Measure of perceived social support during adolescence (APIK)

European Journal of Investigation in Health, Psychology and Education 2019, Vol. 9, Nº 2 (Págs. 83-94) © European j investiga e-ISSN 2254-9625 // www.ejihpe.es doi: 10.30552/ejihpe.v9i2.322 Measure of perceived social support during adolescence (APIK) Iker Izar de la Fuente, Aránzazu Rodríguez-Fernández and Naiara Escalante Mateos University of Basque Country (Spain) Social support has become a subject of great interest in different scientific disciplines, what has generated a great diversity of definitions that have resulted in a high quantity and heterogeneity of measuring instruments. In addition, there are hardly any instruments that differentiate the sources of social support despite it is considered a vital component. The objective of this investigation is the creation and validation of a questionnaire to measure perceived social support (APIK) in Secondary School. 323 students of Secondary Education of a high school of Vitoria-Gasteiz aged between 12 to 18 years old (M=14.41; DT=1.18) participated in the study, being 128 of them boys (40%) and 195 girls (60%). The questionnaires APIK, AFA-R and HBSC were used. The obtained results of the exploratory factor analysis confirm the three-dimensional structure (family, friends and teachers) of the questionnaire, which offers excellent indexes of internal consistency and reliability. In consequence, the new questionnaire manages to resolve the limitation of other measuring instruments that evaluated the sources separately, did not fit with the theoretical model of social support; were too long, etc. Keywords: Perceived social support, validation, questionnaire, exploratory factor analysis. Medida del apoyo social percibido (APIK) en la adolescencia. El apoyo social se ha convertido en un tema de gran interés en distintas disciplinas científicas lo que ha generado una gran diversidad de definiciones que han derivado en una elevada cantidad y heterogeneidad de instrumentos de medida. Además apenas existen instrumentos que diferencien las fuentes de apoyo social pese a considerarse un componente vital. De este modo, el objetivo de esta investigación es elaborar y validar un instrumento para medir el apoyo social percibido en Educación Secundaria. Participaron en la investigación 323 estudiantes de Educación Secundaria de un instituto de Vitoria-Gasteiz con edades comprendidas entre los 13 y 18 años (M=14.41; DT=1.18), de los cuales 128 eran chicos (40%) y 195 (60%) chicas, a los que se aplicaron los cuestionarios: APIK, AFA-R y HBSC. Los resultados obtenidos confirman la estructura tridimensional (familia, amigos y profesorado) del cuestionario, con excelentes índices de consistencia interna y fiabilidad. En consecuencia, el nuevo cuestionario consigue solventar la limitación de otros instrumentos de medida que evaluaban por separado las fuentes, no encajaban con la concepción teórica del apoyo social, eran demasiado extensos, etc. Palabras clave: Apoyo social percibido, validación, cuestionario, análisis factorial exploratorio. Correspondence: Iker Izar de la Fuente. University of Basque Country. C/ Juan Ibáñez de Sto. Domingo, Nº 1. C.P.: 01006. Vitoria-Gasteiz, Álava (Spain). E-mail: [email protected] IZAR et al. APIK measure of social support 84 European j investiga Vol. 9, Nº 2 (Págs. 83-94) Since the nineteen-seventies (Barrón, 1996), when Cassel (1974a, 1974b), Cobb (1976) and Caplan (1974) first consolidated the construct as an object of study, social support has become an issue of great interest (Alemán, 2013), giving rise to a wide variety of definitions and measurement instruments (Azpiazu, 2016). This makes its specific conception difficult (Terol et al., 2004), and although it can be understood generically as the set of resources people provide each other (Cohen, Mermelstein, Kamarck, & Hoberman, 1985), the construct continues to be ambiguous in terms of its definition, composition and evaluation (Sarason & Sarason, 2009). Of the many ways in which social support can be understood, two definitions stand out as being the most widespread: that proposed by Tardy (1985) and that proposed by Lin (1986). Tardy (1985) identified five hierarchically-ordered dimensions: (1) Direction: referring to whether support is being given or