Alejandra Guzman Castillo  ( Department of Basic Sciences, Faculty of Medicine, Catholic University of the Most Holy Concepcion )
Katia Saez  ( Department of Statistics, Faculty of Physical Sciences and Mathematics, University of Concepcion )
Cristhian Perez  ( Department of Medical Education, Faculty of Medicine, University of Concepcion )
Juan Luis Castillo Navarrete  ( Department of Medical Technology, Faculty of Medicine, University of Concepcion )

Objective: To evaluate the factorial structure and internal consistency of SISCO inventory of academic stress in medical students.
Methods: TThe cross-sectional study was conducted at the School of Medicine, University of Concepcion, Chile, in 2014, and comprised second and third year students.Exploratory factor analysis was carried out to determine the reliability of each of the dimensions of SISCO inventory. Descriptive and correlational analyses were also conducted.
Results: Of the 155 students, 121(78.1%)were female. The overall mean age was 20.5+/- 1.55 years (range: 18.9-32.2 years). Of the total, 61(39.4%) students were in the nursing programme, 10(6.5%) in speech therapy, 16(10.3%) in kinesiology, 10(6.5%) in medicine, 20(12.9%) in midwifery and 38(24.4%) in medical technology.For stressor, somatisation and coping dimensions, a factor was identified, with Cronbach’s alpha values of 0.76, 0.86 and 0.52, respectively. Correlation could only be established between the first two dimensions.
Conclusion: Stressors and somatisation dimensions of the SISCO inventory of academic stress were found to be validity and reliabile.
Keywords: Academic stress, SISCO inventory, Psychometrics, Somatisation, Coping. (JPMA 68: 1759; 2018)

Introduction

The psychological, neurological, immune and endocrine systems work collectively as the control mechanism for homeostasis. ‘Stress’ is a neuroendocrine, immune and behavioural response of the body to any demand placed upon it, which arises as a result of an interpretation of threat or danger, enabling the adaptation and survival of the living being.¹The trigger factor of this adaptive response is called a stressor.² Stress is a complex multifactor process in pathophysiological genesis, whose somatic expression can include gastrointestinal (GI) disturbances, obesity, adjustment disorders, depression and others.³When this arises in the context of an educational process, it is referred to as academic stress (AS).⁴ Various studies have shown that AS occurs in elementary students and increases as the student progresses through educational levels, ⁴ reaching its highest levels in college.⁵ Higher education is the point at which AS becomes a response to the higher workload and coincides with the process of separation from the family, joining the job market and adaptation to an unusual environment. The university student population is a group that is subjected to homogenous activity with particularly stressful periods, resulting in exhaustion, reduced interest in studies and difficulties in dealing with their own environmental
challenges. AS has been described as a systematic process of adaptive nature and essentially psychological, presented descriptively in three stages: (i) The student is subjected to a series of demands, which under the assessment of the student, are considered stressors. (ii) These stressors lead to a systemic imbalance (stressful situation), manifested in several symptoms (indicators of imbalance). (iii) This systemic imbalance forces the student to take action to restore the coping system balance.⁶The existence of three procedural systemic AS components have been recognised: (i) stressor stimuli (input), (ii) symptoms suggestive of systemic imbalance, and (iii) coping strategies (output).⁶ Based on this theoretical model, a self-descriptive psychometric instrument of 31 items was developed called the SISCO inventory of academic stress (SISCO-AS).⁷ It was validated in Mexico and then later in Colombia among university populations. 8,9 SISCO-AS has also been used in several studies,^7,10,11showing good reliability with Cronbach’s alpha values between 0.861 and 0.9.However, there are few validation results using this instrument, including in Latin America. In Chile no study has been done on its psychometric properties. When SISCO-AS was presented as a three-dimensional construct of AS from a cognitivist systemic perspective with good psychometric properties,⁷ it supported the idea that each dimension functioned as a factor. Thus, the current study was planned to examine each dimension separately to determine the references to each of the conceptually different variables.This is the first study in Chile to assess the validity and reliability of the construct.

