Section 1: Overview
The following amendments shall be applied to the Amendment to the Employee Wellness Surveys and Pulse Check Surveys PIA & Supplement to Statistics Canada's Generic Privacy Impact Assessment –Employment and Social Development Canada (ESDC) Employee Wellness Survey (EWS) PIA (signed by the Chief Statistician on September 26, 2022):
ESDC has identified that it requires access to the ESDC EWS microdata to perform more complex analysis than what was originally envisioned and beyond the capabilities of the Real Time Remote Access (RTRA)Footnote 1 ,Footnote 2 system (described in the ESDC EWS PIA). As such, Statistics Canada will share with ESDC a sub-set of the Master file (full de-identified microdata of all respondents), in the form of a microdata share file, i.e., an analytical dataset excluding direct identifiers that only includes the information of respondents who provided informed written consent to share their information with ESDC upon completing the survey. A Memorandum of Understanding concerning this sharing of information from the ESDC Employee Wellness Survey (MOU) under the authority of Section 12 of the Statistics Act, encompasses strict conditions that include, but are not limited to:
- ESDC agrees to keep the survey responses confidential and agrees not to match the information that is shared with any other records or data files for any reason, including to identify a person.
- ESDC will use the information for statistical and research purposes only.
- ESDC will take the necessary steps to protect the information in accordance with security requirements that meet or exceed Statistics Canada's for the storage, access (physical and remote) and destruction of shared information.
Additionally, ESDC will be adding a Class of worker variable to the respondent contact information file provided to Statistics Canada for integration into the final analytical results, including those provided in the share file to ESDC. Class of worker will indicate whether the employee is an executive, manager, or employee.
A privacy analysis of these modifications follows, intended to supplement information found in the original Amendment to the Employee Wellness Surveys and Pulse Check Surveys PIA & Supplement to Statistics Canada's Generic Privacy Impact Assessment – Employment and Social Development Canada (ESDC) Employee Wellness Survey (EWS) PIA. Note, only the applicable sections will have changes outlined. For all other sections, please see the original Amendment to the Employee Wellness Surveys and Pulse Check Surveys PIA & Supplement to Statistics Canada's Generic Privacy Impact Assessment – Employment and Social Development Canada (ESDC) Employee Wellness Survey (EWS) PIA.
Section 2: Risk Area Identification and Categorization
The risk area identification and categorization remain unchanged.
Section 3: Analysis of the Personal Information Elements for the Program or Activity
Changes include:
Necessity
The provision of a share file that will only include the information of respondents who consented to sharing their responses with ESDC, is required by ESDC to perform in-depth analysis of the survey results. The RTRA only allows for descriptive statistics (RTRA users can calculate frequencies, means, percentiles, percent distribution, proportions, ratios, and sharesFootnote 3) and does not support complex analysis, such as multivariate linear (i.e., a continuous outcome is predicted by a set of variables) or logistic (i.e., a binary outcome is predicted by a set of variables) regressions as well as path models (i.e., the combination of several multivariate regressions into a single model to test chains of associations involving several predictors and outcomes) that could help ESDC better understand relationships between psychological health and other key variables.
ESDC will benefit from the share file and addition of the Class of worker variable as it will allow them to perform the more complex analyses typical of this area of research into employee wellness (as described in the original amendment) and to maximize the effective use of the survey data, through understanding the survey results by class of worker, among other domains. ESDC employees are also expected to benefit from the outcomes of more in-depth analysis of the survey results which should allow for a better understanding of their work situation and inform ESDC's decision-making process in their efforts to better support employee well-being. For example, additional analyses performed by ESDC might identify other employee groups that are at risk of leaving the department due to unsatisfactory work conditions specific to their own area. As such, ESDC could make a concerted effort to improve the situation and well-being of these specific areas using tailored strategies. Ultimately, Canadians using ESDC's services will benefit by having a more effective and resilient public service.
