Network and AI-Driven Approaches in Psychometrics and Behavioral Science: Concepts and Empirical Case Studies

Department of Social and Behavioural Science, CityU, HK

Jihong Zhang*, Ph.D.

Educational Statistics and Research Methods (ESRM) Program*

University of Arkansas

2025-12-18

About me

 

Let me briefly introduce myself and my current role. My Name is Jihong Zhang.

I am a assistant professor of ESRM program at UArk.

I got my PhD degree in educational measurement & statistics program, focusing on Bayesian psychometric modeling at University of Iowa.

Research Path

My research mainly focus on quantitative methods in psychology, such as psychological network analysis, diagnostic measurement, structural equation modeling, and AI measurement. Besides methodology, my empirical research domains cover students’ learning process, psychopathology topics like emotion regulation or eating disorders, academic motivation, or AI assessment

Presentation Outline

30 minutes

  • Part I: Foundations of network analysis

    • What is network analysis
    • Why use psychological networks
    • Latent factor models vs. Network analysis

  • Part II: Empirical applications

    • EEG-based networks
    • Eating disorder symptom networks

  • Part III: Emerging directions: AI-augmented psychometric research

15 minutes

  • Part IV: Q&A

I’ll walk through four parts: a brief introduction to network analysis, why psychological networks are useful, a case study on EEG network for mediation, eating disorders using longitudinal networks, and then questions.

1 What Is Network Analysis

Network Analysis in General

Network analysis is a broad area. It has many names in varied fields:

  1. Graphical Model (Computer Science, Machine Learning)

    • History: Statistical physics, such as large system of particles (Lauritzen, 1996)
  1. Bayesian Network (Computer Science, Educational Measurement)

  1. Social Network (Sociology, Social psychology)
Figure 1: The Taste for Privacy: An Analysis of College Student Privacy Settings in an Online Social Network (Lewis et al., 2008). Facebook friendship network in a single undergraduate dorm. Red nodes represent students with private profiles; yellow nodes represent students with public profiles.

Note that many fields use network-like models but with different names and goals. The unifying idea is variables as nodes and relationships as edges. Psychological networks are one specific approach within this broader space.

  1. Latent Factor Model or Structural Equation Model (Psychology, Education)
  1. Psychological Network Analysis (Psychopathology, Psychology)

Figure 2: Exploratory factor analysis and psychological network analysis of Big Five personality (Borsboom et al., 2021)

Commonality: Key Components in the network

All network models share common structural elements:

  1. Nodes (vertices): Random variables, individuals, or units of analysis

  2. Edges (links): Relationships or connections between nodes

    • Directed vs. Undirected
    • Weighted vs. Unweighted

  3. Adjacency matrix: Mathematical representation of the network

    • Rows and columns represent nodes
    • Entries indicate presence/strength of edges

  4. Paths and Cycles: Sequences of connected nodes

Before diving into how networks differ, let me clarify the universal building blocks that all network models share:

  1. Nodes (also called vertices) represent the units of analysis. Depending on the network type, nodes can be variables (psychological symptoms, cognitive abilities), individuals (people in social networks), or other entities. The choice of what constitutes a node is theory-driven and depends on your research question.

  2. Edges (also called links or connections) represent relationships between nodes. Edges can be:

    • Directed (arrows, indicating asymmetric relationships like A → B) or Undirected (lines, indicating symmetric relationships like A — B)
    • Weighted (with numerical values indicating strength of relationship) or Unweighted (binary presence/absence of connection)

  3. Adjacency matrix is the mathematical representation of the network. It’s a square matrix where rows and columns represent nodes, and entries indicate whether edges exist (and their weights if applicable). For example, entry (i,j) represents the edge from node i to node j. This matrix is symmetric for undirected networks.

  4. Paths are sequences of nodes connected by edges (e.g., A → B → C). Cycles are paths that return to the starting node (e.g., A → B → C → A). These concepts are important for understanding how effects can propagate through the network and create feedback loops.

