Exploring Predictors of Adolescent Emotional and Behavioral Problems Using Latent Profile Analysis and Machine Learning

2021 ◽  
Vol 26 (3) ◽  
pp. 57-78
Author(s):  
You Rok Do
Author(s):  
Vilmantė Pakalniškienė ◽  
Roma Jusienė ◽  
Sandra B. Sebre ◽  
Jennifer Chun-Li Wu ◽  
Ilona Laurinaitytė

This study explored the profiles of elementary-school-aged children’s Internet use in relation to their emotional and behavioral problems. Participating in this cross-sectional study were 877 child–parent dyads from Latvia, Lithuania, and Taiwan. Children (8–10 years old) provided information on three variables: the amount of time they spent online, frequency of online activities, and knowledge of how to do things online. Latent profile analysis including these three variables provided a four-class solution for child Internet use. A comparison between Latvia, Lithuania, and Taiwan on the percentage of the sample distribution in each class showed that there was no difference between sites for the high class (high ratings on all three variables). The largest differences were for the low and average classes (low and average ratings on all three variables, namely, time online, frequency, and knowledge): the Lithuanian and Taiwanese samples were similar in that a higher percentage of each sample was in the low class, whereas the Latvian sample had children equally distributed between the low class and the average class. Analysis of the data from the entire sample for differences in parent-reported child behavioral difficulties suggested that children in the high class had an elevated level of behavioral problems and compulsive Internet use.


2020 ◽  
Vol 17 (6) ◽  
pp. 517-525 ◽  
Author(s):  
Yunhye Oh ◽  
Jungwon Choi ◽  
Yul-Mai Song ◽  
Kyungun Jhung ◽  
Young-Ryeol Lee ◽  
...  

Objective This study aimed to examine personality profiles and behavioral problems of children with nail biting (NB) to gain insight into the developmental trajectory of pathological NB.Methods 681 elementary school students were divided into non NB (n=436), occasional NB (n=173) and frequent NB group (n=72) depending on the frequency of NB reported in Child Behavioral Checklist (CBCL). Children’s personality was assessed using the Junior Temperament and Character Inventory (JTCI), and behavioral problems were assessed using the CBCL. Latent Profile Analysis (LPA) was performed using JTCI profiles to classify personalities of the children with NB (belonging to frequent and occasional NB group, n=245).Results For subscale scores of CBCL, the total, internalizing, externalizing, anxious/depressed withdrawn/depressed, depression, thought, rule-breaking, and aggressive behavior problems, were most severe in the frequent NB group followed by occasional NB and non NB group. LPA of personality profile in children with NB revealed four classes (‘adaptiveness,’ ‘high reward dependence,’ ‘low self-directedness,’ and ‘maldaptiveness’). The four personality classes demonstrated significant group differences in all of the CBCL subscales. Children who showed low self-directedness and cooperativeness and high novelty seeking and harm avoidance personality profiles demonstrated highest tendency for problematic behavior irrespective of the frequency of NB.Conclusion Children with NB reported significantly more problematic behaviors compared to children without NB. Children with specific personality profile demonstrated higher tendency for problematic behavior irrespective of the frequency of NB. Therefore, accompanying personality profiles should be considered when assessing behavioral problems in children with NB.


2021 ◽  
Vol 13 (1) ◽  
Author(s):  
Shuting Zheng ◽  
Somer L. Bishop ◽  
Tiffany Ceja ◽  
Mona Hanna-Attisha ◽  
Kaja LeWinn

Abstract Objective Children in Flint, Michigan, have experienced myriad sociodemographic adversities exacerbated by the Flint water crisis. To help inform child-focused prevention and intervention efforts, we aimed to describe patterns of neurodevelopmental outcomes among preschoolers who experienced the Flint water crisis before age 2 years. Method Participants were 170 preschoolers who completed a comprehensive neurodevelopmental assessment battery, including directly administered measures of cognitive and executive functioning and maternal-report of adaptive skills and behavioral problems. We used latent profile analysis to derive subgroups. Multivariate multinomial logistic regression was conducted to examine the predictors of profile memberships, including child sex and maternal/family-level factors selected from an array of measured exposures using least absolute shrinkage and selection operator regression. Results Three latent profiles were identified: Profile 1—relative weakness in all domains (50%); Profile 2—normative functioning in all domains (34.1%); and Profile 3—relative strengths in executive function and behavior (15.9%). Profile 1 showed lower scores across cognitive and behavioral domains. Profile 2 demonstrated abilities within the normal range across domains. Profile 3 showed relative strength in executive functioning with few behavior problems, despite lower cognitive performance. Children across all profiles showed adaptive behavior in the adequate range. Child sex and maternal IQ were significant predictors of profile membership. Conclusions Children in Flint demonstrated diverse patterns of development in the face of sociodemographic and environmental adversities. Comprehensive screening and neurodevelopmental profiling of children in this at-risk population are needed to identify areas of needs and inform appropriate service delivery.


Sign in / Sign up

Export Citation Format

Share Document