Subjects and Methods

The cross-sectional study was conducted at the School of Medicine, University of Concepcion, Chile, in the first semester of 2014, and comprised second and third year students who were enrolled following short briefings during a mandatory class. Those who volunteered to participate and signed the consent form were included using convenience sampling. After approval was obtained from the institutional ethics committee,SISCOAS was applied on the sample. SISCO-AS comprises 31
items.⁷ The first is an initial filter, in dichotomous terms (yes-no), for determining whether the respondent will continue or not. The second item, with a single question, aims to identify the overall selfperceived level of AS. The third part has eight items to identify the frequency at which the environmental demands are valued as stressor stimuli. The fourth part consists of 15 items and identifies the frequency of symptoms or reactions to the stressor stimulus. Lastly, the fifth part has six items aiming to identify the frequency with which the individual uses coping strategies. Except the first part, all parts use a fivenumber value system similar to the Likert scale (1–5; 1=never; 2=very rarely; =sometimes; 4=almost always; 5=always). To obtain evidence of validity associated with SISCOASconstruct, the factor structure of each part was analysed separately: stressors, somatisation and coping strategies were taken as different variables that could be used as independent instruments. Exploratory Factor Analysis (EFA) is a statistical methodology of interdependence that aims to establish an underlying structure among the analysis variables, through correlation structures between them, defining groups of variables (factors) that are highly correlated with each other. ¹² For EFA, the principal axis method of factor extraction recommended for this type of instrument was used.^12,13 The adequacy of the EFA in collecting data was analysed using the Bartlett's test sphericity for statistical significance (p<0.05), and the Kaiser-Meyer-Olkin (KMO) measure for sampling adequacy, which should be close to 1.0. Three complementary criteria were used to identify the number of factors: (i) the Kaiser-Guttman criterion, or latent root, and (ii) the contrast fall criterion or Cattell (scree test), which are generally used for this purpose,12 were added to the result of (iii) the Horn's Parallel Analysis (based on 5000 random samples). These are considered the most suitable tools for the purpose of a study like the current one.¹⁴Once the number of factors was established, an array configuration using the EFA with the direct Oblimin rotation method was used to determine the organisation of items, including a factor structure. A value of 0.3 was considered the minimum value to accept as an adequate load factor.^12,15,16
Once the factors were identified, Cronbach’s alpha coefficient reliability was calculated for each to estimate their internal consistency. Subsequently, the discriminative capacity of the items was analysed by calculating the correlation of each with the total fixed factor to which they were assigned. Lastly, the factors were descriptively analysed, evaluating the correlation between them using the Pearson coefficient. P<0.05 was considered statistically significant. For data analysis, SPSS 22.0 was used.

Results

Of the 166 students enrolled, 155(93.3%)completed the instrument. Of them, 121(78.1%)were female. The overall mean age was 20.5±1.55 years (range: 18.9-32.2 years). Of the total, 61(39.4%) students were in the nursing programme, 10(6.5%) in speech therapy, 16(10.3%) in kinesiology, 10(6.5%) in medicine, 20(12.9%) in midwifery and 38(24.4%) in medical technology.For the dimension of stressors, the KMO statistic (0.80) and the Bartlett’s test (p<0.001) showed that EFA was pertinent. Kaiser-Guttman’s criteria identified two factors with eigen values of 3.12 and 1.03, which explains 51.90% of total variance of the items. The scree test showed one factor through sedimentation graph, and the Horn’s parallel analysis indicated one factor with an eigen value of 3.12 which was above the eigen values obtained by 95% (95th percentile) of the random samplesThe Cronbach’s alpha was =0.762, with correlation between the items and the total corrected from r=0.38 (item 3.1) to r=0.56 (item 3.5). For the dimension of somatisation, the KMO statistic (0.87) and the Bartlett’s test (p<0.001) showed that the factor analysis was appropriate for the data. Kaiser-Guttman criterion identified four factors with initial eigen values greater than 1.0 (5.20, 1.32, 1.08, and 1.02) that explains 61.62% total variance of the items. The scree test dentified one factor, and the Horn’s parallel analysis identified one factor with an eigen value of 5.20 which was above the 95th percentile of the values obtained in random samples. The Cronbach’s alpha was =0.858, with correlation between the items and the total corrected from r=0.31 (item without harming others 4.5) to r=0.67 (item 4.9). For the dimension of coping, although the KMO statistic was low (0.55), we proceeded with the EFA given the Bartlett’s test value (p<0.001). Kaiser-Guttman criterion identified two factors with initial eigen values greater than 1.0 (1.68 and 1.08) that explains 45.92% total variance of the items. The scree test identified one factor and the Horn’s parallel analysis identified one factor with eigen value of 1.68 which was above the 95th percentile of the values obtained in the random samples. The factor loadings indicated that all items, except items 5.4 (religious practice - prayer or church attendance) and 5.5 (Information search on the situation), showed a configuration coefficient greater than 0.3. Item 5.4 was eliminated and the analysis was repeated but there was no change in values. For the remaining four items, the analysis was considered plausible (KMO=0.60; p<0.001).The Kaiser-Guttman criterion identified a factor with an initial eigen value higher than 1.0 (1.63) that explains 40.87% total variance of the items. The scree test and the Horn’s parallel analysis identified one factor with an eigen value of 1.63 which was above the 95th percentile of the values obtained in the random samples. The Cronbach’s alpha was =0.516, with correlation between the items and the total corrected from r=0.27 (item 5.1) to r=0.34 (item 5.1) (Table-1).



