling factor in both civil and criminal litigation. Moreover, jurors often apply a significant weight to the value of eyewitness testimony. Thus, it is important for the fire investigator to be able to address its reliability in a manner that is consistent with the scientific method.
While fire investigators are trained on interpretation of burn patterns, they are not generally trained on evaluating the reliability of eyewitness testimony. One of the most common training manuals, referred to as a guide in fire investigation, is NFPA 921. This guide details how to conduct a fire investigation. In determining the fire’s origin, NFPA 921 urges the fire investigator to evaluate witness observations. For instance,
220.127.116.11, “Observations by witnesses are data that can be used in the context of determining the origin.”
NFPA also states,
18.104.22.168, “It is the responsibility of the investigator to evaluate the quality of the data obtained from the witness at the time of the interview”.
This manual has been in existence since 1992, is on its ninth edition, and yet there is no assistance on how to evaluate eyewitness testimony using a scientific approach. Top fire investigation references, e.g. Kirk’s Fire Investigation by DeHaan, Scientific Protocols for Fire Investigation by Lentini, and Fire Investigation by Daeid also do not offer any instruction on how to evaluate eyewitness testimony. And so, the fire investigator is left to their own devices, often relying on personal experiences.
Fortunately, there has been much interest in evaluating the reliability of eyewitness testimony in criminal cases. Beginning in the mid- to late 1970’s, cognitive psychologists have conducted experiments to determine the factors that affect memory and the extent to which errors occur. While these scholarly works offer much help, they are not in a venue frequented by fire investigators. There is no authoritative text that culls together the most pertinent findings related to eyewitness accounts and fire. Moreover, it is important to include tools related to assessing eyewitness credibility involving one of the most influential factors that can affect their perception, i.e. smoke.
This work is aimed at providing the tools that are needed by fire investigators to evaluate eyewitness testimony based on scientific literature rather than just their personal experiences. Peer-reviewed published articles by leaders in the fields of cognitive psychology, neuroscience, and sociology form the foundation for this work. While entire review articles have been written for each topic that will be presented, this review will focus on landmark studies or review articles in attempt to provide an overview and the most likely accepted view amongst scholars in each of these fields.
To start, this work will identify the few studies that have directly examined eyewitness reliability involving fires. Next, the act of witnessing an event that will lead to eyewitness testimony will be divided into three stages: acquisition, storage, and retrieval. At each stage, important factors will be identified and discussed. Overlap of factors into neighboring stages may be possible, i.e. each factor might not exist solely in that stage. Separation of these factors into their respective stages is presented merely as a framework for learning. Some factors might even span all three stages. Lastly, this work identifies successful interview techniques, which can be used at any point following the event in question to assess an eyewitness’s credibility.
Findings & Discussion
Eyewitness Reliability & Fire
There are only a handful of studies that have examined eyewitness reliability involving fires. One of the first, if not the first, was the report from Fennel in 1988 on the King’s Cross Underground Fire (Fennel – 1988). On November 18, 1987, a fire involved an underground escalator in London killing and/or injuring many people. Roberts followed Fennel’s report by examining eyewitness accounts of the fire (Roberts – 1992). He found that eyewitnesses focused on one “frame” of their choosing in the fire development sequence and not one witness was able to give a continuous and coherent account of the sequence (Roberts – 1992).
While Roberts’ work seems to indicate that eyewitness testimony involving fires is unreliable, Geiman and Lord found the opposite in that the majority of accounts corroborated the outcome of the fire investigation (Geiman – 2012). In their investigation of an apartment building fire, the majority of eyewitnesses who observed flames were concentrated in the north wing on the second floor indicating that this is the fire’s area of origin. The results of computer modeling, large-scale testing, and a fire scene examination were consistent with the majority opinion of the fire’s area of origin as determined by eyewitness testimony. Geiman and Lord suggest that eyewitness accounts are complimentary to an investigation and allow the investigator to generate testable fire origin hypotheses.
The number of eyewitnesses was relatively large in the work of Geiman and Lord (96 witnesses) and it is unlikely that many fires will interact with more than a few people. English and Kuzel found that the number of witnesses could play a large role in determining the majority opinion (English – 2014). They analyzed 354 witness statements of the flight 587 crash into the Belle Harbor neighborhood in New York City. The majority of eyewitnesses stated that the aircraft was on fire before impact. However, the geographic location of witnesses along the crash path did not correlate with statements of fire/no fire or smoke/no smoke. An investigation revealed that both engines separated from the aircraft before impact and both showed no signs of fire, indicating that there was no pre-impact fire. They used their large population to infer the results from a smaller sample of eyewitness statements. With a sample of five random witnesses, there is a 24% chance that the majority of them would have reported seeing no fire, which is opposite of the population’s majority opinion. They argue that with a sample of 10 witnesses, the chance decreases to 9% but the fact that sample size has profound implications remains important for consideration.
