Date on Master's Thesis/Doctoral Dissertation


Document Type

Doctoral Dissertation

Degree Name

Ph. D.

Department (Legacy)

Department of Educational and Counseling Psychology

Committee Chair

Hardesty, Patrick H.

Author's Keywords

Adolescent risky behaviors; Attachment-infancy; Childhood loneliness; Structural equation modeling


Attachment behavior in infants; Loneliness in children; Risk-taking (Psychology) in adolescence


The premise that attachment in the early stages of development is influential on future development has been extensively researched. Though research denotes that attachment in early development does influence behavioral outcomes in later stages of development, the path of influence is an indirect one, often involving more temporally proximal mediating variables (i.e. mediating variables that occur in the time between the assessment of attachment in the early stages and the measurement of behavioral outcome variables in the later stages of development). Previous research in this area has identified various mediating variables: relationship variables, individual child characteristics, environmental variables, and behavioral constructs. Of these constructs, internalizing behaviors is not as extensively researched. Also, these mediating variables are generally measured at one time point between the predictor variable and the outcome variable which does not adequately represent the dynamic nature of these constructs. This dissertation extends current research by examining the impact of temporal changes in an often neglected construct within the umbrella of internalizing behaviors, childhood loneliness, on the relationship between attachment in the early stages of development and risk behaviors in the later stages of adolescence. This study used data of 825 participants who participated in three of the four phases of data collection for NICHD Study of Early Child Care and Youth Development that occurred from 1991 to 2007. Data were collected from the child/adolescent in the lab using questionnaires on loneliness (during three different time points of data collection) and risky behaviors (at age 15) and behavioral observations for attachment (at 36 months) as well as the demographic variable of child/adolescent gender. Latent growth curve modeling and structural equation modeling were used to examine the proposed model illustrating the hypothesized relationships. Results from this dissertation indicated that a good model fit for the overall structural model; however, upon closer examination, the relationships between early attachment and loneliness as well as early attachment and adolescent risk behaviors produced nonsignificant path coefficients. The significant relationship within the structural model was between changes in loneliness throughout childhood and adolescent risk behaviors. This significance indicates that participants who reported experiencing greater loneliness through childhood would also report higher levels of engagement in risky behaviors in adolescence with observed gender effects for engagement in risky behaviors (i.e. males were more likely to engage in externalizing risk behaviors). The dissertation concludes with implications of the findings and study limitations.