An Examination of a Functional Mixed-Effects Modeling Approach to the Analysis of Longitudinal Data

Kimberly L. Fine, Hye Won Suk, Kevin Grimm

    Research output: Contribution to journalArticle

    Abstract

    Growth curve modeling is one of the main analytical approaches to study change over time. Growth curve models are commonly estimated in the linear and nonlinear mixed-effects modeling framework in which both the mean and person-specific curves are modeled parametrically with functions of time such as the linear, quadratic, and exponential. However, when more complex nonlinear trajectories need to be estimated and researchers do not have a priori knowledge of an appropriate functional form of growth, parametric models may be too restrictive. This paper reviews functional mixed-effects models, a nonparametric extension of mixed-effects models that permit both the mean and person-specific curves to be estimated without assuming a prespecified functional form of growth. Details of the model are presented along with results from a simulation study and an empirical example. The simulation study showed functional mixed-effects models performed reasonably well under various conditions commonly associated with longitudinal panel data, such as few time points per person, irregularly spaced time points across persons, missingness, and nonlinear trajectories. The usefulness of functional mixed-effects models is illustrated by analyzing empirical data from the Early Childhood Longitudinal Study–Kindergarten Class of 1998–1999.

    Original languageEnglish (US)
    JournalMultivariate Behavioral Research
    DOIs
    StatePublished - Jan 1 2019

    Fingerprint

    Mixed Effects
    Mixed Effects Model
    Longitudinal Data
    Person
    Growth
    Modeling
    Simulation Study
    Trajectory
    Growth Curve Model
    Growth Curve
    Curve
    Panel Data
    Nonlinear Effects
    Parametric Model
    Research Personnel
    Form
    Simulation

    Keywords

    • Functional mixed-effects model
    • longitudinal data analysis
    • nonlinear growth curve modeling

    ASJC Scopus subject areas

    • Statistics and Probability
    • Experimental and Cognitive Psychology
    • Arts and Humanities (miscellaneous)

    Cite this

    An Examination of a Functional Mixed-Effects Modeling Approach to the Analysis of Longitudinal Data. / Fine, Kimberly L.; Suk, Hye Won; Grimm, Kevin.

    In: Multivariate Behavioral Research, 01.01.2019.

    Research output: Contribution to journalArticle

    @article{4789f12414174c7abb577f37b7423897,
    title = "An Examination of a Functional Mixed-Effects Modeling Approach to the Analysis of Longitudinal Data",
    abstract = "Growth curve modeling is one of the main analytical approaches to study change over time. Growth curve models are commonly estimated in the linear and nonlinear mixed-effects modeling framework in which both the mean and person-specific curves are modeled parametrically with functions of time such as the linear, quadratic, and exponential. However, when more complex nonlinear trajectories need to be estimated and researchers do not have a priori knowledge of an appropriate functional form of growth, parametric models may be too restrictive. This paper reviews functional mixed-effects models, a nonparametric extension of mixed-effects models that permit both the mean and person-specific curves to be estimated without assuming a prespecified functional form of growth. Details of the model are presented along with results from a simulation study and an empirical example. The simulation study showed functional mixed-effects models performed reasonably well under various conditions commonly associated with longitudinal panel data, such as few time points per person, irregularly spaced time points across persons, missingness, and nonlinear trajectories. The usefulness of functional mixed-effects models is illustrated by analyzing empirical data from the Early Childhood Longitudinal Study–Kindergarten Class of 1998–1999.",
    keywords = "Functional mixed-effects model, longitudinal data analysis, nonlinear growth curve modeling",
    author = "Fine, {Kimberly L.} and Suk, {Hye Won} and Kevin Grimm",
    year = "2019",
    month = "1",
    day = "1",
    doi = "10.1080/00273171.2018.1520626",
    language = "English (US)",
    journal = "Multivariate Behavioral Research",
    issn = "0027-3171",
    publisher = "Psychology Press Ltd",

    }

    TY - JOUR

    T1 - An Examination of a Functional Mixed-Effects Modeling Approach to the Analysis of Longitudinal Data

    AU - Fine, Kimberly L.

    AU - Suk, Hye Won

    AU - Grimm, Kevin

    PY - 2019/1/1

    Y1 - 2019/1/1

    N2 - Growth curve modeling is one of the main analytical approaches to study change over time. Growth curve models are commonly estimated in the linear and nonlinear mixed-effects modeling framework in which both the mean and person-specific curves are modeled parametrically with functions of time such as the linear, quadratic, and exponential. However, when more complex nonlinear trajectories need to be estimated and researchers do not have a priori knowledge of an appropriate functional form of growth, parametric models may be too restrictive. This paper reviews functional mixed-effects models, a nonparametric extension of mixed-effects models that permit both the mean and person-specific curves to be estimated without assuming a prespecified functional form of growth. Details of the model are presented along with results from a simulation study and an empirical example. The simulation study showed functional mixed-effects models performed reasonably well under various conditions commonly associated with longitudinal panel data, such as few time points per person, irregularly spaced time points across persons, missingness, and nonlinear trajectories. The usefulness of functional mixed-effects models is illustrated by analyzing empirical data from the Early Childhood Longitudinal Study–Kindergarten Class of 1998–1999.

    AB - Growth curve modeling is one of the main analytical approaches to study change over time. Growth curve models are commonly estimated in the linear and nonlinear mixed-effects modeling framework in which both the mean and person-specific curves are modeled parametrically with functions of time such as the linear, quadratic, and exponential. However, when more complex nonlinear trajectories need to be estimated and researchers do not have a priori knowledge of an appropriate functional form of growth, parametric models may be too restrictive. This paper reviews functional mixed-effects models, a nonparametric extension of mixed-effects models that permit both the mean and person-specific curves to be estimated without assuming a prespecified functional form of growth. Details of the model are presented along with results from a simulation study and an empirical example. The simulation study showed functional mixed-effects models performed reasonably well under various conditions commonly associated with longitudinal panel data, such as few time points per person, irregularly spaced time points across persons, missingness, and nonlinear trajectories. The usefulness of functional mixed-effects models is illustrated by analyzing empirical data from the Early Childhood Longitudinal Study–Kindergarten Class of 1998–1999.

    KW - Functional mixed-effects model

    KW - longitudinal data analysis

    KW - nonlinear growth curve modeling

    UR - http://www.scopus.com/inward/record.url?scp=85063143220&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=85063143220&partnerID=8YFLogxK

    U2 - 10.1080/00273171.2018.1520626

    DO - 10.1080/00273171.2018.1520626

    M3 - Article

    JO - Multivariate Behavioral Research

    JF - Multivariate Behavioral Research

    SN - 0027-3171

    ER -