datadub

Blog of an academic data librarian.

[datadub]

The 1%, Census Metropolitan Areas, and our Official Languages

Inspiration for seeking out evidence-based answers to questions can come from almost anywhere. Today I saw a statement written as part of an argument on the wall of a bathroom stall. It was something to this effect:

"Why Occupy? Learn French! Join us 1%ers in Ottawa and stick it to Canada's dumb anglos."

This is referring to the Occupy Movement, whose slogan was "We are the 99%" in reference to the inequality of wealth in the United States, and all over the world. This led me to some questions: Who are the 1% in Canada? Is Ottawa really the richest city? Is the 1% clustered somewhere geographically? Where does language and bilingualism factor into all this?

To address these questions, I hit the data. First I'll mention an interesting blog entry from the Canadian Centre for Policy Alternatives' behindthenumbers.ca. Using tax information from 2009, they indicate that, in terms of income levels, the 1% in Canada would include anyone with an income above $366,717 (Lee, 2011). There is more discussion of income and wealth inequality in Canada and the U.S., but no significant reference to the geographic aspects or to language. 

Next I turned to the Daily. This article shows that Ottawa-Gatineau did in fact have the highest median total family income in Canada in 2009, followed by Calgary and Edmonton. So people generally have higher incomes in Ottawa than other CMAs, but this tells us nothing about the 1%.

To get a closer look, I went into the Census Public Use Microdata Files (PUMFs) to do my own analysis. Here I found that the 1% income figure was lower. According to the PUMF, the 1% includes people with incomes above around $195,000. With a little Excel manipulation, I found that about 1.2% of Ottawa residents would qualify. This is high compared to places like Regina (0.57%) or Winnipeg (0.75%), but not as high a concentration as in places like Toronto (1.5%), or Calgary (2.2%). 

Running a similar analysis, this time against knowledge of official languages, I found that bilingual individuals make up a greater proportion of the 1% than they do the Canadian population in general. So it seems being bilingual does increase the likelihood someone would be a "1%er", as stated on the bathroom wall. Incidentally, speaking only French had the opposite effect, and as we might expect, speaking neither official language almost completely guarantees you are not part of this group. 

So, whoever it was that chose to share their view of this issue on the bathroom wall may have been a little misguided. While becoming bilingual could potentially increase your likelihood of being in the 1%, the data show benefits are far greater for French speakers to learn English than the reverse. Also, moving to Ottawa might increase your income, but if your goal is to join the 1%, you might be better to consider moving to Calgary.

References

Statistics Canada. (2011) "Family Income and individuals income, related variables: Sub-provincial data." The Daily. June 28. http://www.statcan.gc.ca/daily-quotidien/110628/dq110628b-eng.htm

Lee, Marc (2011). "Wealth and income in the top 1%." Behind the Numbers. http://www.behindthenumbers.ca/2011/10/20/wealth-and-income-in-the-top-1/

[datadub]

Strangers Can't be Trusted

There's an interesting collection of variables involving people's level of trust in the General Social Survey, Cycle 23, on victimization. I was keen to see if people trust strangers, and if there are any income effects. What I found was that generally people are not trustful of strangers. On a scale from 1 (cannot be trusted at all) to 5 (can be trusted a lot), 56.2% of Canadians ranked their trust at 1 or 2. Interestingly, the respondent's level of trust seems to rise with income. 

I realized that perhaps being recently victimized would impact how people view strangers, so I decided to control for the number of victimization incidents in the past year. The income effect persisted, but was slightly dampened by the effects of victimization. Distrust of strangers rose with each victimization incident, as expected.

Source: General Social Survey, Cycle 23, 2009 [Canada]: Victimization, Main File.

As an aside, I also discovered that over 40% of people check the back seat of their car for intruders before getting in. Yikes.

[world-shaker-deactivated2013092]

“Why do people suck at Trivial Pursuit?”

[datadub]

Probably not based on real data, but entertaining.

[datadub]

Reasons for Using a Cheque Cashing Service

On the topic of cheque cashing services, I found an interesting survey through ICPSR that addresses the rationale and behaviour of people who do not have bank accounts, or who do not use banks. One variable in the study related to the reasons why people would use a cheque cashing service instead of a legitimate bank. Here is the chart I created using a combination or SPSS and Excel.

Source: United States Department of Commerce. Bureau of the Census, United States Department of Labor. Bureau of Labor Statistics, and Federal Deposit Insurance Corporation (FDIC). Current Population Survey, January 2009: Unbanked/Underbanked Supplement [Computer file]. ICPSR29649-v1. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2011-03-08. doi:10.3886/ICPSR29649.v1

Slightly more than half of the respondents who chose to use a cheque cashing service did so because the service was more convenient than a bank. The other half of respondents were split between a number of other reasons. This further reinforces the point that the absence of banks will serve to aggravate the issue of "fringe" banking behaviour, such as cheque cashing services, in the North End.

[datadub]

Cheque Cashing Services in the North End (cont'd)

Incidentally, the the researcher has a 2005 study published by the Canadian Centre for Policy Alternatives on the issue. I'm considering looking into including the research data they gathered in our institutional repository, to be repurposed by other researchers.

