First off, let me say that I’m a blogging novice, so hopefully I won’t violate too many rules of social networking here! But when Graham Watts (who is Director of Resource & Exchange (REX), Training & Development) asked me to get involved in our team’s contribution to the new blog for our colleagues in the social housing sector, I couldn’t turn him down… turns out Graham is quite persuasive.

Before I go any further, I want to introduce myself to you. I’m Tim Leung, Research Analyst in REX from SHSC. I’ve been in this role for about six months, and prior to that, was a research assistant at the University of Toronto. Before joining SHSC, I held a role in statistical analysis in Charter Cases, and also picked up some interesting knowledge in shelter development in the Third World.

But enough about me… the real reason I’m blogging is to talk about my attendance at the XXV International Methodology Symposium in Gatineau, Quebec from October 26-30, organized by Statistics Canada. To be sure, I am no survey methodologist in the strictest sense (my statistical training was at the analytical end), but I am currently working on satisfaction surveys to evaluate services for the social housing sector (stay tuned).

One way we evaluate our services is through longitudinal data analysis. Why do we care so much about this kind of analysis even if the surveys we conduct do not involve the general public?

The answer?

It matters to us because social housing, as with all sectors, can suffer from “survey overload” and response rates can decline over time (this even happens with big-time statistical agencies like Statistics Canada). For instance, when people feel that their response to a survey produces no value, they are less likely to respond the next time. There are many problems associated with a decline in response rates. Perhaps, the most crucial problem, from an organizational view, is that a decrease in participation can actually increase the costs of the operating cost of conducting a survey (e.g. how many times do we have to call back respondents who did not respond initially or who left out key information?).

When response rates decline in longitudinal surveys about household spending, for instance, one strategy can entail selecting other respondents to be part of the sampling frame. It is harder for us to do this in the social housing sector because we have a smaller pool of respondents – when our “sector” respondents stop responding to surveys, we have nowhere else to go! Moreover, our respondents are pre-determined—they include housing providers and property managers in Ontario—and if we survey the entire group at once and receive a low response rate, there is no replacement!

There are, obviously, possible implications to how we approach surveys and the potential burden it has on busy respondents (e.g. housing providers, property managers). Many social housing providers can recall responding to lengthy surveys over the years as either government or others try to identify issues in the social housing sector and apply common solutions. We acknowledge the importance and the sincerity of these approaches, but at the same time we have to learn from past mistakes and try to create a system of information-gathering that builds on current systems and data to create an ongoing and manageable way of measuring the “state” of social housing. Stay tuned for more as SHSC moves forward with a data collection and inventory plan and evaluates the social housing sector in a “longitudinal” way!