The work of James P. Dillard, John E. Hunter, and Michael Burgoon, Sequential Request Persuasive Strategies -- Meta-Analysis of Foot-in-the-Door and Door-in-the-Face is an excellent approach to looking at the overall results from the studies listed here.
"Results of the initial FITD meta-analysis. The distribution of FITD correlations ranged in value from -.23 to .39. Of the coefficients, 79% were greater than zero. although this was evidence for the existence of the FITD phenomenon, it suggested that the effect is not a particularly strong one. Computation of the mean correlation confirmed this suspicion. The weighted mean correlation was .11, and the corrected standard deviation = .14. Formation of the ratio of the variance due to sampling error over the variance in the observed correlation showed that only 28% of the variance may be attributed to sampling error."
Further inspection of meta-analysis would reveal some increased effect sizes when considered "with an eye toward variables that might raise or lower the size of the effect."
"Results of the initial DITF meta-analysis. Of the coefficients in the DITF distribution, 75% were larger than zero. This gave rise to the same inference that was drawn about the FITD effect: a real but small true effect size. Computations showed a sample-size weighted mean correlation of .08, and a standard deviation adjusted for sampling error of .09. The ratio of error variance to observed variance was .56, an indication that approximately 56% of the variance was accounted for by sampling error."
Comparitavely, these numbers may not be Earth-shattering compared to other numbers generated by Communication research, but increasing compliance roughly one out of ten times can be considered a highly influential and useful tool.