It is generally accepted that soil fertility declines with changes in plant community composition as secondary succession proceeds. However, experimental manipulations of soil fertility by carbon (C) addition in both restoration and basic contexts have had unpredictable results. We conducted a meta-analysis of C addition experiments, using ~50 field and laboratory studies from North America, Europe and Australia. We predicted that the mixed outcomes of these studies were due to many sources of variation, including: rate, timing and form of C application, the number of applications, varying intervals between time of application and sampling, and a wide range of ecosystem types. Our objectives were to identify general patterns of plant and soil response to C addition with respect to those sources of variation. We analyzed plant, microbial, and soil responses separately and calculated Hedge's D and log response ratios. For plant growth responses, there is a larger negative effect of C addition on exotic species compared to native species. Exotic species also show the largest reduction in growth to sucrose addition whereas native species had the largest positive growth response to a sucrose/sawdust mixture. Annual species are negatively affected by C addition overall, but there is a neutral effect on perennial species. We found little evidence for a dose (or rate of C addition) response in plant growth, but a slightly positive one for microbial processes like nitrogen (N) mineralization and immobilization. Responses of soil N pools were negative overall, but did not differ by C type. In most cases, we found significant within-group heterogeneity which underscores the limitations of single-factor analyses. Conducting factorial meta-analysis will allow us to draw further comprehensive insights from the literature about how different sources of variation change the effects of C addition on plant communities.