These universal components provide the language for describing network structure, regardless of what the nodes and edges represent in your specific application.

Uniqueness: Statistical dependence & Purpose

Different network models represent distinct types of statistical dependence:

  1. Bayesian networks (DAG): Conditional dependence P(X | Parents)

  2. Social networks: Observed social relationships among social units

  3. Latent factor models: Marginal dependence via common latent causes

    • Observed variables are conditionally independent given latent factors

  4. Psychological networks: Conditional dependence given all other variables

    • Partial correlations (cross-sectional) or lagged effects (temporal)

This slide highlights the key statistical distinction across network frameworks:

  1. Bayesian networks encode conditional dependence through directed edges. Each variable depends on its parents via conditional probability distributions P(X|Parents). Edges represent direct causal influences in the assumed structure model, making them useful for causal inference.

  2. Social networks model observed relationships among social units (people, organizations) rather than statistical dependence between random variables. Edges represent connections like friendships or collaborations.

  3. Latent factor models explain marginal dependence (correlations) through common latent causes. The crucial property is that observed variables are conditionally independent given the latent factors. We called that local independence. This is the fundamental assumption that distinguishes latent factor models.

  4. Psychological networks capture conditional dependence given all other variables in the system. Edges represent partial correlations (controlling for all other nodes) in Gaussian graphical models, or lagged regression coefficients in temporal models (GVAR). No latent variables are assumed—the network itself is the explanatory model.

The key takeaway: These different dependence structures fundamentally change how we interpret edges and what questions each approach can answer. Choosing the right framework depends on your theoretical assumptions about what generates the observed data.

Psychological Networks and Network Psychometrics

  1. Network psychometrics is a novel psychometric area that represents complex phenomena of measured constructs as sets of elements that interact with each other (Isvoranu et al., 2022).

  2. It is inspired by the mutualism model and research in ecosystem modeling (Kan et al., 2019).

    • The mutualism model proposes that basic cognitive abilities directly and positively interact during development.

  3. In recent years, interest in psychological networks as dynamics or reciprocal causation among variables receive more attention.

Network psychometrics treats symptoms or behaviors as mutually interacting parts of a system. Inspired by mutualism, we look for reinforcing cycles. This viewpoint is useful when feedback loops, not latent traits alone, drive what we observe.

2 Why use psychological networks?

Latent Factor Model VS. Psychological Network Model

  • Latent factor model (common factor model) assumes that associations between observed features can be explained by one or more common factors (e.g., agreebleness, conscientiousness).

  • Psychometric networks, however, assume that associations between observed features are the reason for the development of one system. In this view, “personality” is the network itself.

Exploratory factor analysis and psychological network analysis of Big Five personality (Borsboom et al., 2021)

Contrast: factor models explain covariation via common causes; network models explain it via direct interactions among observed features. I’ll emphasize when each is useful and how the interpretation of “personality” differs between the two views.

Common factor or Mutualism?

  • “Openness” dimension:

    • Q5 (Creativity). … comes up with new ideas: Disagree (1) to Agree (5) 1
    • Q10 (Curiosity). … curious about lots of different things: Disagree (1) to Agree (5)

  • Does “openness” really exist?

Figure 3: Common factor
Figure 4: Mutualism

 

I’ll use “Openness” as an example. Is the shared variance due to a latent factor, or do the items reinforce each other over time? The mutualism perspective suggests growth through interactions among observed skills or traits.

  1. Source: https://arc.psych.wisc.edu/self-report/big-five-inventory-bfi/

Research aims: Density, Centrality, Pathways, and Group Differences

Aim 1. Network Density

Assess overall connectivity and sparsity (system activation likelihood)

Aim 2. Identify Pathways

Map symptom interactions and feedback loops (A ↔︎ B ↔︎ C ↔︎ A)

Aim 3. Intervention Targets

Identify the most central/influential nodes for treatment

Aim 4. Group Differences

Compare network structures across populations

These analytical goals help bridge theory and practice, guiding both our understanding of psychological phenomena and clinical decision-making.

