I have a graph showing bias and sensitivity of a psychometric curve (specifically, as logistic function of the form $\frac{1}{1+\exp(k(x-x_0))}$ , with $k$ = sensitivity and $x_0$ = bias). I am expressing my $x$ as a percentage of a reference stimulus, and therefore my normalized bias ($norm. bias = \frac{bias}{reference} *100$ ) has units of percentage ([%]). What units does my normalized sensitivity have? ($norm. sensitivity = \frac{sensitivity * reference}{100}$). I know they are techinically both unitless, but I am unsure whether in a scientific graph I should specify that the sensitivity value was multiplied by $100$ - something like $\frac{1}{\%} $ or $\%^{-1}$ .
Thanks!
The direct inverse of hundredths (percent when multiplied by a quantity) is hundreds. But I would probably care more, about using
\cdotfor multiplication, and labelling with function names, or formulae instead. That way, people can read it regardless of units. Also, using delimiters to properly show math, might be good. But percentages are relative. Markup percentages, are not the same as margin percentages. At least not when inverting things. Anyways have fun.How to get an accurate result in the following problem?