SciELO - Scientific Electronic Library Online

 
 númeroESPVocational Choice as a source of stressGender and coping in old age índice de autoresíndice de assuntospesquisa de artigos
Home Pagelista alfabética de periódicos  

Serviços Personalizados

Journal

Artigo

Indicadores

  • Não possue artigos citadosCitado por SciELO

Links relacionados

  • Não possue artigos similaresSimilares em SciELO

Compartilhar


Interdisciplinaria

versão On-line ISSN 1668-7027

Interdisciplinaria  n.esp Buenos Aires  2004

 

Interpersonal trust in different ages

Carla Sacchi *

* Psychological Doctor. Independent Researcher of Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET). E-Mail: jsacchi@fibertel.com.ar

Resumen

   El objetivo de este trabajo es presentar diferentes escalas en español para la evaluación de la confianza interpersonal. La calidad de las relaciones establecidas entre los miembros de un grupo social permite el crecimiento de cada integrante y del grupo como conjunto. En la mayoría de los casos, y particular-mente en la infancia, las necesidades solamente pueden ser satisfechas a través de la interacción con los demás; esto implica interdependencia y requiere reciprocidad. Por lo tanto es importante prever cómo actuará la otra persona, para anticipar nuestro comportamiento hacia ella. Las expectativas producen cambios en la atribución, según sea interpretada la actitud del otro como beligerante o cooperativa, y a la vez modifican el comportamiento hacia los demás.

Palabras clave: Confianza interpersonal - interdependencia - reciprocidad - cambios en la atribución.

Abstract

   This presentation aims to review the instruments created for the evaluation of interpersonal trust among various age groups within a Spanish speaking community.
   
The quality of the relationships established between members of any given social group, permits the growth of each person in particular as well as that of the group as a whole. In most cases, and particularly so during infancy, our needs can only be satisfied through interaction with others. This fact necessarily entails interdependence, and requires reciprocity. It is therefore important to be able to predict how the other person will be have in order to anticipate our own behaviour, in order to obtain some advantage or avoid frustration. Expectations cause attributive changes depending on whether the other person's attitude is interpreted as belligerent or co-operative. It also modifies behaviour towards others.

Key words: Interpersonal trust - interdependence - reciprocity - attributive changes.

   This paper intends to go into the evolution of that feeling of trust and assess whether it is a structural trait of personality, or if the peculiar characteristics of each situation bring about a behaviour of trust or mistrust. If we view situations as a set of social factors interacting with personal experience at a given moment, it is unthinkable to trust others at large without attributing a specific meaning that will explain and interpret facts within a given context. We organize our knowledge of the world and react according to personality, language and culture; the situational context taints our interpretation, and, conversely, the way we define or interpret the situation, will influence our behaviour. The interpersonal trust construct would be an important indicator of disposition when re-signifying external data which disrupt internal balance.
    On the other hand, trust would underpin predictability; namely, the capacity to foresee the behaviour of others. Previous research on trust has advanced two divergent conceptualizations of the construct: expectations toward people in general and relationships with specific partners. Trust orientations toward people in general is considered by Rotter (1967, 1971, 1980) to be an expectancy deriving from past experience that an individual can relied on. In this regard, generalized trust may be regarded as something akin to personality trait (Couch, 1994; Couch et al., 1996). Conversely, relational trust refers to the faith that people have that the partner will respond to them with positive feelings. It is also important to differentiate within trust according to the depth or intimacy of the interpersonal relationship (Rempel, Holmes, & Zanna, 1985).
   Trust in may be seen as part of the supporting network of any person; it will resemble attachment in closer relationships (parents, couples).
   Based on these theoretical tenets, we drew up the following scales for children, adolescents, and adults.

