1) This field study examined how intoxicated witnesses' identification performance is affected by identification format (lineup vs. showup).
2) In target-present conditions, alcohol consumption and identification format had no effect on witnesses' ability to identify the perpetrator.
3) In target-absent conditions, identification format affected choosing behavior, with showups increasing the likelihood of choosing, but this did not significantly enhance false identifications.
Field study examines effects of alcohol and ID format on intoxicated witnesses' identifications
1. Show me one or six? Field study examining how
identification format affects real-world intoxicated witnesses
CHRISTOPHER ALTMAN, NADJA SCHREIBER COMPO, KRISTEN SLAPINSKI,
HANNAH LESZCZYNSKI, SEAN REYNOLDS, CHRISTINA KUZMINSKI, ABIGAIL
BRIGGS, & JISELLE CERVERA
2. Intoxicated Witnesses
Police interactions (Evans, Schreiber Compo, Russano, 2009)
Common (53%) or very common (20%)
Most interviewed an intoxicated witness within the past month
≈ .20% had provided an intoxicated witness with a lineup or showup
Jurors opinions (Evans & Schreiber Compo, 2010)
Less credible
More cognitively impaired
Less able to make an accurate identification
Experts opinions (Kassin, Tubb, Hosch, & Memon, 2001)
“Alcohol impairs witnesses’ ability to recall persons and events.” (90%)
Would testify about the phenomenon (61%)
Negative alcohol-memory relationship was common sense (95%)
3. 1) Yuille & Tollestrup (1990)
2) Dysart et al., (2002)
Study Space Analysis
Effect of
alcohol
X
4. 1) Yuille & Tollestrup (1990)
2) Dysart et al., (2002)
3) Hagsand et al., (2013)
4) Harvey et al., (2013)
5) Kneller & Harvey (2015)
6) Flowe et al., (2017)
Study Space Analysis
Ps Sober
During ID
X
X
BAC
Restriction
X
Lab
Study
X
Effect of
alcohol
X
X
X
X
X
Full
Lineup
X
5. 1) Yuille & Tollestrup (1990)
2) Dysart et al., (2002)
3) Hagsand et al., (2013)
4) Harvey et al., (2013)
5) Kneller & Harvey (2015)
6) Flowe et al., (2017)
7) Altman et al., (2018)
Study Space Analysis
Ps Sober
During ID
X
X
BAC
Restriction
X
Lab
Study
X
Effect of
alcohol
X
X
X
X
X
X
Full
Lineup
X
XXX
6. Alcohol & Showups
Showup (Sjӧberg, 2016; Steblay, Dysart, Fulero, & Lindsay, 2003)
Lack of fillers increases likelihood of false identifications
Witness fails to consider alternative s to suspect presented
Has no effect on correct perpetrator identifications
Alcohol (Josephs & Steele, 1990; George, Rogers, & Duka, 2005)
Reduces the ability to consider alternative options (while intoxicated)
Creates a liberal decision-making criteria
7. Current Study
Goal of the present study
Test participants ID performance at elevated BAC levels
Examine whether ID format affects decisions
Lineup vs showup
Hypotheses
Target-present conditions
Alcohol and ID format will have no effect on choosing or identifications
Target-absent conditions
Showup: Alcohol will increase choosing / false identifications
Lineup: Alternatives will lead to rejections/r filler identifications
o No increase in false identifications
8. Participants = 132
Age
Range = 19-66 years
Mean = 38.83 (15.08)
Gender
Male = 53%
Female= 47%
Ethnicity
Caucasian = 129 (97.7%)
African American= 2 (1.5%)
Hispanic = 1 (.8%)
9. Participants BAC Levels
0-
10-
15-
20-
5-
25-
.00 .05 .15 .25.10 .20
N= 132
BAC Range= .000- .240
Legal U.S. limit
BAC= .08
Under the legal
limit
BAC=.00-.079
N=73
Over the
legal limit
BAC >.080
N= 59
BAC Measure
Frequency
10. Materials & Procedure
Research examining cognitive and motor functioning
Taken to a separate testing room
Had their BAC levels recorded
First motor task
Modified version of the game Cornhole
Given 5 bean bags to throw at the Cornhole board
Accumulate high score across 4 rounds
11. Materials & Procedure
Confederate intrusion
Wants to join the Cornhole game
Plays with equipment after told “no”
Stands, yells, flips chair, exits
Intrusion lasted ≈ 1 minute
Identification portion
Bar supplied photos of women who previously caused disturbances
Could you look at them and tell me if the intruder is pictured?
