Did No-Fault Divorce Legislation Matter? Definitely Yes and Sometimes No

Abstract

In our response to Glenn's (1997) first critique of our original article (Nakonezny, Shull, & Rodgers, 1995), we took the middle ground. We did and still do feel that Glenn was correct that the no-fault divorce process included some redundancy, but some cases. In others, both empirical and logical reasons support that no-fault divorce laws explicitly facilitated the divorce process. In our earlier response (Rodgers, Nakonezny, & Shull, 1997), we presented a number of examples of the types of divorces that no-fault laws would be expected to impact most dramatically. The consistent position of Glenn that the redundancy argument applies in most overstates the actual situation, as his own empirical results demonstrate. Glenn's analysis of the results from Table 1 Rodgers et al. (1997) was extremely helpful and insightful. (See Figure 1, Glenn, 1999.) However, we disagree with Glenn's interpretation of his Figure 1. He suggested that the temporal pattern identified that figure was rather bizarre. We believe that this figure clarifies the situation and supports the middle-ground position that we have consistently advocated. The diffusion of any innovation-marketing, educational, or policy-orientedtypically will take time to show an effect. Then the effect of the innovation should be apparent for a number of years. Finally, the size of the effect will decline or disappear-and can even turn slightly negative, especially if the units are not independent. Glenn's Figure 1 graphically shows that process is occurring during the implementation of no-fault divorce laws the SO states. Indeed, Glenn has identified the category of states for which no-fault divorce was largely redundant-the 15 states that implemented no-fault divorce after 1974. This redundancy may be due to a legal system that had already implemented no-fault divorce de facto (as Glenn suggested) or due to the easy availability of no-fault divorces neighboring states. In either case, it appears that by 1974 most of the divorces that would have been difficult earlier had been achieved. We have no problem granting Glenn's position for the states that show zero or negative effect sizes, especially when they cohere a temporal sense, as Glenn's Figure 1 lucidly demonstrates. But what about the 32 states that implemented no-fault divorce during the period of 1965-1974'? We reject Glenn's repeated implication that these results should be downplayed because they occurred only during the first 10 years of the divorce boom, Two thirds of the states fall into this period, and this is when most of the legislative action was occurring. Glenn's Figure 1 shows much larger effect sizes these 32 states than those from our earlier estimates. We would prefer simply to let the data speak this case, and we are comfortable with a reinterpretation of our previous findings light of Glenn's analysis. For most of the states after 1974, the implementation of no-fault divorce was largely redundant. For most of the 32 states that implemented no-fault divorce during the divorce boom (1965-1974), both our results and Glenn's results support the interpretation that no-fault laws resulted a substantial number of divorces that would not have occurred otherwise. We now wish to address Glenn's criticism of the projection method Rodgers et al. (1997). His suggestion that our regression estimates substantially overestimated the effect size is an insightful suggestion. But again, he jumps too quickly to a negative and contentious conclusion. Glenn suggested that the potential artifact resulting from the use of regression is automatic. He states that linear projections from nonlinear trends [are] rarely a useful procedure (p. 800). That simple assertion might be correct or might be completely wrong. We argue that this is an empirical question to be resolved analytically, and we have run several analyses that address this issue. …