{"id":276701,"date":"2019-05-30T13:28:00","date_gmt":"2019-05-30T17:28:00","guid":{"rendered":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/?post_type=decoded&p=276701"},"modified":"2025-10-09T10:38:18","modified_gmt":"2025-10-09T14:38:18","slug":"working-with-geography-in-survey-data","status":"publish","type":"decoded","link":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/decoded\/2019\/05\/30\/working-with-geography-in-survey-data\/","title":{"rendered":"Working with geography in survey data"},"content":{"rendered":"\n
\"\"<\/figure>\n\n\n\n

Survey researchers frequently explore differences in public opinion by demographic group \u2014 how men\u2019s views compare with those of women, for example, or how younger people compare with older people. Often, it\u2019s also possible to look at differences by survey respondents\u2019 geographic location.<\/p>\n\n\n\n

Yet, when geographic information is available in survey data, it\u2019s not always at the level researchers are looking for. At Pew Research Center, for instance, we typically ask our phone survey respondents for their ZIP code, primarily so we can accurately identify their state and Census region for the purposes of weighting our surveys<\/a>. On their own, ZIP codes are of little value to researchers interested in doing geographic analyses, in part because there are over 30,000 ZIP codes in the United States.<\/p>\n\n\n\n

However, researchers can<\/em> use ZIP codes to determine location at a more granular level and do geographic analyses that use spatial relationships. Using ZIP codes, it\u2019s possible to locate respondents in a specific place with some degree of accuracy \u2014 specifically, the latitude and longitude at the centroid (or geographical center) of the Census\u2019s ZIP code tabulation area.<\/p>\n\n\n\n

In this post, I\u2019ll walk through an example from early 2018 that used ZIP codes to conduct geographic analysis. The analysis in question<\/a> was based on a survey question we asked about public support for offshore drilling. It seemed reasonable to explore whether respondents\u2019 proximity to a coastline was associated with their attitudes toward offshore drilling.<\/p>\n\n\n\n

Geolocating respondents<\/h4>\n\n\n\n

The first step in this kind of analysis is geolocating survey respondents. In this data, the best we can do is locate people within their ZIP codes.<\/p>\n\n\n\n

There are a couple of things to keep in mind when working with ZIP codes. First, we have to determine how to translate the ZIP codes into geographic coordinates. In this analysis, we\u2019ll assign respondents to the centroids of their ZIP code tabulation areas. (One great resource for georeferenced Census units is GeoCorr<\/a>, and you can download the ZIP code data here<\/a>)<\/p>\n\n\n\n

First, we\u2019ll read in the data (you can find the survey data here<\/a>) and merge it together, as shown below. (A side note on working with ZIP codes: Always check the format. ZIP code is stored as a factor variable in the survey data but as a character in the GeoCorr data. To avoid problems merging the two together, make sure they are in the same format. Additionally, with ZIP codes in particular, always be mindful of leading zeros.)<\/p>\n\n\n\n

