User's guide for cross-sectional public-use microdata file
Survey of Labour and Income Dynamics (SLID)
Information identified as archived is provided for reference, research or recordkeeping purposes. It is not subject to the Government of Canada Web Standards and has not been altered or updated since it was archived. Please "contact us" to request a format other than those available.
Reference year 2008
Using the record layouts, data dictionary and frequencies
Guidelines for applying weights
Guidelines for release (data quality and rounding)
Confidentiality of the Public-Use Microdata
SLID Content, Notes and Definitions, Methodology
Related Products and Services
Questions and Comments
The cross-sectional public-use microdata file for the Survey of Labour and Income Dynamics (SLID) is a collection of income, labour and family variables on persons in Canada and their families. SLID is an annual household survey covering the population of the 10 Canadian provinces with the exception of Indian reserves, residents of institutions and military barracks.
The Survey of Labour and Income Dynamics began collecting data for reference year 1993. Initially, SLID was designed to be, first and foremost, a longitudinal survey, with primary focus on labour and income and the relationships between them and family composition. Then, the decision was made to extend the objectives of SLID to be the primary source of cross-sectional household income data.
For many years, the Survey of Consumer Finances had provided public-use microdata files (PUMFs) to meet the needs of cross-sectional household income data users. Survey of Consumer Finances (SCF) PUMFs were released up to and including reference year 1997. For the purpose of standard publications, Statistics Canada has made the transition from SCF to SLID between 1995 and 1996. Therefore, SLID cross-sectional PUMFs are being made available beginning with reference year 1996. The SLID files have been designed to be analogous to those produced for the SCF. The type of income data collected by SLID was identical to that of the former household income survey SCF, with the distinction that SLID respondents had the choice of a traditional income interview or granting permission to Statistics Canada to use their T1 income tax data.
To find more information on comparability between SLID and SCF data please consult the two following documents: Bridging two surveys - An integrated series of income data from SCF and SLID - 1989-1997 and Comparison of income estimates from the survey of consumer finances and the survey of labour and income dynamics (see also section Related products and services)
How to cite SLID in publications For publication of any information based on the SLID microdata files on CD-ROM (75M0010XCB), the following form of accreditation is recommended: "This analysis is based on Statistics Canada's Survey of Labour and Income Dynamics Public Use Microdata, which contains anonymized data collected in the Survey of Labour and Income Dynamics. All computations on these microdata were prepared by (Name of user). The responsibility for the use and interpretation of these data is entirely that of the author(s)".
Although often referred to as one file, the SLID cross-sectional PUMF is four separate flat files: key, person, economic family and census family. To a large extent, the file structure used for SCFPUMFs has been maintained.
On the person file, there is one record per person in the sample aged 16 and over. Job characteristics such as industry, wage rates and work schedule are included on the person file and relate to the person's main job during the reference year (the job at which the most hours were worked during the year). Although SLID collects data on all jobs held during the year by each person under 70 years old, the characteristics of all other jobs are not included on the SLIDPUMFs. The person file does contain identifiers that allow a researcher to group persons into households, economic families and census families.
The key file contains one record per person in the sample including children under the age of 16. Basic demographic information is available on this file and the necessary identifiers in order to match to the files for person, economic and census families. The economic and census family files contain one record per family. Both files include a demographic summary, income information and labour characteristics.
|Files||Number of Records||Number of Variables||Record length|
|Economic family file||26,068||81||463|
|Census family file||29,025||71||424|
Additional information files are provided to assist users of the SLID public-use microdata files. For each of the four data files (key, person, economic family and census family), record layout, data dictionary and univariate distributions are provided. These information files are organized by content themes and in some cases sub-themes.
The following describes the structure of the additional information files:
A. The columns of the record layout file
Variable name. This is the variable name assigned for the public-use microdata file.
Type. Indicates whether the variable is numeric (in the sense that it can logically be used in mathematical operations) or character.
Size. Indicates both the number of spaces including the decimal point if there are decimal places and the number of decimal places, if any. For example, a variable which can have values of zero (00.0) to 99.9 would have a format expressed as: 4.1. A variable which can have values of zero (00) to 99 would have a format expressed as: 2.
Sequence. Indicates the order of variable appearance.
Start position. This shows the location of the variable on the public use microdata file.
Long variable name. A standardized name which can be used to quickly identify variables, to label tables, and so on. Although still rather cryptic, it is considerably more revealing than the variable name. However, this longer name obviously excludes a lot of important information contained in the variable description shown in the data dictionary. In short, analysts are warned against making assumptions about the variable definition based on the long variable name.
Number of categories. Shows the number of categories in the value set for the variable in question. Applies only to "character" variables. Numeric variables have ranges, which are specified in the data dictionary.
