The laboratory scale tracer experiments on homogenous and heterogeneous porous media have been increasingly used to study the dispersion phenomenon of contaminants in saturated and un-saturated groundwater system. The present study is concerned with the estimation of scale effect of dispersion coefficient of conservative tracer solute. Tracer test in soil column is one of the reliable techniques for the estimation of dispersion coefficient. Soil samples were collected from different locations in a contaminated area and five different sets of experiments were conducted by varying the dimensions of the column and head in-order to arrive at optimal size of column and the minimum head required. The raw textile effluent considered as conservative tracer was characterized. During the column study, the effluent samples were collected in regular intervals at the bottom of column and its quality parameters were tested. The results reveal that the absorbance rate of various salts by the various soil columns goes on decreasing with length and at a certain point the concentration becomes negligible or zero. The Break through Curve (BTC) was obtained for the concentration difference with time. The longitudinal dispersion coefficient (αL) and transverse dispersion coefficient (αT) are estimated using the sigmoidal curve method. It is found that αT is one tenth of αL and was verified by summation distribution graph approach and the results are found to be more or less same. The scale effect of dispersion coefficient was estimated for different length versus diameter (L/D) ratios. In this paper, a new approach of estimation of dispersion coefficient using BTC has been proposed and evaluated its estimation with that of estimate from Summation Distribution (SD) and Sigmoidal Curve (SC) methods, and these scale effect has been quantified by all these methods. It is found that the estimation of dispersion coefficient and its scale effect from all these methods were matching well. The results also established that the scale effect of dispersion coefficient was predominantly due to the inappropriate L/D ratio. It is recommended that the optimal L/D ratio of 4 to 20 is acceptable without any pronounced scale effect. © 2015 CAFET-INNOVA TECHNICAL SOCIETY. All rights reserved.