The focus of this study is to refine the current methodology used in the Rankings model. In our opinion, three important conceptual assumptions of the County Health Rankings & Roadmaps (CHR&R) construction require validation.
First, by the use of the arithmetic mean the CHR&R assumes perfect compensation between its dimensions. It is doubtful that a county’s poor performance in one of the CHR&R dimensions can be fully compensated for by a good performance in another dimension. Hence, we argue that methods involving different levels of compensation should be investigated for the final aggregation formula.
The second assumption relates to the sensitivity of the CHR&R to changes in its dimensions (Health Factors (HF) and Health Outcomes (HO)), components, and sub-components. The CHR&R – again by using arithmetic average – assumes equal sensitivity. We question this assumption as we argue that, for example, an expansion of afterschool services can improve well-being and health of working parents more in counties where these services are poor, than in counties where they are already well developed. Thus, we want to investigate whether a non-linear relationship between the CHR&R and its dimensions, components, and sub-components should be adopted, favoring greater improvements in those ranking components which are underdeveloped.
Third, the set of nominal weights assigned to components of the HO and the components of the HF translates into implicit assumption of the relative importance of the various factors affecting HO and HF.
This study will provide answers to the following questions regarding construction of the CHR&R: (1) whether the theoretical framework of the CHR&R is supported by data, i.e. if relationships between indicators justify current grouping into sub-components/components/dimensions of the CHR&R; (2) if indicators populating the framework contribute to the CHR&R, i.e., indicators/sub-components are not lost in the aggregation process, (3) if weighting scheme corresponds to empirical importance of elements of the CHR&R. Based on the results, the CHR&R may be refined (if needed and desired). The methodology – presenting consequences of the subjective methodological choices (unavoidable in the computation process) – will be better justified.
For more information about county health rankings, visit www.countyhealthrankings.org