Hierarchical Linear Modeling

Conceptual Reasons to Use HLM

Conceptually, communication scholars are generally interested in relationships, interactions, social norms, and other phenomena that involve more than one person. HLM lets researchers examine multiple parts of a group or system, as opposed to focusing on a single part. Researchers can use it to identify aspects of the context or relationship that are important but might be overlooked if scholars focused on the individual, as opposed to the couple or group. For example, HLM can be used to study how a person’s conflict style affects his or her partner’s relationship satisfaction, or how a team leader contributes to his or her team’s effectiveness. Attention to nesting also allows researchers to develop theories that identify and specify effects at both the individual and group levels. Effects might differ between the individual and group levels of analysis (e.g., team members who talk more might look at their teammates less, but teams with more talking will have more eye contact overall), so the inclusion of both levels is important to communication theory development.

Attention to nesting structure is also important because it can help researchers identify causal mechanisms and design interventions. For example, both individuals in a married couple might report negative conflict strategies and dissatisfaction, but ignoring the relationship between the spouses would make it impossible to know whether spouses’ dissatisfaction is the result of their own conflict strategies or their partner’s conflict strategies. Without such information, it would be harder to make recommendations about how to increase satisfaction in that relationship.

Ignoring the nested structure of data can cause researchers to draw incorrect conclusions or commit ecological or atomistic fallacies. An ecological fallacy is committed when associations between variables are observed at the group level of analysis and are assumed to also occur at the individual level of analysis. For example, a researcher trying to understand conflict styles might examine levels of demand and withdrawal behaviors in romantic couples and find that the more demand behaviors occur within the couple, the more withdrawal behaviors also occur. However, using this finding to conclude that a person who demands frequently also withdraws frequently would be incorrect. A more likely explanation is that the more one partner demands, the more the other partner withdraws (and vice versa).

...