Is the unit symbol written twice when using the +\- symbol? When notating error using the $\pm$ symbol, are the units only ever included at the end? For example:

10.2  $\pm$ 3.2 m

rather than   

10.2 m $\pm$ 3.2 m

This seems to be correct though I wonder if that also applies to writing an angle in degrees.  In this case, it seems better to repeat it as in,

36.8$^\circ \pm$ 1.3$^\circ$

rather than 

36.8 $\pm$ 1.3$^\circ$

I am not sure where to find any kind of standard related to this.
 A: It appears that none of Physical Review Style and Notation Guide, the  AIP Style Manual, the IAU Style Manual, or The ACS Style Guide: A Manual for Authors and Editors, weigh in at all on this matter, so I would say it is to some extent up to personal taste. 
On the other hand, the NIST Manuscript Checklist does take a position:

$$\begin{array}{rl}
\text{proper:} 
& \text{123 g ± 2 g or (123 ± 2) g }\\
& \text{70 % ± 5 % or (70 ± 5) %  }\\
& \text{240 × (1 ± 10 %) V  }\\ \\
\text{improper:}
& \text{123 ± 2 g }\\
& \text{70 ± 5 % }\\
& \text{240 V ± 10 % (one cannot add 240 V and 10 %) }\\
\end{array}$$

Similarly, the IUPAC guide Quantities, units and symbols in physical chemistry explicitly recommends the form $(100.021\:47\pm0.000\:70)\:\mathrm g$.
Nevertheless, actual usage is very forgiving, and you end up seeing all forms to some extent. Some recent examples from the literature: 


*

*$6356 ± 8 \:\mathrm{keV}$

*$75.3a_0^3 ± 0.4a_0^3$

*$65±4\:\mathrm{fb}$

*$0.083±0.011(\mathrm{stat.})±0.007\mathrm{(syst.)\:ps}^{−1}$

*$(5.7 ± 0.6) × 10^{−6}\:\mathrm{cm^{−2}s^{−1}sr^{−1}}$

*$(89.8 ± 2.9)\:\mathrm{GeV}$

*$2\% ± 4\%$

*$ 200 ± 10\%$ (for a magnification)

*$178.4°±0.03°$

*$3.16(±0.10) × 10^{−2}\:\mathrm{cm^2W^{−1}min^{−1}}$
As you can see, pretty much anything goes as long as it is clear. A lotof people do tend to use $$10.2±3.2\:\mathrm m$$ and that is perfectly clear (even though slightly discouraged). However, if there is some ambiguity, such as with $36.8°± 1.3°$, it doesn't hurt to include the extra degree symbol, or to use parenthesis as in $(10.2±3.2)\:\mathrm m$. In general, I find repeated units to just make the text wordier without doing all that much to clarify any ambiguity.
The aim of scientific and technical writing is to communicate clearly and unambiguously. As long as it's immediately clear what you mean, you're fine, and if on top of it you're concise and readable then all the better.
