The notions of cause and effect are critically important to science. In the behavioral science we are interested in assessing the existance of cause and effect relationships among variables. There is a need in behavioral research for a technique to assess the direct causal contribution of one variable to another 'when experimental control are not applied. Sewall wright first brought path analysis to the attention of researchers in biological sciences in 1918. He has subsequently written three substantial surveys of the technique (1921. 19 34, 1954) for the purpose of clarifying the representation and demonstrating the range of application. The technique of path analysis consists of constructing a diagram to represent the cause and effect relationships hypothesized by the investigator and estimating the parameters from this model. The path coefficient which Wright symbolized by P, denotes the direct causal contribution of one variable to another. The basic assumption of this path model is recursive equation and linearity. An excellent critical survey of the technique was made by Tukey(l954). A very clear summary of the technique has been written by Tuner and Stevens(1959). Path analysis has lately received much attention from sociologists and other field researchers.(Blalock, 1963, 1964, 1965, 1971 : Duncan, 1966: Goldberg & Duncan, 1973: Land, 1969) The purpose of this study is to interpret and discuss the theoretical and statistical assumptions of path analysis and to apply the technique to a set of variables which are supposed to affect language development of a child. From the results obtained by the research, we find that path analysis is a powerful tool for the analysis of causal structures underlying a set of variables. A drawback to path analysis, however, is that there is no way to get a test of the over-all goodness of fit of the path analysis model hypothesized.