Evaluating the feasibility of using multivariate meta-analysis to examine the association of traits and communication patterns Traits have long been fruitfully examined as potential predictors of communication patterns. Indeed, a plethora of meta-analytical studies have examined the role of personality traits in producing one or more communication outcomes (e.g. Schrodt, Witt, & Shimkowski, 2014, Le, Dove, Agnew, Korn, & Mutso, 2010, Li & Chan, 2012, etc.). From meta-analytical studies, we know personality factors can be significantly associated with demand/withdrawal behaviors among couples(Schrodt, Witt, & Shimkowski, 2014), and that individuals higher in attachment avoidance or anxiety may show a small, but significant association with non-marital romantic dissolutions (Le, Dove, Agnew, Korn, & Mutso, 2010). We also know that high attachment anxiety is usually related to conflicts in romantic behaviors (Li & Chan, 2012). Indeed, the technique of synthesizing and combining research outcomes of interests into effect sizes have been providing powerful means to understand the general pictures of human communication. In standard meta-analysis, authors usually choose to focus on a single communication outcome (e.g., xxx). For example, to understand the associations between attachment styles and conflict romantic behaviors, Li & Chan (2012) designed codebooks to combine conflict patterns such as consistently using dominating and avoiding strategies as one general indicator of negative communication. This was to reduce the excessive weight that was to be assigned to a study if dominating and avoiding were entered as two separate outcomes for a single study (Borenstein et al., 2011). Indeed, in a meta-analysis, the long-recommended approach is to use a synthesized effect size for all outcomes. It can solve the problem of assigning weight to studies with more than one outcomes, it can also account for the problem of non-independent information: if more than more outcomes were retrieved from the same sample of population, those outcomes are inherently correlated with each other, then to include those outcomes separately into a meta-analysis is also problematic because we are ignoring the correlations between those outcomes (Borenstein et al., 2011; Mavridis & Salanti, 2013; Ian R. White, 2015). However, this approach has its drawbacks: it losses meaningful information by reducing various outcomes observed in those studies to a combined single effect size (Borenstein, Hedges, Higgins, & Rothstein, 2011). What if a research would want to evaluate, in great nuance, the relationship between attachment styles and specific conflict patterns? What if the research question is not about whether attachment anxiety is positively or negatively related to negative communication, but to evaluate what specific conflict strategy is problematic for people who are highly anxious? Do anxious people indicate a pattern of using avoiding strategies during a conflict, meanwhile, do they also tend to use other strategies like to oblige to others' goals? With detailed results regarding specific communication problems associated with those individuals, interventions with accurate behaviors of interests can be provided. Indeed, can meta-analytical approach leverage our needs for more detailed associations between communication variables? With sufficient amount of studies documenting those detailed relationships, and with the development of meta-synthesis approach: multivariate meta-analysis, we now may be able to produce meta-analysis that can satisfy both the need of researchers in getting more nuanced associations and the methodological requirement of meta-analysis by not assigning extra weight and to account for the covariance between different outcomes on the same sample of participants (Mavridis & Salanti, 2013; I. White, 2011; I. R. White, 2009). In the following sections, this essay is to review the traditions regarding methods and problems of combining effect sizes under the paradigms of univariate meta-analysis, and to present details regarding assumptions of multivariate meta-analysis, this study also uses a set of studies that measure attachment styles and multiple conflict behaviors to discuss the feasibility and challenges that are associated with using multivariate meta-analysis to provide more insightful information for the research need of communication scholars.