Running head: HEADER TEXT/SHORTENED TITLE, e.g. Face Inversion 1

Title of Your Lab Report, e.g. ‘The effect of inversion on the identification of familiar faces’

by

Catherine Chauder

200901948

Liam Walsh
200704453

A laboratory report

presented to R.McInnis

in Psychology 225

Sensation and Perception

Department of Psychology

St. Francis Xavier University

November, 29, 2012

Abstract[PD1]

The purpose of the present study was to investigate how [SOME INDEPENDENT VARIABLE (S)] affect [SOME DEPENDENT VARIABLE(S)].[PD2]This was done ...
[PD3]It was predicted that…. [PD4]As expected, …. [PD5]We conclude that …[PD6]

Title of your study

Face perception is an important process through which the brain and mind can understand and interpret human faces. The process is essential for recognizing familiar faces, interpreting human expressions which outwardly display emotions, and basic social interaction. Facial expressions are used as a silent language, communicating our current thoughts or feelings with those around us. Our study, however, focuses on face identification pulling from memory cues as soon as our eyes scan familiar features. It’s easy to recognize a familiar face, but sometimes it’s hard to remember under which circumstances you know the face or in the case of our study, match it with a name. Our study focuses not simply on facial recognition, but facial identification. For the most part, humans are experts at upright facial recognition, but inverted faces are significantly harder. When it comes to face identification face orientation is everything.

The aim of our experimental study is to identify the effects face orientation and order of face stimuli have on the participants’ accuracy, as well as their reaction time. Specifically, will investigate how the inverted face effect will vary the accuracy and reaction time in comparison to the accuracy and reaction time of identifying upright faces. We will also investigate how the order of face stimuli presented (upright first or inverted first) will affect the reaction time and accuracy of face identification by comparing results between participants of each order (A or B). In order to fully understand the face-inversion effect and how it affects our ability to identify faces, we must explore the notion that there is a process of facial scanning that takes place during the course of face recognition.

Peter Hills of Anglia Ruskin University, and David Ross and Michael Lewis of Cardiff University (2011) believe that there is a hierarchy of facial features in terms of saliency for face recognition. They state that when upper facial features (such as the eyes) are concealed, face recognition is more heavily affected than if lower facial features are concealed. According to their study the scanning of eyes is critical in face recognition and when the eyes are the first fixed focal point the inversion effect may be smaller (Hills, Ross, Cardiff, 2011). However, Hills and his colleagues (2011) continue to say that there is evidence that supports the conclusion that eyes are typically less scanned in an inverted face and that the scanning performance follows a more random pattern for inverted faces. Through 3 separate experiments they tested the effects of cueing (eyes, mouth, or no cueing at all) on facial recognition. They expected and hypothesized that eye cueing during facial inversion would implement a higher accuracy of facial recognition than no cueing, and that cueing the mouth in both upright and inverted faces would implement a lower accuracy than no cueing at all. The results of their experiments demonstrated that the diagnostic utility of eyes in face perception and the location of the first focal point have an effect on the accuracy of facial recognition, which could potentially underpin the face-inversion effect (Hills et al, 2011).
In our experiment it is impossible to tell how participants implemented cueing, if at all. Therefore, since Hills and his colleagues (2011) stated that the scanning performance follows a random sequence during inverted facial recognition (Hills et al, 2011), we can assume that the majority of the participants in our study cued the mouth during scanning, or did not cue any focal point at all. This means that if our participants cued the eyes during inverted facial scanning, would we expect our mean accuracy of participants to be much higher.

GalitYovel and colleagues (Yovel et al, 2012) set out to find if there were a difference in orientation between different race-faces and same-race faces in the perception of neonatology nurses. They found that poor different race-face recognition is usually credited to little experience of the visual system with different race-faces. Recognition of passive exposure to same race-faces is often representedby attempts to individuate same race-faces at a sub categorical level. They conducted two experiments. First they had participants view three sets of six face photos. Of the six photos were three infants and three adults. Presented in random order the identification of adult faces score was higher. In the next phase each of the identified faces was contrasted with five distractor faces. The participants were asked to identify one of the six faces. Massive exposure to different race-faces in participant neonatology nurses shows better ability to individuate faces from a control group of same-race adult faces than that of infant faces. Despite seeing massive amounts of different-race faces over years neonatology nurses more readily identified same race-faces consistent with the hypotheses quality of faces is more salient in face recognition in the visual system than passive but massive exposure to different race-faces.

