Question
A francophone student has been asked to give a conference presentation on changes to urban aboriginal women’s housing (a fair bit of migration within the city). He wants to know if he should make the presentation or not, as he is worried about the validity of his results re: data quality and CT-level non-response rates.
As his research was at the CT level, I gave him two sources to help assess his data quality. Did I miss any technical documentation? The first source is a StatCan web page, and I compiled the second source from the NHS profile (for which there may already be a list, but I wasn’t able to find it today).
1) Statistique Canada (2013), ENM Liste des secteurs de recensement (SR) non diffusés.
http://www12.statcan.gc.ca/nhs-enm/2011/ref/sup_CT-SR-fra.cfm
2) Global Response Rate – Census Tracts – NHS 2011 / Taux global de réponse - Secteurs de recensement - ENM 2011
The attached table lists the non-response rates of non-suppressed Census Tracts from the National Household Survey, in other words, those with a non-response rate of less than 50 %.
Answer
I consulted the subject matter division regarding your questions, please find their responses below:
As far as information about Census Tracts and the 2011 NHS, you have provided all of the correct links, I would just add to that the Aboriginal Peoples Technical Reportand the NHS User Guide.
French version of the Technical Report: http://www12.statcan.gc.ca/nhs-enm/2011/ref/reports-rapports/ap-pa/index-fra.cfm
And the NHS User Guide: http://www12.statcan.gc.ca/nhs-enm/2011/ref/nhs-enm_guide/index-fra.cfm
In particular this note from the Aboriginal Peoples Technical Report would be of use:
Classement recoupé des variables relatives au logement
Les variables relatives au logement sont souvent croisées avec d'autres variables dans un tableau pour permettre l'analyse plus approfondie d'un sujet donné. Les utilisateurs de données doivent noter que les estimations seront susceptibles de présenter une plus grande variabilité attribuable à l'erreur d'échantillonnage lorsqu'ils examinent de petites populations, soit en sélectionnant des régions géographiques de petite taille ou en croisant plusieurs variables
