Research ArticleSCIENTIFIC COMMUNITY

Spatial inequalities leave micropolitan areas and Indigenous populations underserved by informal STEM learning institutions

See allHide authors and affiliations

Science Advances  09 Oct 2020:
Vol. 6, no. 41, eabb3819
DOI: 10.1126/sciadv.abb3819
  • Fig. 1 Landscape of ILIs in the United States (density per 1000 km2).

    (A) Kernel density surface of all ILIs displayed as six quantiles (cell size = 1000 m2; μ = 6.5 ILI per 1000 km2). (B) Mean density of ILIs per 1000 km2 at each type (abbreviations follow) with SE bars [colors correspond to (A) and (C) to (H)]. (C) National Park Service (NPS) lands. (D) Biological field stations and marine laboratories (FSMLs). (E) Zoos, aquariums, and wildlife conservation (ZOO). (F) Science museums, children’s museums, and planetariums (MUS). (G) Libraries (LIB). (H) Botanical gardens, arboretums, and nature centers (BOT). All densities are displayed as six quantiles, and values associated with each break are in table S4. Mean values for the bar plot are provided in table S5. All ILI point data with associated kernel density values are in data file S1.

  • Fig. 2 Distribution of ILIs among the U.S. population summarized at the county level.

    (A) SD of ILI residuals with the darkest purple indicating the fewest number of ILIs (σ < −2.5) and the darkest green indicating the most ILIs (σ > 1.5) relative to the number expected. Orange counties have no ILIs. (B) ILI residuals grouped by RUCCs with SE bars. (C) ILI residuals grouped by poverty categories with SE bars. ILI residuals are from a spatially corrected regression between log ILI density and the interaction of log population density and poverty percentage. Residual values associated with each SD unit are in table S6, and county data are in data file S2.

  • Fig. 3 Counties that are the most underserved by ILIs.

    (A) Counties that do not have ILIs. (B) Counties with ILI residuals in the lowest 0.5% (σ < −2.5). (C) Non-metro, non-adjacent counties with urban populations over 20,000 (RUCC 5) with colors indicating standard deviations of log ILI residuals and the interaction of log population density and poverty percentage as in Fig. 2A. (D) Racial and ethnic percentages in underserved counties. U.S. population is the percentage of the general population in each underserved group of counties for comparison with the racial and ethnic groups. Bar plot values are in table S1.

  • Fig. 4 ILI density from Fig. 1A with underserved counties from Fig. 3 (A to C) highlighted in white.

  • Table 1 Summary of the generalized linear model that included log population density, poverty percentage, and their interaction as factors with a rational quadratic correlation structure to account for spatial autocorrelation (R2McFadden = 0.322).

    FactorCoefficientt valueP value
    Intercept−7.63−74.9<0.01
    Log population
    density
    0.46826.9<0.01
    Poverty
    percentage
    −0.014−4.35<0.01
    Log population
    density ×
    poverty
    percentage
    0.0066.46<0.01

Supplementary Materials

  • Supplementary Materials

    Spatial inequalities leave micropolitan areas and Indigenous populations underserved by informal STEM learning institutions

    Rachel A. Short, Rhonda Struminger, Jill Zarestky, James Pippin, Minna Wong, Lauren Vilen, A. Michelle Lawing

    Download Supplement

    The PDF file includes:

    • Tables S1 to S6
    • Legends for data files S1 and S2

    Other Supplementary Material for this manuscript includes the following:

    Files in this Data Supplement:

Stay Connected to Science Advances

Navigate This Article