study to replace immigration questions with administrative ... old definitions... · compare data...

Delivering insight through data, for a better Canada

Study to replace immigration questions with administrative data for the 2021

Census of Population

Rose Evra

Athanase Barayandema

Statistics CanadaIFMS, January 2020


Project overview

Replace immigrant status and year of immigrationquestions on the 2021 Census with administrative data

2


Objectives of study

Reduce response burden

Improve data quality

Potentially add new content to the 2021 Census from administrative data

3


Steps

Assess and compare limitations of census questions and administrative data

Create test environment based on 2016 Census

Replace responses with linked values and run E&I

Compare data quality of test vs. 2016 Census results

Identify issues with replacement

Identify and test solutions to address known issues

Prepare report with final recommendations

4


Immigration data in Canada

Census of Population

Key data source on the socio-economic outcomes of immigrants in Canada

Immigration questions: immigrant status, year of immigration, place of birth, place of birth of parents and citizenship

Immigrant Non-immigrant

Canadian by birth Canadian by birth

Canadian by naturalization Canadian by naturalization

Not a Canadian citizen Permanent resident Non-permanent resident


Immigration data in Canada (2)

Immigration, Refugees and Citizenship Canada (IRCC) administrative data

Immigration landing file (1980 to present)

Historical landing file (1952 to 1979)

Asylum claims, study and work permits (1980 to present), visitor records (2006 to present), and citizenship records (2005 to present)

IRCC administrative data was used to enhance 2016 Census

Used to improve 2016 Census data processing

Used to add admission category variables onto the 2016 Census


Limitations of census questionsPercentage of census respondents linked to IRCC immigrant records who responded that they are not immigrants by years since admission

7.6% of respondents who were linked to administrative immigration records (for immigrants admitted since 1980) responded “No” to the immigrant status question. This proportion increases the longer they have been in Canada. These cases are largely corrected during processing based on other responses.

Source: Statistics Canada, 2016 and 2006 Censuses linked to IRCC administrative immigrant data since 1980 and 2011 National Household Survey (NHS) linked to IRCC administrative immigrant data since 1980


Limitations of census questions (2)

Match rate for census reported year of immigration by linked administrative value

Over 21% of immigrants had a year of immigration response which did not match their linked administrative value. The longer immigrants have been in Canada, the less likely they are of accurately reporting their year of immigration.

Source: Statistics Canada, 2016 and 2006 Censuses linked to IRCC administrative immigrant data since 1980 and 2011 National Household Survey (NHS) linked to IRCC administrative immigrant data since 1980


Limitations of administrative data

Administrative data are not updated to reflect deaths or emigration

The data need to be linked to estimate the current population of immigrants

No records for immigrants who landed prior to 1952

111,000 immigrants or 1.5% of immigrants in 2016

Weaker linkage quality for the pre-1980 file

Especially records from 1961-1972 due to incomplete date of birth (only year of birth is available)

No records for dependants of non-permanent resident permit holders, visitors, or other non-permit non-permanent residents


Limitations of administrative data (2)Percentage of 2016 Census immigrants linked to administrative immigration records by 2016 Census year of immigration


How to replace immigrant status and year of immigration

Record linkage to administrative data identifies immigrants and non-permanent residents

Year of immigration for immigrants

Edit and imputation processes to address missing links or inconsistencies between linked values and other census responses


Challenges

Limitations of administrative data

No immigrant records before 1952

Lower linkage rates for immigrants between 1961-1972

Coverage of non-permanent residents in initial linkage restricted to asylum claimants since 2004 and temporary residents with a valid permit in 2016

Comparability with past census results

Year of immigration will be more reflective of administrative value vs. perceived value


Proposed solutions

No immigrants before 1952

Use past census responses to supplement administrative data for years prior to 1952 (2011, 2006, or 2001)

Lower linkage rates for 1961-1972

Use of multipliers to assign a year of immigration in this period for a specified number of unlinked records

Multipliers are the ratios of predicted probabilities of being linked if DOB was not missing


Proposed solutions (2)

Lower coverage of non-permanent residents

Linking to permits which were no longer valid in 2016 and refugee claimants with claims before 2004 Supplementing 2016 linkage results using SDLE results

Linking to visitors records A number of visitors respond to Census questionnaire

Dependents of non-permanent resident permit holders

“Super-Visas”


Test results

Table 1

Immigrant status and citizenship, 2016 Census and test results

Source: Statistics Canada, 2016 Census and administrative data replacement results

