1{2 "record_id": "1cd473df8f393883694de4530b0d1c667382890c",3 "eligible": true,4 "exclusion_reason": null,5 "study": {6 "title": "Exploring a multilevel approach with spatial effects to model housing price in San Jose, Costa Rica",7 "year": 2021,8 "doi": "10.1177/23998083211041122",9 "country": "Costa Rica",10 "region_city": "San Jose Metropolitan Region",11 "sample_start": null,12 "sample_end": null,13 "property_type": "single_family",14 "price_type": "listing",15 "data_source": "1,264 real estate listings of detached houses from an aggregator website (year of listings not stated in main text; covariates from 2018 and Census 2011)",16 "document_type": "journal",17 "estimators": [18 "other"19 ],20 "spatial_method": "Bayesian multilevel random intercepts (166 districts) with intrinsic CAR spatial prior (CARM); MCMC in WinBUGS",21 "time_fe": false,22 "location_fe": true,23 "se_type": "unknown",24 "n_obs_total": 1264,25 "ai_methods": [],26 "notes": "Environment and Planning B (accepted manuscript version). Dependent variable = natural log of listing price per m2 of living area (US$/m2), so living area is not a regressor. Three models with covariates (GaLM, MLRE, CARM); VPM model has no covariates and is not extracted. Estimator recorded as 'other' = Bayesian hierarchical (MCMC). Sample years of listings not stated. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."27 },28 "models": [29 {30 "model_id": "T3-c1",31 "table": "3",32 "column": "1",33 "page": 10,34 "dep_var": "ln housing price per m2 of living area (US$/m2), from listings",35 "dep_transform": "log",36 "estimator": "other",37 "n_obs": 1264,38 "r2": null,39 "se_or_t_in_parentheses": "none",40 "sample_subset": "Linear model without multilevel effects (Bayesian, WinBUGS); posterior means; dependent variable is log of price per m2"41 },42 {43 "model_id": "T3-c2",44 "table": "3",45 "column": "2",46 "page": 10,47 "dep_var": "ln housing price per m2 of living area (US$/m2), from listings",48 "dep_transform": "log",49 "estimator": "other",50 "n_obs": 1264,51 "r2": null,52 "se_or_t_in_parentheses": "none",53 "sample_subset": "Multilevel random-intercept model (district level); posterior means; dependent variable is log of price per m2"54 },55 {56 "model_id": "T3-c3",57 "table": "3",58 "column": "3",59 "page": 10,60 "dep_var": "ln housing price per m2 of living area (US$/m2), from listings",61 "dep_transform": "log",62 "estimator": "other",63 "n_obs": 1264,64 "r2": null,65 "se_or_t_in_parentheses": "none",66 "sample_subset": "Multilevel random intercepts with intrinsic CAR spatial effect at district level; posterior means; dependent variable is log of price per m2"67 }68 ],69 "coefficients": [70 {71 "model_id": "T3-c1",72 "var_name_printed": "beta1 (ln house plot area)",73 "taxonomy_id": "lot_size",74 "x_transform": "log",75 "x_unit": "m2",76 "beta": 0.0785,77 "se": null,78 "t": null,79 "p": null,80 "p_stars": "",81 "ci_lo": 0.0633,82 "ci_hi": 0.0936,83 "page": 10,84 "confidence": "high",85 "notes": "ln plot area (m2); lower-level covariate. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."86 },87 {88 "model_id": "T3-c2",89 "var_name_printed": "beta1 (ln house plot area)",90 "taxonomy_id": "lot_size",91 "x_transform": "log",92 "x_unit": "m2",93 "beta": 0.072,94 "se": null,95 "t": null,96 "p": null,97 "p_stars": "",98 "ci_lo": 0.0559,99 "ci_hi": 0.088,100 "page": 10,101 "confidence": "high",102 "notes": "ln plot area (m2); lower-level covariate. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."103 },104 {105 "model_id": "T3-c3",106 "var_name_printed": "beta1 (ln house plot area)",107 "taxonomy_id": "lot_size",108 "x_transform": "log",109 "x_unit": "m2",110 "beta": 0.0696,111 "se": null,112 "t": null,113 "p": null,114 "p_stars": "",115 "ci_lo": 0.0538,116 "ci_hi": 0.0854,117 "page": 10,118 "confidence": "high",119 "notes": "ln plot area (m2); lower-level covariate. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."120 },121 {122 "model_id": "T3-c1",123 "var_name_printed": "beta2 (ln travel time to CBD)",124 "taxonomy_id": "cbd_distance",125 "x_transform": "log",126 "x_unit": "minutes",127 "beta": 0.613,128 "se": null,129 "t": null,130 "p": null,131 "p_stars": "",132 "ci_lo": 0.426,133 "ci_hi": 0.801,134 "page": 10,135 "confidence": "high",136 "notes": "ln estimated travel time to CBD San Jose (min); quadratic in logs with beta3. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."137 },138 {139 "model_id": "T3-c2",140 "var_name_printed": "beta2 (ln travel time to CBD)",141 "taxonomy_id": "cbd_distance",142 "x_transform": "log",143 "x_unit": "minutes",144 "beta": 0.506,145 "se": null,146 "t": null,147 "p": null,148 "p_stars": "",149 "ci_lo": 0.264,150 "ci_hi": 0.756,151 "page": 10,152 "confidence": "high",153 "notes": "ln estimated travel time to CBD San Jose (min); quadratic in logs with beta3. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."154 },155 {156 "model_id": "T3-c3",157 "var_name_printed": "beta2 (ln travel time to CBD)",158 "taxonomy_id": "cbd_distance",159 "x_transform": "log",160 "x_unit": "minutes",161 "beta": 0.383,162 "se": null,163 "t": null,164 "p": null,165 "p_stars": "",166 "ci_lo": 0.118,167 "ci_hi": 0.647,168 "page": 10,169 "confidence": "high",170 "notes": "ln estimated travel time to CBD San Jose (min); quadratic in logs with beta3. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."171 },172 {173 "model_id": "T3-c1",174 "var_name_printed": "beta3 (ln travel time to CBD squared)",175 "taxonomy_id": "cbd_distance",176 "x_transform": "other",177 "x_unit": "minutes",178 "beta": -0.113,179 "se": null,180 "t": null,181 "p": null,182 "p_stars": "",183 "ci_lo": -0.148,184 "ci_hi": -0.0779,185 "page": 10,186 "confidence": "high",187 "notes": "Square of ln travel time to CBD. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."188 },189 {190 "model_id": "T3-c2",191 "var_name_printed": "beta3 (ln travel time to CBD squared)",192 "taxonomy_id": "cbd_distance",193 "x_transform": "other",194 "x_unit": "minutes",195 "beta": -0.0945,196 "se": null,197 "t": null,198 "p": null,199 "p_stars": "",200 "ci_lo": -0.142,201 "ci_hi": -0.0491,202 "page": 10,203 "confidence": "high",204 "notes": "Square of ln travel time to CBD. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."205 },206 {207 "model_id": "T3-c3",208 "var_name_printed": "beta3 (ln travel time to CBD squared)",209 "taxonomy_id": "cbd_distance",210 "x_transform": "other",211 "x_unit": "minutes",212 "beta": -0.0723,213 "se": null,214 "t": null,215 "p": null,216 "p_stars": "",217 "ci_lo": -0.125,218 "ci_hi": -0.0198,219 "page": 10,220 "confidence": "high",221 "notes": "Square of ln travel time to CBD. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."222 },223 {224 "model_id": "T3-c1",225 "var_name_printed": "beta4 (ln distance to major road)",226 "taxonomy_id": "highway_distance",227 "x_transform": "log",228 "x_unit": "m",229 "beta": -0.0111,230 "se": null,231 "t": null,232 "p": null,233 "p_stars": "",234 "ci_lo": -0.0207,235 "ci_hi": -0.00147,236 "page": 10,237 "confidence": "high",238 "notes": "ln Euclidean distance to nearest major road (m). Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."239 },240 {241 "model_id": "T3-c2",242 "var_name_printed": "beta4 (ln distance to major road)",243 "taxonomy_id": "highway_distance",244 "x_transform": "log",245 "x_unit": "m",246 "beta": -0.012,247 "se": null,248 "t": null,249 "p": null,250 "p_stars": "",251 "ci_lo": -0.0224,252 "ci_hi": -0.00165,253 "page": 10,254 "confidence": "high",255 "notes": "ln Euclidean distance to nearest major road (m). Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."256 },257 {258 "model_id": "T3-c3",259 "var_name_printed": "beta4 (ln distance to major road)",260 "taxonomy_id": "highway_distance",261 "x_transform": "log",262 "x_unit": "m",263 "beta": -0.0113,264 "se": null,265 "t": null,266 "p": null,267 "p_stars": "",268 "ci_lo": -0.0218,269 "ci_hi": -0.000723,270 "page": 10,271 "confidence": "high",272 "notes": "ln Euclidean distance to nearest major road (m). Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."273 },274 {275 "model_id": "T3-c1",276 "var_name_printed": "beta5 (ln distance to bus route)",277 "taxonomy_id": "transit_distance",278 "x_transform": "log",279 "x_unit": "m",280 "beta": 0.0112,281 "se": null,282 "t": null,283 "p": null,284 "p_stars": "",285 "ci_lo": 0.00463,286 "ci_hi": 0.0178,287 "page": 10,288 "confidence": "high",289 "notes": "ln Euclidean distance to nearest bus route (m). Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."290 },291 {292 "model_id": "T3-c2",293 "var_name_printed": "beta5 (ln distance to bus route)",294 "taxonomy_id": "transit_distance",295 "x_transform": "log",296 "x_unit": "m",297 "beta": 0.0084,298 "se": null,299 "t": null,300 "p": null,301 "p_stars": "",302 "ci_lo": 0.00183,303 "ci_hi": 0.015,304 "page": 10,305 "confidence": "high",306 "notes": "ln Euclidean distance to nearest bus route (m). Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."307 },308 {309 "model_id": "T3-c3",310 "var_name_printed": "beta5 (ln distance to bus route)",311 "taxonomy_id": "transit_distance",312 "x_transform": "log",313 "x_unit": "m",314 "beta": 0.00806,315 "se": null,316 "t": null,317 "p": null,318 "p_stars": "",319 "ci_lo": 0.0016,320 "ci_hi": 0.0145,321 "page": 10,322 "confidence": "high",323 "notes": "ln Euclidean distance to nearest bus route (m). Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."324 },325 {326 "model_id": "T3-c1",327 "var_name_printed": "gamma2 (ln crime rate)",328 "taxonomy_id": "crime",329 "x_transform": "log",330 "x_unit": "count",331 "beta": 0.000414,332 "se": null,333 "t": null,334 "p": null,335 "p_stars": "",336 "ci_lo": -0.00336,337 "ci_hi": 0.00419,338 "page": 10,339 "confidence": "high",340 "notes": "ln homicides per 100,000 residents in district (2018), 0.001 added; district-level covariate. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."341 },342 {343 "model_id": "T3-c2",344 "var_name_printed": "gamma2 (ln crime rate)",345 "taxonomy_id": "crime",346 "x_transform": "log",347 "x_unit": "count",348 "beta": -6.36e-05,349 "se": null,350 "t": null,351 "p": null,352 "p_stars": "",353 "ci_lo": -0.00642,354 "ci_hi": 0.00641,355 "page": 10,356 "confidence": "high",357 "notes": "ln homicides per 100,000 residents in district (2018), 0.001 added; district-level covariate. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."358 },359 {360 "model_id": "T3-c3",361 "var_name_printed": "gamma2 (ln crime rate)",362 "taxonomy_id": "crime",363 "x_transform": "log",364 "x_unit": "count",365 "beta": 0.000859,366 "se": null,367 "t": null,368 "p": null,369 "p_stars": "",370 "ci_lo": -0.00455,371 "ci_hi": 0.00622,372 "page": 10,373 "confidence": "high",374 "notes": "ln homicides per 100,000 residents in district (2018), 0.001 added; district-level covariate. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."375 },376 {377 "model_id": "T3-c1",378 "var_name_printed": "gamma3 (ln number of jobs in district)",379 "taxonomy_id": "employment_access",380 "x_transform": "log",381 "x_unit": "count",382 "beta": 0.0188,383 "se": null,384 "t": null,385 "p": null,386 "p_stars": "",387 "ci_lo": 0.00934,388 "ci_hi": 0.0283,389 "page": 10,390 "confidence": "high",391 "notes": "ln number of jobs in district (2018); district-level covariate. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."392 },393 {394 "model_id": "T3-c2",395 "var_name_printed": "gamma3 (ln number of jobs in district)",396 "taxonomy_id": "employment_access",397 "x_transform": "log",398 "x_unit": "count",399 "beta": 0.0203,400 "se": null,401 "t": null,402 "p": null,403 "p_stars": "",404 "ci_lo": 0.00406,405 "ci_hi": 0.0367,406 "page": 10,407 "confidence": "high",408 "notes": "ln number of jobs in district (2018); district-level covariate. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."409 },410 {411 "model_id": "T3-c3",412 "var_name_printed": "gamma3 (ln number of jobs in district)",413 "taxonomy_id": "employment_access",414 "x_transform": "log",415 "x_unit": "count",416 "beta": 0.0108,417 "se": null,418 "t": null,419 "p": null,420 "p_stars": "",421 "ci_lo": -0.00437,422 "ci_hi": 0.0261,423 "page": 10,424 "confidence": "high",425 "notes": "ln number of jobs in district (2018); district-level covariate. Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."426 },427 {428 "model_id": "T3-c1",429 "var_name_printed": "gamma4 (ln poverty proxy)",430 "taxonomy_id": "income",431 "x_transform": "log",432 "x_unit": "count",433 "beta": -0.0195,434 "se": null,435 "t": null,436 "p": null,437 "p_stars": "",438 "ci_lo": -0.0346,439 "ci_hi": -0.00427,440 "page": 10,441 "confidence": "high",442 "notes": "ln number of households with unsatisfied housing needs (Census 2011), poverty proxy; district-level covariate; closest leaf income (sign negative = poorer). Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."443 },444 {445 "model_id": "T3-c2",446 "var_name_printed": "gamma4 (ln poverty proxy)",447 "taxonomy_id": "income",448 "x_transform": "log",449 "x_unit": "count",450 "beta": -0.0259,451 "se": null,452 "t": null,453 "p": null,454 "p_stars": "",455 "ci_lo": -0.0542,456 "ci_hi": 0.00214,457 "page": 10,458 "confidence": "high",459 "notes": "ln number of households with unsatisfied housing needs (Census 2011), poverty proxy; district-level covariate; closest leaf income (sign negative = poorer). Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."460 },461 {462 "model_id": "T3-c3",463 "var_name_printed": "gamma4 (ln poverty proxy)",464 "taxonomy_id": "income",465 "x_transform": "log",466 "x_unit": "count",467 "beta": -0.0117,468 "se": null,469 "t": null,470 "p": null,471 "p_stars": "",472 "ci_lo": -0.0349,473 "ci_hi": 0.0114,474 "page": 10,475 "confidence": "high",476 "notes": "ln number of households with unsatisfied housing needs (Census 2011), poverty proxy; district-level covariate; closest leaf income (sign negative = poorer). Bayesian posterior mean with 2.5% and 97.5% percentiles (95% posterior interval) recorded as ci_lo/ci_hi; no SE or stars printed."477 }478 ]479}