A 2x2 mixed design (within and between subjects) was employed in which the two independent variables were: face orientation (upright or inverted) and order (A – upright presented first, or B – inverted presented first). The 56 participants of the study were paired up into groups and assigned an order in terms of face stimuli. By doing this both levels of upright and inverted face orientation were shown to each participant of each group. The participants assigned A would view the upright face stimuli first, and the participants assigned B would view the inverted face stimuli first. Since conditions were assigned, a measure of control was established in the form of counterbalancing to ensure consistency across situations. The face stimuli were counterbalanced as each face appeared upright or inverted in a different but equalorder across all groups.Participants were shown the face stimuli on a computer screen controlled by their partners (playing the roles of the experimenter) whom used a keyboard. Participants were timed on the length of their reaction to face stimuli and were required to write their response on a sheet of paper. Assigning an order to each participant is important for contrasting results between both groups. After the experiment we should be able to verify the advantages or effects that each order has on reaction time and accuracy. Face orientation plays a very important role in face identification, and our study intends to emphasize our ability to recognize faces when they are presented to us as altered or unaltered entities.

It is expected and hypothesized that participants will have a slower reaction time to stimuli caused by the inverted faces, and that participant accuracy will also be affected by the inversion effect causing the percentage correct to be lower.In short, face orientation (independent variable 1) and order (independent variable 2) will have a heavy effect on participant reaction time and participant accuracy.To address whether these potential effects will occur an experiment with 56 participants was conducted aiming to test these hypotheses.

Method

Participants

Participants were 56 students from a sensation and perception psychology class at Saint Francis Xavier University, who participated as part of their lab requirement.
Apparatus

Participants used a computer during the study. The computer presented stimuli comprised of 18 celebrity faces in total (9 were presented normally and 9 were presented in an inverted form). The partner (experimenter) of the participant used the keyboard to change the stimulus when the participant indicated that they were ready for a new stimulus to be presented. A printed Participant Response Sheet was used by the participant to write and record their answers. Finally, a time keeping instrument was used by the experimenter (instrument varied per group; phone, stop watch, wrist watch etc.) to record the reaction time of the participant in seconds.

Procedure

The experiment required the 56 participants to formulate and work in pairs (the odd number of students was compensated by the instructor pairing up with a student). Each participant was assigned a letter within their group (A or B) and sat in front of one computer with a slideshow presented on the computer screen. The slide show consisted of 18 celebrity face (the stimuli of the experiment); 9 were presented normally and 9 were presented in an inverted form. The subject to undergo the experiment first was decided amongst the pair, leading the partner to take on the role of experimenter.

Two experiments occurred per group, ensuring that each participant played the role of both the subject and the experimenter. During the experiment the participant was presented with 18 celebrity faces in a different order which depended on their assigned letter. Order A presented 9 faces in a non-inverted manner, followed by 9 faces in an inverted manner. Order B presented 9 faces in an inverted manner, followed by 9 faces in a non-inverted manner. However, for both orders the organization of celebrity faces stayed the same; meaning the faces in Face Identification Part 1 and Part 2 were identically organized for both Participant A and B, but they were presented in the opposite orientation. The participant to go first was presented with the first 9 faces and they recorded their written answer on the printed Participant Response Sheet. While this was happening, the experimenter changed the stimulus by using the keyboard when the subject indicated they were finished recording their response. The experimenter timed the subject using a time keeping device and recorded their reaction time in seconds once the subject had completed the first portion of the experiment. When both the subject and experimenter were ready, 9 more faces were presented in the opposite orientation and the recording proceeded. After experiment 1 was finished, the participants reversed roles and conducted experiment 2. Finally, after both experiments were conducted the participants recorded how many facial recognitions they had recorded successfully and how many faces they recognized as familiar.
Participants were given 75 minutes to set up and complete the experiment.

Results

From the data collected it can be seen that for Upright Face Orientation, Group A (participants who viewed upright first) on average obtained a mean accuracy score of 77.3, whereas Group B (participants who viewed inverted first) on average obtained a mean accuracy score of 61.6. For Inverted Face Orientation, Group A on average obtained a mean accuracy score of 38.9, whereas Group B on average obtained a mean accuracy score of 59.6.In terms of Latency for Upright Face Orientation, Group A on average obtained a mean reaction time of 47.3 seconds, whereas Group B on average obtained a mean reaction time of 49.1 seconds. For Inverted Face Orientation, Group A on average obtained a mean reaction time of 61.2 seconds, whereas Group B on average obtained a mean reaction time of 50.3 seconds.

