Interest in plant responses to rising atmospheric CO₂ is clearly demonstrated by the sharply increasing number of research reports published in this area. Strain and Cure (1994) found an average of 90 elevated CO₂ papers per year published between 1989 and 1992. In 1997, approximately 300 papers were published, and the total number of reports to date exceeds 3000. Many review articles are published yearly. The quantitative methods for combining data across studies that were used in the great majority of the reviews, principally vote counting techniques and averaging of unweighted response ratios, while useful in some instances, are flawed statistically and are much less powerful analytically than review methods widely used in other fields.

Many traditional narrative reviews of the elevated CO₂ literature have been insightful and have proven influential in furthering our understanding of plant and ecosystem responses to elevated CO₂. However, in attempting quantitative integration of research results across studies, reviewers of the CO₂ literature have worked entirely outside established statistical methods for such integration. A central tenet of meta-analytical methods is that reviewers of published studies should be held to the same standards as primary researchers whose results form the basis of their reviews with regard to techniques of data sampling, statistical analysis, and hypothesis testing.

Quantitative meta-analysis involves the calculation of a treatment effect size, with known statistical properties, that can be averaged across independent studies, giving an overall mean effect size (Arnqvist and Wooster 1995). Variation in mean effect size can then be partitioned within and among categorical groups. Meta-analysis is not a substitute for well-designed, multi-factorial experiments, and it is important to acknowledge the limits to establishing causal relationships from meta-analytic results where categorical groups created by the meta-analyst were not randomly assigned treatments within the primary studies (Miller and Pollock 1994). What meta-analysis offers is an objective and statistically rigorous methodology for integrating primary research results with the goal of estimating the magnitude of treatment effects within and among categorical groups. We suggest that it is important from a scientific as well as a policy perspective that elevated CO₂ research results continue to be integrated using these tools.