Of course we can directly test it!
And I don't mean the mechanisms (natural selection, genetic drift, gene flow, mutations) for microevolution ... these are obviously testable. The domestic breeding of animals and crops that is the backbone of the invention of agriculture, would not be possible if differential reproduction didn't create heritable change.
I also mean that what creationists (inaccurately) call "macroevolution", including the concept of common ancestry between living species, is also "directly tested" every day.
To see this, we need to understand what "test" means in science.
A "test" in science is a *predicted observation*.
An observation does NOT have to be in the form of an "experiment". For example, the entire science of Astronomy is not based on "experiment", but purely on *observation*.
The only requirement is that the predicted observation be in the form "if this prediction is not verified, then this disproves the theory."
I'll give an example from physics, and then an equivalent one for evolution.
When Einstein posited the theory of relativity, it *predicted* that light would be affected by gravity. The *test* of this prediction was to carefully observe a distant star just to see if its position appeared to "shift" as its light passed within the large gravitational field of the sun. We had to wait for a solar eclipse to do the observation. If there is no shift, then the prediction fails, and the theory fails its test.
But the predicted observation was verified ... so that aspect of the theory of relativity was *tested* ... and passed.
With evolution, one of the *predictions* of the claim that humans and apes share a common ancestor, is that there would be signs of chromosome fusion in human DNA to explain why the other apes have one more pair of chromosomes than we do. The *test* of this prediction was to look for tell-tale signs of chromosomal fusion in our chromosomes ... such as telomeres (DNA code normally at the tips of chromosomes) in the middle of a chromosome. If there are no signs of fusion, then the prediction fails, and the theory fails its test.
But the predicted observation was verified (we found markers of fusion in Human chromosome #2) ... so that aspect of the theory of evolution was *tested* ... and passed. [1]
For more, consider fossils. Evolution *predicts* that certain fossils representing certain developments, would appear in certain layers, and certain geographical locations. This is precisely how fossil-experts KNOW WHERE TO LOOK for these fossils. This can range from knowing there won't be trilobite fossils in the same layer as a T. rex, to knowing were to look for the first fishes transitioning to land, or feathered dinosaurs, or certain hominids with certain bipedal structure and brain capacity. The discoveries of tiktaalik, archaeopteryx, D. masillae, A. afarensis, and hundreds of others were based on these *predictions*.
So EVERY SINGLE NEW FOSSIL DISCOVERY is a *test* of evolutionary theory! One trilobite in the wrong layer ... one hominid in a Devonian layer ... one dinosaur in a Cambrian layer ... and evolution *FAILS THE TEST*.
So of course evolution can be directly tested! It is tested every single day!
---- {edit: Response @Questioner} ----
First, notice that he does not actually refute my Chromosome #2 or fossil examples ... he just dismisses it a "sanitized fan-boy story". But then IN THE VERY SAME SENTENCE accuses others of just "sweeping things under the rug".
[Edit: He later added a "rebuttal", saying "All that it shows (at most) is humans in the past had 24 pairs." ... as did the apes, and still do! That's a "rebuttal"? But more importantly, do you think he noticed that he used the phrase "in the past" in a conclusion we might draw from evidence? Oops.]
But second, and most amusingly, he seems to be answering both yes and no to your **question** ... is evolution "testable"? At first he seems to be arguing that evolution is not testable ("Any argument or idea that makes claims about the unrepeatable past requires belief.") ...
... but then immediately after arguing that evolution is not testable, he gives a long list of "tests" that evolution has supposedly failed!
The fact that I can rebut every one of his "tests" (if my answer were not already too long) ... and in fact they have been rebutted over and over again [2] ... is not my point.
It's that his mere eagerness to engage in arguments about evidence in the first place, means that Questioner admits that evolution is "testable" after all!
{edit} He even links to a creationist web site that actually uses phrases like "predicted by" and "should be found if" in reference to evolution.
So YES, EVEN CREATIONISTS ACCEPT THAT EVOLUTION CAN BE TESTED. They are admitting this every time they make statements of the form "if evolution is true, then we should see X."
---