There does not appear to be a consensus as to whether or not eyewitness testimony involving fires is reliable. Some studies suggest that eyewitness accounts are reliable while others suggest the opposite. It is more likely that this determination should be made on a case-by-case basis, rather than a blanket statement one way or another. Thus, it is important to have a solid background on the elements of eyewitness testimony and the important factors influencing its reliability to be able to evaluate eyewitness testimony on its merits.
Witnessing an event
Loftus describes witnessing an event as consisting of three stages: acquisition, retention, and retrieval (Loftus – 2013). First, the witness experiences sensation and perception of the event. This information then enters their memory and together this process is termed acquisition. The second stage is retention, where some amount of time passes before the witness attempts to remember the event. The last stage consists of the witness trying to recall the stored information, termed retrieval. The following sections identify important factors that influence each stage and summarize the research regarding each factor.
The first stage of witnessing an event is acquisition, which begins with sensation and is followed by perception. Sensation is the process whereby our sensory organs detect some external stimuli and transform the stimulus into electrical impulses that is interpreted by the brain. Perception refers to the subjective experience for which the brain organizes, identifies, and interprets a particular sensation. For the brain to process each sensation, it must do this based on individual experience, memory, emotion, and expectation. Because perception is based on the individual, it is not infallible. Optical illusions are but one example, where the visual interpretation of the illusion appears different from reality. That is not to say sensation is infallible either. It too is affected by individual traits and also on the environment that is producing the sensation.
Vision is usually an important piece of sensory information in eyewitness accounts. On a basic level, the eye contains light sensitive photoreceptor cells called rods and cones. Cones are primarily located in the central portion of the retina while rods dominate the periphery. Compared to cones, rods are much more sensitive to light. As such, rods mediate nighttime vision and cones mediate day time vision.
A witness often encounters an event where the lighting changes dramatically during the event. One example could be turning on a light in a dark room to find an intruder in your home. Depending on how long the witness was exposed to dark conditions, it may take a finite amount of time to be able to discern the intruder’s face. This example demonstrates the concept of light adaptation, which occurs when the eye is exposed to light after previous exposure to darkness. Pitts found that light adaptation is mostly complete after a few seconds (Pitts – 1982) while Loftus explains that the cones take approximately 15 seconds to complete their recovery (Loftus 2013).
The opposite of light adaptation is dark adaptation, which occurs when the eye is exposed to a dark environment after previous exposure to light. Pitts found that dark adaptation is complete after 40 minutes for all ages (Pitts – 1982). Robins found that cone adaptation is complete within about 10 minutes, which means that we see better in the central portion of the visual field during the transition from light to dark conditions. (Robins 2009).
Other information that can be useful when assessing the reliability of eyewitness testimony include basic medical information about the witness. For example, do they need eyeglasses and were not wearing them or do they suffer from near/far-sightedness, glaucoma, macular degeneration, color-blindness, night-blindness, and/or cataracts?
A person’s age has been shown to affect both vision and memory. Pitts found that visual acuity and contrast sensitivity decline with age as lens opacity increases and the pupil diameter decreases, resulting in less light reaching the retina (Pitts – 1982). While his work indicated that dark adaptation is complete after 40 minutes for all ages, the threshold light intensity is over 10000 times greater for someone who is 80-89 than for someone who is 20-29 (Pitts – 1982). This means that the time required to see in the dark after being exposed to light is independent of age but elderly adults are less likely to see well in the dark even after adaptation is complete. Pitts also found that Snellen decimal acuity is relatively constant up to age 50 and steadily declines thereafter (Pitts – 1982). The more familiar designation of 20/20 is associated with the Snellen decimal equivalent of 1.0. For example, a person with 20/40, or 0.5 Snellen decimal equivalent, vision would mean that their acuity is half as good as a normal person.
In terms of memory, Kausler and Wiley reported that elderly subjects performed worse on recognition and temporal information memory tests than young adults (Kausler – 1990). In their study, thirty-two young adults between the ages of 18-24 and 32 elderly adults between the ages of 63-82 participated in series of tests on memory. Subjects performed a series of actions, e.g. ringing a bell, cutting a string, etc., and were asked to pick from a list the actions that they just performed. This test was termed a recognition memory test. They were also asked to reconstruct the order of the actions performed, termed a temporal memory test. They then came back the following day and were asked to complete both tests again. The results show that elderly adults could not recognize the tasks or remember their order nearly as well as young adults both immediately after the actions were performed and after the 24 hour retention interval.
Similarly, Uttl and Graf found an age-related decline in episodic spatial memory performance and that this decline starts in a person’s 60’s (Uttl – 1993). Episodic memory pertains to a personally experienced event while semantic memory reflects general knowledge of facts. In their study, 302 subjects participated with at least 30 subjects in each age range (15-24, 25-34, etc.). Subjects entered a room containing paintings, photographs, panels, and table-top exhibits. Items in the room served as targets to be remembered and each item was photographed and put into a booklet for use during the test. Subjects were given a blank floorplan of the room and were asked to locate each item using the booklet. A one-way analysis of variance showed a significant overall effect from the age group.
