MODELLI STATISTICO-PSICOMETRICI NELLA REVISIONE E COSTRUZIONE DI STRUMENTI DI ASSESSMENT NEL DISTURBO DELLO SPETTRO DELL’AUTISMO: IMPLICAZIONI PER LA PRATICA CLINICA

In recent years, research has been focusing on identifying reliable and valid assessment and evaluation methods to increase knowledge and symptom characterization of Autism Spectrum Disorder (ASD). Although there is well-established evidence of the need to develop valid and reliable instruments, with the introduction of measures there has also been the introduction of methodological problems and limited psychometric quality; for example, reviews of the literature report that only 7% of the studies reviewed included information on the psychometric properties of measures for ASD. Another aspect that should be taken into consideration when choosing instruments for use in clinical settings is the length and complexity of test administration. This thesis, therefore, aimed to highlight the importance of interdisciplinary approaches, such as the use of solid statistical and psychometric methodologies in the field of clinical psychology, particularly for ASD.This approach plays an important role when using tools for diagnosis or early identification of ASD-related symptoms. One example is the use of the Toddler Module of the Autism Diagnostic Observation Schedule (ADOS-2), which is used to estimate a range of risk for ASD in children aged 12 to 30 months. At present, there is no evidence about diagnostic agreement of Toddler Module with the other Modules of the ADOS-2 with which the diagnosis is, instead, placed; moreover, studies investigating its use in clinical practice are limited. The use of statistical and psychometric methodologies also plays an important role in developing and revising test instruments to investigate the core features of the condition, such as the construct of Social Cognition (SC), that complex set of cognitive functions and processes that underlie the ability to understand and act appropriately within social situations.The introduction of SC measures, which are valid and provide normative values, within diagnostic assessments for ASD is crucial not only in terms of accuracy of diagnosis but also for the implementation of individualized treatments that take into account the specific developmental stage of the subject and the abilities to be enhanced. Based on these assumptions, four research studies have been conducted, in which statistical- psychometric methods are applied in the analysis, review, and development of diagnostic and SC measures. Specifically, we investigated the diagnostic agreement between two measures used in early diagnosis, namely the Toddler Module and Module 1 of ADOS-2; furthermore, through the use of psychometric networks, particularly using Exploratory graph analysis (EGA), we attempted to replicate the latent structure of the Toddler Module, with the aim to analyze the connection patterns between items, assessing the internal consistency of the Toddler Module, and identifying items that might not be in line with the general structure of the test.Through the application of EGA, we have, in addition, constructed and proposed a new, easy- to-administer nonverbal test, the Intentions Attribution Comic Strip Task.Psychometric properties and clinical validity were evaluated using the ROC curve, proving to be a valid test to be introduced in clinical settings, and in the assessment pathway for ASD, particularly for people with limited verbal abilities.With a view to use an interdisciplinary approach and to highlight the importance of integrating SC measures alongside the “gold standard” instruments in clinical practice, we reviewed three SC measures, Advanced-Theory of Mind, Emotion Attribution Task and Social Situation Test, according to the Item Response Theory model.The results have highlighted the importance and need for the use of statistical and psychometric methodologies in the field of clinical psychology, with the aim to develop solid psychometric tools, given the impact they can have on the individual in accurate clinical framing and in structuring personalized interventions.