[datadub]

Cheque Cashing Services in the North End

An inquiry from a faculty member has prompted me to explore the Canadian Financial Capability Survey from 2008. This survey examines Canadians' knowledge and behaviour concerning financial decisions.

This made me think of a news story I watched recently that discussed the widespread closures of financial institutions in Winnipeg's North End neighbourhood over that past 15 years. The implication was that the lack of availability of legitimate financial institutions forces many to use various cheque cashing and pay-day loan services. 

The North End is a neighbourhood with a large Aboriginal population (29% overall, with some areas as a high as 80%), and a high instance of low incomes, with many families depending on social assistance (Source: City of Winnipeg Community Profiles 2006). Although the CFCS doesn't give us a very detailed geography, ( I would love, for instance, to see exactly what this survey reveals about the North End), we can control for income levels and Aboriginal identity. 

In analyzing the survey, we see that 8.9% of Aboriginal respondents had used a cheque cashing service at least once in the past year. 4.1% had used this service more than three times. These figures compare to the 1.7% of non-Aboriginal respondents who had used the service. 6.6% of the Aboriginal respondents had used a payday loan one or more times. 

In terms of income, those with less than $20,000 in annual income were the most likely to have used a cheque cashing service (3.1%), and more than twice as likely as those with an income between $40,000 and $60,000 (1.4%). Of the respondents who used a cheque cashing service in the past year, 20% were receiving social assistance or disability payments. 

One of the things this survey reveals is that economically disadvantaged populations are more likely to use these "fringe" banking services. The disappearance of legitimate banking options from underprivileged neighbourhoods will only serve to exaggerate this problem. 

[datadub]

There are some really intriguing visualizations in here. Great inspiration for creating unique maps.

[datadub]

ACCOLEDS 2011 - Day Three

Day three was the last day of this year's training. We covered a lot of very useful stuff over the three days. Here are today's highlights:

- We started with a session on aboriginal data and statistics. The breadth of information available is a little overwhelming. I've spent some time looking at these types of data queries (aboriginal and indigenous studies are important at my institution) and yet still there are many sources I haven't used.

- Anna Bombak presented some of the unique data sets available on the University of Alberta Mirror Site.

- The next presenter was Valerie from vital statistics. She explained that the purpose of vital statistics is to examine these four major life events: hatch, hitch, ditch, and dispatch. That is, vital statistics looks at birth, marriage, divorce, and death. There ware loads interesting tables, particularly with reference to cause of death.

- Last, we were given a presentation of ICPSR (Inter-University Consortium of Political and Social Research). Historically, ACCOLEDS is linked to the ICPSR data, as these were the only data sets available when the concept of Data Librarianship first emerged. Since then, Canada's data has been largely liberated. I'll take a closer look at the ICPSR stuff in the near future, and I'll certainly consider taking the trip to their training, whenever it's available.

So another Data Librarian training session has come and gone. I've been pretty interested in the idea of mapping social data for a while now and I think I have the tools to move forward on that. The ACCOLEDS folks are amazing. They're really supportive and each has a unique skill-set. Another great year. Special thanks to Kwantlen Polytechnic University for hosting.

[datadub]

ACCOLEDS 2011 - Day Two

Today was largely focused on the PCCF, which relates to some of yesterday's activities using postal codes. There was also the AGM and an afternoon session on non-DLI data. Here are the highlights: - Russell Wilkins described how reserchers will often want to merge data they've gathered with census information by dissemination area (DA). This can be done using postal codes, as we witnessed yesterday. If the data are 'clean', this works like a charm. Unfortunately, some of the postal codes gathered by researchers may be problematic. Imagine postal codes for prisons, hospitals, universities, etc., that would skew the aggegates. This is really just the tip of the iceberg. We used SAS and PCCF+ in the lab to identify potentially problemtic postal codes in a data set. - Mary, from UBC, showed us some reference questions she addressed using data outside the DLI. Most interesting was looking at the Canadian Institute for Health Information (CIHI) website. Here we found a range of administrative data for Canadian health. It seems to me CIHI serves as the administrtive counterpart to the CCHS. - We did a walking tour of Steveston, an old fishing village here in Richmond. Some interesting history and development. Dinner at Steveston Seafood was fantastic. More sessions tomorrow. I'll provide updates whenever I can.

[datadub]

ACCOLEDS 2011 - Day One

Today was the first day of the ACCOLEDS training in Richmond, B.C. Here are some of the highlights: - The opening welcome from Jane Fee, Associate VP Academic of Kwantlen Polytechnic, was brief but interesting. Most notably, she told an anecdote of a women who took the only living copy of a records management system with her into retirement, causing the organization much distress. This is a good illustration of the power of controlling information. - The DLI update was encouraging. CANSIM will be free come February. E-Stat will be deactivated by July, forcing some to retool access and instruction. - A GIS lab showed us how to plot researcher data on a map using Postal Code information. As a bonus, we were also shown how to create a coloured map of average income by dissemination area. This will be extremely useful. - David Wrate gave a detailed presentation of Data B.C. - We finished the day with another lab, this time opening a .dat file in SPSS, effectively creating a working syntax file in the process. This is an absolutely fundamental skill for data librarians, so I'm really glad it was covered in detail. - The reception was informal, with some decent local wine. I'm going to try to post my notes each day, if I remember.