library(foreign)\n###Read in the datasets\ndat <- read.spss(\"small_jan_data.sav\",\nto.data.frame = TRUE)\n##zip code data from geocorr for all zip code centroids\nzip <- read.csv(\"geocorr2018.csv\", as.is = TRUE)\nhead(zip)\n#remove the first row\nzip <- zip[-1,]\nzip$lon <- as.numeric(zip$intptlon)\nzip$lat <- as.numeric(zip$intptlat)\n##format zipcode as string variable\ndat$zipcode <- as.character(dat$finalzip)\n###merge data\ndat <- merge(dat, zip[,c(\u201czcta5\u201d,\u201dlat\u201d,\u201dlon\u201d)], by.x = \u201czipcode\u201d, by.y = \u201czcta5\u201d, all.x = TRUE)\n###Plot the zip code coordinates\nwith(dat, plot(lon, lat, pch = 20))<\/code><\/pre>\n\n\n\n
\"\"<\/a>
Plot of ZIP code centroids<\/figcaption><\/figure>\n\n\n\n
Calculating distance from coast<\/h4>\n\n\n\n
The next step is to bring in the map of coastlines and find the distance between respondents\u2019 locations and the coastline.<\/p>\n\n\n\n
##bring in coastline shapefile\nlibrary(sp)\nlibrary(rgeos)\nlibrary(rgdal)\n##availble from here: http:\/\/www.soest.hawaii.edu\/pwessel\/gshhg\/\nmap <- readOGR(\u201ccoastline map\/GSHHS_l_L1.shp\u201d)\n####extract the coordinates from the map object\npolys <- map@polygons\n##container for coordinates\nall.coords <- NULL\nfor (j in 1:length(polys)) { ##loop through polygons\n##find coordinate slots\nif (is.element(\u201ccoords\u201d, slotNames(polys[[j]]))) {\ncoords <- polys[[1]]@coords\nall.coords <- rbind(all.coords, coords)\n}\n##get polygon slots for more complicated geographies\nif (is.element(\u201cPolygons\u201d, slotNames(polys[[j]]))) {\np <- polys[[j]]@Polygons\n##extract coordinates from these more complex objects\ncoords <- NULL\nfor (k in 1:length(p)) {\ncoords <- rbind(coords, p[[k]]@coords)\n}\nall.coords <- rbind(all.coords, coords)\n}\n}\nplot(all.coords, pch = \u2018.\u2019)<\/code><\/pre>\n\n\n\n
\"\"
Plot of the points that make up the coastline shapefile<\/figcaption><\/figure>\n\n\n\n
###select N. America coast line\nx <- which(all.coords[,1] > -180 & all.coords[,1] < -45)\ny <- which(all.coords[,2] > 20)\nwh <- intersect(x, y)\nplot(all.coords[wh,], pch = \u2018.\u2019)\n###cut out NE Canadian coastline\nx <- which(all.coords[,1] > -120)\ny <- which(all.coords[,2] > 55)\nwh2 <- intersect(x, y)\nwh <- setdiff(wh, wh2)\nus <- all.coords[wh,]\nplot(us, pch = \u2018.\u2019)\npoints(zip$lon, zip$lat, pch = \u2018.\u2019, col = \u2018red\u2019)<\/code><\/pre>\n\n\n\n
Calculating distance is a little more involved than simply using Euclidian distance. Given that we are dealing with a coordinate system on a globe, we\u2019ll use a great circle distance approximation:<\/p>\n\n\n\n
####Functions for calculating great circle distances in miles\n####slightly modified from this post: https:\/\/www.r-bloggers.com\/great-circle-distance-calculations-in-r\/\n####convert degrees to radians for distance calculation\ndeg2rad <- function(deg) return(deg*pi\/180)\n# Calculates the geodesic distance between two points specified by radian latitude\/longitude using the\n# Spherical Law of Cosines (slc)\ngcd.slc <- function(long1, lat1, long2, lat2) {\n##convert coordinates to radians\nlat1 <- deg2rad(lat1)\nlong1 <- deg2rad(long1)\nlat2 <- deg2rad(lat2)\nlong2 <- deg2rad(long2)\nR <- 3959 # Earth mean radius [miles]\n##container for distances\nd_vec <- rep(NA, length(long2))\nfor (i in 1:length(long2)) {\n##find distances from each point\nd_vec[i] <- acos(sin(lat1)*sin(lat2[i]) +\ncos(lat1)*cos(lat2[i]) * cos(abs(long2[i]-long1))) * R\n}\nreturn(d_vec)\n}\n##column for distance to coast\ndat$dist_to_coast <- NA\nfor (j in 1:nrow(dat)) {\ndist <- gcd.slc(dat$lon[j], dat$lat[j],\nus[,1], us[,2])\ndat$dist_to_coast[j] <- min(dist)\n}\n##create a 3-way distance variable\ndat$dist3 <- 99\nw <- which(!is.na(dat$dist_to_coast))\ndat$dist3[w] <- 3\nw <- which(dat$dist_to_coast < 300)\ndat$dist3[w] <- 2\nw <- which(dat$dist_to_coast < 25)\ndat$dist3[w] <- 1\ndat$dist3 <- factor(dat$dist3, labels = c(\u201cLess than 25 miles\u201d,\n\u201c25\u2013300 miles\u201d, \u201cMore than 300 miles\u201d, \u201cMissing\u201d))<\/code><\/pre>\n\n\n\n
Close reading of this code shows that we\u2019re actually calculating distance to the nearest points that make up the coastline. It would be better to calculate the distance to the nearest point on the line segments that make up the coastline, but this would be a much more involved calculation. However, the large number of points that make up the coastline means the error is pretty minimal. A larger source of error is the fact that respondents are located at the centroids of their ZIP codes rather than their actual addresses.<\/p>\n\n\n\n
Examining the results<\/strong><\/h4>\n\n\n\n
Given the distance measure, we can now examine attitudes toward offshore drilling by respondents\u2019 proximity to the coast:<\/p>\n\n\n\n