B. Data Dictionary
The data dictionary presents the complete information about each survey variable on each of the four files. For each variable in the record layout the following information is shown: the variable name, the description or definition, code lists with descriptions or alternatively the range of values that the variable can take on, the variable type, its length (or format), and the population to which the variable pertains, i.e. for whom it is applicable (See Appendix 1).
These distributions are provided to allow users of the public use microdata files to verify totals that they produce. These distributions relate to the public-use files and not to the internal database; the distributions will be similar but not identical due to confidentiality processing procedures used to produce the public-use files.
For character variables, the description, unweighted and weighted frequencies and weighted percentages for each code, including reserved codes (see below), are provided. For numeric variables, the values are broken into several ranges and show the description, unweighted and weighted frequencies and weighted percentages.
Missing Values and Reserved Codes
There are a few types of missing values on the public use files. SLID has adopted standard codes which have a particular meaning. It is important to account for reserved codes in any analysis, particularly with numeric variables. If your calculation of means or aggregates seems too high, check to ensure that you have excluded reserved codes from the calculation. With only a few exceptions, the reserved codes are the highest four values permitted according to the length of the variable. A brief explanation of reserved codes is provided below.
If the coverage of a variable does not extend to a certain population sub-group, then there are no valid values for that sub-group and the values (reserved codes) that do appear are in the form 9, 99, 9.9 and so on, which indicates that the variable is not applicable. The coverage of each variable on the file is referred to in the data dictionary as the "population".
For certain records, no valid value is available, although the value is applicable. Possibly, the respondent did not provide the information or it failed an edit in processing and the value was not imputed. Such missing values appear with a reserved code such as 7, 97, 9.7 and so on depending on the format. For certain variables, the number of missing values has been reduced through imputation. Missing values for the income variables have been entirely imputed, but most other variables may have missing values.
Finally, a few values may have been coded as 8, 98, 9.8, etc. These represent refusals to particular items in the interview. The approach for dealing with missing values of this last kind depends on the type of analysis being carried out and the extent of missing data. Although the end solution may be to exclude the records with missing values from the analysis, a review should first be carried out to assess the impact of missing values on the overall represent ativeness of the data. Is it possible that a bias results from the missing data? For example, are the (other) characteristics of the people with missing values different from those of the observed part of the sample? It may be necessary to take into account the possible impact in some way. In all cases, analysts should note exclusions of records with missing values in their published results.
The microdata on the public use files are unweighted. It is the responsibility of data users to apply the appropriate weights in any estimates they wish to produce. If proper weights are not used, the results derived from the microdata cannot be considered to be representative of the survey population, and will not correspond to those that would be produced by Statistics Canada. The weights are provided as variables under "Sample control". On the SLID PUMF, the weight variable is named WTCSLD26.
Microdata users should apply the rules for assessing data quality (see below) to all estimates they produce, and retain only those that satisfy the release criteria. Estimates that do not satisfy the release criteria are not reliable.
The guidelines for release and publication make use of the concept of sampling variability to determine whether the estimates obtained from the microdata are reliable. Sampling variability is the error in the estimates caused by the fact that we survey a sample rather than the entire population. The concept of standard error and the related concept of coefficient of variation and confidence interval provide an indication of the magnitude of the sampling variability.
The standard error and coefficient of variation do not measure any systematic biases in the survey data which might affect the estimate. Rather, they are based on the assumption that the sampling errors follow a normal probability distribution.
Subject to this assumption, it is possible to estimate the extent to which different samples that have the same design and the same number of observations would give different results. This indicates the margin of error that is likely to be included in the estimates derived from our single sample.
For a more complete description of the measures of sampling variability, see A. Satin and W. Shastry, Survey Sampling: A Non-Mathematical Guide, Statistics Canada, Catalogue 12-602E.
Minimum sizes of estimates for release
Suppression rules, or data reliability cut-offs, are currently established based on the sample size that underlies the estimate. In general, a sample size of 25 observations is required for the estimate to be published. Depending on the type of estimate, this rule can vary slightly. These rules help protect the confidentiality of survey respondents and ensure the reliability of estimates.
|Percentage, distribution, proportion/shares:|
||Denominator1 sample size < 25
Denominator1 sample size < 100 and numerator sample size < 5
||Numerator sample size < 25
Denominator sample size < 25
|Quintiles (shares, means and upper income limits)|
||sample in all quintiles/5 < 25
upper income limit for upper income quintile or total of quintiles
||sample < 25|
|1. The denominator sample size refers to the sample size of the total estimate from which the distribution, percentage, proportion or share is derived.|
Estimates of provincial aggregates and means
When producing estimates for provincial aggregates and means it should be noted that for a small number of records, province of residence has been suppressed. This will result in a small bias in provincial estimates.