Psychological network analysis serves multiple applied and theoretical goals:

  1. Identify pathways: We can map how symptoms or behaviors influence each other over time, including feedback loops (e.g., insomnia → fatigue → worry → insomnia). Understanding these pathways reveals mechanisms of symptom maintenance.

  2. Group differences: We can compare network structures across different populations (e.g., by gender, diagnosis, treatment response) to understand how symptom dynamics differ and tailor interventions accordingly.

  3. Intervention targets: By calculating node centrality measures, we identify which symptoms are most influential in the network. Targeting these central nodes in treatment may have cascading positive effects on the entire system.

  4. Network density: A denser network (more edges) indicates stronger interconnections among symptoms. Individuals with denser networks may be more vulnerable because activating one symptom more easily triggers others, increasing risk for comorbidity.

  5. Symptom communities: Clustering algorithms can identify groups of symptoms that tend to co-occur, revealing distinct sub-syndromes within broader diagnostic categories. This can inform dimensional approaches to psychopathology.

Data types: Cross-sectional vs. Longitudinal network

  1. Cross-sectional network

    • Data: Multivariate cross-section data
    • Model: Gaussian graphical model (GGM)
    • Dependence: The variables depend on other variables at the same time
    • Edge Statistics: Partial Correlation (r)
    • Methodology: Correlation analysis

  1. Longitudinal network (today’s focus)

    • Data: Multivariate time-seires data
    • Model: Graphical vector autoregressive (VAR) model
    • Dependence: The values of the variables at time t depend linearly on the values at time t−1
    • Edge Statistics: Lag-1 regression coefficient (β)
    • Methodology: Granger-causality

3 Case Study 1: A longitudinal Electroencephalography (EEG) network analysis

Brain Activity Networks

The human brain is one of the most complex networks in the world.

– Farahani et al. (2019)

studies on its static and dynamic properties have undergone explosive growth in recent years

Research background

  • The team utilize a meditation training called Focused Attention (FA) meditation to improve participants’ meditation quality.

  • Key question:

    • What is the relationship between brain activities and participants’ meditation outcomes.

  • Why use network analysis for EEG data?

  1. Functional connectivity: Brain regions interact and coordinate during cognitive processes
  2. Brain dynamics: Alpha and theta rhythms reflect neural communication across brain regions
  1. Topographical patterns: Multiple brain regions (frontal or posterior) form interconnected systems
  2. Temporal changes: Allow us to track Longitudinal effects which indicates brain network activities developing with meditation training

Network analysis is well-suited for EEG data in meditation research for several reasons:

  1. Functional connectivity: The brain operates as a complex network where different regions interact and coordinate during cognitive processes like meditation. Network models can capture these dependencies among brain regions, revealing how they work together to produce meditative states.

  2. Brain dynamics: EEG measures brain activity (alpha, theta waves). These rhythms don’t occur in isolation—they reflect neural communication across brain regions. Network analysis can model these patterns and identify which regions are most strongly coupled with meditation process.

  3. Topographical patterns: We’re measuring activity across multiple scalp regions simultaneously (frontal, temporal-central, posterior). Rather than analyzing each region independently, network analysis captures the system-level connectivity, such as which regions is the functional hubs, how information flows between regions, and which connections are strongest.

  4. Temporal changes: With 24 sessions over 8 weeks, we can use longitudinal network methods (like GVAR) to examine how brain network structure evolves with meditation training.

By treating EEG signals from different brain regions as nodes in a network, we can identify the functional architecture of meditation and how it develops with practice.

Data

  • Sample: 19 participants completed 8 weeks of FA training with 3 visits per week, resulting in 24 sessions.

  • Training: in each session, they completed 20-minutes of standardized audio-guided Focus Attention meditation practice. EEG data was collected

  • Measure: Self-reported State Mindfulness Scale (SMS_Mind) and body (SMS_Body)

Self-reported State Mindfulness Scale Scores over 24 sessions

Alpha Power (8-12 Hz)

State: Relaxed wakefulness, quiet contemplation, drowsiness, eyes closed.