Since this experiment was conducted in the field and suffered from typical problems of uncontrolled variables, e.g. how crowded the room was or how much time each person spent in the room, the authors conducted a similar study in a more controlled laboratory environment. Regardless of where the experiment was performed, the older adults scored lower than young adults (Uttl – 1993).
Older adults also appear to be more susceptible to false memories than healthy young adults (Balota – 2000). False memories are apparent recollections of events that did not occur and will be discussed in more detail below. Norman and Schacter (1997) and Balota et al. (1999) have suggested that the increased false memories are consistent with age related decreased efficiency of frontal lobe functioning.
Depending on how long the eyewitness has been awake, the reliability of their testimony could be impaired due to a concept called sleep inertia. Jewett et al. found that sleep inertia dissipated in an asymptotic manner and took 2–4 hours to near the asymptote (Jewett – 1999). In other words, it took 2-4 hours for subjects to be at full mental capacity. Jewett el al. also fit exponential functions to the sleep inertia data and found time constants of 0.67 hours for alertness and 1.17 hours for cognitive performance. In their study, alertness was self-reported on a continuous scale and cognitive throughput was measured by asking subjects to complete sums of two 2-digit numbers as quickly and as accurately as possible in 2 minutes. They also found that alertness and cognitive throughput were impaired upon awakening regardless of whether subjects got out of bed, ate breakfast, showered and were exposed to ordinary indoor room light. There does not appear to be a way to speed up the process of sleep inertia dissipation. Lastly, there was no effect of sleep stage at the time of awakening.
The effect of training and/or experience on eyewitness reliability does not appear to be universal. Some believe that police and other trained individuals are better at managing stress during emotion-eliciting events and thus will be better at encoding event details. They argue that police might experience lower physiological or psychological arousal during stressful events than citizens. If this is true, it may be because police work might attract people who are less reactive to stressful events, which does not bode well for the argument that training and/or experience enhances eyewitness reliability. Another explanation may be that police learn to use processing strategies that allow them to manage their stress and still engage in adequate encoding, which would support the idea that training and/or experience are beneficial in terms of eyewitness reliability. The studies outlined below show that police experience either more or less stress than citizens, depending on the study. These studies also show that trained individuals perform either better or the same compared to citizens on recall of event details, again depending on the study. With no clear consensus but still providing valuable information, these studies present the current state of knowledge on the effect of training and/or experience on eyewitness reliability.
Keating and Loftus found that arson investigators, a type of trained individual, were no more accurate in recalling nontechnical details than college-aged students but were more accurate on technical details related to firefighting (Keating – 1981).
Yuille found that police recalled more correct descriptive facts than citizens. In this study, 119 police trainees and 132 university students were shown a slide sequence that depicts an automobile/pedestrian accident and were then asked to provide a written description of what they had just seen. In similar experiment involving 62 experienced police officers with an average of 8 years of experience, he found that they performed similarly to the trainees. Yuille questions if the results were motivation driven, offering the possibility that the police had more motivation to perform than the students. He also acknowledges that the viewing slides weren’t very stressful. As a result, the slides may not have activated the stress-controlling training that police potentially possess (Yuille – 1984).
Lindholm, Christianson, and Karlsson conducted an experiment where 81 police officers and 56 civilians watched a film of a simulated, violent robbery (Lindholdm – 1997). They found that police recalled more details about the perpetrator than did citizen eyewitnesses, offering the explanation that “police experience may provide domain-specific knowledge that facilitates encoding and storing of new information.”
Stanny and Johnson found the opposite result compared to Lindholm, Christianson, and Karlsson (Stanny – 2000). Their work concluded that police and citizen eyewitnesses did not differ appreciably in recall accuracy. The experiment that they employed involved 13 students and 16 police officers who participated in the Firearms Training System, a computerized, interactive weapon-use scenario. Subjects participated (police) or witnessed (citizens) a domestic disturbance call and an attempted abduction and were presented with a shoot and no-shoot scenario for each event. Stanny and Johnson found that police and bystander witnesses did not differ in their ability to recall details. In these experiments access to information was the same between the two groups, meaning that this variable is similar as in a real-world scenario. They do acknowledge that their sample size was not large, which could limit its impact. A major finding was that police experienced more stress than the citizens in the work from Stanny and Johnson while police experienced less stress than citizens in work by Lindholm, Christianson, and Karlsson. Stanny and Johnson argue that stress could be the reason for the disagreement between the two studies.
Loftus concludes that mundane facts or details are not typically better remembered by those with specialized training but certain unique details are remembered more readily by a trained individual (Loftus – 2013, pg. 44).
Research has shown that the way people perceive events is affected by expectations. In the context of eyewitness testimony, the way they perceive the fire event could be influence by news articles, talking with friends and coworkers, or by past experiences with their personal property. In the seminal work by Bugelski and Alampay, one group was shown a series of drawings of human faces and another was shown animals (Bugelski – 1961). Subjects were then shown an ambiguous image that could be either a bald man in profile wearing glasses or a rat. Perceptions of this image depended almost entirely on which drawings they were shown previously.
