Table 3 Validation of Infutor and MIGRATE estimates over time for selected representative metrics

From: Inferring fine-grained migration patterns across the United States

Root mean squared error (RMSE) compared to Census data

Quantity

Dataset

2011

2012

2013

2014

2015

2016

2017

2018

2019

CBG population

Infutor

–

–

–

507

519

531

556

581

–

MIGRATE

–

–

–

80

91

91

91

88

–

State-to-state movers

Infutor

3031

3684

3549

3372

3607

3315

2801

2518

2431

MIGRATE

365

403

400

399

434

410

343

361

383

State in-migration rate (per 1000)

Infutor

12.4

14.9

14.4

13.4

14.1

12.4

9.77

6.78

5.56

MIGRATE

1.68

1.73

1.92

1.48

1.29

1.18

1.07

1.03

1.12

Pearson correlation compared to Census data

Quantity

Dataset

2011

2012

2013

2014

2015

2016

2017

2018

2019

CBG population

Infutor

–

–

–

0.824

0.826

0.827

0.822

0.818

–

MIGRATE

–

–

–

0.996

0.995

0.995

0.996

0.996

–

State-to-state movers

Infutor

0.950

0.956

0.953

0.951

0.954

0.950

0.948

0.936

0.919

MIGRATE

0.998

0.997

0.997

0.998

0.997

0.998

0.998

0.998

0.998

State in-migration rate

Infutor

0.895

0.916

0.911

0.884

0.885

0.896

0.871

0.863

0.809

MIGRATE

0.980

0.981

0.976

0.987

0.991

0.992

0.993

0.994

0.992

  1. Year corresponds to year of the Census data release. 5-Year estimates of Census Block Group Populations are only validated against matrices from 2014 to 2019, which span the full 5-year period. Metrics for in-migration rates are weighted by population area. More details on computation of these metrics can be found on our validations section.
Back to article page