library(survey)\ndesign <- svydesign(id=~1, weights=~weight, data=dat)\nsvyby(~q90, ~dist3, design = design,\nFUN = svymean, keep.names = FALSE, na.rm = TRUE)\ndist3                 q90Favor  q90Oppose\n1 Less than 25 miles  0.3457668 0.5562020\n2 25\u2013300 miles        0.4507683 0.4783657\n3 More than 300 miles 0.4603733 0.4967827\n4 Missing             0.3720786 0.4979092<\/code><\/pre>\n\n\n\n
Indeed, there appears to be a significant difference in attitudes toward offshore drilling between those who live nearer to a coastline and those who live farther away. People who live within 25 miles of the coast were about 10 percentage points less likely to say they favor offshore drilling than those who live more than 25 miles from the coast. However, Democrats and Democratic-leaning independents live nearer to the coast, on average, than Republicans and Republican leaners do:<\/p>\n\n\n\n
svyby(~dist_to_coast, ~partysum, design = design,\nFUN = svymean, keep.names = FALSE, na.rm = TRUE)\npartysum         dist_to_coast se\n1 Rep\/lean Rep   256.6374      12.05583\n2 Dem\/lean Dem   205.2576      10.72659\n3 DK\/Ref-no lean 216.7199      28.74201<\/code><\/pre>\n\n\n\n
In a multivariate framework, there seems to be no relationship between attitudes toward offshore drilling and proximity to the coast after controlling for partisanship:<\/p>\n\n\n\n
summary(svyglm(q90 == \"Favor\" ~ dist3 + partysum,\n design = design, family = 'quasibinomial'))\nCall:\nsvyglm(formula = q90 == \"Favor\" ~ dist3 + partysum, \n    design = design, \n    family = \"quasibinomial\")\nSurvey design:\nsvydesign(id = ~1, weights = ~weight, data = dat)\nCoefficients:\n                         Estimate Std. Error t value Pr(>|t|)    \n(Intercept)                0.6551     0.1519   4.314 1.71e-05 ***\ndist325-300 miles          0.2855     0.1719   1.661    0.097 .  \ndist3More than 300 miles   0.2284     0.1823   1.253    0.210    \ndist3Missing               0.1270     0.3351   0.379    0.705    \npartysumDem\/lean Dem      -2.0681     0.1494 -13.844  < 2e-16 ***\npartysumDK\/Ref-no lean    -1.4222     0.2350  -6.051 1.81e-09 ***\n---\nSignif. codes:  0 \u2018***\u2019 0.001 \u2018**\u2019 0.01 \u2018*\u2019 0.05 \u2018.\u2019 0.1 \u2018 \u2019 1\n(Dispersion parameter for quasibinomial family taken to be 0.999249)\nNumber of Fisher Scoring iterations: 4<\/code><\/pre>\n\n\n\n
Notes<\/strong><\/h4>\n\n\n\n
As mentioned above, there is some error with using the centroid of a ZIP code instead of an actual address. If more granular data (e.g., exact addresses) are available, that would involve translating those addresses into exact coordinates rather than the ZIP code centroid. There are many ways to translate street address into geographic coordinates \u2014 for example, the ggmap package in R interfaces with the GoogleMaps API to extract latitude and longitude coordinates for address data.<\/p>\n\n\n\n
If you\u2019re interested in trying your own geographic analysis of survey data, it\u2019s possible to get ZIP code data for a subset of Pew Research Center general public surveys conducted by phone, once those datasets have already been publicly released. Because respondent privacy is of utmost concern, we do request additional information from researchers and require them to subscribe to additional data usage agreements. If you have a request along these lines, please send an email to info@pewresearch.org<\/a> and include the specific survey name and dates (or link to the survey). Please note that detailed geographic variables (like ZIP code) are not available for surveys of rare populations or surveys conducted through the Center\u2019s American Trends Panel<\/a>. (Due to the longitudinal nature of the panel, there is a great deal of information available about panelists so we need to take extra safeguards to protect their confidentiality.)<\/p>\n","protected":false},"excerpt":{"rendered":"
Survey researchers frequently explore differences in public opinion by demographic group \u2014 how men\u2019s views compare with those of women, for example, or how younger people compare with older people. Often, it\u2019s also possible to look at differences by survey respondents\u2019 geographic location. Yet, when geographic information is available in survey data, it\u2019s not always […]<\/p>\n","protected":false},"author":658,"featured_media":276722,"comment_status":"closed","ping_status":"closed","template":"","meta":{"sub_headline":"","sub_title":"","_prc_public_revisions":[],"_ppp_expiration_hours":0,"_ppp_enabled":false,"ai_generated_summary":"","relatedPosts":[],"_prc_fork_parent":0,"_prc_fork_status":"","_prc_active_fork":0,"datacite_doi":"","datacite_doi_citation":"","_prc_seo_qr_attachment_id":0,"spoken_article_player_enabled":true,"displayBylines":true,"footnotes":"","prc_watchers":[]},"categories":[],"bylines":[905],"collection":[],"_post_visibility":[],"decoded-category":[531,530,532],"formats":[454],"_fund_pool":[],"languages":[],"regions-countries":[515],"research-teams":[],"workflow-status":[],"class_list":["post-276701","decoded","type-decoded","status-publish","has-post-thumbnail","hentry","bylines-bradley-jones","decoded-category-coding-how-to","decoded-category-data-science","decoded-category-survey-methods","formats-decoded","regions-countries-united-states"],"label":"Decoded","post_parent":0,"word_count":1666,"canonical_url":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/decoded\/2019\/05\/30\/working-with-geography-in-survey-data\/","art_direction":{"A2":{"id":276722,"rawUrl":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp","url":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp?w=268&h=151&crop=1","width":268,"height":151,"caption":"","chartArt":false},"XL":{"id":276722,"rawUrl":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp","url":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp?w=720&h=405&crop=1","width":720,"height":405,"caption":"","chartArt":false},"A3":{"id":276722,"rawUrl":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp","url":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp?w=194&h=110&crop=1","width":194,"height":110,"caption":"","chartArt":false},"A4":{"id":276722,"rawUrl":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp","url":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp?w=268&h=151&crop=1","width":268,"height":151,"caption":"","chartArt":false},"A1":{"id":276722,"rawUrl":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp","url":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp?w=564&h=317&crop=1","width":564,"height":317,"caption":"","chartArt":false},"social":{"id":276722,"rawUrl":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp","url":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-content\/uploads\/sites\/20\/2025\/10\/d_19.11.9_geography_featured.webp?w=1200&h=628&crop=1","width":1200,"height":628,"caption":"","chartArt":false}},"_embeds":[],"watchers":[],"table_of_contents":[],"datacite_doi":"","prc_seo_data":{"title":"Working with geography in survey data","description":"Survey researchers frequently explore differences in public opinion by demographic group \u2014 how men\u2019s views compare with those of women, for example, or how younger people compare with older people.…","og_title":"Working with geography in survey data","og_description":"","schema_type":"Article","noindex":false,"canonical_url":"","primary_terms":{"regions-countries":515},"custom_schema":[],"og_image":276722,"indexnow_submitted_at":null,"gsc_index_status":null},"prepublish_checks":{},"jetpack_sharing_enabled":true,"relatedPostsOrdered":[],"bylinesOrdered":[{"key":"_50fg5nmin","termId":905}],"acknowledgementsOrdered":[],"_links":{"self":[{"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/decoded\/276701","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/decoded"}],"about":[{"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/types\/decoded"}],"author":[{"embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/users\/658"}],"replies":[{"embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/comments?post=276701"}],"version-history":[{"count":1,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/decoded\/276701\/revisions"}],"predecessor-version":[{"id":276726,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/decoded\/276701\/revisions\/276726"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/media\/276722"}],"wp:attachment":[{"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/media?parent=276701"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/categories?post=276701"},{"taxonomy":"bylines","embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/bylines?post=276701"},{"taxonomy":"collection","embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/collection?post=276701"},{"taxonomy":"_post_visibility","embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/_post_visibility?post=276701"},{"taxonomy":"decoded-category","embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/decoded-category?post=276701"},{"taxonomy":"formats","embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/formats?post=276701"},{"taxonomy":"_fund_pool","embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/_fund_pool?post=276701"},{"taxonomy":"languages","embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/languages?post=276701"},{"taxonomy":"regions-countries","embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/regions-countries?post=276701"},{"taxonomy":"research-teams","embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/research-teams?post=276701"},{"taxonomy":"workflow-status","embeddable":true,"href":"https:\/\/alpha.pewresearch.org\/pewresearch-org\/wp-json\/wp\/v2\/workflow-status?post=276701"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}