Function: Associated with suppressing irrelevant sensory input, internal focus, proactive cognitive control, and memory encoding/retrieval.

Changes: Decreases (desynchronizes) during focused attention or demanding tasks, increases when relaxed.

Theta Power (4-8 Hz)

State: Drowsiness, light sleep, REM sleep, deep relaxation, creativity.

Function: Linked to memory formation, navigation, mental effort, executive functions, and threat prediction.Changes: Increases with cognitive effort, drowsiness, and during sleep; inversely related to alpha in some tasks

Nodes

  • 2 State Mindfulness Nodes

    • SMS_Mind: the subjective quality of mindfulness of mind (e.g., thoughts, emotions, mental acuity)

    • SMS_Body: the subjective quality of mindfulness of body (e.g., physical and bodily sensations)

  • 6 EEG Nodes: Alpha (8-12 Hz; relaxed wakefulness) and Theta (4-8 Hz; light sleep) power separated across respective frontal, temporal-central, and posterior regions

    • frontal alpha
    • frontal theta
    • temporal-central alpha
    • temporal-central theta
    • posterior alpha
    • posterior theta

For EEG nodes, there are 3 (region) X 2 (brain waves) = 6 nodes.

Results: regularization

Unregularized psychological network

Regularized psychological network

Alpha power; Theta power; Self-report mindfulness

Pink nodes represent Alpha waves, blue nodes represent theta wave, green nodes represent outcomes. We start from original network which including all longitudnal effects between brain activities with meditation outcomes. In the original form, we see they are interconnected with each other. Utilizing regularization technique, we can then extract key paths from the network.

Then how to interpret the network.

Discussion

  • Brain is dynamical

    1. e.g., Posterior alpha can enhance frontal theta over time

  • Brain region (topography) matters

    1. e.g., Frontal alpha power enhances mindfulness, while posterior alpha power suppresses mindfulness of body and mind over time.

    2. e.g., Posterior theta power enhances mindfulness, while temporal-central theta power suppresses mindfulness

  • Identify ideal brain states

    1. e.g., According to centrality measures, frontal alpha and posterior theta are most important EEG signal for FA meditation training.

  1. Alpha wave and theta wave may change over time, which means your brain is active during the meditation.

  2. Brain wave detected from different brain regions play different roles. Some may decrease the quality and some may increase. They may conflict with each other.

  3. Last, We also identify ideal brain status for meditation. If you have high frontal alpha and posterior theta, you tend to have better meditation experience.

Conclusion

  • In neuroscience, psychological network could be very useful to understand brain activity patterns and their relationships with external behaviours.

  • But, you may wonder if it can be used to analyze the complex mental disorders?

Due to the time limit, I did not show you the full details of the research. But you can tell that, in neurosicence.

4 Case Study 2: Gender Differences for Eating Disorder Symptom Networks

Research background

  • The present study examined sex-specific, symptom-level relationships among emotion regulation (ER), interpersonal problems (IP), and eating disorder (ED) psychopathology in a large sample of Chinese adolescents (Zhang et al., 2024).

  • Background: Eating disorders are serious issues for college students. The eating disorders symptoms are dangerous and co-occur with other psychological issues.

  • Motivation: Given the complex relationships between eating disorders with other risky behaviors, we need a novel model to untangle those interplay which can help with further intervention.

  • Assumption: Eating disorders combined with other risky factors be considered a network?

This project is a collaborative project with Prof. He from CUHK ShenZhen. They have a longitudinal data about eating disorder symptoms of chinese adloscents. They suspected that ED symptoms are not isolated and highly relevant to other problems such as social issues, or emotion regulation.

Some who have eating disorders often have bad emotion contorls and unstable friendship.

So, that is the initial motivation… We want to undertand…

Eating Disorder Network?

