Part 4: Use of Statistics in Assessing Systemic Discriminatory Effects
Four: The Need to Collect Data Indicating Status in a Protected Group
pool and Dan: for the moment, you are both right - we cannot yet operate blind to one's group membership. In fact, since the playing field is not yet level, operating blindly will result in adverse impact and de facto discrimination.
I do not, at this time, advocate or support the adoption of any policy that will eliminate collecting data that reflects an individual's membership in protected group. In fact, I believe that no one should be allowed to opt out of such data collection (which is currently allowed) since this will render the resulting tally inaccurate and, therefore, interfere with our ability to accurately determine impact, set goals, or measure social change over time. (As it happens, I was a consultant to the committee responsible for introducing questions into the 1994 census designed to collect this data from all residents of Canada - adopted specifically to ensure that these essential base population numbers were available to set goals and measure institutional compliance with the Federal Employment Equity Act and its two associated components: the Legislated Employment Equity Program - LEEP, and the Federal Contractors Program - FCP.)
Further, these numbers are essential to the statistical control of the impact of group membership in the mathematical analysis of data relevant to an examination of allegations of systemic discrimination or other Human Rights violations that are based on differential group outcome. The only way to determine that protected group membership has no influence on outcome, is to first account for (statistically eliminate) its effects, and then examine the remainder for similarily across groups.[/b] This is a specific case where sameness does imply equality, because group effects and individual differences are first removed.
Finally, it is noteworthy that the “colour-blind” approach is usually suggested/advocated by the right wing in backlash to advancements in Human Rights legislation as an effective tool in eliminating historical disadvantage and establishing equity for all members of protected groups. It is forwarded to sabotage this necessary analysis – with the desired effect to render the mathematical analysis and tracking of the impact of systemically discriminatory activities impossible.
For example, those who suggest that race not be measured know that without this base-line census figure the statistical analysis forwarded by theories of distributive justice cannot be conducted: the population measure, required to determine proportional representation of protected group members in a context, and essential as a “known variable” (as opposed to an “unknown variable” or the thing that is to be determined through solving the equation) in the mathematical formula that allows comparison to outcome, will be unavailable. This (is likely) why non-collection is suggested.
The proposal to eliminate data collection is commonly heard where racial profiling is called into question as the basis for (alleged) disproportional enforcement in policing and security and severity of punishment/criminal sentencing as a function of membership in a protected group defined by visible minority status (i.e., being non-white). Without knowing the racial or visible minority status of individuals, the donsrable adverse impact (if any) of profiling on these measures as a function of race, is impossible.
Four: The Need to Collect Data Indicating Status in a Protected Group
pool and Dan: for the moment, you are both right - we cannot yet operate blind to one's group membership. In fact, since the playing field is not yet level, operating blindly will result in adverse impact and de facto discrimination.
I do not, at this time, advocate or support the adoption of any policy that will eliminate collecting data that reflects an individual's membership in protected group. In fact, I believe that no one should be allowed to opt out of such data collection (which is currently allowed) since this will render the resulting tally inaccurate and, therefore, interfere with our ability to accurately determine impact, set goals, or measure social change over time. (As it happens, I was a consultant to the committee responsible for introducing questions into the 1994 census designed to collect this data from all residents of Canada - adopted specifically to ensure that these essential base population numbers were available to set goals and measure institutional compliance with the Federal Employment Equity Act and its two associated components: the Legislated Employment Equity Program - LEEP, and the Federal Contractors Program - FCP.)
Further, these numbers are essential to the statistical control of the impact of group membership in the mathematical analysis of data relevant to an examination of allegations of systemic discrimination or other Human Rights violations that are based on differential group outcome. The only way to determine that protected group membership has no influence on outcome, is to first account for (statistically eliminate) its effects, and then examine the remainder for similarily across groups.[/b] This is a specific case where sameness does imply equality, because group effects and individual differences are first removed.
Finally, it is noteworthy that the “colour-blind” approach is usually suggested/advocated by the right wing in backlash to advancements in Human Rights legislation as an effective tool in eliminating historical disadvantage and establishing equity for all members of protected groups. It is forwarded to sabotage this necessary analysis – with the desired effect to render the mathematical analysis and tracking of the impact of systemically discriminatory activities impossible.
For example, those who suggest that race not be measured know that without this base-line census figure the statistical analysis forwarded by theories of distributive justice cannot be conducted: the population measure, required to determine proportional representation of protected group members in a context, and essential as a “known variable” (as opposed to an “unknown variable” or the thing that is to be determined through solving the equation) in the mathematical formula that allows comparison to outcome, will be unavailable. This (is likely) why non-collection is suggested.
The proposal to eliminate data collection is commonly heard where racial profiling is called into question as the basis for (alleged) disproportional enforcement in policing and security and severity of punishment/criminal sentencing as a function of membership in a protected group defined by visible minority status (i.e., being non-white). Without knowing the racial or visible minority status of individuals, the donsrable adverse impact (if any) of profiling on these measures as a function of race, is impossible.
Last edited:






