Absentee (information about this variable and its quality)

Stats NZ initially provided one rating of very poor for the absentees suite of variables. Each variable has now been reassessed independently and number of census night absentees has been been given a rating of moderate quality.

The Data quality processes section has more information.

Further definitions for absentee variables:

• ‘number of census night absentees' is the total number of absentees for each household
• 'absentee in New Zealand on census night' provides information on whether absentees were elsewhere in New Zealand or were overseas on census night
• 'absentee – time away from New Zealand' collects information on how long altogether an absentee, who is overseas on census night, is away from New Zealand
• a 'New Zealand resident temporarily overseas' is an absentee who is overseas and away from New Zealand for less than 12 months.

Priority level

Priority level 1

We assign a priority level to all census variables: Priority 1, 2, or 3 (with 1 being highest and 3 being the lowest priority).

Priority levels are set at the start of the census cycle and absentee variables were rated priority 1. The new methodology used to create the 2018 Census dataset has reduced the importance of these variables and absentee data was not used for population counts for the 2018 Census. The absentee variables should therefore be reconsidered as priority 2.

The priority level for the absentee variables in 2013 was 2 (previously known as a ‘defining’ variable).

Quality Management Strategy and the Information by variable for absentee (2013) have more information on the priority rating.

Overall quality rating for 2018 Census

• Number of census night absentees: Moderate quality
• Absentee in New Zealand on census night: Very poor quality
• Absentee time away from New Zealand: Very poor quality

Data quality processes section below has more detail on the rating for this variable.

The External Data Quality Panel has provided an independent assessment of the quality of this suite of variables and has rated it as very poor quality for the suite. The External Data Quality Panel has not given a quality rating to each of the variables within the absentee suite. 2018 Census External Data Quality Panel: Assessment of Variables has more information.

Subject population

Subject populations for the absentee variables are:

• ‘number of absentees’: occupied private dwellings
• ‘absentee in New Zealand on census night’: absentees from occupied private dwellings
• ‘absentee – time away from New Zealand’: absentees from occupied private dwellings who are away from New Zealand on census night.

‘Subject population’ means the people, families, households, or dwellings to whom the variable applies.

How this data is classified

Census number of absentees classification V2.1.0

Census number of absentees is a flat classification with the following categories:

0000 No absentees

0001 1 absentee

0002 2 absentees

0003 3 absentees

… …

… …

0098 98 absentees

0099 99 absentees or more

9999 Not elsewhere included

’Not elsewhere included’ contains the residual categories of ‘response unidentifiable’ and ‘not stated’.

In New Zealand on census night V1.1.0

Absentee in New Zealand on census night is a flat classification with the following categories:

01 Absentee in New Zealand on census night

02 Absentee not in New Zealand on census night

99 Not elsewhere included

’Not elsewhere included’ contains the residual categories of ‘response unidentifiable’ and ‘not stated’.

Census Time Away from New Zealand Classification V2.1.0

Absentee – time away from New Zealand is a flat classification with the following categories:

01 Away from New Zealand for less than 12 months

02 Away from New Zealand for 12 months or more

99 Not elsewhere included

’Not elsewhere included’ contains the residual categories of ‘response unidentifiable’ and ‘not stated’.

Number of absentees can be grouped for output in single categories from 0–4 absentees, and 5 or more absentees.

The classification of the absentee variables in the 2018 Census is consistent with the classifications used in the 2013 and 2006 Censuses

2013 information by variable – absentee has more information

The Standards and Classifications page provides background information on classifications and standards.

Question format

Absentee data is derived from absentee indicator, absentee number, and absentee relationship to reference person, on the household setup form (questions 18, 19, 20 on the paper dwelling form).

Stats NZ Store House has samples for both the individual and dwelling paper forms.

There were changes to the question wording and format from the 2013 Census:

• the 2018 paper dwelling form allowed up to four absentees to be noted in comparison with 2013 which allowed up to five absentees.

There were differences between the wording or question format in the online and paper versions of this question:

• the 2018 household setup form allowed up to five absentees to be noted for a dwelling in comparison to four on the paper dwelling form
• the household setup form asked the respondent “Are there any people who usually live at [insert address of dwelling] who are away on the night of Tuesday 6 March?”. The paper version (question 20) asked “Will everyone who usually lives in this dwelling be staying here on the night of Tuesday 6 March 2018?”
• the household setup form asked the respondent to “Please enter the following details for all the people who usually live at [insert address] but are away on the night of Tuesday 6 March”. On the paper form this question (20) was asked as follows: “Please list everyone who won’t be filling in an Individual Form here in this dwelling (and won’t have one filled in for them here), and answer the questions about them.”
• there was no ‘number of absentees’ question on the household setup form. The respondent was asked to confirm that the number of absentees (generated from the details provided) was correct on completion of the form. When forms were completed on paper, question 19 asked the respondent to write in the number of people who were absent on census night.

There were also differences in the way a person could respond:

On the household setup form:

• the question “Are there any people who usually live at [insert address] who are away on the night of Tuesday 6 March?” allowed only one response ‘yes’ or ‘no’ to be selected. If a further response was selected, the first response given disappeared.
• if the respondent ticked ‘Yes’ to the absentee question on the household setup form, the absentee details section automatically displayed and it was mandatory to give at least one name, but optional to provide further details such as their sex, their relationship to the person completing the form, whether they were in New Zealand on census night or not, and if not in New Zealand how long they were away for.

On the paper dwelling form:

• non-response, responses outside the valid range and multiple responses were possible. Edits were used to resolve these responses in the data.

How this data is used

Absentee data informs family and household derivations which are widely used both within Stats NZ and externally.

2018 data sources

No alternative data sources or imputation was used to replace missing responses or responses that could not be classified for the absentee variables.

Number of absentees is derived from the count of absentees at a particular dwelling. Therefore it has no non-response.

The ‘no information’ percentage is where we were not able to source absentee data for a person in the subject population.

Note: All records in the subject population for time away from New Zealand were edited to non-residual codes.

Missing and residual responses

Absentee in New Zealand on census night

‘No information’ in the data sources table, is the percentage of the subject population coded to ‘not stated’. In previous censuses, non-response was the percentage of the subject population coded to ‘not stated.’

Percentage of ‘not stated’ for absentee in New Zealand on census night – absentees from occupied private dwellings:

• 2018: 41.4 percent
• 2013: 9.4 percent
• 2006: 7.5 percent.

Responses that could not be classified or did not provide the type of information asked for, such as response unidentifiable, remain in the data where we have been unable to find information from another source. In the 2018 Absentee in New Zealand on census night – Data Sources table, these residuals are included in the ‘Response from 2018 Census’ percentage.

For output purposes, as with the 2013 and 2006 Censuses, these residual category responses are grouped with ‘not stated’ and are classified as ‘not elsewhere included’.

Due to the small number of residuals other than ‘not stated’ in 2006, the ‘not elsewhere included’ percentage was previously the same as ‘not stated’.

Percentage of ‘not elsewhere included’ for absentee in New Zealand on census night - absentees from occupied private dwellings:

• 2018: 41.5 percent
• 2013: 9.5 percent
• 2006: 7.5 percent.

Absentee – time away from New Zealand

For the 2018 Absentee – time away from New Zealand subject population all records were edited to non-residuals codes which meant that not stated and not elsewhere included percentages are zero.

Percentage of ‘not stated’ for absentee – time away from New Zealand - absentees from occupied private dwellings, who were away from New Zealand on census night.

• 2018: 0.0 percent
• 2013: 4.4 percent
• 2006: 4.0 percent.

Percentage of ‘not elsewhere included’ for absentee – time away from New Zealand - absentees from occupied private dwellings, who were away from New Zealand on census night.

• 2018: 0.0 percent
• 2013: 4.6 percent
• 2006: 4.1 percent.

2013 Census data user guide provides more information about non-response in the 2013 Census.

Data quality processes

• Number of census night absentees: Moderate quality
• Absentee in New Zealand on census night: Very poor quality
• Absentee time away from New Zealand: Very poor quality

Data was evaluated to assess whether it meets quality standards and is suitable for use.

Three quality metrics contributed to the overall quality rating:

• data sources and coverage
• consistency and coherence
• data quality.

The lowest rated metric determines the overall quality rating.

Data quality assurance for 2018 Census provides more information on the quality rating scale.

Data sources and coverage:

• Number of census night absentees: High quality
• Absentee in New Zealand on census night: Very poor quality
• Absentee time away from New Zealand: Very poor quality

We have assessed the quality of all the data sources that contribute to the output for the variable. To calculate a data sources and coverage quality score for a variable, each data source is rated and multiplied by the proportion it contributes to the total output.

The rating for a valid census response is defined as 1.00. Ratings for other sources are the best estimates available of their quality relative to a census response. Each source that contributes to the output for that variable is then multiplied by the proportion it contributes to the total output. The total score then determines the metric rating according to the following range:

• 98–100 = very high
• 95–<98 = high
• 90–<95 = moderate
• 75–<90 = poor
• <75 = very poor.

Number of census night absentees is a derived variable and does not have a data sources and coverage calculation table. All number of absentee information is sourced from the 2018 Census forms, with no admin data or imputation. This variable was rated as high quality.

For absentee in New Zealand on census night, the low proportion of data from received forms resulted in the score of 0.58, determining the very poor quality rating.

The subject population for absentee time away from New Zealand is comprised of absentees who were marked “no” to the absentee in New Zealand on census night question. Therefore, this variable is also rated as very poor quality for data sources and coverage.

Consistency and coherence:

• Number of census night absentees: Moderate quality
• Absentee in New Zealand on census night: Poor quality
• Absentee time away from New Zealand: Poor quality

Number of census night absentees data is mostly consistent with expectations across consistency checks. There is an overall difference in the data compared with expectations.

Absentee in New Zealand on census night, and absentee time away from New Zealand data is not consistent overall with expectations across one or more consistency checks. There is an overall difference in the data compared with expectations and benchmarks. Where this difference occurs, this cannot be fully explained through likely real-world change, incorporation of other sources of data, or a change in how the variable has been collected.

• Absentee in New Zealand on census night: there was a significant decrease in the percentage of absentees in New Zealand on census night compared to 2013 and 2006 Censuses. This was due to the high percentage of ‘not stated’ responses to this question.
• Absentee time away from New Zealand: the low response rate to ‘Absentee in New Zealand on census night’ is likely to have impacted the subject population for ‘Absentee time away from New Zealand’.

Data quality:

• Number of census night absentees: High quality
• Absentee in New Zealand on census night: Very poor quality
• Absentee time away from New Zealand: Very poor quality

The data quality checks for the absentee variables included edits for consistency within the dataset and cross-tabulations to the SA2 geographic level.

The data quality of number of census night absentees was assessed by determining the percentage of households that may have missing people not accounted for by absentees. The percentage of households with missing people was estimated to be about 1.8%, determining the high quality rating.

Absentee in New Zealand on census night has very high non-response and is not representative of the real world. Absentee time away from New Zealand is potentially impacted by the non-response to absentee in New Zealand on census night.

Recommendations for use and further information

The quality of number of census night absentees is moderate and comparable with 2013 and 2006 data.

Absentee in New Zealand on census night and absentee time away from New Zealand are very poor quality. Data for these variables is not available for use.

Contact our Information Centre for further information about using this variable