Scale of Interpersonal Trust for Children

   Interpersonal trust in infancy originates in attachment: a system based on love, whose aim is to provide for the immature members of the species (Ainsworth, Salter, & Bowlby, 1991).
   Social expectations and skills are developed within the scope of the mother and the nuclear family. Later on, they are extended to the wider social group: neighbourhood, school, and society at large.
   On the other hand, Mitchell (1990) has found that the children whose parents fulfilled their promises in the past, generally trust other authority figures. This shift entails better interpersonal relationships, since trusting children are more optimistic and lively. They are also less prone to cheating in situations that are controlled without their knowledge. From trusting, they become trustworthy. Social sanction modifies expectations and adapts them to the social context.
   We drew up a list of situations implying trust, both positive and negative, towards the people concerned. These persons were parents, friends, teachers, work or studies colleagues, the media and people at large. The situations were those normally encountered by each age group. A jury of three psychologists chose the most adequate questions in the list. Items were evaluated on a YES/NO basis. Half the items were negative statements, so values were inverted in the analysis. Eight items of social desirability were added: they were taken from Marlowe Crowne's Social Desirability Scale. Pilot studies were carried out in all scales, to check that the items were properly worded and relevant.
   The final scale is made up of 24 items administered to a group of third grade children (9-10 years old) and to a group of seventh grade children (11-12 years of age). The lists were given out to the whole group, they were self-administered and written. All the tests were given out in the classroom, in municipality-run, full day schools attended by the children.
   The final sample was formed by 626 children (348 boys, and 278 girls), between 9, and 13 years of age: 9 years old (n =184), 10 years old (n = 140), 11 years old (n = 10), 12 years old (n =142), and 13 years old (n =160).
   Because the age distribution, we presumed that younger children would show greater dependence (trust, in this case) on parents and authority figures. But older children, according to the same hypothesis, would depend more on their freinds, and would be more critical of authority figures.
   This led us to analyze each age sub-sample independently. Items were assessed for their discrimination power: those that were answered in the same way by 75% or more respondents were deemed non-discriminative.
   The sub-sample of older children was finally composed of 299 children (167 boys, and 132 girls), between 11, and 13 years of age, distributed as follows: 11 years old (n = 9), 10 years (n = 140), 12 years (n =142), 13 years (n =150).
   Items 2, 4, 6, 8, 15, and 31 proved non-discriminative. Items 5, and 28 show a high percentage of missing.
    In order to study the reliability of the scale, we calculated Cronbach's alpha index, which reached a .70 value.
    Factorial validity was studied with the principal axes method, Promax rotation. The data matrix proved liable to factorization since KMO index attained a .70 value, and Bartlett's Sphere Test was p = .00. Anti-image correlation values fluctuated between 630, and 773.
   Three factors liable of interpretation were found. This accounts for the 18% variance. Factor 1 accounts for 11.57% variance; Factor 2, 4.16%, and Factor 3, 2.63%.
    Factor 1 inverted includes items 6; 11 ;13; 20; inverted 26; inverted 27; inverted 31. Factor 2 includes items 8; inverted 11; 13; 18; 21; 24; inverted 26; inverted 27; 13; 21; inverted 29; inverted 30; and Factor 3 includes inverted 2; 8; 9; 13; 20; inverted 27 (see Table 1).

   The sub-sample of younger children was finally formed by 327 children (146 boys, and 132 girls), between 9, and 10 years of age, distributed as follows: 9 years, n =184; 10 years, n = 142. Items 15, and 20 proved non-discriminative.
    Reliability was tested through Cronbach's alpha index, which reached a .61 value; with item 8 being the most unstable.
    Factorial validity was studied with the principal axes method, Promax rotation. The data matrix proved liable to factorisation since the KMO index attained a .70 value, and Bartlett's Sphere Test was p = .00. Anti-image values ranged between 595, and 740.
   Three factors liable of interpretation were found. This accounts for the 20.84% total variance. Factor 1 accounts for 11.05% variance; Factor 2, 6.30%, and Factor 3, 3.48%.
   Factor 1 includes items 3; 6; 8 (negative); 13; 15; 20; 21. Factor 2 includes inverted 11; inverted 16; inverted 26; inverted 27; inverted 29. Factor 3 includes items 3 (negative); 8; 4; 15 (negative); 31 (negative).