  1. Interrelationships among components—emotion regulation, interpersonal problems, and eating disorders:

    • Emotion regulation theory suggests that difficulties in emotion regulation can result in ED behaviors.

    • Interpersonal psychotherapy theory posits that interpersonal problems may exacerbate ED (Murphy et al., 2012).

    • Empirical studies consider these three to constitute an “ecosystem” (Ambwani et al., 2014). Emotion regulation and interpersonal functioning exhibit reciprocal effects on the maintenance of ED.

  • Advantages of longitudinal network

    1. It considers symptoms cause each other over time.
    2. It allows us to identify some symptoms may play the most important roles across time points
    3. It allows sex-specific developmental patterns of eating disorders

The motivation is to obtain a holistic picture of the eating disorders ecosystem. However, the symptom-level dynamics of eating disorders have not been well investigated.

I’ll justify ED as an ecosystem involving emotion regulation and interpersonal problems. Theory and evidence suggest reciprocal effects. We want to see symptom-level dynamics rather than only broad trait scores.

Longitudinal psychological networks fit this problem well. They treat ED behaviors and related traits as interrelated nodes and help us find impactful symptoms while separating temporal from concurrent effects.

Group Comparison and Research Questions

  • In network analysis, groups can be compared from three aspects:

    1. Network structure (e.g., some nodes connected in group A but not in group B).

    2. Node-level measures: node centrality (importance) or node bridging strength (e.g., some nodes may be more connected to other communities).

    3. Network edge weights (e.g., node 1 and node 2 may have a strong relationship in group A but a weaker relationship in group B).

  • Research Questions:

    1. Are there gender differences in the network structures of eating disorder longitudinal networks?

    2. Are there gender differences in the network importance of longitudinal networks?

    3. Are there gender differences in key paths of longitudinal networks?

Group comparisons can target structure, node-level importance, and specific edges. I’ll preview how we test these differences and what they imply for tailored interventions.

We ask three questions: do networks differ by gender in overall features, in which nodes connect, and in which nodes are most influential or act as bridges?

Data & Measures

  • Sample

    • Four waves of data were collected over 18 months.

    • For each wave, demographic information and self-reports on three questionnaires (emotion regulation, interpersonal problems, and eating disorders) were collected from 1,652 high school students in China.

    • After data cleaning, N = 1,540 remained, including 53.9% girls and 46.1% boys.

    • Ages ranged from 11 to 17 years, with a mean of 15.2 years.

  • Measure

    1. Emotion regulation: Difficulties in Emotion Regulation Scale (DERS-18). Six subscales measure different aspects of emotion dysregulation: Awareness, Clarity, Goals, Non-acceptance, Impulse, Strategies.
    2. Interpersonal problems: Inventory of Interpersonal Problems—Short Circumplex (IIP-SC). Eight subscales measure varied aspects of interpersonal problems (e.g., domineering, cold, avoidant).
    3. Eating disorders: 12-item short form of the Eating Disorder Examination Questionnaire (EDE-QS). Twelve items measure different disordered eating behaviors.

We have 26 nodes in the initial networks.

The study follows high school students across four waves over 18 months. After cleaning, we analyze 1,540 students with a balanced gender split and typical adolescent ages.

We include emotion regulation subscales, interpersonal problem subscales, and ED items. This mix lets us see how emotional and social factors interact with specific ED behaviors.

Results

Temporal Network Structure (left: boys; right: girls)

Emotion Disorder; Interpersonal Problems; Eating Regulation

Boys’ temporal network

Girls’ temporal network

These plots show longitudinal effects among interpersonal issues (blue), emotion regulation issue (yellow), and eating disorders (orange) . I’ll point out shared patterns and key differences by gender, setting up the summary tables on the next slides.

Overall Network Structure

Correlation Stability

  1. The multigroup network stability statistics were acceptable.

Boys

  1. Sparsity: 8.06% of non-zero edges (temporal)
  2. Strength: Mean (SD) of edge weights is 0.127 (0.094).

Girls

  1. Sparsity: 10.60% of non-zero edges (temporal)
  2. Strength: Mean (SD) of edge weights is 0.128 (0.102).

Contemporaneous and between-person layers are stable; temporal effects are moderately stable. I’ll note where to be cautious when interpreting lagged edges.