Keep in mind she may or may not be there
Lineup: 6-person simultaneous TA or TP
Showup: TA or TP
Selection, rejection, ‘not sure’
18. Analyses
Separated based on target-presence
Logistic regressions
BAC level (continuous)
ID format (showup vs. lineup)
Interaction
Choosing behavior
Did the participant make an identification
Accuracy
Target-present: Perpetrator identifications
Target-absent: False identifications
19. Choosing Behavior (TP)
Model: 2(3) = 2.28, p = .52
ID Format: p = .87
BAC level: p = .22
Interaction: p = .80
No Selection
Made Selection
71%
65%
31%
35%
0
10
20
30
40
50
60
70
80
90
100
Lineup Showup
27. Discussion
Target- present condition
No effect of alcohol or ID format
Target-absent condition
No effect of alcohol
ID format affected choosing behavior
Did not enhance false identifications
Filler similarity
Dysart et al. (2002): “highly similar filler “
Present study: Aggregate of all fillers
WCS suspect was identified 50% of the time
VS.
28. Take-Aways
1) BAC level does not have a significant impact on witnesses’
identification decisions
2) Exception could be when a showup procedure is used and the
innocent suspect is highly similar to the witness
132 participants were recruited for this study. These participants ranged in age from 19-66 and were evenly split between males and females. Given the demographic makeup of the area they were also predominately Caucasian.
I don’t like much about this slide, it feels very busy and the photo doesn’t really help display what I’m saying.
Here you can see witnesses choosing behavior in the target absent lineup condition across the various BAC levels recorded
Below that is a graph displaying witnesses choosing behavior in the showup condition (bottom graph appears)
In the lineup condition 19 participants made a selection while 13 did not (top selection panel appears)
In the showup condition 9 made a selection and 30 did not (bottom selection panel appears)
Our predictors were able to produce a model that accurately classified participants choosing behavior (regression panel appears)
However, identification format was the only significant predictor
Contrary to what we predicted, participants in the lineup condition made more selections than those in the showup condition
Important to note however is that alcohol had no effect on participants choosing behavior
As you can see from the various choosers and non-choosers scattered throughout the graph
I don’t like much about this slide, it feels very busy and the photo doesn’t really help display what I’m saying.
Here you can see witnesses choosing behavior in the target absent lineup condition across the various BAC levels recorded
Below that is a graph displaying witnesses choosing behavior in the showup condition (bottom graph appears)
In the lineup condition 19 participants made a selection while 13 did not (top selection panel appears)
In the showup condition 9 made a selection and 30 did not (bottom selection panel appears)
Our predictors were able to produce a model that accurately classified participants choosing behavior (regression panel appears)
However, identification format was the only significant predictor
Contrary to what we predicted, participants in the lineup condition made more selections than those in the showup condition
Important to note however is that alcohol had no effect on participants choosing behavior
As you can see from the various choosers and non-choosers scattered throughout the graph
I don’t like much about this slide, it feels very busy and the photo doesn’t really help display what I’m saying.
Here you can see witnesses choosing behavior in the target absent lineup condition across the various BAC levels recorded
Below that is a graph displaying witnesses choosing behavior in the showup condition (bottom graph appears)
In the lineup condition 19 participants made a selection while 13 did not (top selection panel appears)
In the showup condition 9 made a selection and 30 did not (bottom selection panel appears)
Our predictors were able to produce a model that accurately classified participants choosing behavior (regression panel appears)
However, identification format was the only significant predictor
Contrary to what we predicted, participants in the lineup condition made more selections than those in the showup condition
Important to note however is that alcohol had no effect on participants choosing behavior
As you can see from the various choosers and non-choosers scattered throughout the graph
I don’t like much about this slide, it feels very busy and the photo doesn’t really help display what I’m saying.
Here you can see witnesses choosing behavior in the target absent lineup condition across the various BAC levels recorded
Below that is a graph displaying witnesses choosing behavior in the showup condition (bottom graph appears)
In the lineup condition 19 participants made a selection while 13 did not (top selection panel appears)
In the showup condition 9 made a selection and 30 did not (bottom selection panel appears)
Our predictors were able to produce a model that accurately classified participants choosing behavior (regression panel appears)
However, identification format was the only significant predictor
Contrary to what we predicted, participants in the lineup condition made more selections than those in the showup condition
Important to note however is that alcohol had no effect on participants choosing behavior
As you can see from the various choosers and non-choosers scattered throughout the graph