The effects of agricultural best management practices and in-stream restoration on suspended-sediment concentrations, stream habitat, and benthic macroinvertebrate assemblages were examined in a comparative study of three small, rural stream basins in the Piedmont and Blue Ridge Physiographic Provinces of North Carolina and Virginia between 2004 and 2007. The study was designed to assess changes in stream quality associated with stream-improvement efforts at two sites in comparison to a control site (Hogan Creek), for which no improvements were planned. In the drainage basin of one of the stream-improvement sites (Bull Creek), several agricultural best management practices, primarily designed to limit cattle access to streams, were implemented during this study. In the drainage basin of the second stream-improvement site (Pauls Creek), a 1,600-foot reach of the stream channel was restored and several agricultural best management practices were implemented. Streamflow conditions in the vicinity of the study area were similar to or less than the long-term annual mean streamflows during the study. Precipitation during the study period also was less than normal, and the geographic distribution of precipitation indicated drier conditions in the southern part of the study area than in the northern part. Dry conditions during much of the study limited opportunities for acquiring high-flow sediment samples and streamflow measurements. Suspended-sediment yields for the three basins were compared to yield estimates for streams in the southeastern United States. Concentrations of suspended sediment and nutrients in samples from Bull Creek, the site where best management practices were implemented, were high compared to the other two sites. No statistically significant change in suspended-sediment concentrations occurred at the Bull Creek site following implementation of best management practices. However, data collected before and after channel stabilization at the Pauls Creek site indicated a statistically significant (p<0.05) decrease in suspended-sediment discharge following in-stream restoration. Stream habitat characteristics were similar at the Bull Creek and Hogan Creek reaches. However, the Pauls Creek reach was distinguished from the other two sites by a lack of pools, greater bankfull widths, greater streamflow and velocity, and larger basin size. Historical changes in the stream channel in the vicinity of the Pauls Creek streamgage are evident in aerial photographs dating from 1936 to 2005 and could have contributed to stream-channel instability. The duration of this study likely was inadequate for detecting changes in stream habitat characteristics. Benthic macroinvertebrate assemblages differed by site and changed during the course of the study. Bull Creek, the best management practices site, stood out as the site having the poorest overall conditions and the greatest improvement in benthic macroinvertebrate communities during the study period. Richness and diversity metrics indicated that benthic macroinvertebrate community conditions at the Hogan Creek and Pauls Creek sites declined during the study, although the status was excellent based on the North Carolina Index of Biotic Integrity. Experiences encountered during this study exemplify the difficulties of attempting to assess the short-term effects of stream-improvement efforts on a watershed scale and, in particular, the difficulty of finding similar basins for a comparative study. Data interpretation was complicated by dry climatic conditions and unanticipated land disturbances that occurred during the study in each of the three study basins. For example, agricultural best management practices were implemented in the drainage basin of the control site prior to and during the study. An impoundment on Bull Creek upstream from the streamgaging station probably influenced water-quality conditions and streamflow. Road construction in the vicinity of the Pauls Creek site potentially masked changes related to stream-improvement efforts. In addition, stream-improvement activities occurred in each of the three study basins over a period of several years prior to and during the study so that there were no discrete before and after periods available for meaningful comparisons. Historical and current land-use activities in each of the three study basins likely affected observed stream conditions. The duration of this study probably was insufficient to detect changes associated with agricultural best management practices and stream-channel restoration.