Guidelines for the use of structural versus regression analysis in geomorphic studies
Simple-regression and structural analysis are similar methods of developing a linear relation from a bivariant group of data. Regression analysis is a useful curve-fitting technique, but often is misapplied to geomorphic data sets. When error components can be identified for both variables, the statistical technique of structural analysis is preferred. If regression results are available, conversion to a structural analysis can be made either manually or by computer.
Use of computer-generated data sets permits the construct ion of curves relating variation between regression and structural analyses to the range of data of the independent variable. The data have randomly imposed error components of specified standard deviation and a slope of the lin-ear relation that simulates gradient-discharge relations of natural alluvial streams. The empirically developed curves can be used to determine the need for structural analysis of real geomorphic data sets.
Citation Information
Publication Year | 1978 |
---|---|
Title | Guidelines for the use of structural versus regression analysis in geomorphic studies |
DOI | 10.3133/wri78135 |
Authors | W. R. Osterkamp, Jesse M. McNellis, Paul Robert Jordan |
Publication Type | Report |
Publication Subtype | USGS Numbered Series |
Series Title | Water-Resources Investigations Report |
Series Number | 78-135 |
Index ID | wri78135 |
Record Source | USGS Publications Warehouse |