Balcetis and Dunning found that participants who were shown an ambiguous figure tended to report seeing the interpretation that assigned them to outcomes they favored (Balcetis – 2006). In their work, 88 students were told that they were participating in a taste-testing experiment. A computer would display either a number or letter that corresponded to which of two beverages the student would taste. One beverage was desirable and the other was much less desirable. Students reported seeing the letter/number that corresponded with the more desirable beverage. Their next study examined whether subjects saw both interpretations and simply chose to report the one that resulted in a more favorable outcome. They found that participants truly saw the interpretation they preferred rather than simply reporting the preferred interpretation.
Exposure time is defined as the amount of time that a stimulus is physically available to a witness. In an effort to understand the effect of exposure time on eyewitness reliability, among other things, Shapiro and Penrod conducted a meta-analysis of 128 studies that involved 16,950 subjects (Shapiro – 1986). Their results regarding exposure time weren’t exactly clear. While exposure time had a large effect size for hits, as the duration of exposure per face increased so did the false-alarm rate. Shapiro and Penrod thought that this was anomalous and could be due to a confounding variable that was not been coded.
Bothwell argues that longer exposure durations permit more efficient processing of faces to which eyewitnesses have been exposed (Bothwell – 1987). But, Wells adds that the amount of time a culprit’s face is in view is not as critical for eyewitness identification accuracy as the type or amount of attention given by the witness (Wells – 2003). For example, Leippe exposed unsuspecting people to a staged theft of a package (Leippe – 1978). Some were led to believe that the package contained a trivial item and other others were led to believe that the package contained a valuable item. In addition, some learned of the package’s value before the theft and others learned it afterwards. Although all had the same opportunity to view the thief, the witnesses who knew the value of the item beforehand were significantly more accurate at identification than the other three groups.
Weapon focus refers to the concentration of a witness’s attention on a weapon during a crime, leaving less attention available for viewing other items. Loftus, Loftus, and Messo conducted an experiment where 36 subjects viewed a series of slides in which a customer moves through a line at a fast-food restaurant (Loftus – 1987). In one version, the customer either points a gun at the cashier and receives money in return or hands the cashier a check and again receives money in return. Following a 15-minute retention interval, subjects were asked 20 multiple-choice questions regarding the details of the scene. These researchers found that subjects made more eye fixations on the weapon versus the non-weapon and the duration was longer. This study was replicated but with more (80) subjects and only a 7-item questionnaire about just the customer. The results indicated that the memory of subjects about the customer in the weapon condition was poorer than the non-weapon condition.
Simons explains that without attention, we may not even perceive objects, termed inattentional blindness (Simons – 1999). The likelihood of noticing an unexpected object depends on the similarity of that object to other objects in the display and on how difficult the monitoring task is. He also explains that the spatial proximity of the critical unattended object to attended locations does not appear to affect detection, suggesting that observers attend to objects and events, not spatial positions.
Recently, Fawcett, Russell, Peace, and Christie performed a meta-analysis on 28 studies involving weapon focus and argue that the weapon focus effect should be called the ‘salient feature focus effect’ since unusual objects create an analogous effect as a weapon (Fawcett – 2013).
There have been many studies that show that stress, anxiety, arousal, or negative emotionality have all yielded effects on memory. Negative emotional events refer to new, unexpected, and potentially threatening events that can cause emotional stress. Other researchers have used words like violent, shocking, or traumatic to describe these events. Emotional stress is a state that follows these events.
Christianson argues that eyewitness memory for stressful emotional events should be understood in terms of complex interactions between type of event, type of detail information, time of test, and retrieval conditions (Christianson – 1992). The event type can be emotional or neutral, the type of detail information can be central or peripheral, and the time of test could be immediate or delayed. Retrieval conditions can be in the presence of sufficient retrieval cues, mood, or other context cues, repeated testing, and so on. Christianson believes that there is not a simple relationship between intense emotion and memory, that is, the more negative the emotion or stress, the poorer the memory. Central, critical detail information of the negative emotion-eliciting event are retained relatively well – see the section on weapon focus. Also, this information is less susceptible to forgetting compared to similarly detailed but neutral information.
Deffenbacher et al claim that Christianson (1992) reviewed many studies that elicited orienting responses, rather than defensive responses, which would generate facilitation of memory for central details (Deffenbacher – 2004). An orienting response is a physiological reaction that results in a deceleration of heart rate, lowered blood pressure and muscle tone, and an increase in skin conductance. This type of response has been linked to tasks involving simple perceptual intake, i.e. a non-threatening event. Instead, Deffenbacher et al (2004) focused on experimental manipulations productive of defensive responses to stimulating conditions. A defensive response is the physiological response that results in acceleration in heart rate, increased blood pressure and muscle tone and is characterized by a readiness for action. They performed a meta-analysis on 36 independent tests of the effects of stress on eyewitness recall of details associated with the crime and found that high levels of stress negatively impact accuracy of crime-related details.
