The Effect of Backloading Instructions on Eyewitness Identification from Simultaneous and Sequential Lineups

Abstract

The sequential lineup is multifaceted, including serial presentation of faces, multiple decisions, and often backloading (indicating to an eyewitness that a lineup contains more photos than there actually are). We evaluated the effect of backloading instructions on response bias and sensitivity with an eyewitness identification paradigm. Importantly, we included an 'undisclosed' condition that provided no information to participants about the number of lineup members to expect. Experiment 1 (N = 780) tested sequential lineups; Experiment 2 (N = 532) tested simultaneous lineups. As predicted, signal detection analysis showed that backloading induced participants to be more conservative in choosing from both lineup types, but did not affect d′. We conclude that the criminal justice system should be mindful of this shift in response bias, as it has implications for both guilty and innocent suspect identifications.

Do sequential lineups impair underlying discriminability?

“…Consistent with previous studies (Carlson et al, 2016;Clark, 2012;Dobolyi & Dodson, 2013;Gronlund et al, 2009;Meissner et al, 2005;, we found that sequential presentation led to more conservative responding. This conforms to the original intention behind the introduction of sequential lineups, to reduce false alarms.…”

“…Model-fit statistics and parameter values for each dataset are available in Table S2. SDT-MAX fit 12 of 13 simultaneous datasets at α = .05, failing to fit the backloaded simultaneous condition of Carlson et al (2016). SDT-SEQ fit 10 of 13 sequential datasets at α = .05, failing to fit the sequential data from Sučić et al (2015), the sequential weapon present plus distinctive feature condition from Carlson and Carlson (2014) and the sequential data from Pozzulo et al (2013).…”

“All I remember is the black eye”: A distinctive facial feature harms eyewitness identification

“…However, this prediction derived from the hybrid‐similarity model is best tested with an analysis of theoretical discriminability between target and fillers rather than empirical discriminability between guilty and innocent suspects (see Wixted & Mickes, 2018, for an extensive discussion of the differences between empirical and theoretical discriminability). The compound‐decision SDT model (SDT‐CD; Duncan, 2006) has been used in the eyewitness ID literature to test for differences in theoretical discriminability (e.g., Carlson, Carlson, Weatherford, Tucker, & Bednarz, 2016; Palmer & Brewer, 2012; Smalarz, Kornell, Vaughn, & Palmer, 2019; Smith, Wells, Smalarz, & Lampinen, 2018), and we will follow suit here. Specifically, as recommended by Wixted, Vul, Mickes, and Wilson (2018), we applied the independent observations version of the SDT‐CD model.…”

Line-Up Image Position in Simultaneous and Sequential Line-Ups: The Effects of Age and Viewing Distance on Selection Patterns

“…However, some findings suggest that the shift from conservative to liberal only occurs when participants are aware of the number of images in the lineup; if they are not aware of this, then, the "pressure" to select toward the end of the line-up does not increase (Horry et al, 2012). These results have found support in later studies (Carlson et al, 2016), but findings are not uniform, and position effects have also been found in line-ups where no information has been given (Carlson et al, 2016;Meisters et al, 2018). Overall, there is little agreement regarding position effects in sequential line-ups, and there is an ongoing debate regarding how these effects should be evaluated (e.g., Wilson et al, 2019).…”

“…We have already discussed some of the limitations of our current design in an earlier publication (Nyman et al, 2019b); however, limitations that pertain to the current design are that there is evidence to suggest that in sequential line-ups, prior knowledge of the number of photos to be shown can result in an increase in choosing bias earlier on in the line-up (Horry et al, 2012;Carlson et al, 2016), but the same effect has been found when no information was given (Carlson et al, 2016;Meisters et al, 2018). As the present study is based on four trials per person, we could have purposefully neglected to mention the number of photos to be shown in the sequential line-ups, thus, enabling a comparison between the first trial where the participants were unaware versus the subsequent trials where they were aware of the number of photos to be shown.…”