Mapping Community Disadvantage to Better Target Home Visiting Services in Montana

Background and Purpose. Home-visiting programs have impacts on a wide-range of indicators of family and child well-being including better cognitive development and fewer behavioral problems for children and less use of welfare and fewer pregnancies for families (Olds et al., 2004). Until recently, Montana used county-wide disadvantage measures to target home-visiting services.  However, there is great variation across counties that are masked when using this geographic level.  The challenge of using such a large Census unit is heterogeneity as internal homogeneity is assumed (Charnock, 1982). However, this assumption is challenged in rural regions like Montana, as Census units are geographically larger where populations are smaller. In a geographically large unit, variation will be high and variation between spatial units may be limited, contextual effects will be harder to detect, and neighborhood effects may be underestimated (O’Campo, 2003). Thus, we used GIS methods to map disadvantage at a lower level, the Census block group, to examine how well served families map onto need and to help the state better target their services. 

Methods. To map disadvantage, we used five Census variables from the American Community Survey (ACS) estimates from 2010-2014 at the block group level (Census, 2008).  For the ACS, 5 years of data are pooled together for reliable estimates of characteristics, particularly useful for geographic areas with smaller populations. Based on Brody et al. (2001), the following five variables were combined in an index using principal components analysis to compute composite disadvantage scores for Montana: 1) Proportion of female-headed households; 2) Average per capita income; 3) Proportion of households below the poverty line; 4) Proportion of residents receiving public assistance; and 5) Proportion of unemployed residents. We then created heat maps for each individual indicator and the disadvantage score overlaid with the location of current program participants.  Key landmarks were also mapped including major cities and American Indian reservation boundaries.

Results. The maps show that the highest need census block groups are particularly concentrated on reservation lands including the Flathead, Crow, Fort Belknap, and Northern Cheyenne reservations.  However, the majority of the families served reside in the more populous, but lower-need census tracts in Missoula, Billings, and Bozeman. There is a cluster of participating families adjacent but not residing on the Flathead reservation. Detailed maps will be displayed.  

Conclusions and Implications. The finer-grained mapping of disadvantage at the Census block group demonstrates that the current county method used by the state to target services to families is not optimal, missing families on the reservations and in the more rural settings in a highly rural state where only 55.9% of residents live in metro areas (Census, 2010). In comparison, 95% of California residents live in metro areas. In addition, although there are other data sources that may be utilized, the systematic way in which Census data are collected and the availability of a wide variety of data, particularly for socioeconomic composition, make these constructs a valuable data source (Diez Roux, 2001). Similar methods could be employed for targeting services beyond home-visiting.