Reisberg and Hertel suggest that people may differ in how they react to trauma, which could explain the mixed results (Reisberg – 2004). For example, persons who are anxious and preoccupied as measured by standard tests, performed more poorly on a test measuring eyewitness accuracy (Siegel – 1978).
Levine and Edelstein summed up the current state of knowledge, offering that people typically show good memory for central features of emotional events and poorer memories for peripheral features, termed “memory narrowing” (Levine – 2009). However, defining “central” is the key issue. Possible definitions include a) information that captures an emotionally aroused person’s attention, b) information that constitutes an integral part of an emotional stimulus, or c) information relevant to goals. Memory narrowing as a result of emotion, and violations of the memory-narrowing pattern can best be explained by the view that emotion enhances memory for information relevant to currently active goals, or definition c). They do acknowledge, however, that difficulty lies in determining a person’s goals a priori.
Length and time estimates
Eyewitnesses are often asked to estimate their proximity to certain objects or the time it took for events to transpire. As example of estimating distances, Harte found that subjects grossly underestimated the length of both the guidelines and the spaces between guidelines issued on highways as tested by memory and under actual driving conditions (Harte – 1975). Loftus, Schooler, Boone, and Kline tested the ability of people to estimate time durations during a stressful event (Loftus – 1987). In their experiment, 469 subjects watched a short videotape of a bank robbery and later estimated the duration of the event. The results indicated that subjects invariably overestimated the duration. They also found that a more stressful version of the event produced greater overestimates than a less stressful version.
Speed may also be a quantity that an eyewitnesses is asked to estimate. Foster and Naylor conducted a study with 50 adults (Group 1) that were shown a short video of traffic movement (Foster – 1999). On a 2-lane roadway, there were two vehicles traveling at 37 MPH towards a fixed camera and one vehicle that traveled away. There was pedestrian movement around a parked red van on the opposite side of the road. Another set of 50 adults (Group 2) were shown the same video but with an additional 3 seconds where the camera is turned 180 degrees and shows the last vehicle skidding to a halt at an unusual angle near the curb. In addition, an object can be seen on the pavement slightly ahead of the stationary car. Each group was asked a series of questions immediately after the video. Ninety percent of Group 1 reported that the last vehicle seen was traveling at a similar speed to the other vehicles (10% had no opinion) while 89% of Group 2 reported that the last vehicle seen (the one skidding to a halt) seemed to be going faster than the other vehicles. Foster and Naylor argue that any ‘adjustment’ in the speed estimate of a driver who is considered by the lay-witness to be at fault is upwards rather than downwards.
This concludes the ‘acquisition section’ that may be categorized as stemming from pure psychology studies. In other words, the factors discussed so far have not been a product of fire. Rather, these factors are not unique to fires though there is no reason to believe they cannot be applicable to situations involving fire. Next, we will look at studies of how smoke and other combustion products can affect the acquisition phase of witnessing an event.
During the early stages of a compartment fire, when burning is localized, the combustion products (smoke) will be progressively diluted as they rise vertically in a buoyant plume until deflected by the ceiling. The hot smoke will then flow as a ceiling jet until it encounters a barrier, e.g. a wall. The smoke layer will deepen at a rate dependent on the rate of burning and also on the amount of air entrained into the plume before it enters the layer (Drysdale – 2011). This smoke layer will obscure the witness’s field of view and may impair their ability to discern objects or estimate distances.
There are two reasons for the decreased visibility through smoke: (1) luminous flux from a source are interrupted by smoke particles, and (2) luminous flux scattered from general lighting by smoke particles is superimposed on the reduced fluxes mentioned in (1) (Jin – 2008). The human eye can distinguish a luminous source from the background in smoke only when the difference in intensity between the flux from the source and that from the background is larger than some threshold value. However, the threshold value varies with the intensity of luminous flux from the background and the properties of smoke.
It may be useful to estimate the visibility in a smoky environment to assess the reliability of a particular eyewitness. Visibility can be estimated using the following procedure as outlined by Drysdale (Drysdale 2011). First, identify the principal fuel involved. Next, look up the smoke potential (units are ob m3/g) for the principal fuel involved under flaming or non-flaming conditions. Then, estimate the mass flow rate of smoke as a function of the heat release rate and the height above the fire. Lastly, calculate the optical density per meter (ob)
=QċD0ρ∆Hcṁand use the correlation from Butcher and Parnell (1979) to relate optical density per meter to visibility (m). It’s also important to note that irritant smoke decreases visibility compared to non-irritant smoke at same smoke density (Jin – 1985).
While these models may be useful, they only show whether or not an object is visible in the smoke. We could not identify any research that addressed depth perception in smoke. Depth perception is dependent on monocular and binocular cues and of the binocular cues, stereopsis has been shown to be affected by luminance (Reynaud – 2013). This means that smoke reduces luminance, impairs our stereopsis, and thus reduces our depth perception but the extent of this effect or functional relationship remains unknown.