[datadub]

Sunscreen Exposure Survey

I got playing around with an old survey from 1996 about sun exposure. I was using a tool called Nestar which allows me (and only me at my institution) to access the DLI data and even do some analysis. I usually use SDA, but I thought I'd try something new. 

The sun exposure survey covers a wide range of issues related to exposure to and protection against the sun. Here's what I discovered:

Men are (or were in 1996) much more likely to NEVER wear sunscreen, on both their faces (48.1%) and their bodies (46.3%). This compared to women, 23.4 and 24.2% of whom don't ever wear sunscreen on their face and bodies, respectively.

The prairie region boasts the highest level of sunscreen non-use (41.9%). Ontarians were the most likely to always wear sunscreen on their bodies (19.2%).

Nearly 60 percent of people (59.2%) who earn less than $10k a year will never put on sunscreen. The likelihood that someone would wear sunscreen on their body seems to increase with their income (see chart), so that at $80k+ people are slightly more likely  to wear sunscreen all the time than to never wear it at all.

[datadub]

This is a pretty neat tool. You can collect and visualize any data you like. There are literally millions of uses for this app. It's free, but to enjoy privacy of data you need to subscribe to DAYTUM+.

[datadub]

This is a really neat tool. I've been pondering this idea for a while now. Good to see this type of thing moving forward.

[datadub]

SPSS to Excel

There was a question this morning about opening raw data files in Microsoft Excel. Most of the microdata available through Statistics Canada can be opened in SPSS by use of a syntax file. Even this process has given me headaches in the past. It just seems very inconsistent. I discovered this morning that rather than it being my lack of experience with the software, it's my workstation that is effectively allergic to SPSS. It just doesn't read the file path correctly, regardless of what I try. This might be a Mac issue.

When I transported the data to an older PC, the data opened seamlessly. I saved it as .sav and brought it back. I was then able to view the data file without any problems. I'll need to sort this out eventually, but for the moment I'll just walk the data back and forth via USB key.

In terms of opening the data in Excel, the old PC had some older version which couldn't open all the variables. My Mac has Excel 2007, which managed to save and open the whole file.

One significant limitation was how the data were displayed in Excel. Rather than translating the numeric data into the corresponding text, only the numeric values were displayed. This means that were the researcher to use Excel, they would need to constantly consult the Codebook to match the values with their meanings. It seems to me this would be an unwelcome extra step.

For the time being, even with some of the difficulties I'm having, I'll keep working with SPSS. Excel just doesn't seem to be an effective replacement when working with large data sets.

[datadub]

This is actually a very sophisticated study. It incorporates aspects of humanities and social science / financial data analysis. The results are very interesting.

[datadub]

Privacy Examined

This article was brought to my attention by a colleague, which has opened a figurative can of worms. This relates only partially to library data services, but the proposed Bills (C-50, C-51, and C-52), have prompted me to take a closer look at the issue of privacy.

After learning a little about the Canadian legislation regarding privacy at LEGISInfo, I found some very interesting opinion articles, and even a CBC podcast about the bills. The main concerns are that legal authorities will be able to obtain access to personal data about their online activities, without a warrant, from Internet Service Providers (ISPs) and Telecommunications Providers.  This basically represents a slackening of regulations on how the government goes about getting information about suspected criminals, or even those not suspected of committing crimes. 

The most interesting thing I found today was an article from the Chronicle of Higher Education, outlining the importance of privacy. This article is written in response to common "I have nothing to hide..." argument. Very interesting.  

Privacy is rarely lost in one fell swoop. It is usually eroded over time, little bits dissolving almost imperceptibly until we finally begin to notice how much is gone. (Daniel J. Solove, CHE, May 15, 2011)

Incidentally, there is an online petition here.

[datadub]

Campus Researchers and the DLI

I recently had a query about accessing DLI data by a researcher here on campus. The Prairie Women's Health Centre for Excellence is a campus research group that has as its mission "improving the health status of Candaian Women by supporting policy-oriented, and community-based research and analysis on the social and other determinants of women's health." 

It was unclear at first whether the researchers qualified as authorized users of DLI data. I consulted the DLI License, the Licensing Flowchart, and examples of prior licensing questions and decisions. Because the PWHCE is on campus and the researchers are University of Winnipeg staff members, I decided they qualified as authorized users. They may publish or share the results of their research and analysis, however the raw data must not be shared with non-authorized researchers. 

This is the second or third time I've dealt with somewhat peculiar licensing issues. Not all requests for data are as cut and dry as student/faculty/staff access. I have yet to have any issue so strange as to not be explained using the three documents linked above. If these documents didn't shed enough light on the license to help me make a decision, I would have to turn to the DLI staff for guidance.