Escala de Avaliação de Papéis Familiares: assessment of psychometric properties

Theoretical framework: The involvement of family members in the provision of informal care makes it important to assess the distribution of family roles. Objectives: To develop and validate a Escala de Avaliação de Papéis Familiares (Family Role Assessment Scale). A convenience sample (n=909) and a questionnaire were used. The information was coded and processed using the SPSS software, version 18.0. Methodology: A descriptive exploratory study was carried out. Results: The following eight factors were created: papel de prestador cuidados à criança (child-caregiver role) (α=0.963); papel de dona de casa (housekeeper role) (α=0.923); papel de dona de casa cuidados externos (housekeeper role, care outside the home) (α=0.624); papel organizador de atividades recreativas (recreational organiser role) (α=0.837); papel terapêutico (therapeutic role) (α=0.811); papel sexual (sexual partner role) (α=0.839); papel provedor de família (family provider role) (α=0.695), and papel socializador (socialiser role) (α=0.719). Each factor was assessed using subscales, resulting in a KMO=0.916 and c2=35065.895, with a degree of freedom of 2278 and α=0.894 for all items. Conclusion: Results showed satisfactory psychometric properties. This instrument has the potential to be applied to research activities and the supervision of family roles.


Introduction
As a profession in the health area, the purpose of Nursing is to provide nursing care to healthy or ill human beings, throughout their life cycle, and the social groups to which they belong so that they maintain, improve and recover their health projects. Therefore, it becomes important to understand how family roles evolve and their impact on family functioning. To this end, it is important for nurses to assess the functional category of the various roles targeting the family and not only the individual ( Wrigth & Leahey, 2002;Ordem dos Enfermeiros, 2012). Given the lack of an original measurement instrument for the Portuguese population that assessed the distribution of the different family roles, an assessment scale was developed for this purpose. Thus, the objective of this study was to develop an instrument to assess family roles and verify its psychometric properties when applied to families with only healthy members, regardless of the family typology. The development of the Escala de Avaliação de Papéis Familiares is an added value to Nursing Science to better identify the nurses' foci of intervention among families. The recognition of this need is at the basis of a major transformation of paradigm in the field of Nursing care: from a care perspective with the family as context to a paradigm with the family as an intervention system. The need was felt to develop this instrument because families are being asked to assume more responsibilities in health care. They are expected to care for their healthy family members and/or those with acute and chronic diseases.

Background
The family is currently seen as a group of human beings considered as a social unit or a collective whole, composed of people connected by consanguinity, emotional affinity or legal kinship, including people who are important to the individual (Conselho Internacional de Enfermeiros, 2006). The family is a social system composed of one or more persons who coexist within a given context with some expectations of reciprocal affection, mutual responsibility and temporary duration. It is characterised by commitment, joint decision-making and sharing of goals. This way of perceiving the family allows including the different settings and compositions of families that exist in contemporary society. Understanding the structure, role and processes of each family unit is of extreme importance to characterise the family's health as well as the contributions to the health of both individuals and groups. Contemporary families have undergone several changes, but the most visible one has been in its structure, which is defined as "the ordered set of relationships within the family, and between the family and other social systems" (Hanson, 2005, p. 86).
To determine the family structure, it is first necessary to identify the individuals making up the family, their relationship to each other, and the relationships between the family and other social systems. Hanson advocates that the organisational patterns of a family are more or less stable and some of its changes throughout its lifecycle may be predictable. Even nowadays, society expects the family to play specific roles to meet the needs of the family itself. From an international perspective and within the scope of health sciences, specifically Nursing, it is important to understand both the family functions and processes. These represent a type of Nursing phenomenon with specific characteristics. They are based on the ongoing positive or negative interactions and patterns of relating among family members (Conselho Internacional de Enfermeiros, 2006). Other authors suggest that the family process is an ongoing interaction between family members through which instrumental and expressive tasks are accomplished. Hanson (2005) believes that the process is what makes every family unique and, even though families may have similar structures and functions, they may interact differently, thus promoting the singularity of each family. For the family to survive, family members develop expectations about how each member should behave while interacting with each other. To this end, they take on different roles and adopt unique and distinct behaviours within each family unit. In social psychology and sociology, the concept of role is characterised by gender-related attitudes and behaviours, thus expressing a normative dimension (Amâncio & Oliveira, 2002). Talcott Parsons was one of the first sociologists to use the concept of role associated with the individual's that promotes a satisfactory relationship between both partners. The therapeutic role involves types of help such as sharing of concerns, willingness to listen to others, active participation in problemsolving and emotional support; it also includes health promotion and prevention activities, as well as rehabilitation activities in case of disease. The recreational organiser role covers the planning and implementation of leisure and free time activities, as well as events with family members and others. Finally, the kinship role involves keeping in touch with other family members and friends.