Other effects of smoke
Jin attempted to study the emotional state of subjects as the amount of smoke increased in the room that the subjects were placed. Their emotional state was measured by a steadiness tester (cite). He found that most subjects began to be emotionally affected when the smoke density reached 0.1 1/m (Jin – 1981). Jin and Yamada tested the emotional stability and cognitive abilities of subjects as they were affected by heat and smoke (Jin 1989). They found that walking speed and mental arithmetic capability decreased as smoke density and radiant heat exposure increased.
Inhalation of combustion products
Inhalation of combustion products, namely carbon monoxide, causes neurobehavioral and cognitive effects that have been attributed at least in part to CO’s binding with hemoglogin. Carbon monoxide’s affinity for hemoglobin is more than 200 times that of oxygen, thus forming carboxyhemoglobin, COHb, quite readily. The formation of COHb creates hypoxia, i.e. oxygen deficiency, impairing red blood cells’ capability to carry and release oxygen. COHb, expressed as a percentage, is commonly used as an index to quantify CO exposure. Predictions for COHb from CO exposure can be found using models by Peterson and Stewart, for example (Peterson – 1975).
Laties and Merigan performed a literature review on the effects of CO exposure and concluded that there were no auditory or visual changes in subjects up to 12% COHb (Laties – 1979). Of the visual characteristics tested, visual acuity, depth perception, and color vision were not affected up to 12% COHb. They also found that dark adaptation was not affected at levels less than 20% COHb. However, they noted that some behavioral changes, e.g. vigilance, driving, and tracking, seem to be associated with COHb as low as 5-8%.
Fifteen years later, Benignus performed a meta-analysis on the effects of CO exposure and concluded that there is no significant impairment of visual or other behavioral functions in healthy young sedentary subjects at COHb levels below 18-25%. (Benignus – 1994).
Time since event/ Forgetting curve
Atkinson and Shiffrin describe that there are three stages of memory – sensory, short-term, and long-term (Atkinson – 1968). Information starts out in sensory memory, moves to short-term memory, and then transitions to long-term memory. Not all information experiences all three stages though. Unless information is attended to and processed sufficiently most is forgotten. A major focus in memory research has been on characterizing the lifetime of information in each stage, termed the forgetting curve.
Sperling found that information in sensory memory is quickly forgotten (Sperling – 1960). In his experiments, he showed subjects a display of letters in rows for only about 50 ms. He then asked them to recall all of the letters they could remember, which was on average a quarter of the letters shown. Subjects stated they had seen more letters but could not recall them. In his next study, he showed the same display of letters and asked them to recite only one row in particular. Subjects were able to recall almost all of the letters in that row, suggesting that we have what is known as sensory memory. Sperling describes sensory memory as a persistence of the sensation resulting from the stimulus. Results from his work indicate that information in sensory memory is available for less than one second.
With regard to short- and long-term memory, Rubin and Wenzel attempted to establish a retention function that could quantitatively describe the regularities of human memory and forgetting (Rubin – 1996). They assembled 210 published data sets that recorded the amount remembered versus time and fit hyperbolic, exponential, logarithmic, and power functions to the data.
While the exact function that best describes forgetting is under contention, one fact remains clear: large amounts are forgotten early on and less so as time progresses. For example, less than 25% is retained after 10 seconds for short-term memory and less than 25% is retained after 10 years for long-term memory.
Deffenbacher, Bornstein, McGorty, and Penrod argue that while people are focused on the rate of decay, which is important, it is also important to know the initial memory strength or “starting point” on the ordinate of the forgetting function in order to specify in absolute terms how much memory strength remains. (Deffenbacher – 2008).
Post-event information consists of specific misinformation that may be mistakenly incorporated unchanged into memory, thereby adding false information, or replacing/distorting memories (Davis – 2007). Researchers have hypothesized that misinformation alters the original memory, renders it more difficult to retrieve the original memory, or influences the reports of subjects who never encoded the original memory. Sources of misinformation include the media, co-witnesses, and updating/reevaluating memories.
As an example on the effect of post-event information, Loftus, Miller, and Burns (1978) showed a traffic accident to a group of participants followed by a series of questions. One of the questions incorrectly referred to a stop sign as a yield sign. Later, when asked if they had seen a stop or yield sign, misled subjects were more likely than control subjects to select the suggested (or incorrect) sign.
In the work by Zaragoza and Lane, subjects were asked specific questions about their memory for the source of the suggested items. (Zaragoza – 1994). Surprisingly, they found that misled subjects can come to believe they remember actually seeing the items that were merely suggested to them, which is called “source misattribution”. The tendency to commit source misattribution errors, however, depends heavily on the reflective and elaborative thought processes subjects engage in when encoding the suggestions, not just the content of the suggestions themselves.
Skagerberg and Wright found that instead of memory compliance, the process where a person could either respond with what they had originally seen or with what the other person saw, they observed memory blending (Skagerberg – 2008). In their work, 92 subjects were shown one of two versions of 6 different videos. Subjects were paired with someone who watched the other version and they discussed the videos together. Individuals then answered questions about the videos that they had seen. Their responses indicated that some of the memories were blends of the original and post-event information.
