Page 127 - Demo

125Can third-party observers detect attraction?6Table 2. Overview of all accuracy predicting models (1–3) for Experiment2.Predictors Accuracy (Median estimate of the coefficient with 95% HDI)Model 1 Model 2 Model 3β(95% HDI) β(95% HDI) β(95% HDI)Intercept -0.01 [-0.10, 0.08] -0.02 [-0.11, 0.07] -0.02 [-0.11–0.07]Age Group -0.05 [-0.14, 0.04] -0.05 [-0.15, -0.04]Attracted to Partner 0.25 [0.16, 0.34]Age Group × Attracted to Partner 0.09 [0.00, 0.18]Random EffectsVar(Participant) 0.00 0.00 0.00participants could not reliably detect attraction or its absence in the datingvideos. Interestingly, we also replicate the effect that participants detectedattraction somewhat more accurately (56%) when the person depicted wasattracted to their partner than not (44%).Experiment 3In Experiment 3, we manipulated the phase and length of the presentedvideo segment. We used muted videos from the Verbal Interaction (VI)phase of Prochazkova et al.’s study (2021) and varied their lengths (i.e., 3,6, and 9 s). Furthermore, to probe whether the observed accuracy was dueto a general low emotion recognition accuracy, we included an additionalEmotion Recognition Task (ERT). Low scores in the ERT would indicatethat participants could not detect basic emotional expressions and might ex,plain the low accuracy in our task of primary interest (AJT). Also, to ensurethat the low accuracy was not due to potential individual differences thatmight influence emotion detection accuracy, we collected information usingthe Autism-Spectrum Quotient (AQ; Baron-Cohen, Wheelwright, Skinner,Martin, & Clubley, 2001) and Beck-Depression Inventory (BDI-II; Beck,Steer, & Brown, 2011). Participants also indicated whether they were in arelationship or not and its duration. Because in Study 2 there were no dif,ferences between children and adults in the accuracy of detecting attraction,for feasibility, we decided to recruit adults only.MethodsDue to the restrictions because of the COVID-19 pandemic, data collec,tion took place online using the Gorilla platform (Anwyl-Irvine, Massonnié,Flitton, Kirkham, & Evershed, 2020).Table 2. Overview of all accuracy predicting models (1–3) for Experiment2.Predictors Accuracy (Median estimate of the coefficient with 95% HDI)Model 1 Model 2 Model 3β(95% HDI) β(95% HDI) β(95% HDI)Intercept -0.01 [-0.10, 0.08] -0.02 [-0.11, 0.07] -0.02 [-0.11–0.07]Age Group -0.05 [-0.14, 0.04] -0.05 [-0.15, -0.04]Attracted to Partner 0.25 [0.16, 0.34]Age Group × Attracted to Partner 0.09 [0.00, 0.18]Random EffectsVar(Participant) 0.00 0.00 0.00participants could not reliably detect attraction or its absence in the datingvideos. Interestingly, we also replicate the effect that participants detectedattraction somewhat more accurately (56%) when the person depicted wasattracted to their partner than not (44%).Experiment 3In Experiment 3, we manipulated the phase and length of the presentedvideo segment. We used muted videos from the Verbal Interaction (VI)phase of Prochazkova et al.’s study (2021) and varied their lengths (i.e., 3,6, and 9 s). Furthermore, to probe whether the observed accuracy was dueto a general low emotion recognition accuracy, we included an additionalEmotion Recognition Task (ERT). Low scores in the ERT would indicatethat participants could not detect basic emotional expressions and might ex,plain the low accuracy in our task of primary interest (AJT). Also, to ensurethat the low accuracy was not due to potential individual differences thatmight influence emotion detection accuracy, we collected information usingthe Autism-Spectrum Quotient (AQ; Baron-Cohen, Wheelwright, Skinner,Martin, & Clubley, 2001) and Beck-Depression Inventory (BDI-II; Beck,Steer, & Brown, 2011). Participants also indicated whether they were in arelationship or not and its duration. Because in Study 2 there were no dif,ferences between children and adults in the accuracy of detecting attraction,for feasibility, we decided to recruit adults only.MethodsDue to the restrictions because of the COVID-19 pandemic, data collec,tion took place online using the Gorilla platform (Anwyl-Irvine, Massonnié,Flitton, Kirkham, & Evershed, 2020).Table 2. Overview of all accuracy predicting models (1–3) for Experiment2.Predictors Accuracy (Median estimate of the coefficient with 95% HDI)Model 1 Model 2 Model 3β(95% HDI) β(95% HDI) β(95% HDI)Intercept -0.01 [-0.10, 0.08] -0.02 [-0.11, 0.07] -0.02 [-0.11–0.07]Age Group -0.05 [-0.14, 0.04] -0.05 [-0.15, -0.04]Attracted to Partner 0.25 [0.16, 0.34]Age Group × Attracted to Partner 0.09 [0.00, 0.18]Random EffectsVar(Participant) 0.00 0.00 0.00participants could not reliably detect attraction or its absence in the datingvideos. Interestingly, we also replicate the effect that participants detectedattraction somewhat more accurately (56%) when the person depicted wasattracted to their partner than not (44%).Experiment 3In Experiment 3, we