Effectiveness - working assumptions
The share file will allow ESDC to perform more in-depth analysis than via RTRA, as the ability to perform more advanced statistical functions such as multivariate linear and logistic regression, factor analyses (i.e., whether the survey questions that were intended to measure a factor did so well enough across different subgroups of employees), path analysis (i.e., the combination of several multivariate regressions into a single model to test chains of associations involving several predictors and outcomes), dominance analysis (i.e., a technique used to establish an order of relative importance when using multiple predictors in a regression analysis), mediation (i.e., identification of intervening variables that carry the effects from the predictors to the outcomes) and moderation (i.e., identification of variables that impact the association between the predictors and the outcomes) analyses, or latent profile analysis (i.e., the classification of employees into distinct groups based on their similarities) should allow for more targeted analysis relating to the well-being of the employees. These more complex analyses may result in more nuanced insights that will enable the identification of key drivers to better delineate and understand the data, providing for more informed decisions, leading to more tailored strategies and effective interventions to improve employee well-being and ultimately a more effective workforce (e.g. lower absenteeism rates) that is better able to serve Canadians. The additional variable (class of workers) will allow for the identification of differences between executive, manager and employee groups, as there may be significant nuances that require distinct interventions.
Proportionality
In addition to Statistics Canada providing a summary of aggregated results and ESDC having the option to use the RTRA service to submit queries to build basic data tables as described in the original PIA, the provision of a share file will allow for a wider range of analysis and generate further value from the survey responses beyond what was pre-planned for dissemination, while continuing to respect the privacy and confidentiality of respondents by sharing only the de-identified data of those who provided informed consent to share their information with ESDC for statistical and research purposes only. These more insightful data analyses will enable ESDC to enact more effective policy changes that more proportionately impact specific groups or sub-groups to improve overall employee well-being across the department. For example, , it might be better to focus on minimizing work-life interference for some organizational units with ESDC, while for others, the reduction of their workload might be a more effective way of improving their well-being and psychological functioning at work, or perhaps it might be necessary to put more emphasis on clarifying the roles and expectations related to their positions instead.
As with other variables supplied by ESDC to create the frame of respondents, there is a proportional reduction in respondent burden by using additional ESDC administrative information to append the Class of worker variable to the share file, rather than conducting a new survey to collect the information directly. The inclusion of this variable is expected to be of benefit to understanding well-being at different levels of the worker hierarchy at ESDC. The additional variable represents a minimal and proportional intrusion when weighed against the benefits of being able to better consider the survey results disaggregated to this level for all segments of the workforce. ESDC has done analyses based on the class of workers in the past and found important differences between the executives, managers and employees with respect to their well-being. These findings mirror results from other employee (well-being) surveys that have also found differences between executives and non-executives in terms of their psychological health. Adding this variable to the data file will allow them to disaggregate the results by class of worker and allow them to tailor their interventions and policies to specific areas. Using this tailored process, ESDC may be in a better position to meet the specific needs of these groups, improve the overall well-being of its workforce, and be able to better serve Canadians.
Alternatives:
Without a share file, the only alternatives to allow ESDC access to the microdata in a manner that still permits the calculation of quality indicators would be RTRA (discussed in the original ESDC EWS PIA). However, this method would not support the more complex analyses typical for these kinds of wellness surveys. Providing ESDC with access to the Master file in the RDCs is also not an option, because some respondents did not consent to share their information with ESDC; a custom file containing only data from respondents that consented to sharing their data with ESDC must therefor be created.
The only alternative to appending the Class of worker variable to the respondent contact information file would be a direct collection of this information from respondents, which could be overly burdensome due to response fatigue and risk sacrificing statistical quality and accuracy.
Section 7: Summary of Analysis and Recommendations
The summary of analysis and recommendations remains unchanged.
Section 8: Breach protocol
The breach protocol remains unchanged.
Section 9: Supplementary Documents List
The supplementary documents list remains unchanged.