A direct correlation and statistical significance was found, with medium effect size between the dimensions of somatization and the stressors (r(152)=0.58; p<0.01). The dimension of coping did not show significant correlations with any of the other measurements (Table-2).



For interpretative purposes, a score below the 25th percentile on any dimension was considered as a low level of AS, a value equal to or greater than the 25th percentile and under the 75th percentile as a medium level; and a score equal to or greater than the 75th percentile as a high level (Table-3).



Discussion

AS corresponds to a systemic response of the organism to adapt to the demands originating in an educational context. Research has shown AS in elementary students,¹⁷ and that increases as the students progress through educational levels.⁴ This study, the first of its kind in Chile, presents information on some of the psychometric properties of the SISCO-AS and its dimensions (as it refers to conceptually different variables), as well as information on its structure, and evidence of construct validity and reliability among college students pursuing health-related degrees in Concepcion, Chile. The dimensions of stressors and somatisation presented a well-defined factor whereby all items reached significant loads. Furthermore, they had good reliability¹⁸and itemsdemonstrated adequate discriminative ability (r>0.30) (Table-1). The dimension of coping, presented challenges, as the EFA required elimination of two items successively, as significant factor loadings were not reached. Additionally, the resulting factor with the remaining four items showed questionable reliability.¹⁸15 the results contrasted with those reported in Colombian university students in a study which showed the imension of coping as having good reliablity (Cronbach’s alpha=0.7).⁹ It is interesting that the differences between the Colombian and the Chilean populations are not only geographical, but also idiosyncratic and socio-cultural, as reflected in the items eliminated during the analysis process (religious practice and searching for information about their situation). Most importantly, coping is a multidimensional construct, which is not a constituent part of AS, but an associated factor that enables moderation, and has been defined as "those constantly changing cognitive and behavioural efforts that develop to handle specific external or internal demands which are evaluated as excessive or overflowing the individual's resources".¹⁹ It is a dynamic process that involves continuous assessment by the individual strategies aimed to the "action" (actions to modify or alter the problem) and to the "emotion" (actions that help to regulate the emotional responses to the existing problem). This demonstrates that coping is a different construct and supports its exclusion by the second order factor analysis and the absence of correlation with the dimensions of stressors and somatisation. Furthermore, coping, as a multidimensional construct, is difficult to adequately measure with four or six questions, and the existing instruments specific for coping, such as the inventory of coping strategies,has 40 items,20 the coping scale for extreme risks has 36 items,21 and the coping strategies questionnaire has 42 items. ²² Thus, although the factor structure and reliability of the stressor and somatisation dimension are theoretically plausible and support the use of the SISCO-AS, we propose the possibility of dispensing with the dimension of coping, which will require further study. The dimensions of stressors and somatisation demonstrated good reliabilities, but not that of coping. The overall reliability of the instrument, taking into consideration the dimension of coping, was very good,17 and was similar to that reported by studies.^7-9,23 However, it is important to consider that the second order FA showed that the three dimensions of SISCO-AS contributed to a common measurement (Table-1), and that the high reliability observed of the instrument may be due to the effect of a greater number of items that artificially overestimate the Cronbach’s coefficient.¹⁵ As was expected, the dimensions of stressors and somatisation show direct correlations with one another and correlate with the overall stress question of the SISCO-AS and thus coincide with the general appreciation of the subject. Consequently, the dimensions of somatisation and stressors of SISCO-AS constitute instruments with evidence of validity and reliability for their use in research and diagnostic approaches. However, their use as instruments for monitoring and determining AS requires more stringent psychometric standards. ^15,24 It is recommended that a overall score of this instrument should not be drawn. The use of coping dimension, for its unreliability and lack of relation with the question of global assessment of stress is discouraged. Instead, the use of specific tools is recommended to measure the variable coping. To facilitate and define a cut-off point in this study, a ranking of the scores of the subjects was based on 25 and 75 percentiles. These percentile scales were proposed as only a reference, leaving it as a line of future research, the construction of scales in a broader and more heterogeneous sample. This study has several limitations. First, the sample consisted primarily of female students (78.1%), limiting extrapolating to the university community. Second, the sample size, which did not allow the baremation nor allowed for a confirmatory factor analysis of the SISCO-AS.Third, the data comes from a self-reported type of instrument, which, although designed as such, has the potential for errors in the reporting or biased responses of the individuals.

Conclusion

The validity and reliability of the dimensions of stressors and somatisation of SISCO-AS was established, but not that of the coping dimension.

