Statistical Politics

N.B. You may have to click on some figures to view them properly in your browser.

Yes, those are probably two of the most boring, dry words for some – but, for me, statistics and politics are fun!

Let’s take a look at a recent attempt I saw to consider the effectiveness of what’s now referred to in Australian political circles as the Pacific Solution.

Image

For my international audience, the context is essentially to do with what are often dubbed ‘boat people’ but are essentially just asylum seekers who make their way to Australia from nearby Indonesia. The journey is dangerous and the payments made to smugglers often fund criminal activity. However there is ongoing debate about the factors which influence people to brave the journey; are they ‘pushed’, i.e. by violence or discrimination in their country of origin, or are they ‘pulled’, i.e. drawn to Australia for economic reasons and/or influenced by domestic immigration policies.

The above graph attempts to draw a comparison between asylum seeker applications and the number of Irregular Maritime Arrivals (IMAs). However, as a scientist, I was immediately put off by the lack of statistical rigour, and with what I saw as clear outliers/trends which seemed better explained by policy changes.

Now, I will at this point state that I do not have a strong position on this issue currently, mainly because I have only just begun to delve into its complex history. However, I will say that humane treatment of our fellow human beings should always be our first priority.

So, motivated by my disappointment with the above graph, I went about collating and crunching the numbers myself. Someone had beaten me to it some years ago, but as you’ll see, I conclude something very different here.

As a good scientist, I started by formulating my hypotheses:
– null hypothesis: neither global asylum seeker application numbers nor domestic immigration policy changes in Australia influence IMAs;
– alternate hypothesis A: global asylum seeker application numbers affect IMAs;
– alternate hypothesis B: domestic immigration policy changes in Australia influence IMAs;
– alternate hypothesis C: a combination of global asylum seeker application numbers and domestic immigration policy changes in Australia influence IMAs.

I collected data from various official sources, including:
– the UNHCR’s Population Statistics Database and Statistical Yearbooks for asylum seeker application numbers;
– a Parliament of Australia’s Research Publication for IMA and boat arrival figures; and
– a comprehensive list of media and government reports on deaths of asylum seekers who attempted to sail to Australia.

As the first major immigration policy change was enacted by the Keating (Labor) government in 1992 (see Mandatory Detention), and because data becomes scarcer or less reliable the earlier one wishes to start from, I decided to begin my analysis at the year 1988. Due to some domestic data for the year 2012 being incomplete or forthcoming, I decided to finish my analysis at the year 2011. This gave a date-range of 23 years, from 1988 to 2011.

Major immigration policy changes in Australia were then noted and used to break up the date-range according to domestic policy. (N.B. If a policy was enacted between January 1 and June 30 in a given year, a break was made at the beginning of that calendar year, whereas if a policy was enacted between July 1 and December 31, the break was made at the beginning of the next calendar year.)

This resulted in the following date-ranges and their associated policy differences:
– 1988-1991: Pre-Mandatory Detention
– 1992-2001: Mandatory Detention
– 2002-2007: Pacific Solution
– 2008-2011: Suspension of Pacific Solution

Next, one-way ANOVAs with post-hoc Tukey’s honestly significant difference tests were conducted on the collected data with groupings as per the above date-ranges (Figure 1).

Figure 1

This revealed significant differences between the Pacific Solution and Pre-Mandatory Detention eras with the Suspension of Pacific Solution era, which had a significant increase in boat arrivals and IMAs. However, during the same eras, no significant differences were found between the numbers of asylum seeker application for 32 UN-selected developed countries (see Appendix I for a list of these countries, and refer to UNHCR’s Statistical Yearbooks for rationale and discussion).

But surely the number of asylum seekers are associated to some degree with boat arrivals or IMAs? Figure 2 claims not (at least for the overall date-range – both non-transformed and ranked).

Figure 2

Startled, I then looked at each policy era separately, conducting the same analysis (Figure 3).

Figure 3

However the relationships appears inconsistent and largely insignificant. And where significance was gained or approached, it was mild enough (esp. considering the context of the other date-ranges) that it was likely spurious.

As a point of interest, I also plotted the number of IMAs with the number of boat arrivals (Figure 4) to see if even greater overcrowding was occurring at any point. It appears not, fortunately, as the average number of 53 people per vessel (excluding crew) appears to remain fairly stable.

Figure 4

But, as the great tragedy of IMAs (in my opinion) is the loss of life which has and does occur when vessels capsize or sink, I also plotted ranked IMAs against the ranked number of deaths.

Rank - IMAs vs deaths

So, what do we make of all this? Honestly, I’m not sure.

But because I’m interested in the truth, I’m posting up the original data as a Google Spreadsheet. Check my numbers! Check the sources for yourselves! And, if you can, do the analyses!

A few of my initial impressions are, though:
– In criticism and to contrast my results with Possum Comitatus’ article: (a) no statistical test was used to determine the significance of the regressions whereas they are tested here, (b) the Possum article focused on transformed data, which loses a lot of detailed, absolute information (though I have included ranked-transformed data here for comparison), (c) the article falsely assumes Australia and New Zealand are completely comparable in terms of regional immigration (Australia is far easier to access by boat from Asia) and destination desirability (partly joking), and (d) I consider a larger data-range spanning more policy differences here.
– The UNHCR data is awful to sift through. I could well have taken the wrong numbers by accident somewhere – so, please, check!
– To repeat, I am in no way suggesting or implying the morality of any of these policies; I am simply attempting to address the above-mentioned hypotheses (speaking of which, I’m forced to hesitantly and tentatively conclude alternate hypothesis B until I get more time to think about the data or add to it).  I am personally yet to make up my mind what is the right course (and, like any rational thinker, I’ll likely change my mind in the future, slightly or fully in response to evidence or situation).

Appendix I: List of 32 developed countries as

Australia
Austria
Belgium
Bulgaria
Canada
Czech Republic
Denmark
Finland
France
Germany
Greece
Hungary
Iceland
Ireland
Italy
Japan
Liechtenstein
Luxembourg
Netherlands
New Zealand
Norway
Poland
Portugal
Romania
Slovakia
Slovenia
South Africa
Spain
Sweden
Switzerland
United Kingdom of Great Britain and Northern Ireland
United States of America

Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s