To ensure that estimates from this microdata file intended for publication or any other type of release correspond to estimates that would be obtained by Statistics Canada, we strongly recommend that users comply with the following guidelines for rounding estimates:
- Estimates in the body of a statistical table must be rounded to the nearest hundredth using the traditional rounding technique, i.e., if the first or only number to be eliminated is between 0 and 4, the preceding number does not change. If the first or only number to be eliminated is between 5 and 9, the value of the last number to be retained increases by 1. For example, when using the traditional technique of rounding to the nearest hundredth, if the last two numbers are between 00 and 49, they are replaced by 00 and the preceding number (denoting hundredths) stays as is. If the last two numbers are between 50 and 99, they are replaced with 00 and the preceding number increased by 1.
- Total partial sub-totals and total sub-totals in statistical tables must be calculated using their unrounded corresponding components, then rounded in turn to the closest hundredth using the traditional rounding technique.
- Means, ratios, rates and percentages must be calculated using unrounded components (i.e., numerators and/or denominators), and then rounded to a decimal using the traditional rounding technique.
- Totals and differences in aggregates (or ratios) must be calculated using their corresponding unrounded components, then rounded to the nearest hundredth (or decimal place) using the traditional rounding technique.
- If, due to technical or other limitations, a technique other than traditional rounding is used, with the result that the estimates to be published or released differ in any form from the corresponding estimates that would be obtained by Statistics Canada using this microdata file, we strongly advise users to indicate the reasons for the differences in the documents to be published or released.
- Unrounded estimates should not under any circumstances be published or released. Unrounded estimates give the impression that they are much more precise than they actually are.
Hypothesis tests provided by statistical software packages
Microdata users should be aware that the results of hypothesis tests (such as the p values accompanying t statistics or Pearson statistics) that are provided automatically by most standard statistical software packages are incorrect for data provided by surveys with a complex survey design, such as SLID. Such packages calculate these test results under the assumption of simple random sampling. That is, they do not take into account the special sample design features of SLID such as stratification, clustering, and unequal selection probabilities. While many of the standard packages can account for the unequal selection probabilities in the production of estimates by allowing the use of weights, these packages do not properly take the sample design into account when producing variance estimates that form part of most test statistics.
To perform hypothesis tests, a two-step method can be employed with the standard statistical software to form the test statistics. First, estimate the characteristics of interest (total or mean) using the weights provided on the microdata file. Second, obtain approximate variance estimates of these characteristics by rerunning the same software procedure as that used for producing the characteristic estimates but using a scaled weight that consists of the original weight divided by the average of the original weights of all the observations being used in your computations. The standard error can be derived by using the estimate and the rough estimate of the variance. These quantities (estimate, variance, standard error) can then be combined to form test statistics. It must be noted that this method provides only rough approximations to the variance.
It should be noted that users of the SLID PUMF cannot readily obtain better design-based variance estimates through the use of statistical software specifically designed for survey data. This is because the design information required by these software packages is not currently available on the SLID data file due to confidentiality considerations. However, better variance estimates can be produced by Statistics Canada on a cost-recovery basis.
The production of a public-use microdata file includes many safeguards to prevent the identification of any one person. Longitudinal surveys are faced with an extra challenge when it comes to ensuring confidentiality, because data are collected for the same people for several years. For this reason, Statistics Canada plans to release only cross-sectional files from SLID. The number of topics covered in SLID also contributes to the amount of processing required specifically to ensure confidentiality. Confidentiality of the public-use file is ensured mainly by reducing information, i.e. deleting whole variables or suppressing or collapsing some of their detail.
SLID uses a number of techniques to ensure confidentiality:
- The SLID public-use file is comprised of a sample of the households randomly selected from the full SLID sample.
- All the variables which would permit direct identification of individuals are, of course, deleted from the file. This includes name, telephone number, and other data used for collection purposes;
- Collapsing categories. This is applied to categorical (i.e. qualitative) variables such as the variable 'size of area of residence'.
- Top and bottom coding. Very high and very low values usually are rare or unique in the population. Such extreme values are replaced with the value of an upper or lower limit.
- Rounding. Some variables, particularly monetary values, are rounded.
- Suppression and modification of characteristics was done while preserving integrity of the file for the purpose of producing precise and accurate statistics.
- Imputed records and variables on the file are not identified as such.
- Addition of "noise" (perturbation). Numeric values may have been raised or reduced by unequal amounts and proportions in a random-like fashion (addition of "noise"), while maintaining data integrity for the purpose of producing precise and accurate statistics.
See the appropriate section in Survey of Labour and Income Dynamics (SLID) - A survey overview
See the appropriate section in Survey of Labour and Income Dynamics (SLID) - A survey overview
If you have any questions or comments about the data in this CD-ROM product, you can contact the Income Statistics Division.
Telephone: 1-888-297-7355 or 613-951-7355
Facsimile Number: 613-951-3012
Income Statistics Division
150 Tunney's Pasture Driveway