received; (2) Disposition: the assessment of the availability of the support, as well as real support; (3) Description/evaluation: how satisfied people feel with their support; (4) Content: the type of support, with emotional (providing empathy, affection, etc.), instrumental (helpful behaviors), informational (giving advice) and appraisal (feedback) support being the most common; and (5) Network: i.e. the different sources providing support. However, it is the second definition, proposed by Lin (1986), that will be taken as a reference in this study, since it is one of the most comprehensive (Ramos, 2015), understanding social support as the "perceived or actual instrumental and/or expressive provisions supplied by the community, social networks, and confiding partners" in both everyday situations and moments of crisis (Lin, 1986). This definition rests on three key independent pilars: -Support provided and support perceived: A distinction is made between the actual support that is offered to someone and the support perceived by the recipient (Herrero, 2004). -Context in which the support occurs: Lin (1986) refers to the help that may be supplied in each context: (1) Community: superfluous relationships. (2) Social networks: the most common social contacts (teachers, colleagues, etc.). (3) More intimate and trust-based relationships: family and close friends. -Types of support: A distinction is made between support to achieve a specific goal (instrumental) and support as an end in itself, to satisfy emotional needs (expressive) (Lin, 1986). This distinction is not the most widely accepted, with that proposed by Schaefer, Coyne and Lazarus (1981) proving more popular. According to these authors' proposal, there are three types of support: (1) emotional support, linked to the affective sphere (attachment and comfort); (2) tangible support, or the material help or actions aimed at resolving problems (donations, care, etc.); and (3) informational support, or giving advice and information such as feedback. The majority of studies do not distinguish between the sources of social support within the dimensions of the construct (Hombrados & Castro, 2013), despite this factor being IZAR et al. APIK measure of social support European j investiga Vol. 9, Nº 2 (Págs. 83-94) 85 considered a vital component during adolescence (Musitu & Cava, 2003). It is therefore important to study these sources separately (Mendoza, Carrasco, & Mendoza, 2000) in order to understand the real experience of support (Hombrados & Castro, 2013). Although some studies have focused on family and friends (Ramos, 2015), ecological models also highlight the school as a key social context (Hombrados & Castro, 2013). Adolescents spend a large percentage of their time at school with their teachers, and are heavily influenced by them (Eccles & Roeser, 2003). This line of thinking is also consistent with that advocated by Vieno, Santinello, Galbiati, and Mirandola (2004), who argue that it is the social support provided by teachers, family and peers that is most important during adolescence. Social support measurement instruments During adolescence, social support is a strong protector against a wide range of adversities, and as such is necessary for ensuring healthy development (Demaray & Malecki, 2002). It is therefore vital to have valid, reliable measurement instruments that guarantee a precise evaluation of social support during this development period, as well as in the educational environment. Unfortunately, following an analysis of available questionnaires, the conclusion was reached that such an instrument does not yet exist. The disparity of criteria (Laireiter & Baumann, 1992) regarding social support has given rise to a large number and variety of questionnaires (Alemán, 2013), with only a minority being designed specifically for adolescents (Pastor, Quiles, & Pamies, 2012). The first problem is that the majority of the instruments examined fail to analyze the source of the support provided (Barrera, 1980; Barrera, Sandler, & Ramsay, 1998; Bridges, Sanderman, & van Sonderen, 2002; Cutrona & Russell, 1987; Kliem et al., 2015; Kempen & Van Eijk, 1995; Trujillo et al., 2012). Secondly, many were developed for very specific contexts, such as the clinical practice (Holden, Lee, Hockey, Ware, & Dobson, 2014; Moser, Stuck, Silliman, Ganz, & Clough-Gorr, 2012; Richman, Rosenfeld, & Hardy, 1993; Sherbourne & Stewart, 1991) and sport (Freeman, Coffee, & Rees, 2011). Of those questionnaires which do identify sources, some analyze each source individually, measuring the support provided by family separately from that provided by friends (Procidano & Heller, 1983; Domínguez, Salas, Contreras, & Procidano, 2011) or focusing only on the support supplied by teachers (Moreno et al., 2012). Others analyze only two sources together: family and friends (Alemán & Calvo, 2006; González & Landero, 2014; Hardesty & Richardson, 2012; Macdonald, 1998). Of those which focus on the family, some instruments include siblings (Scholte, van Lieshout, & van Aken, 2001) or romantic partners (Calvo & Díaz, 2004; Furman & Buhrmester, 2009; Zimet, Dahlem, Zimet, & Farley, 1988), despite the fact that these figures are not considered principal sources of support during this life stage (Hombrados & Castro, 2013). Moreover, most are too long (Gordon-Hollingsworth et al., 2015). IZAR et al. APIK measure of social support 86 European j investiga Vol. 9, Nº 2 (Págs. 83-94) A few of the questionnaires analyzed measure all three dimensions of social support together (family, friends and teachers), but they do so along with other sources (classmates, grandparents, etc.). This either renders the questionnaire overly cumbersome (Malecki & Demaray, 2002) or prevents the items from correctly fitting into the theoretical conception of social support (Furman & Buhrmester, 1985; Harter, 1985; Pastor et al., 2012). In light of these limitations, a need was identified for a new instrument to evaluate perceived social support among school-going adolescents aged between 12 and 18 years. The questionnaire should divide the construct into three dimensions (support from family, friends and teachers), each comprising three subscales (emotional, tangible and informational support). The aim of this study was therefore to develop and validate such an instrument for measuring perceived social support in Secondary Education. METHOD Participants The initial sample comprised 331 students, of which 8 were eliminated due to a failure to respond to more than 1% of the items, and 6 were eliminated due to inconsistent response patterns. The final sample therefore comprised 323 secondary school students aged between 13 and 18 (M=14.41; SD=1.18). All were from a semi-private school (i.e. a private school which receives some state funding) in the province of Álava with a medium-high socioeconomic level. As regards the sex of the sample group, 60.4% were girls and 39.6% boys, with both groups being evenly balanced (  ²=4.87; p.05). The sample was selected using an incidental procedure. Measurement instruments Participants were asked to complete three questionnaires, all with 5-point Likert-type response scales. To evaluate perceived social support, two additional questionnaires were administered alongside the one being validated in this study. Firstly, to assess perceived social support from family and friends, the Perceived Social Support from Family and Friends (AFA-R) questionnaire by González & Landero (2014) was used. This instrument evaluates two dimensions (support from family and support from friends) and is made up of 15 items with an internal consistency of α=.87 for this study. Secondly, the teacher support subscale of the HBSC questionnaire (Moreno et al., 2012) was used to measure perceived social support from teachers. This scale comprises 8 items and had an internal consistency of α=.92 for this study. Procedure A meeting was held with the management team of the school, and once official agreement to participate had been received, the authors went to the classrooms to administer IZAR et al. APIK measure of social support European j investiga Vol. 9, Nº 2 (Págs. 83-94) 87 the questionnaires to students. All students signed an informed consent form prior to participating. Questionnaires were administered individually and collectively by the authors in class time and in the classrooms themselves. The battery had previously been tested to ensure that it took no longer than 30 minutes to complete. Data analysis Subjects who failed to respond in over 1% of the items were eliminated, along