Laboratorio de Realidad Virtual de la Universitat de Barcelona

Ostracism is important in the field of psychology because of the often aversive consequences from a psychological perspective that may cause to people who suffers it. It has been defined as “being ignored and excluded, and it often occurs without excessive explanation or explicit negative attention”. It has been detected to be responsible of many negative reactions caused on individuals such as lower the sense of belonging, control, self-esteem and meaningful existence; increased emotion-focused coping responses; dehumanization; lessen the will to cooperate in groups, anxiety or aggression. Negative effects are appreciated even when ostracized by despised outgroups or when the individual observes another person being socially excluded.

Ostracism has been found to be related with other social phenomena such as religiousness increase of non-sedentary activity among children and imitation in the social behavior.

Across ages, some studies have been carried out for examining possible differences in the effects of an act of social exclusion. It found out higher anxiety among as well as other negative reactions for adolescents in peer rejection in comparison with adults. Another study comparing older individuals with adults and young adults found a decrease of ostracism-induced needs for satisfaction and negative affectivity the older the age group was. Experiments involving a specific population of old people and evaluate the effects induced by ostracism seem to be scarce.

Techniques for the study of ostracism developed in the domain of psychology have been several such as Life Alone or Get Acquainted paradigms. In virtual reality environments (VRE), the most used paradigm for the study of ostracism is the software Cyberball developed by Christopher Cheung, Wilma Choi and Kipling D. Williams (2000). A ball-tossing game where the participants interact with two or three computer-run characters. In spite of the poor graphics and the simplicity of the game it has been proven to be quite effective for creating a sense of social exclusion in individuals and detecting the emotional and social results enhanced by it.. However, the use of a more graphically developed program may help to find out some aspects of ostracism in VRE that are not seen in simpler softwares such as Cyberball.

The goal in the study is to evaluate the anxiety-induced by ostracism among old people and the possible differences compared to younger individuals using a tridimensional program quite similar to Cyberball but more graphically realistic for trying to improve the immersion effect in the VRE activity. We also want to study the difference that may be in overall anxiety levels as it has been suggested that anxiety is reduced by age.

The main hypothesis is that old people are going to suffer less ostracism-induced aversive effects in anxiety as in the traits studied. The study by Sebastian et al. (2010) which suggest that anxiety itself decrease in adults compared to adolescents also may be an indication that young adults will suffer more anxiety in a socially-based VRE than the elderly.

The experiments has been carried out in a semi-spherical virtual dome Immersadome 3.0, 3.22x2.37x1.5m and a projector XGA 2.500 lumens using a PC Dell T7400 in the virtual reality lab in the faculty of psychology of the University of Barcelona. The program has been developed in the same faculty using the software Unity 4 v4.3.4f1 by United Technologies Aps and adapted for being watched in a dome (Figure 1).

Figure 1. Images of the program “PasaPelota”.

For measuring anxiety we used Spielberger’s State Trait Anxiety Inventory-State version questionnaire adapted to Spanish language. Reliability of the test has shown an internal consistency of .86 to .95 and a test-retest .65 to .75 whose validity has been tested in both adolescents and adults (N=>10.000) with strong correlations with the Taylor Manifest Anxiety Scale and Cattell and Scheier’s Anxiety Scale Questionnaire. The State version has been validated with subjects in stressful environments. In the elderly has proven to be able to discriminate between people with or without anxiety disorders.

The sample was composed by 13 individuals whose age was over 55 years old (9 females and 3 males) for being compared later with a sample collected before of 40 youths (20 females and 20 males) between 21 to 28 years old. Both, young and old, held university degrees or were undergraduates. They were recruited in the university through acquaintances or were known by the experimenters. They passed the experiment voluntarily without any significant reward for participating. When it comes to the older people they were also taught about several virtual reality programs that we used in the laboratory or were designed for the use in psychological therapy as an incentive to come and learn more about this field.

Each subject had to complete four different phases in a repeated measures experiment, the virtual reality was presented at random. The phases consisted in two ostracism phases and two non-ostracism phases. Within these phases, one of them had females as avatars and the other had males.

In the elderly sample the outcomes showed no significant differences between ostracism and non-ostracism phases (p=.246) with an F=1.500 (Table 1).

Table 1. Results for ostracism and non-ostracism in the old sample

In ostracism phases had mean of 24,04 (SD=5.19) and in non-ostracism phases of 24.54 (SD=4.99) (Table 2). This occurred when we only took into account same-sex or opposite-sex avatars. For the interaction with same-sex avatars the results were unsignificant (F=0.34 p=.857) with a mean of 24.08 (SD=5.616) for ostracism phases and 24 (SD=5.752) for non-ostracism (Table 3) . Interaction of the two independent variables was not significant with an F=1,910 and p=0,173.

Table 2. Mean and standard deviation for the samples of old participants

Table 3. Results for ostracism and non-ostracism in the young sample

In contrast, the sample of young men and women showed a clear significant difference between both phases (F=17.073 p<0.001). The means for ostracism phases were 15.83 (SD=10.43) and 12.38 (SD=8.48) for social inclusion (Table 4).

Table 4. Mean and standard deviation for the samples of young participants.

The differences of ostracism-induced anxiety between age groups was significant with an F=5.8822 (p=0.02). Overall ostracism was also significant (F=4.354, 0.042) but was just because of the larger sample of young participants (Table 5).

Tabla 5. Results for ostracism and non-ostracism in the old sample.

Curiously, what we can subtract from this analysis is that old people were more anxious in all the situations than the young despite they were unaffected by the stigmatization of the avatars in the game, with a total mean of 13.98 (SD=9.06) for the young sample and 24.17 (SD=24.17) for the elderly one with an F=13.623 (p=0.001) for intergroup differences.

Our hypothesis that the elderly suffer less aversive effects than the young people is confirmed.

The results show a clear different reaction in ostracism between both samples. That goes in concordance with the study of Hawkley et al. (2011) that finds lower feelings of ostracism-induced needs for satisfaction and negative affectivity in the older participants. However, it is not just a decrease in anxiety that we found here. The old participants don’t seem to feel affected by ostracism at all. One possible explanation for this is that anxiety measured by STAI is a stronger feeling that the items they use for belonging, self-esteem control and meaningful existence which would make an anxiety measure as irrelevant unlike some not-as-bad feelings.

The fact that the program used here had better graphic quality didn’t seem to alter the results. This is one of the shortcomings that one may find in programs like Cyberball that could affect some results. In this case, a more detailed VRE didn’t get significant scores for old people. The low standard deviation also shows that there is no big dispersion in the data of the subjects within the sample. However, the experience of the participants with computers and virtual reality was not tested and that could be a confounding variable to take into account.

A more interesting finding is the remarkably higher levels of anxiety of the older group contrary to evidence presented which shows a decline in anxiety as a trait among older people, First, we have to take into account that we passed the state version of STAI so that would mean that we were not using a completely good way for measuring the average anxiety. However, what it is obvious is that old people had higher feelings of anxiety at this moment and it did not decrease after the passing of the experiment. A possibility is that old people would feel more anxious while passing a VRE experiment. Further investigation is needed.

Another possible shortcoming in the experiment is that the sample of the old people (N=13) was quite low. A larger one would be necessary for the replication of some of the results to shed more light on this.