Methodology
A descriptive exploratory cross-sectional study was carried out with the purpose of designing the Escala de Avaliação de Papéis Familiares (EAPF) (Family Roles Assessment Scale) and validating its psychometric properties. Data were collected using a questionnaire. A convenience sample (n=909) was selected consisting of family members residing in a district in northern Portugal. In fact, this sampling technique is particularly appropriate when the researcher has a prior knowledge of the population homogeneity. Sample size was calculated based on Anastasi's recommendations (1990), i.e. 10 participants per scale item.

Participants -inclusion criteria
The inclusion criteria were households in which no member had functional or cognitive limitations. Similarly to other studies, all household members aged 12 or more could participate in the study (Onso, 1988;Derogatis, 1993).

Data collection instrument
Based on Hanson's theoretical framework (2005) and the qualitative analysis of interviews with families and experts in family Nursing (a study which was previously conducted and is not presented in this article), the Escala de Avaliação de Papéis Familiares (EAPF) was designed, i.e. its dimensions were operationalised and items were set out. The first version of the scale included eight dimensions (to assess family roles) in a total of 74 items. All items were assessed on an ordinal scale (a Likert-type scale) ranging from 1 to 5 (Never, Rarely, Often, Always and Does not apply). gender, conceiving it from a functionalist perspective at both the family structure and socialisation process levels. The author reminds us that, from the very beginning, women were socialised to play an expressive role as leaders within the family, which guaranteed the well-being of the social unit. On the contrary, men were socialised to act as providers and meet the family's needs (Amâncio & Oliveira, 2002). The concept proposed by Parsons in the 1950s evolved and has now a central position in the theory of social sciences. It was intensely worked upon in the 1980s within the field of gender studies and is now a key concept to analyse family-related roles, as it is also within the family that gender roles are produced, reproduced and manifested. The literature search confirmed the evolution of the concept of role throughout the years. According to Burr (1998), the concept of role corresponds to the set of behaviours, duties and expectations attached to occupying a particular position in the social hierarchy. The author adds that applying such concept to sexual roles corresponds to a set of behaviours, expectations and duties attached to belonging to a specific group. The latest conceptions show that family roles may be understood as established patterns of behaviour of family members based on the goals set out by the family itself (Wright & Leahey, 2002). A role is a constant behaviour in a given situation; it is developed in the interaction between individuals and influenced by different standards, beliefs and values. Family members take on different roles throughout the individual and family lifecycle. Despite the lack of studies in the area of family roles, Hanson (2005) identifies eight roles associated with the spouse/partner: provider, housekeeper, child-caregiver, socialiser, sexual partner, therapist, recreational organiser and kinship member. The provider role, or head of family role, relates to the need to protect and ensure the necessary income to meet the family's needs. The housekeeper role relates directly to all types of household chores, including house maintenance and gardening. The child-caregiver role entails caring for one's children to satisfy their basic human needs for safety and entertainment. The socialiser role implies the interaction between family members and external members. The sexual partner role is characterised by a set of actions such as sharing of affection, emotional support and interest in sex life -everything The scale included two groups of questions: Group I -socio-demographic variables; and group II -family role variables.

Procedures
Following the development of the questionnaire, the empirical phase of the study started. To implement this scale and aiming at accessing a large number of participants, collaboration and consent were obtained from a group of students attending a higher education institution. After institutional permission was granted, we met with the students and explained them the objectives and purpose of the study, as well as the organisations and institutions involved. They were also informed of their right to refuse to participate. Therefore, the access to the families for completion of the data collection tool was facilitated through the higher education institution and, consequently, the students, who acted as a link between us and their families. A total of 950 questionnaires were distributed to several households, with 918 questionnaires being returned. Of these, nine were removed, because they had not been completely filled in. Thus, a total of 900 questionnaires were used. The collected data were codified, stored and, subsequently, processed. Data were analysed using the statistical software SPSS, version 18.0. In a first stage, data were analysed using descriptive statistics, namely measures of central tendency, dispersion and frequency. The psychometric characteristics of the EAPF were calculated using the Principal Components Analysis with orthogonal Varimax rotation. The number of factors was selected following the guidelines recommended by Polit and Hungler (1997): (1) eigenvalues >1; (2) the exclusion of factor loadings less than 0.30; (3) the application of the principle of discontinuity. The internal consistency of the scale was calculated using Cronbach's alpha value and the test-retest method.