with those with aberrant response patterns detected using the clusters test. Missing values (less than 1%) were replaced using the linear trend estimation at the point. Next, a normality analysis was conducted, revealing some normal and some nonnormal data. Although this was taken to indicate a violation of normality, the skewness and kurtosis values were not too far from normal and, moreover, parametric tests are sufficiently robust to this degree of violation (Chok, 2010; Edgell & Noon, 1984; Schmider, Ziegler, Danay, Beyer, & Bühner, 2010). The decision was therefore made to use the data. To validate the results, a statistical analysis of the items was conducted, along with an exploratory dimensionality study (factorization of the main axes iterated with oblique rotation). The internal consistency of the instrument was calculated (Cronbach's alpha) and, finally, to calculate the external validity, the APIK questionnaire was correlated with the other two questionnaires which evaluate the same variables (AFA-R and HBSC). RESULTS Validation of the measurement instrument The questionnaire was found to have adequate psychometric characteristics in terms of factorality, internal consistency and validity. To determine its dimensionality, diverse data relating to the psychometric characteristics of the 27 items of the APIK questionnaire were analyzed: minimum and maximum, central tendency, variability, skewness, kurtosis and the alpha value if the item was eliminated (see table 1). The minimum and maximum data were correct, since they oscillated between 1 and 5. The mean scores for the items were between M=3.04 and M=4.60, with the arithmetic means being above the recommended halfway point of the scale (Nunnally & Bernstein, 1995). The standard deviations oscillated around the recommended value of 1 (Carretero-Dios & Pérez, 2005), exceeding this value in various cases and revealing reduced values for only six items (i6, i16, i20, i23, i27 and i36). In general, the recommended range is [-2, 2] for skewness (Bandalos & Finney, 2010) and [-5, 5] for kurtosis (Bentler, 2005). These conditions were met in all cases. Finally, the elimination of none of the items was found to improve the Cronbach’s alpha value (α = .95), which is a good indication. IZAR et al. APIK measure of social support 88 European j investiga Vol. 9, Nº 2 (Págs. 83-94) Table 1. Descriptive characteristics of the APIK items Min. Max. Mean SD skewness Kurtosis Alfa if the item is eliminated APIK06_FAM_EMO 1 5 4.60 .759 -2.000 3.941 .954 APIK07_FAM_EMO 1 5 4.47 .811 -1.553 1.915 .954 APIK09_FAM_EMO 1 5 4.37 .999 -1.644 1.994 .954 APIK11_FAM_MAT 1 5 4.24 .808 -.845 .296 .955 APIK13_FAM_MAT 1 5 4.14 1.062 -1.022 .153 .955 APIK16_FAM_MAT 1 5 4.50 .793 -1.685 2.863 .955 APIK20_FAM_INF 1 5 4.58 .706 -1.862 3.684 .955 APIK21_FAM_INF 1 5 4.39 .854 -1.415 1.600 .955 APIK23_FAM_INF 1 5 4.54 .739 -1.637 2.394 .955 APIK27_AMI_EMO 1 5 4.53 .777 -1.807 3.041 .955 APIK29_AMI_EMO 1 5 4.51 .805 -1.740 2.568 .955 APIK30_AMI_EMO 1 5 4.47 .801 -1.565 2.265 .955 APIK36_AMI_MAT 1 5 4.38 .767 -1.173 1.234 .955 APIK38_AMI_MAT 1 5 4.23 .923 -1.087 .620 .955 APIK39_AMI_MAT 1 5 4.13 .967 -1.140 1.029 .955 APIK41_AMI_INF 1 5 4.32 .868 -1.396 1.961 .955 APIK45_AMI_INF 1 5 4.33 .826 -1.255 1.447 .955 APIK46_AMI_INF 1 5 4.13 .894 -.870 .271 .955 APIK48_PROF_EMO 1 5 3.45 1.043 -.343 -.279 .954 APIK50_PROF_EMO 1 5 3.21 1.133 -.217 -.601 .954 APIK53_PROF_EMO 1 5 3.04 1.192 -.117 -.741 .954 APIK56_PROF_MAT 1 5 3.43 1.197 -.533 -.566 .954 APIK59_PROF_MAT 1 5 3.75 1.029 -.651 .099 .954 APIK60_PROF_MAT 1 5 3.50 1.082 -.447 -.391 .954 APIK62_PROF_INF 1 5 3.74 1.013 -.779 .389 .954 APIK66_PROF_INF 1 5 3.66 .924 -.406 -.124 .954 APIK68_PROF_INF 1 5 3.42 1.106 -.386 -.402 .954 Total Alfa .955 Note. Legend: Min.=minimum; Max.