   - When comparing Factorial Analyses of both sub-samples, we notice that in Factor 1, both groups show common values for items 6, 13, and 20. In younger groups, items 3, 8 (negative); 15, 21, and 28 are added;and in the older group, items inverted 11; inverted 26, and inverted 27.

   - In Factor 2, both groups show common values for items inverted 11, inverted 26, inverted 27, inverted 29. In younger group, item inverted 16 is added, and in the older group, 13, 20, 21, 24, inverted 29, inverted 30.

   - In Factor 3, both groups show common values 4, and 8. The younger add 3 (negative),15 (negative), and inverted 3, and the older, items inverted 2, 13, 20, 21, inverted 27 (see Table 2).

   Because the little interpretability of factor, it was decided to merge both samples in one formed by 626 children, 348 boys, and 278 girls. In order to study the reliability of the scale, we calculated Cronbach's alpha index, which reached a .61 value. The most unstable items were 21, 23, 28 and 32.
   Items 15, 23, 27, and 31 proved non-discriminative. Items 5, and 28 had a high percentage of missing.
   Factorial validity was studied with the principal axes method, Promax rotation. The data matrix proved liable to factorization since KMO index attained an adequate level (.72), and Bartlett's Sphere Test was p = .00. Three factors liable of interpretation were found. This accounts for the 20% variance, as seen in the table. Factor 1 accounts for 12.19% variance; Factor 2 , 9.2%, and Factor 3, 6.1%.
    Factor 1 includes items 6, 9, 13, 15, 20, 21, 24, inverted 26, inverted 27, and 28. Factor 2 include items 3 (negative), 4, 6 (negative), 8, 15 (negative), and 31 (negative). Factor 3 includes all inverted items: 11, 14, 16, 26, 27, and 30 (see Table 3).

Analysis of results and conclusions on the scale for children

   The first two factors, which we termed trust factors, include items implying trust as well as mistrust. Factor 3, on the other hand, is the clearest as regards its components since it only includes mistrust items.

   - Factor 1 includes items 6, 9, 13, 15, 20, 53 which imply trust in parents, teachers and people at large. This factor also includes items of mistrust towards parents, teachers and other children.

   - Factor 2 includes items 3 (negative), 4, 6 (negative), 8, 15 (negative), and 31 (negative) that express trust in politicians, companions, and parents, and lack of trust in parents and other children.

   - Factor 3 includes items 11, 14, 16, 26, 27, 29, and 30: lack of trust in teachers, people at large, parents and other children. Is a clear factor of mistrust, there are only mistrust items.

   The scale shows an acceptable level of reliability. Nevertheless, the study of factorial validity does not provide clear results due to the introduction of items of mistrust in trust factors. The mistrust factor, on the other hand, is usually clear.
   Secondly, we found significant differences when the group was divided by age. This leads us to presume there are different ways of coping with the evaluation of trust on others.
    When studying non-discriminative items (15, 23, 27, and 31) as well as those with a high level of missing (5.28) or unstable in Cronbach's alpha study (21, 23, 28, 32) we may come to the conclusion that parent -and teacher- figures are hardly ever criticized; while nobody wants to pass judgement on his/her companions' behaviour.
   The fact that the study of factorial validity was not adequate, since the factors obtained do not coincide with the tenets underpinning the choice of items, calls for further studies of the test, with a substitution of the items called into question.

Scale of Interpersonal Trust for Adolescents

   The scale for adolescents was made up of 32 items. The sample studied included 671 secondary school students (273 boys, and 398 girls) aged 13-16. The scale was administered to different groups, at the youngsters' school. Anonymity, and secrecy were strictly kept.
    Analysis of reliability showed a .62 Cronbach's alpha index. The most unstable item was 31. Though none of the items proved non-discriminative, addition of values 1, and 2 surpassed 75% in items 10, 15, 22, 23, and 28. On the other hand, value 3 (neuter) attained or surpassed 30% in items 1, 4, 6, 11, 12, 13, 17, 18, 19, 20, 24, 25, and 27. Lastly, items 10, 12, 23, and 1 show a high missing percentage.
    Data matrix was deemed liable of factorization since the KMO index reached a .75 value and Bartlett's Sphere Test,p = .00. Anti-image values range from a minimum .548 to a maximum .843. Three factors proved open to interpretation, which accounts for the 22% total variance, as can be seen in the Table 4. Factor 1 accounts for an 11.23% variance; Factor 2, 6.34%, and Factor 3, 4.86%.