Principal Components Analysis
The Principal Components Analysis (PCA) assesses item/factor loadings (item loading on each factor, thus indicating the item-factor covariance). A total of 14 factors (roles) were identified through this statistical procedure; however, as Hanson (2005) advocates, an 8-factor solution was forced for conceptual reasons. The same sphericity coefficient (KMO) was obtained with both 8-factor and 14-factor solutions. A KMO of 0.916 (p=0.000) was obtained, which implies a significant sphericity for our factors. As a result, the chi-square increases, which implies a strong relationship between the variables in the factors. With regard to factor analysis, and our concern being the development of a reliable instrument to assess the different family roles, we decided to delete all items loading less than 0.3, i.e. five items were deleted. Subsequently, a new PCA was performed (Table 1). A KMO = 0.916; c 2 = 35065.895 and p=0.000 were obtained, as well as a Cronbach's alpha of 0.894 for all items, which is excellent according to Hill and Hill (2000). Thus, after applying the abovementioned tests, the KMO value reinforced a very good correlation between the different variables, while the Bartlett's test of sphericity showed a very significant correlation between most variables of the EAPF scale. Internal consistency was also very good in each factor, namely: Factor 1 with a Cronbach's alpha of 0.963; Factor 2 with a Cronbach's alpha of 0.923; Factor 3 with a Cronbach's alpha of 0.837; Factor 4 with a Cronbach's alpha of 0.811; Factor 5 with a Cronbach's alpha of 0.839; Factor 6 with a Cronbach's alpha of 0.695, Factor 7 with a Cronbach's alpha of 0.719, and Factor 8 with a Cronbach's alpha of 0.624. Thus, after several simulations were performed using factor analysis, internal consistency and intraclass correlation coefficient (consistency and absolute agreement) and after adjustments have been made, the final version was reduced to 68 items with the following characteristics: The EAPF measures different family roles. The scale has eight subscales to assess the various roles: provider or head of family (6 items); Housekeeper (10 items); Child-caregiver (14 items); Socialiser (4 items); Sexual partner (6 items); Therapist (11 items); Recreational organiser (14 items); Outside home care (3 items). The internal consistency was assessed through Cronbach's alpha coefficient. This statistical method is indicated for Likert-type scales and assesses whether the total test variance is associated with the sum of the individual item variance. The internal consistency of the scale was alpha = 0.894. All subscales demonstrated adequate internal consistency values: the lowest value was 0.624 and concerned the care outside the home role (3 items), and the highest value was 0.963 and referred to the childcaregiver role (14 items). A Cronbach's alpha >0.80 is an indicator of good internal consistency, with values greater than 0.60 being considered acceptable in scales with small numbers of items (Ribeiro, 1999;Freire & Almeida, 2001). The mean score was calculated for each subscale of the EAPF, i.e. the sum of the subscale was divided by the number of applicable items (Table 2). According to Table 2, factors 7 and 8 obtained the lowest weighted scores, which indicate that these roles were hardly visible in the families under study, contrary to factors 3 and 1, which were highly visible roles in the same context.

Conclusion
With this study, we aimed at describing the development and validation of a scale that assessed the various family roles, which may be performed by different members of a family unit. Theoretical, empirical and analytical procedures were used. At the theoretical level, it was important not to proceed with the design and validation processes without first duly explaining the underlying theory. At the empirical level, the plan, the various stages and the measures to be considered in the different moments or stages of the implementation and validation of the scale were explained. Finally, the statistical analytical procedures highlighted the sensitivity, accuracy and validity of the instrument. It was also our purpose to include in our sample only those individuals who showed no physical or cognitive dependence. The findings in this study allowed us to conclude that the EAPF meets validity and reliability criteria. The instrument is well structured, with clear terminology and, though being somewhat lengthy, it was well accepted by participants. As regards the item contents, the study of the inter-subscale correlations revealed no atypical or contradictory associations. It is an easy-to-apply instrument, which makes it particularly suitable to be used in populations with low levels of education and facilitates its possible use in oral format. As the promotion of family health is a challenge for Nursing, the strategy to (re)focus on the distribution of family roles may be considered as essential for the families' well-being and to achieve healthy families and, therefore, healthy communities. Thus, we suggest the implementation of the EAPF as an inclusive assessment instrument that is part of the clinical records of families registered with Family Health Units and/or Health Care Centres. The data obtained through the application of this scale may be of interest to the decision-making process on the provision of care to the families. We recommend that further studies on the use of the EAPF should take into account different types of families and compare the different realities experienced in both healthy families and families in which some of the members have chronic diseases.