Pfaddiagramm

Pfaddiagramm is a German term used for path diagrams, a graphical representation of causal relationships among variables in path analysis and structural equation modeling (SEM). In these diagrams, variables are depicted as nodes, with directed arrows indicating direct causal effects from one variable to another. Double-headed arrows are used to convey covariances or correlations between variables or error terms. Latent variables are usually shown as circles or ellipses, while observed variables are often represented by rectangles. Measurement models connect latent constructs to their observed indicators with arrows pointing from the latent variable to the indicators.

Historically, path diagrams originate in Sewall Wright’s path analysis from the early 20th century and were subsequently extended into the broader framework of SEM by researchers such as Jöreskog and others in the 1970s. The Pfaddiagramm serves as a compact visual encoding of a system of structural equations that describe the relationships among variables.

Notational and interpretative aspects are central to Pfaddiagramm. Path coefficients attached to directed arrows quantify the strength and direction of causal effects, and can be expressed as standardized or unstandardized estimates. Exogenous variables are those without incoming directed edges, while endogenous variables have at least one antecedent. Model specification, identification, and assessment of fit are essential, as the diagram encodes an entire system of equations that must be estimable from data.

Applications include testing theoretical models of mediation and moderation, estimating direct and indirect effects, and comparing competing structures. Software such as AMOS, LISREL, Mplus, and lavaan in R are commonly used to draw Pfaddiagramm and estimate the associated parameters. Limitations include potential misspecification, reliance on causal assumptions, and sensitivity to sample size and distributional conditions.