=maximum; SD=standard deviation Before conducting the exploratory factor analysis, steps were taken to verify the adequacy of the KMO index (.90) and Bartlett's test (χ² [351]=4538.97; p<.001) (Kaiser, 1970; Montoya, 2007). A 3-factor solution was extracted with a total percentage of explained variance of 50.53% and adequate values of around .50 for commonalities (see table 2) (Hair, Anderson, Tatham, & Black, 2004). To test the internal consistency of the APIK questionnaire, Cronbach’s alpha was used, obtaining a value of α=.90 for the total questionnaire, a value well above the recommended minimum of .70 (Nunnally & Bernstein, 1995). External validity was tested through convergent validity, based on an analysis of relations between the study test (APIK) and others with the same measurement objective: the AFA-R (perceived support from family and friends) and HBSC (perceived support from teachers). IZAR et al. APIK measure of social support European j investiga Vol. 9, Nº 2 (Págs. 83-94) 89 Table 2. APIK’s commonalities and factorial weights Commonalities EFA 1 2 3 APIK06_FAM_EMO .583 -.755 APIK07_FAM_EMO .599 -.768 APIK09_FAM_EMO .486 -.664 APIK11_FAM_MAT .323 -.530 APIK13_FAM_MAT .349 -.558 APIK16_FAM_MAT .516 -.715 APIK20_FAM_INF .597 -.798 APIK21_FAM_INF .570 -.764 APIK23_FAM_INF .535 -.719 APIK27_AMI_EMO .431 .652 APIK29_AMI_EMO .396 .616 APIK30_AMI_EMO .609 .775 APIK36_AMI_MAT .414 .623 APIK38_AMI_MAT .397 .616 APIK39_AMI_MAT .365 .598 APIK41_AMI_INF .419 .650 APIK45_AMI_INF .614 .795 APIK46_AMI_INF .451 .626 APIK48_PROF_EMO .529 .700 APIK50_PROF_EMO .566 .697 APIK53_PROF_EMO .536 .723 APIK56_PROF_MAT .396 .631 APIK59_PROF_MAT .546 .734 APIK60_PROF_MAT .739 .846 APIK62_PROF_INF .510 .730 APIK66_PROF_INF .559 .749 APIK68_PROF_INF .610 .767 Note: 3=family; 2=friends; 1=teachers All items saturated within the factor for which they were designed with factor loadings far above the minimum value of .40 (Table 2) (Lozano & De la Fuente, 2009; Morales, 2011), reaching more restrictive minimum values of .70. Table 3. Correlations among APIK, AFA-R and HBSC AFA-R Family AFA-R Friends AFA-R Total HBSC Teachers APIK FAMILY Pearson .804*** .333*** .703*** .328*** P .000 .000 .000 .000 APIK FRIENDS Pearson .302*** .692*** .569*** .195*** P .000 .000 .000 .000 APIK TEACHERS Pearson .372*** .123* .308*** .842*** P .000 .027 .000 .000 APIK TOTAL Pearson .658*** .483*** .688*** .680*** P .000 .000 .000 .000 Note. *p<.05; **p<.01; ***p<.001 The APIK family support dimension (Table 3) was found to correlate closely and positively with the same dimension of the AFA-R questionnaire (r=.804). The APIK support from friends dimension also correlated closely and positively with the same dimension of the AFA-R (r=.692). This trend was also observed in the case of the support from teachers IZAR et al. APIK measure of social support 90 European j investiga Vol. 9, Nº 2 (Págs. 83-94) dimension and the HBSC (r=.842). Moreover, the general APIK support scale was found to correlate closely and positively with the general AFA-R scale (r=.688) and the HBSC scale (r=.680). Finally, the correlations between scales evaluating different sources of support were much weaker. DISCUSSION AND CONCLUSIONS The aim of this study was to develop and validate an instrument for measuring perceived social support in Secondary Education. The results reveal adequate psychometric properties with excellent internal consistency and validity values. The exploratory factor analysis of the APIK questionnaire supported the threedimensional structure of support from family, friends and teachers hypothesized in other studies (Hombrados & Castro, 2013). Consequently, the new APIK questionnaire overcomes the limitations of other measurement instruments which evaluate sources separately or focus on only one of them (Domínguez et al., 2011; González & Landero, 2014; Hardesty & Richardson, 2012; Moreno et al., 2012; Procidano & Heller, 1983). Moreover, the new questionnaire focuses solely on the most relevant sources for adolescents (Hombrados & Castro, 2013), without analyzing others which are not considered relevant (Furman & Buhrmester, 2009; Scholte et al., 2001; Zimet et al., 1988), and having only 27 items also overcomes the limitations of other questionnaires, which are too long (Gordon-Hollingsworth et al., 2015; Malecki & Demaray, 2002). Finally, unlike those of other questionnaires (Furman & Buhrmester, 1985; Harter, 1985; Pastor et al., 2012), the items of the APIK questionnaire correctly fit the theoretical conception of social support. Nevertheless, the study does have some limitations. Participation in the study was limited to a single school with a medium-high socioeconomic level. This may result in a possible variation in the results if the instrument is administered in other schools