manipulated the phase and length of the presentedvideo segment. We used muted videos from the Verbal Interaction (VI)phase of Prochazkova et al.’s study (2021) and varied their lengths (i.e., 3,6, and 9 s). Furthermore, to probe whether the observed accuracy was dueto a general low emotion recognition accuracy, we included an additionalEmotion Recognition Task (ERT). Low scores in the ERT would indicatethat participants could not detect basic emotional expressions and might ex,plain the low accuracy in our task of primary interest (AJT). Also, to ensurethat the low accuracy was not due to potential individual differences thatmight influence emotion detection accuracy, we collected information usingthe Autism-Spectrum Quotient (AQ; Baron-Cohen, Wheelwright, Skinner,Martin, & Clubley, 2001) and Beck-Depression Inventory (BDI-II; Beck,Steer, & Brown, 2011). Participants also indicated whether they were in arelationship or not and its duration. Because in Study 2 there were no dif,ferences between children and adults in the accuracy of detecting attraction,for feasibility, we decided to recruit adults only.MethodsDue to the restrictions because of the COVID-19 pandemic, data collec,tion took place online using the Gorilla platform (Anwyl-Irvine, Massonnié,Flitton, Kirkham, & Evershed, 2020).Table 2. Overview of all accuracy predicting models (1–3) for Experiment2.Predictors Accuracy (Median estimate of the coefficient with 95% HDI)Model 1 Model 2 Model 3β(95% HDI) β(95% HDI) β(95% HDI)Intercept -0.01 [-0.10, 0.08] -0.02 [-0.11, 0.07] -0.02 [-0.11–0.07]Age Group -0.05 [-0.14, 0.04] -0.05 [-0.15, -0.04]Attracted to Partner 0.25 [0.16, 0.34]Age Group × Attracted to Partner 0.09 [0.00, 0.18]Random EffectsVar(Participant) 0.00 0.00 0.00participants could not reliably detect attraction or its absence in the datingvideos. Interestingly, we also replicate the effect that participants detectedattraction somewhat more accurately (56%) when the person depicted wasattracted to their partner than not (44%).Experiment 3In Experiment 3, we manipulated the phase and length of the presentedvideo segment. We used muted videos from the Verbal Interaction (VI)phase of Prochazkova et al.’s study (2021) and varied their lengths (i.e., 3,6, and 9 s). Furthermore, to probe whether the observed accuracy was dueto a general low emotion recognition accuracy, we included an additionalEmotion Recognition Task (ERT). Low scores in the ERT would indicatethat participants could not detect basic emotional expressions and might ex,plain the low accuracy in our task of primary interest (AJT). Also, to ensurethat the low accuracy was not due to potential individual differences thatmight influence emotion detection accuracy, we collected information usingthe Autism-Spectrum Quotient (AQ; Baron-Cohen, Wheelwright, Skinner,Martin, & Clubley, 2001) and Beck-Depression Inventory (BDI-II; Beck,Steer, & Brown, 2011). Participants also indicated whether they were in arelationship or not and its duration. Because in Study 2 there were no dif,ferences between children and adults in the accuracy of detecting attraction,for feasibility, we decided to recruit adults only.MethodsDue to the restrictions because of the COVID-19 pandemic, data collec,tion took place online using the Gorilla platform (Anwyl-Irvine, Massonnié,Flitton, Kirkham, & Evershed, 2020).Table 2. Overview of all accuracy predicting models (1–3) for Experiment2.Predictors Accuracy (Median estimate of the coefficient with 95% HDI)Model 1 Model 2 Model 3β(95% HDI) β(95% HDI) β(95% HDI)Intercept -0.01 [-0.10, 0.08] -0.02 [-0.11, 0.07] -0.02 [-0.11–0.07]Age Group -0.05 [-0.14, 0.04] -0.05 [-0.15, -0.04]Attracted to Partner 0.25 [0.16, 0.34]Age Group × Attracted to Partner 0.09 [0.00, 0.18]Random EffectsVar(Participant) 0.00 0.00 0.00participants could not reliably detect attraction or its absence in the datingvideos. Interestingly, we also replicate the effect that participants detectedattraction somewhat more accurately (56%) when the person depicted wasattracted to their partner than not (44%).Experiment 3In Experiment 3, we manipulated the phase and length of the presentedvideo segment. We used muted videos from the Verbal Interaction (VI)phase of Prochazkova et al.’s study (2021) and varied their lengths (i.e., 3,6, and 9 s). Furthermore, to probe whether the observed accuracy was dueto a general low emotion recognition accuracy, we included an additionalEmotion Recognition Task (ERT). Low scores in the ERT would indicatethat participants could not detect basic emotional expressions and might ex,plain the low accuracy in our task of primary interest (AJT). Also, to ensurethat the low accuracy was not due to potential individual differences thatmight influence emotion detection accuracy, we collected information usingthe Autism-Spectrum Quotient (AQ; Baron-Cohen, Wheelwright, Skinner,Martin, & Clubley, 2001) and Beck-Depression