   - Factor 1 is defined by variables 10, 15, 16 negative, 22, 23 negative, 24, 25 negative, 27 negative, 28, 32.
   - Factor 2 is formed by items 3, 9, 16 negative, 17, 18, 19, 23 negative, 30, 31.
   - Factor 3 is formed by items 1, 5, 6, 7, 8, 16 negative, 21, 23 negative, 24, 25, 26, 27 negative, 29 (see Table 4).

   Items 2, 4, 11, 12, 13, 14, and 20 have no weight in any factor. Factor 1 is defined by items of trust in parents, friends and partner. There are items of mistrust too, in parents, partner and friends.
   All items in Factor 2 are of mistrust: in people at large, politicians, unknown persons. Items 16, and 23 are of trust as regards parents. Item 23 proved non-discriminative.
   Factor 3 includes trust in people, in justice and mistrust of friends, parents, and partner. Mistrust items are included in the non-discriminative group.
   It is worth noting that the questions with the highest rate of missing values were those related to partners. We believe that at this age, youngsters do not usually think in terms of intimacy with a sexual partner. Their sexual relationships are normally connected to defiance in boys, and romance in girls. One of the trust factors includes people who are close to them, and the other one is formed by people at large, politicians, etc. Unlike the other two groups studied (children and adults), trust factors in adolescents include mistrust items, while the mistrust factor is clearly cut.

Scale of Interpersonal Trust for Adults

   This is a modified version of the first scale, which was drawn up to study moderating variables in stress (Sacchi, 1995). Deficiencies perceived in that first version were connected to the fact that the questions did not contemplate specific situations concerned with more emotional contents (attachment) in close relationships. These were modified in the current version.
    The scale was administered to 134 adults: 59 men and 74 women, aged between 25, and 55. It was individual and voluntary. Most scales were administered within the context of work evaluation.
    Analysis of reliability showed a .61 Cronbach's alpha index. Item 7 proved non-discriminative. In items 4, 8, 11, 12, 14, 16, 17, 18 the addition of values 1 and 2 surpassed 75%. On the other hand, value 3 (neuter) attained or surpassed 30% in items 13 and 22. Missing percentage is very low.
    Data matrix was deemed liable of factorisation since the KMO index reached a .60 value, and Bartlett's Sphere Test, p = .00. Anti-image values range from a minimum .461 to a maximum .720.
    Three factors proved open to interpretation, which accounts for the 22% total variance, as can be seen in the Table 5. Factor 1 accounts for 10.10% variance; Factor 2, 7.47%, and Factor 3, 4.66%.

   - Factor 1 is defined by variables 2 negative, 4, 7 negative, 14, 15 negative, 16, 18, 20 negative, 21 negative, 24 negative, 25, 27 negative.
   - Factor 2 is defined by variables 1, 2, 5, 9, 19, 25, 26 negative, and 28.
   - Factor 3 is composed of items 10, 11, 13, 15, 20, 22(see Table 5).