In Figure A & B the effect of Face Orientation (independent variable X) and Order (independent variable Y) on the dependent variables are shown. It can be noted that Face Orientation is a within-subject variable and Order is a between subject variable. Figure A. demonstrates the effect of Face Orientation and Order on the accuracy (percentage correct) of the participants.Each darkened bar represents the average of 28 participants (Group A) for upright and inverted face orientation, while each lightened bar represents the average of 28 participants (Group B) for upright and inverted face orientation. A trend in the data shows that those who were exposed to upright stimuli first (Order A) had a higher accuracy value for upright faces than those who were exposed to inverted stimuli first (Order B). Furthermore, those who were exposed to inverted stimuli first (Order B) had a higher accuracy value for inverted faces than those who were exposed to upright stimuli first (Order A). Another trend in the data shows that the average accuracy for both Group A & B was higher for upright face orientation than inverted face orientation.

Figure B. demonstrates the effect of Face Orientation and Order on the reaction time (in seconds) of the participants. Again, each darkened bar represents the average of 28 participants (Group A) for upright and inverted face orientation, while each lightened bar represents the average of 28 participants (Group B) for upright and inverted face orientation. A trend in the data shows that order had the same effect on accuracy that it did on reaction time. Another trend in the data shows that the average reaction time for both Group A & B was higher for upright face orientation than inverted face orientation

Discussion[PD7]

The results of the current study …[PD8]

The present study lends support to the ideas proposed by …[PD9]

It is possible that the present study is limited by …

[PD10]Future research would benefit from…[PD11]

There are some broader implications of the currents study with regards to ….……………. ……………. ……………. ……………. [PD12]

The experiment that was conducted reports results that demonstrate the face-inversion effect and its influence on reaction time and accuracy of face identification. The experiment established primarily that inverted face stimuli were harder to identify and the process of recognizing the inverted face had a longer duration. Specifically, when compared to upright face identification, inverted face stimuli were identified less accurately and reaction time to the inverted face stimuli was much longer than that of the reaction time to the upright face stimuli. These findings support our first hypothesis suggesting that face orientation, specifically the face-inversion effect would have a negative influence on our dependent variables, reaction time and accuracy. Furthermore, it was discovered that in terms of order, whichever face orientation stimuli (whether it be upright or inverted) was presented to the participant first, on average had an advantage on their face identification results. In other words, participants who were first exposed to inverted faces, on average, tended to do better on inverted face stimuli identification, in terms of both accuracy and reaction time, than their partner. The same can be said for the opposite face orientation; participants who were first exposed to upright faces, on average, tended to do better on upright face identification, in terms of both accuracy and reaction time, than their partner.

It is easy to understand why upright face stimuli were identified more easily and faster, but it was unexpected that the order of face stimuli would have such a drastic effect. One explanation can be related back to Hills and his colleague’s (2011) study on feature diagnosticity in the face-inversion effect. In their study, the emphasized over and over again the importance of cueing in face recognition, and that the process for upright face recognition and inverted face recognition differed. Inverted face recognition process tended to be more random, without cueing, while the upright face recognition process tended to take cues from the eyes and the mouth (Hills et al, 2011). Perhaps the first exposure to face orientation set the tone for cue patterns in each participant. For example, the cue patterns Participant A developed from the first portion of the experiment (in their case upright face stimuli first), may have carried over to the second portion of the experiment and were implemented in identifying inverted face stimuli. This resistance to changing the scanning process during facial recognition could have negative effects on both the accuracy and reaction time of the participant.

It is possible that the present study is limited by the methodology of the experiment. Due to the condition that all the participants were located in the same room it was easier for them to hear the verbal answers of the other participants. The participant to undergo the experiment second within their group was also at an advantage. Since the second participant had been previously exposed to the face stimuli during the first participant’s trial, they could rely more heavily on their memory to make an accurate guess as opposed to solely relying on their ability to recognize the face stimuli.

References

Hills, P. J., Ross, D. A., & Lewis, M. B. (2011). Attention misplaced: The role of diagnostic features in the face-inversion effect. Journal Of Experimental Psychology: Human Perception And Performance, 37(5), 1396-1406. doi:10.1037/a0024247

Yovel, G., Halsband, K., Pelleg, M., Farkash, N., Gal, B., & Goshen-Gottstein, Y. (2012). Can massive but passive exposure to faces contribute to face recognition abilities?. Journal Of Experimental Psychology: Human Perception And Performance, 38(2), 285-289. doi:10.1037/a0027077