Disclaimer: None.
Conflict of Interest: None.
Source of Funding: Office of the Vice Rector of Research and Development, University of Concepcion and Research and Innovation Direction, Catholic University of the Most Holy Concepcion, Chile.

References

1. Selye H. The Evolution of the Stress Concept: The originator of the concept traces its development from the discovery in 1936 of the alarm reaction to modern therapeutic applications of syntoxic and catatoxic hormones. Am Scientist 1973; 61: 692-9.
2. Selye H. The stress of life. New York: McGraw-Hill, 1956.
3. Tafet G, Nemeroff C. The Links Between Stress and Depression: Psychoneuroendocrinological, Genetic, and Environmental Interactions. J Neuropsychiatry Clin Neurosci 2016; 28: 77-88.
4. Putwain D. Researching academic stress and anxiety in students: some methodological considerations. Br Educ Res J 2007; 33: 207-19.
5. Dyson R, Renk K. Freshmen adaptation to university life: Depressive symptoms, stress, and coping. J Clin Psychol 2006; 62: 1231-44.
6. Barraza-Macias Arturo. A conceptual model for the study of academic stress. [Online] 2006 [Cited 2017 Dec 11]. Available from: URL: http://www.psicologiacientifica.com/estresacademico-modelo-conceptual.
7. Barraza-Macias Arturo. Psychometric properties of the SISCO Inventory of academic stress. [Online] 2007 [Cited 2017 Dec 11]. Available from: URL:http://www.psicologiacientifica.com/siscopropiedades-psicometricas/.
8. Barraza-Macias Arturo. Psychometric analysis of the SISCO Inventory of academic stress in young adults of the University of Sinu. In: Barraza-Macias Arturo; Jaik Dipp, Adla, eds. Estres, burnout y bienestar subjetivo, Mexico: Instituto Universitario Anglo Espanol Red Durango de Investigadores Educativos AC. 2011; pp 92-107.
9. Malo D, Caceres G, Ballesteros G. Validation of the SISCO inventory of academic stress and comparative analysis in young adults of the Industrial University of Santander and Universidad Pontificia Bolivariana, Bucaramanga sectional, Colombia. Prax. Investig. ReDIE 2010; 2: 26.
10. Mejia-Rubalcava C, Alanis-Tavira J, Argueta-Figueroa L, Legorreta-
Reyna A. Academic stress as a risk factor for dental caries. Int Dent J 2012; 62: 127-31.
11. Olivan B, Boira S, Lopez del Hoyo Y. Stress and other psychological factors associated with physiotherapy students. Fisioterapia 2011;33: 19-24.
12. Hair J, Black W, Babin B, Anderson R. Multivariate data analysis. 6th ed. Austin: Prentice Hall, 2010.
13. Tabachnick B,Fidell L. Using Multivariatestatistics. New Delhi: Pearson, 2012.
14. Buja A, Eyuboglu N. Remarks on Parallel Analysis. Multivariate Behav Res 1992; 27: 509-40.
15. Martinez R, Hernandez M, Hernandez MJ. Psychometry. Madrid: Alianza, 2006.
16. Perez C, Ortiz L, Fasche E, Parra P, Matus O, Maccoll P,et al. Psychometricpropertiesof a questionnaire to assess academic expectations in first-year medical students. Rev Med Chile 2015; 143: 1459-67.
17. Connor M. Pupil stress and standard assessment tasks (SATs) An update: An update.Emotional and Behavioural Difficulties 2003; 8: 101-7.
18. George D, Mallery P. In: SPSS for Windows step by step: A simple guide and reference. 11.0 4th ed. Michigan: Allyn & Bacon, 2003.
19. Lazarus R, Folkman S. Stress and cognitive processes, Barcelona: Martinez Roca, 1991.
20. Martinez J. Spanish adaptation of the Inventory of Coping Strategies. Actas Esp Psiquiatr 2007; 35: 29-39.
21. Lopez-Vazquez E, Marvan M. Validation of a coping scale in the face of extreme risks. Salud Publica Mexico 2004; 46: 216-21.
22. Sandin B, Chorot P. Stress Coping Questionnaire (SCQ): Development and preliminary validation. Rev Psicopatologia Psicol Clinica 2003; 8: 39-53.
23. Jaimes R. Validation of the SISCO inventory of academic stress in
young adults of the Universidad Pontificia Bolivariana, Bucaramanga Sectional. [Online] 2008 [Cited 2017 Dec 11]. Available from: URL: https://goo.gl/G2auLi.
24. Nunnally D, Bernstein I. Psychometric theory. New York: McGraw- Hill, 1994.