   Factor 1 includes items 4, 14, 16, and 18, which may be assessed as scarcely discriminative since over 75% of subjects expressed agreement. Item 7 proved non discriminative: 75% of respondents disagreed. When the items where done in general terms, that is trust would not be attributed to any subjects, there are high level of missing.
   Factor 2 items expressed trust in people in general.
   In Factor 3, items 13, and 22, connected to false accusations of politicians and scarce reliability of sports competitions, obtained very high levels of neuter answers (neither agreement nor disagreement), which we can classify, up to a certain point, as missing values.
   The wording of items 3, 6, 8, 12, and 17 is done in general terms; consequently, trust would not be attributed to any subject (actor). Respondents would therefore not have been able to make up their minds, since they may not have known how to apply the contents of the item to their own experience.
   As shown in the table, Factor 1 is connected to trust in the partner, offspring, companions, people in general. Factor 2 is related to people at large, teachers, etc., and Factor 3 is connected to mistrust in people in general, politicians, and it includes item 10, trust in one's partner. These items deal with friends' opinion, partner's trustworthiness, help received by parents from their own children and co-operation among colleagues at work.

Conclusion

   The aim of this research was to explore the differential experience of trust in a broader range of interpersonal relationships. As proposed by Rotter (1971), we chose items based in trust, or mistrust, generated by parents, friends, partner, teachers, colleagues, and people in general.
   The results show that this difference had not appeared, but we can better interprete the Factorial Analysis results differentiating Relational Trust and General Trust (Couch et al., 1996). In Relational Trust we find feelings of confidence in close relationship, based on emotion-laden interactions, and in general or global trust, that assesses trust in novel situation, and in the belief that the others are honest.
   In children sample we find:
   - Factor 1: Relational trust and mistrust (parents and teachers).
   - Factor 2: General trust, that includes companions and impersonal expressions.
   - Factor 3: Mistrust, general and relational.
   In adolescent sample we find:
   - Factor 1: Relational trust and mistrust.
   - Factor 2: General mistrust and Relational trust parents (only one item).
   - Factor 3: General trust.
   In adults sample we find:
   - Factor 1: Relational trust, partner, friends.
   - Factor 2: General trust.
   - Factor 3: Mistrust, general and relational. In all the samples the Mistrust factor was more clear than the Trust Factor.
   The purpose of this work was to develop a comprehensive, reliable and valid measure of trust as it is experience in differing types and levels of relationships.
   This difference on General and Relational is clearly seen in adults, however in children and adolescents Relational Trust and Mistrust are interwoven. In early adolescence, the bond with the parental figure is as intense and ambivalent as in childhood.

References

1. Ainsworth, M., Salter, D., & Bowlby, J. (1991). An ethological approach to personality development. American Psychologist, April, 333-341.         [ Links ]

2. Couch, L.L. (1994). The development of the Trust Inventory. Unpublished Master's Thesis. University of Tennessee, Knoxville.         [ Links ]

3. Couch, L.L., Jeffrey, A.M., & Jones, W.H. (1996). Measuring level of trust. Journal of Personality Assessment, 67(2), 305-323.         [ Links ]

4. Mitchell, C. (1990). Development of restoration of trust in interpersonal relationship during adolescence and beyond. Adolescence, 25, 847-854.         [ Links ]

5. Rempel, J.K., Holmes, J.G., & Zamma, M.P. (1985). Trust inclose relationships. Journal of Personality and Social Psychology, 49, 95-112.         [ Links ]

6. Rotter, J. (1967). A new scale for the measurement of interpersonal trust. Journal of Personality, 35, 651-665.         [ Links ]

7. Rotter, J. (1971). Generalized expectancies for interpersonal trust. American Psychologist, 26(5), 443-451.         [ Links ]

8. Rotter, J. (1980). Interpersonal trust, truthworthiness and gullibility. American Psychologist, 35(1), 1-7.         [ Links ]

9. Sacchi, C. (1995). Evaluación de la confianza interpersonal [Interpersonal trust evaluation]. Interdisciplinaria, 12(2), 65-72.         [ Links ]

Centro Interdisciplinario de Investigaciones en Psicología Matemática y Experimental (CIIPME) Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) Tte. Gral. Perón 2158 (C1040AAH) Buenos Aires – Argentina

Received: March 7, 2001
Accepted: November 9, 2001

Creative Commons License Todo o conteúdo deste periódico, exceto onde está identificado, está licenciado sob uma Licença Creative Commons