Long-term, stable marriages of midlife Americans were characterized by a linear increase in relationship satisfaction over 20 years & a linear decline in sexual satisfaction in the same time frame

Relationship and sexual satisfaction: A developmental perspective on bidirectionality. Christopher Quinn-Nilas. Journal of Social and Personal Relationships, September 19, 2019. https://doi.org/10.1177/0265407519876018

Abstract: Researchers have investigated the directionality between relationship and sexual satisfaction; however, there remains no definitive conclusion. Previous longitudinal studies have not conceptualized relationship and sexual satisfaction as systematic developmental processes and have focused on predicting scores at later time points. Instead, researchers should be concerned with understanding how relationship and sexual satisfaction change together over time. The objective of this study was to use longitudinal data from midlife American marriages to test the directionality of the association between relationship satisfaction and sexual satisfaction. Multivariate latent growth curve modeling of 1,456 midlife Americans married for 20 years from the Midlife in the United States study was used to compare directionality models. Findings support that long-term, stable marriages of midlife Americans at the sample level were characterized by a linear increase in relationship satisfaction over 20 years and a linear decline in sexual satisfaction during the same time frame. A co-change model, wherein relationship and sexual satisfaction changed together over time, fit the data best. Trajectory correlations showed that changes in relationship and sexual satisfaction were strongly interconnected. High initial levels of sexual satisfaction protected against declines in relationship satisfaction over 20 years. Results support that relationship and sexual satisfaction change together over time and highlight that the longitudinal association between these outcomes is dynamic rather than static.

Keywords: Marriage, midlife Americans, MIDUS, multivariate growth curve, relationship satisfaction, sexual satisfaction

---
Strengths
The MLGCM methodology is a key strength of this study because it allowed the examination of bidirectional associations between sexual and relationship satisfaction growth factors and allowed each outcome to be estimated as its own developmental process. Furthermore, this study is novel because it approached the question of directionality between relationship and sexual satisfaction by systematically testing competing unidirectional and bidirectional models using characteristics of change trajectories rather than static scores. This is partly buttressed by the large sample drawn from a nationally representative study that meets standards for statistical power (Hertzog et al., 2006). The 20-year longtime horizon underlying the MIDUS study has made possible inferences covering a substantial amount of time.

Limitations
Despite the noted strengths, several limitations of the study are noteworthy. This study used single indicator measurement to assess relationship and sexual satisfaction. Additionally, several limitations are introduced by having only three time points. Having three time points does not allow for the thorough testing of nonlinear or more complex trends which could be assessed using a daily diary format (e.g., Day et al., 2015).Even though the duration of the MIDUS study was long, which allowed for long-term inferences, this structure neglects the more idiosyncratic microdevelopments and microchanges that may occur ona smallertime scale (i.e.,day to day,weektoweek, month to month, yeartoyear).
The original MIDUS data were from a nationally representative sample obtained with random-digit dialing, but it should be noted that there may be a selection effect inherent to the current study’s inclusion criteria (participants who stayed married to the same person for the entire duration). In the current study, the regressive bidirectional model (Model 8) was discarded because the design of the MIDUS study did not contain a meaningful intercept point. Although this is true in the present analyses, in other study designs that contain substantive intercepts (i.e., studies that begin at the start of a marriage), the regressive model need not be discarded. In such a case, a more thorough assessment can be made between the correlated model and the regressive model. Another limitation concerns the wording of the sexual satisfaction question, which asked about sexual satisfaction broadly rather than specifically about the marital relationship.