Paradoxically, some researchers have found that by simply retrieving a memory, false recognition can increase. St. Jacques and Schacter studied 42 people who conducted a self-guided museum tour while wearing a camera that automatically took photos at certain stops (St. Jacques – 2013). After 48 hours, some participants were shown photos of their tour in the correct temporal order and some participants were shown photos in the incorrect temporal order. All participants were then shown a photo from a different tour and were asked if it was related to the other photos. Forty-eight hours after this questionnaire, some participants were shown two photos of their tour that they had seen before, some participants were shown one photo of their tour and one photo from a different tour (that they had seen before), and others were shown two photos from their tour that they had not seen before. All participants were asked if the two photos showed stops that had been experienced together. If participants were shown photos that they had seen before, regardless of whether or not the photos were from their tour, hit rates slightly increased (0.77 and 0.71 compared to 0.65). However, false alarm rates increased significantly (0.64 and 0.59 compared to 0.38). In other words, reactivation of memory increased false recognition of photographs depicting stops that were not experienced during their tour.
Wording of questions
Loftus and Palmer found that questions asked subsequent to an event can cause reconstruction in one’s memory of that event leading to inaccuracies (Lofus – 1974). Subjects were shown films of automobile accidents and then answered questions about the events that occurred. When subjects were asked how fast the cars were going when they smashed into each other, estimated speeds were much higher than if the words collided, bumped, or contacted were used. Moreover, when subjects were asked if there was broken glass, those who received the word smashed were more likely to respond “yes”, even though there was no broken glass.
There are several findings in this field that are not intuitive. For example, it might be reasonable to think that the quicker someone responds, the more likely that they are correct since they have less time to be deceptive. However, Smith found that response time is more correlated with confidence than accuracy (Smith – 1989).
Deffenbacher reviewed a collection of studies done over the previous 80 years to assess the accuracy-confidence relation. He concluded that confidence as a predictor of accuracy was not supported by the evidence (Deffenbacher – 1980). There are those, however, who believe that that the predictability of accuracy from overtly expressed confidence varies directly with the optimality of information-processing conditions during encoding of the witnessed event, memory storage, and testing of the witness’s memory. In other words, the magnitude of the accuracy-confidence relation is moderated by external factors. For example, poor viewing conditions leads to a weaker accuracy-confidence relation than do good viewing conditions (Bradfield – 2002) and longer exposure durations permit more efficient processing of faces (Deffenbacher – 1980). Additionally, confirming feedback inflated retrospective certainty more for inaccurate witnesses than for accurate witnesses, which has the effect of significantly reducing the accuracy-confidence relation. As the optimality hypothesis would predict, such efficiencies of processing allow for greater predictability of recognition memory from confidence. Bothwell, Brigham, and Deffenbacher revisited this optimality hypothesis by performing a meta-analysis of 35 staged-event studies. They found that this new data was supportive of the original findings of Deffenbacher (Bothwell – 1987).
Clearly, witnesses who are confident in their testimony are not substantially more accurate than those who are not. However, Brewer argues that the accuracy-confidence correlation does not provide a comprehensive picture of the accuracy-confidence relation. By charting the proportion of accurate responses for each confidence level to assess the accuracy-confidence relation across a variety of stimuli, exposure and attention conditions, and retention intervals, he showed that confidence does provide a meaningful guide as to the likely accuracy of decisions made by adult eyewitnesses but only when measured immediately after identifying a perpetrator (Brewer – 2011). In many cases, however, it is difficult if not impossible to obtain eyewitness statements immediately after the event.
Discerning between true/false memories
Loftus and Pickrell found that subjects can “remember” a false event (Loftus – 1995). In their work, 24 subjects were given three true stories and one false story from their childhood, which were provided by a family member. Subjects wrote about each event and then rated their clarity for the memory and rated for confidence that they could remember more given extra time. Subjects wrote about each event one more time, were told that one event was false, and were asked to identify which one was false. Loftus and Pickrell found that 30% of subjects “remembered” the false event. Their results indicated that subjects used more words to describe the true events, clarity ratings were lower for the false event, and confidence ratings were lower for the false event. They also found that 25% of subjects chose a true event as the false one. Loftus and Palmer argue that this does not mean that the ones who chose correctly didn’t at least partially embrace the false event and while some subjects correctly chose the false event, they did this through process of elimination.
To assess the veracity of an eyewitness account, Schooler and Loftus explain that suggested-memory descriptions are longer and contain more hedges, more reference to cognitive operations, and fewer sensory details than real-memory descriptions (Schooler – 1986). Similarly, Bernstein and Loftus suggest that on average, real memories have more sensory and conceptual information, including visual, auditory, and olfactory details and spatial and temporal details (Bernstein – 2009). They also argue that suggestion and imagination can make a false memory feel and appear real, making it impossible to tell if it is a true or false memory.