Boys and girls have similar average temporal strengths and density.

Network Hub (Important Symptoms) differences between sex groups

Centrality measures: Boys vs. Girls
  1. Long periods without eating (EDE-WE) and Food preoccupation (EDE-FP) have significant sex differences in node out-strength.
  2. Weight/shape control by vomiting or taking laxatives (EDE-VT) has significant sex differences in node in-strength.

The table compares centrality by sex. Bootstrap tests indicate meaningful differences for specific ED items. I’ll highlight how out-strength and in-strength map to potential symptom drivers and receivers.

Network conncetor (Bridge Symptoms) differences between sex groups

Bridge strength measures how strongly a node connects to nodes in different communities or clusters within the network.
  1. Awareness (Awr) and Goals (Gls) have significant sex differences.
  2. Weight/shape preoccupation (EDE-WP) and Binge eating episode (EDE-BE) have significant sex differences.

Target nodes for intervention on comorbidity.

Bridge strength findings identify nodes that connect ED with emotional and interpersonal domains. Differences suggest gender-specific pathways to comorbidity and potential targets for reducing cross-domain activation.

Discussion

Group Commonalities

  1. Emotion dysregulation has consistent interconnections with eating disorders and interpersonal problems across genders.

Group Differences

  1. Overall network structures of boys and girls are significantly different.
  2. Overall, for boys, the most important bridging nodes were awareness and nonacceptance of the DERS, while weight/shape preoccupation and domineering/controlling emerged as the most central nodes.
  3. For girls, weight/shape dissatisfaction was identified as the most central symptom and the strongest bridging node.

Common patterns: similar overall temporal impact, strong coupling among ED behaviors, and central roles for weight/shape concerns. These are consistent targets across genders.

Differences: denser girls’ networks and different bridge nodes imply higher comorbidity risk for girls and distinct maintenance pathways for boys. I’ll connect these to tailored clinical strategies.

Conclusion

  • Network psychometrics seems very promising to understand complex psychological phenomenon.

  • Is there any other novel psychometric areas? My answer is AI psychometrics.

5 Further studies: AI psychometrics

AI-enhanced psychometrics is becoming more and more important.

AI data augmentation for test development

  • One of my recent project let AI participants can serve as the data points in psychometric assessment given enough accurate human information.

  • Interview-informed large language models can align with real human responses regarding survey responses very well.

Behavioral Regulation in Exercise Questionnaire

In the future, AI will be invoved into psychometric studies, such as AI scoring, item generation, automate assessment. One key challenge of psychometrics, sometime we do not have enough samples with high quality. AI agents with simulated persona may help me.

We are still at the very early stage of AI psychometrics. A lot challeges exists, such as ethnic issues, validity of AI. However, I think as AI grows, we will see more AI-drive methods in psychometrics.

Wrap-up

  1. We talk about the network analysis framework

  2. Two case studies showing how network analysis can be applied to different research areas

  3. In the future, we need novel AI-driven methodology in AI era because AI can serve as research tools or data collection tools.

6 Q&A

Thank you.

Let me know if you have any questions.

You can also contact me via jzhang@uark.edu

Thank you for your attention. You can reach me by email for follow-ups. I’m happy to take questions now.

Reference

References are provided for methods, measures, and prior findings mentioned today. I’m happy to share code and additional materials upon request.