from more diverse regions. To this end, future research should seek to repeat the study with more varied samples. Finally, this avenue of research could be further developed in the future by conducting a confirmatory factor analysis to verify the model obtained here. Thus, structural equations could be used to confirm a hypothetical model which combines two levels and different hierarchized dimensions: firstly, sources of support and, within these, types of support provided. This would provide a more global, hierarchized and multidimensional view of the construct. Acknowledgment This piece of research forms part of the work carried out by the Basque University System's Consolidated Research Group IT934-16, the results of the PPG17/61 research IZAR et al. APIK measure of social support European j investiga Vol. 9, Nº 2 (Págs. 83-94) 91 project at the University of the Basque Country and the project EDU2017-83949-P of the State subprogram of Knowledge Generation of the Ministry of Economy, Industry and Competitiveness. REFERENCES Alemán, I.L. (2013). Desarrollo y validación del cuestionario de apoyo social V.I.D.A. (vínculos interpersonales de apoyo). Doctoral thesis. Las Palmas of Grand Canary: University of Las Palmas of Grand Canary. Alemán, I., & Calvo, F. (2006). Validación del cuestionario de apoyo social (CAS): Estudio preliminar. Ansiedad y Estrés, 12(1), 63-74. Azpiazu, L. (2016). El ajuste escolar: Un modelo explicativo en función de variables contextuales y personales. Doctoral thesis. Lejona: UPV/EHU. Bandalos, D.L., & Finney, S.J. (2010). Factor Analysis: Exploratory and Confirmatory. In G.R. Hancock & R.O. Mueller (Eds.), Reviewer’s guide to quantitative methods (pp. 93-114). Routledge: New York. Barrera, M. (1980). A method for the assessment of social support networks in community survey research. Connections, 3(3), 8-13. Barrera, M., Sandler, I., & Ramsay, T. (1981). Preliminary development of a scale of social support: Studies on college students. American Journal of Community Psychology, 9(4), 435-447. Barrón, A. (1996). Apoyo social: Aspectos teóricos y aplicaciones. Madrid: Siglo XXI. Bentler, P.M. (2005). EQS 6 Structural equations program manual. Encino: Multivariate Software. Bridges, K.R., Sanderman, R., & van Sonderen, E. (2002). An English language version of the social support list: Preliminary reliability. Psychological Reports, 90(3), 1055-1058. Calvo, F., & Díaz, M.D. (2004). Perceived social support: Psychometric properties of CASPE questionnaire in urban not institutionalized elderly population. Psicothema, 16(4), 570-575. Caplan, G. (1974). Support systems and community mental health: Lectures in concept development. New York: Behavioral Publications. Carretero-Dios, H., & Pérez, C. (2005). Normas para el desarrollo y revisión de estudios instrumentales. International Journal of Clinical and Health Psychology, 5(3), 521-551. Cassel, J.C. (1974a). Psychiatric epidemiology. In G. Caplan (Ed.), Americans handbook of psychiatry (pp. 401-411). New York: Basic Books. Cassel, J.C. (1974b). Psychosocial process and “stress”: Theoretical formulations. International Journal of Health Services, 4(3), 471-482. Chok, N.S. (2010) Pearson’s versus Spearman’s and Kendall’s Correlation Coefficients for Continuous Data. Doctoral thesis. Pittsburgh: Pittsburgh University. Cobb, S. (1976). Social support as a moderator of life stress. Psychosomatic Medicine, 38(5), 300-314. Cohen, S., Mermelstein, R., Kamarck, T., & Hoberman H.M. (1985). Measuring the functional components of social support. In I.G. Sarason & B.R. Sarason (Eds.), Social support: Theory, research and applications (pp. 73-94). Dordrecht: Martinus Nijhoff Publishers. Cutrona, C., & Russell, D. (1987). The provisions of social relationships and adaptation to stress. In W. Jones, & D. Perlman (Eds.), Advances in personal relationships (pp. 37-67). Greenwich: JAI Press. Demaray, M., & Malecki, C., (2002). Critical levels of perceived social support associated with student adjustment. School Psychology Quarterly, 17(3), 213-241. Domínguez, A., Salas, I., Contreras, C., & Procidano, M. (2011). Validez concurrente de la versión mexicana de las escalas de Apoyo Social Percibido de la Familia y los Amigos (PSS-Fa y PSS-Fr). Revista Latinoamericana De Psicología, 43(1), 125-137.