Gay Suicide ‘Epidemic’: Trope on a Rope

On August 31, the left media broke into applause at news of support for same-sex marriage from an unexpected quarter. At their annual general conference over the preceding weekend, NSW Young Nationals voted in favour of ‘marriage equality’ and followed up with a press statement circulated widely to the media. Fairfax journalist Phil Coorey, amongst others, posted it on Twitter.

Claiming the mantle of ‘first youth division of a conservative political party to [endorse same-sex marriage] in Australia’, the Young Nationals go on to assert (emphasis added) that ‘a young homosexual man living in regional NSW is six times more likely to commit suicide than members of the community as a whole’. This is footnoted to a 2003 paper on ‘rural suicide and same-sex attracted youth’ by K. Quinn of Deakin University.

Although Quinn’s paper is readily available online, the Young Nationals don’t appear to have read it. Contrary to the press statement, Quinn concludes after reviewing the available literature that any link between rural living and higher rates of suicide for SSAY (same-sex attracted youth) is only conjectural. There is no evidence to support it, as the paper’s concluding section makes clear:

Although same-sex attraction is not in itself a cause for rural youth suicide it may be a contributing factor; significant evidence is not yet available to support this hypothesis.

On page 2, the paper does say ‘it is suggested that SSAY are six times more likely to commit suicide than the population as a whole’, but that refers to SSAY generally, not just those living in rural areas. In any event, it is footnoted to an obscure reference which turns out to be a 2002 press release quoting Labor politician John Thwaites, then Victorian Health Minister. Thwaites’ comment that SSAY “are six times more likely to try to take their own life than for the population as a whole” is not accompanied by any academic, statistical or other evidence.

The link between ‘six times more likely’ to commit suicide and SSAY appears to have murky origins.

A Fact Sheet on ‘Suicide in Rural Australia’ issued by the National Rural Health Alliance in May 2009, post-dating the Quinn paper, explains that:

At any given time, rates of suicide tend to increase with increasing rurality. The most recent AIHW [Australian Institute of Health and Welfare] Mortality Report shows that 15-24 year old males in regional areas are 1.5–1.8 times more likely to end their life by suicide than their urban counterparts. The incidence is up to six times higher in very remote areas.

Regional young males are generally more likely to commit suicide than urban young males. The Fact Sheet says young males are generally six times more likely to commit suicide in ‘very remote areas’, but that appears to be distinguishable from ‘regional areas’. The question is whether it’s substantially more likely in the case of same-sex attracted youth. AIHW Mortality Report data is concrete in the sense that it records actual suicides, while the likelihood of suicide amongst SSAY has been measured by specialized studies of suicide attempts (it’s difficult to classify actual suicides by sexual orientation, raising the problem of how to be sure an attempt was genuine). Hence the Fact Sheet has a section on ‘Same Sex’ which doesn’t present any statistical quantification.

Accordingly, the ABS collects statistics on actual suicides broken down by factors like age and gender, but not sexual orientation.

Despite these difficulties, the claim that SSAY are six times more likely to commit suicide continues to have currency.

A 2013 Briefing Paper from the National LGBTI Health Alliance, for instance, says on page 3 that suicide ‘rates are 6x higher for same-sex attracted young people (20-42% cf. 7-13%)’, which is footnoted to a 2003 report on SSAY suicide data by Sue Dyson and others at La Trobe University. But that report casts doubt on the percentage figures, explaining that ‘[US] studies that have consistently found the rate of attempted suicide among gay and lesbian young people to be in the range of 20 to 42 per cent have also been criticized (page 16)’. However it does contain the following passage (page 11):

Research in Australia has identified that same-sex attracted young people may be up to six times more likely to commit suicide than the population in general, with those in rural areas being particularly at risk … (Nicholas 1998)

The referenced 1998 study was authored by Jonathan Nicholas and John Howard, who consider their results broadly consistent with the results, though not the methods, of American research suggesting a range of 7 to 13 per cent and 20 to 42 per cent incidence of suicide attempts for young heterosexuals and homosexuals respectively. Their formal finding appears in this opening abstract:

Previous research of gay youth has estimated the rate of gay attempted suicide at between 20% and 42%. This research has been criticized as being mostly uncontrolled and biased. The study reported here, however, uses a matched sample of 57 gay and 54 straight males [in Greater Sydney] and finds gay youth are 3.7 times more likely to attempt suicide.

Ultimately the ‘six times more likely’ claim has its origins in the old pre-1998 American studies which received qualified support from Nicholas and Howard. Ever since, writers have conveniently and arbitrarily chosen the low-end of the heterosexual range, of 7 per cent, and the high-end of the homosexual range, of 42 per cent, to come up with ‘six times more likely’ (42 divided by 7 equals 6).

Given that the figures are ranges, however, it would be just as valid to choose any other combination of numbers. If the high-end of the heterosexual range of 13 per cent was chosen with the low-end of the homosexual range of 20 per cent, the result is only 1.5 per cent more likely. This is lower than the estimate of 1.8 per cent more likely for regional youth generally in the AIHW Mortality Report.

So a search for substantiation of the claim that ‘a young homosexual man living in regional NSW is six times more likely to commit suicide than the community as a whole’ leads to a 17-year-old study, conducted on a small sample of subjects living in an urban environment, counting attempted rather than actual suicides, producing a lower estimate than six times, and which may or may not be relevant to social conditions in 2015.