Immigration status and citizenship 2016 Census Admin. data replacement

Difference

Number %

Canadian citizens by birth 26,412,610 26,413,060 450 0.0

Immigrants, total 7,540,830 7,549,380 8,550 0.1

Permanent residents 1,918,855 1,925,325 6,470 0.3

Canadian citizens by naturalization 5,621,975 5,624,055 2,080 0.0

Non-permanent residents 506,625 497,625 -9,000 -1.8


Single year of immigration, 2016 Census and Test results

-

50,000

100,000

150,000

200,000

250,000

300,0001

93

0

19

33

19

36

19

39

19

42

19

45

19

48

19

51

19

54

19

57

19

60

19

63

19

66

19

69

19

72

19

75

19

78

19

81

19

84

19

87

19

90

19

93

19

96

19

99

20

02

20

05

20

08

20

11

20

14

Test (with multipliers) Test (without multipliers) 2016 Census



Test results

Table 2

Period of immigration, 2016 Census and test results


Period of immigration

2016 CensusAdmin data replacement

Difference

N % N % N P.P.*

Before 1980 1,837,220 24.4 1,821,835 24.1 -15,385 -0.2

1980-1990 1,019,840 13.5 969,420 12.8 -50,420 -0.7

1991-2000 1,486,665 19.7 1,556,430 20.6 69,765 0.9

2001-2005 928,940 12.3 928,560 12.3 -380 0.0

2006-2010 1,056,085 14.0 1,058,560 14.0 2,475 0.0

2011-2016 1,212,070 16.1 1,214,560 16.1 2,490 0.0


Retention/survival rates of immigrants by year of immigration, 2016 Census and Test results

0%

20%

40%

60%

80%

100%

120%

19

52

19

54

19

56

19

58

19

60

19

62

19

64

19

66

19

68

19

70

19

72

19

74

19

76

19

78

19

80

19

82

19

84

19

86

19

88

19

90

19

92

19

94

19

96

19

98

20

00

20

02

20

04

20

06

20

08

20

10

20

12

20

14

20

16

Test (with multipliers) Test (without multipliers) 2016 Census

Note: Retention/survival rates are ratios between 2016 Census/test estimates and IRCC (denominator)

Source: Statistics Canada, 2016 Census and administrative data replacement results, and IRCC data


Single year of immigration, 2016 Census, Test results, IRCC and IMDB Data

-

50,000

100,000

150,000

200,000

250,000

300,000

350,000

Test (with multipliers) 2016 Census IRCC IMDB_2015

Note: IRCC and IMDB data provide information on immigration in Canada, they were used for certification of the 2016 Census.Source: Statistics Canada, 2016 Census and administrative data replacement results, IMDB and IRCC data


Summary of test results

Improved quality of year of immigration since 1980 Better coherence with other data sources such as IMDB

Notable shift in early 1990s

Slightly lower estimates for years prior to 1952 after supplement

Mitigation of lower linkage quality for years 1961-1972

Lower estimate of non-permanent residents

Resulting higher estimate of permanent residents


Next steps

Improve administrative data: Integrate new administrative data to improve E&I processes

Additional data (more visitor records) for non-permanent residents


IRCC-Statistics Canada collaboration

Immigration, Refugees, and Citizenship Canada

Several groups from Statistics Canada Methodology

Sean Crowe, Caroline Cauchon, Lyne Guertin

Subject-matter Scott McLeish, Kathryn Spence, Tristan Cayn, Eric Mongrain

Census Operations


Contact information

Rose Evra

Analyst

Diversity and Sociocultural Statistics

[email protected]

Athanase Barayandema

Analyst

Diversity and Sociocultural Statistics

[email protected]

mailto:[email protected]

mailto:[email protected]


Response mobility by immigrant status, 2016 Census and Test results

15.5%

6.6%

-17.3%

-6.3%

0.0%0.0%

-1.8%

0.3%

-25.0%

-20.0%

-15.0%

-10.0%

-5.0%

0.0%

5.0%

10.0%

15.0%

20.0%

Canadian citizens by birth Canadian citizens bynaturalisation

Non-permanent residents Permanent residents

Gain Loss Response mobility



Measurement Risk

Table 3Measurement risk for year of immigration, 2016 Census and test results

*Differences between responses and linked administrative values

Note: All figures are based on years of immigration from 1952 to 2016


Measurement risk 2016 Census Admin data replacement

Response mismatch* 29.1%

False positive linkage rate 0.3%

Imputation rate 9.3% 14.2%

study to replace immigration questions with administrative ... old definitions... · compare data...

Documents