Inventory (BDI-II; Beck,Steer, & Brown, 2011). Participants also indicated whether they were in arelationship or not and its duration. Because in Study 2 there were no dif,ferences between children and adults in the accuracy of detecting attraction,for feasibility, we decided to recruit adults only.MethodsDue to the restrictions because of the COVID-19 pandemic, data collec,tion took place online using the Gorilla platform (Anwyl-Irvine, Massonnié,Flitton, Kirkham, & Evershed, 2020).ParticipantsOne hundred and seventy-six (N�=�176) adults were recruited using socialmedia platforms and the university psychology student pool, 13 of whomdid not complete the study. Therefore, the final sample consisted of 163participants (age range: 18–66; M age: 27.69, SD�=�13.20; 95 female). Allparticipants provided informed consent and were informed that they couldwithdraw their participation with no adverse consequences as according tothe Declaration of Helsinki. Participants were not remunerated for their participation except for course credits. The study was approved by the LeidenUniversity Psychology Ethics Committee (CEP 2020–02-27-M.E. Kret-V2-2192). Differences in participants’ age and gender between Experiment 1,Experiment 2, and Experiment 3 are reported in the Supplementary Material. Participants’ emotion recognition was good (75% correct) and in linewith previous studies (e.g., Akdag, 2020).StimuliRegarding the Attraction Judgment Task (AJT), to examine whether theoverall low mean accuracy observed in Experiment 1 and Experiment 2was due to either the brief duration of the stimuli or the interaction phaseemployed (i.e., first-impression phase; FI), in Experiment 3, we manipulatedthe video segment in two ways: length and interaction phase. Specifically,we used the following segments: a) 3-s FI segments (as in Experiment 1 andExperiment 2); b) 3-s; c) 6-s; and d) 9-s segments from the verbal interaction(VI) phase.Experimental TaskThe AJT was the same as in Experiment 2. Participants were assigned inthe stimulus condition in a counterbalanced order.ProcedureAfter participants provided informed consent, they were asked to providedemographic information (i.e., age, gender, sexual orientation, nationality,and educational level). Next, participants were informed that they wouldview a series of videos and, they should indicate whether the person depictedwould like to go on another date with their partner and their level of certainty regarding their judgement. Participants were instructed to respondas fast and accurately as possible. Participants were prompted to indicatewhether the person would like to go on another date with their partner bypressing the corresponding keyboard key (y yes, and n for no); followed bytheir certainty regarding their decision from 0 (not at all) to 6 (very) withTable 2. Overview of all accuracy predicting models (1–3) for Experiment2.Predictors Accuracy (Median estimate of the coefficient with 95% HDI)Model 1 Model 2 Model 3β(95% HDI) β(95% HDI) β(95% HDI)Intercept -0.01 [-0.10, 0.08] -0.02 [-0.11, 0.07] -0.02 [-0.11–0.07]Age Group -0.05 [-0.14, 0.04] -0.05 [-0.15, -0.04]Attracted to Partner 0.25 [0.16, 0.34]Age Group ⇥ Attracted to Partner 0.09 [0.00, 0.18]Random EffectsVar(Participant) 0.00 0.00 0.00participants could not reliably detect attraction or its absence in the datingvideos. Interestingly, we also replicate the effect that participants detectedattraction somewhat more accurately (56%) when the person depicted wasattracted to their partner than not (44%).Experiment 3In Experiment 3, we manipulated the phase and length of the presentedvideo segment. We used muted videos from the Verbal Interaction (VI)phase of Prochazkova et al.’s study (2022) and varied their lengths (i.e., 3,6, and 9 s). Furthermore, to probe whether the observed accuracy was dueto a general low emotion recognition accuracy, we included an additionalEmotion Recognition Task (ERT). Low scores in the ERT would indicatethat participants could not detect basic emotional expressions and might ex,plain the low accuracy in our task of primary interest (AJT). Also, to ensurethat the low accuracy was not due to potential individual differences thatmight influence emotion detection accuracy, we collected information usingthe Autism-Spectrum Quotient (AQ; Baron-Cohen, Wheelwright, Skinner,Martin, & Clubley, 2001) and Beck-Depression Inventory (BDI-II; Beck,Steer, & Brown, 2011). Participants also indicated whether they were in arelationship or not and its duration. Because in Study 2 there were no dif,ferences between children and adults in the accuracy of detecting attraction,for feasibility, we decided to recruit adults only.MethodsDue to the restrictions because of the COVID-19 pandemic, data collec,tion took place online using the Gorilla platform (Anwyl-Irvine, Massonnié,Flitton, Kirkham, & Evershed, 2020).Iliana Samara 17x24.indd 125 08-04-2024 16:36