Conclusion

The often cited number of 45% suicide rate for transgender women is not generalize-able beyond the sample used in the study by \cite{Haas}.     The statement that 45% of transgender women attempt suicide has to mean that, In a given year one either tries to kill themselves or they do not, or it is a meaningless statistic.  Saying that 45% of transwomen try to kill themselves at some point in their lives has to mean that in an average year that many try to kill themselves.  That is what is implied by the misuse of the findings of   \cite{Haas} by people like   \cite{Shapiro2016}.     A simple mathematical extrapolation for what that would do to the population over time shows it cannot be naively generalized to all transgender people.     To do so is like scooping a glass of water out of the ocean, finding no fish in it and concluding that the ocean is, therefore, devoid of life. 

The UCLA study was done well

We wish to be clear the study honestly reported the suicide attempt rate based on their survey.  For their survey and their sample it is valid.  However, the mathematics reinforce the following limitations of the study which are stated in the paper  \cite{Haas} .
The study was conducted as a survey distributed online and on paper.  It is based on self reports.  \cite{Haas} states that while they took precautions that surveys which ask people if they've attempted suicide will often inflate the rate through people reporting other types of self harm  as an attempt.  The survey also did not explore the mental health status of the people sampled. 
Lastly and most importantly the sample used was one of convenience not a selected and randomized study with proper control groups.  Therefore the people who answer the survey are those who are most motivated to answer the survey.  
Commentator Comments are Not Facts
The pure mathematics, facts of the strongest kind, show that given the actual population data, the second strongest kind of fact,  that the statistic based on a sample is not actually a fact outside of that sample.     
The above three facts need to be mentioned when anyone such as \cite{Shapiro2016} or others who wish to make an anti transgender rights case.  This UCLA study does not support their conclusion which is based on a simple over generalization of the findings of that study by commentators who are able to give the impression of intellectual rigor to the masses.