Thoughts from a reflective educator.
Every form of assessment of learning has bias. This bias may be hidden, or it may be quite obvious. As Cathy O'Neil points out, assessment is a proxy for what we want to measure - learning. We cannot measure the building of connections between neurons that is happening in the brain directly (or even potentially understand what that growth even means) so we use a proxy in the form of an assessment of the externally visible signs of learning.
The director of our senior school, Brad Smith, is doing a workshop today in our staff meeting on assessment for learning. He's found the following quotes, which he wants to use in his presentation. I'm helping him edit his presentation (since I have time, and he does not), and I'm hoping to find some photos or other prompts which describe these statements to include in his presentation.
See this piece of paper?

(Image credit: D Sharon Pruitt)
Throw it away.
Imagine the limitations of the piece of paper shown above do not influence how you share the record of learning your students have done, with their parents, and the wider community.

(Image credit: Dilbert comics)
I think this comic speaks for itself. How do your assessments fit into the big picture? Is this clear to your students? (I doubt many educators are giving students assessments that a monkey could be taught to do.)
Here a few experiments in assessment I'm considering for next year.
The objective of traditional grading is to compress information teachers have gathered about a student down into a single score to make understanding the information easier. One of the original reasons for this compression was the limitation on how much information could be shared on a single piece of paper.
The typical report card looks like this (click to embiggen):
This xkcd comic demonstrates a big problem with averages.

By the way, this same problem occurs when you average grades as well.
When I started working at my current school, I realized that they do something in assessment which is very rare.
I read this article by Alan Stange on assigning penalties to students who hand in work late. He makes the point at the end of his blog post, "There is in fact relatively little significance to learning to complete on time." I agree with this statement and I'm going to expand upon it.