Ambwani, S., Slane, J. D., Thomas, K. M., Hopwood, C. J., & Grilo, C. M. (2014). Interpersonal dysfunction and affect-regulation difficulties in disordered eating among men and women. Eating Behaviors, 15(4), 550–554. https://doi.org/10.1016/j.eatbeh.2014.08.005
Farahani, F., Karwowski, W., & Lighthall, N. (2019). Application of Graph Theory for Identifying Connectivity Patterns in Human Brain Networks: A Systematic Review. Frontiers in Neuroscience, 13, 585. https://doi.org/10.3389/fnins.2019.00585
Isvoranu, A.-M., Epskamp, S., Waldorp, L., & Borsboom, D. (Eds.). (2022). Network Psychometrics with R: A Guide for Behavioral and Social Scientists. Routledge. https://doi.org/10.4324/9781003111238
Kan, K.-J., Maas, H. L. J. van der, & Levine, S. Z. (2019). Extending psychometric network analysis: Empirical evidence against g in favor of mutualism? Intelligence, 73, 52–62. https://doi.org/10.1016/j.intell.2018.12.004
Lauritzen, S. L. (1996). Graphical Models. Clarendon Press. https://books.google.com?id=mGQWkx4guhAC
Lewis, K., Kaufman, J., & Christakis, N. (2008). The Taste for Privacy: An Analysis of College Student Privacy Settings in an Online Social Network. Journal of Computer-Mediated Communication, 14(1), 79–100. https://doi.org/10.1111/j.1083-6101.2008.01432.x
Murphy, R., Straebler, S., Basden, S., Cooper, Z., & Fairburn, C. G. (2012). Interpersonal Psychotherapy for Eating Disorders. Clinical Psychology & Psychotherapy, 19(2), 150–158. https://doi.org/10.1002/cpp.1780
Zhang, J., Cui, S., Zickgraf, H. F., Barnhart, W. R., Xu, Y., Wang, Z., Ji, F., Chen, G., & He, J. (2024). A Longitudinal Network Analysis of Emotion Regulation, Interpersonal Problems, and Eating Disorder Psychopathology in Chinese Adolescents. International Journal of Eating Disorders, 57(12), 2415–2426. https://doi.org/10.1002/eat.24292
1 / 41
Network and AI-Driven Approaches in Psychometrics and Behavioral Science: Concepts and Empirical Case Studies Department of Social and Behavioural Science, CityU, HK Jihong Zhang*, Ph.D. Educational Statistics and Research Methods (ESRM) Program* University of Arkansas 2025-12-18

  1. Slides

  2. Tools

  3. Close
  • Network and AI-Driven Approaches in Psychometrics and Behavioral Science: Concepts and Empirical Case Studies
  • About me
  • Research Path
  • Presentation Outline
  • 1 What Is Network Analysis
  • Network Analysis in General
  • Latent Factor Model...
  • Commonality: Key Components in the network
  • Uniqueness: Statistical dependence & Purpose
  • Psychological Networks and Network Psychometrics
  • 2 Why use psychological networks?
  • Latent Factor Model VS. Psychological Network Model
  • Common factor or Mutualism?
  • Research aims: Density, Centrality, Pathways, and Group Differences
  • Aim 3. Intervention...
  • Data types: Cross-sectional vs. Longitudinal network
  • 3 Case Study 1: A longitudinal Electroencephalography (EEG) network analysis
  • Brain Activity Networks
  • Research background
  • Data
  • Nodes
  • Results: regularization
  • Discussion
  • Conclusion
  • 4 Case Study 2: Gender Differences for Eating Disorder Symptom Networks
  • Research background
  • Eating Disorder Network?
  • Group Comparison and Research Questions
  • Data & Measures
  • Results
  • Overall Network Structure
  • Network Hub (Important Symptoms) differences between sex groups
  • Network conncetor (Bridge Symptoms) differences between sex groups
  • Discussion
  • Conclusion
  • 5 Further studies: AI psychometrics
  • AI data augmentation for test development
  • Wrap-up
  • 6 Q&A
  • Thank you.
  • Reference
  • f Fullscreen
  • s Speaker View
  • o Slide Overview
  • e PDF Export Mode
  • r Scroll View Mode
  • b Toggle Chalkboard
  • c Toggle Notes Canvas
  • d Download Drawings
  • ? Keyboard Help