The limitations of a standard 1-D inversion applied to multidimensional (synthetic) data are investigated. Simple correction procedures for interpreting field data distorted by 3-D structures are suggested. Two different transmitter/receiver configurations of the transient electromagnetic (TEM) sounding method are examined: a central loop configuration for the near zone (sometimes called short offset) and a fixed transmitter/moving receiver configuration for the far zone (long offset). The 3-D models are structural depressions and highs in both resistive and conductive basements. The fixed transmitter (grounded dipole) in the long offset TEM configuration is located at a distance significantly greater than both the size and depth of the structure.In all cases, 1-D interpretation of the central loop soundings recovers geoelectric parameters of the section with good reliability, although fictitious layers may appear near vertical boundaries. The 1-D interpretation of long offset soundings does not, in most cases, show the actual structures. The data along various profiles are interpreted in terms of a two-layer model without the structures. In some cases 1-D interpretation does show the structure, but the errors are far greater than those obtained in the inversion of central loop soundings.In all cases considered, the distortion of central loop soundings caused by 3-D effects is systematic and could, therefore, be corrected by simple procedures. These procedures permit interpretation of real field data that were previously abandoned owing to the strong distortions by lateral inhomogeneities.