Recent research has shown success of an interview technique known as the cognitive interview (Geiselman – 1984). The cognitive interview involves encouragement to mentally reconstruct the physical and personal context that existed at the time of the event. This could involve asking witnesses about their activities that day or the feelings they had throughout their activities. Note, however, that the interviewer should not rely on a self-assessment of emotional state (Reisberg – 2007). Rather, the interviewer should ask questions that can be used to assess a witness’s emotional state. The cognitive interview also involves asking participants to report everything they can recall even if it is partial or incomplete. Even a trivial detail could act as a trigger for more key information. Another aspect of the cognitive interview involves instructing witnesses to recall from a variety of perspectives, describing what they think other witnesses would have seen. Lastly, the cognitive interview encourages retrieval attempts in a different temporal order. Witnesses may remember more recent events better than earlier events and should be encouraged to recall backwards from the end towards the beginning. A meta-analysis on the cognitive interview revealed that there was a large overall effect size for the increase in correctly recalled details (Kohnken – 1999) and an updated meta-analysis revealed that older adults beneﬁt even more from the cognitive interview than younger adults in generating correct details with no observed differences for incorrect details (Memon – 2010).
Articles that assess the reliability of eyewitness testimony involving fires are scarce and not in agreement. Thus, it is vital for fire investigators to be able to assess reliability on a case-by-case basis. The goal of this work was to provide the tools that are needed by fire investigators to evaluate eyewitness testimony based on scientific literature. The majority of the literature employed can be found in peer-reviewed journals in the fields of cognitive psychology, neuroscience, and sociology. Assessment of eyewitness reliability begins by identifying influencing factors for each stage, i.e. acquisition, retention, and retrieval, of witnessing an event.
- Light adaptation is mostly complete after a few seconds while dark adaptation is complete after 40 minutes.
- Visual acuity and contrast sensitivity decline with age. Snellen decimal acuity is relatively constant up to age 50 and steadily declines thereafter. In terms of memory, elderly subjects performed worse on recognition and temporal information memory tests than young adults. The age-related decline in episodic spatial memory performance starts in a person’s 60’s. Older adults also appear to be more susceptible to false memories than healthy young adults.
- Upon awakening, it took 2-4 hours for subjects to be at full mental capacity.
- Mundane facts or details are not typically better remembered by those with specialized training but certain unique details are remembered more readily by a trained individual.
- The way eyewitnesses perceive the fire event could be influence by news articles, talking with friends and coworkers, or by past experiences with their personal property. Participants who were shown an ambiguous figure tended to report seeing the interpretation that assigned them to outcomes they favored.
- Although all people who witnessed a staged theft had the same opportunity to view the thief, the witnesses who knew the value of the item beforehand were significantly more accurate at identification than the other three groups.
- Weapon focus refers to the concentration of a witness’s attention on a weapon during a crime, leaving less attention available for viewing other items. The weapon focus effect should be called the ‘salient feature focus effect’ since unusual objects create an analogous effect as a weapon.
- Subjects who watched a short videotape of a bank robbery and later estimated the duration of the event invariably overestimated the duration.
- People typically show good memory for central features of emotional events and poorer memories for peripheral features, termed “memory narrowing”. Memory narrowing as a result of emotion, and violations of the memory-narrowing pattern can best be explained by the view that emotion enhances memory for information relevant to currently active goals. The difficulty lies in determining a person’s goals a priori.
- Smoke could obscure a witness’s field of view and irritant smoke decreases visibility compared to non-irritant smoke at the same smoke density. In addition, walking speed and mental arithmetic capability decreased as smoke density and radiant heat exposure increased. Lastly, there was no significant impairment of visual or other behavioral functions in healthy young sedentary subjects at COHb levels below 18-25%.
- Sensory memory is available for less than one second. Less than 25% is retained after 10 seconds for short-term memory and less than 25% is retained after 10 years for long-term memory.
- Post-event information consists of specific misinformation that may be mistakenly incorporated unchanged into memory, thereby adding false information, or replacing/distorting memories. Sources of misinformation include the media, co-witnesses, and updating/reevaluating memories.
- Questions asked subsequent to an event can cause reconstruction in one’s memory of that event leading to inaccuracies. Response time is more correlated with confidence than accuracy. Witnesses who are confident in their testimony are not substantially more accurate than those who are not. Confidence does provide a meaningful guide as to the likely accuracy of decisions made by adult eyewitnesses but only when measured immediately after. In many cases, however, it is difficult if not impossible to obtain eyewitness statements immediately after the event.
- Subjects can “remember” a false event. Suggested-memory descriptions are longer and contain more hedges, more reference to cognitive operations, and fewer sensory details than real-memory descriptions. On average, real memories have more sensory and conceptual information, including visual, auditory, and olfactory details and spatial and temporal details. Suggestion and imagination can make a false memory feel and appear real, making it impossible to tell if it is a true or false memory.
- The cognitive interview involves encouragement to mentally reconstruct the physical and personal context that existed at the time of the event, asking participants to report everything they can recall even if it is partial or incomplete, instructing witnesses to recall from a variety of perspectives, and encouraging retrieval attempts in a different temporal order.