The mistakes that were made on this flight were all a matter of pilot error.
1. Decision to penetrate thunderstorms.
2. Decision to land with winds above AA flight manual limits
3. Failure to arm spoilers
As with most mishaps (crashes), there is a chain of events, and if you break one link, the crash would have been avoided.
So, WHY did this crew make the mistakes? Well, the Captain died in the crash, so it's hard to know what he was thinking, it's all imputed from the CVR and the FO testimony. Now, I'm not picking on the FO when I say this, but he's alive and his job is on the line, so, what he says, and what he remembers may not be 100% accurate. For example, he remembers arming the spoilers, yet they clearly weren't, so, it's hard to say that his recollection is perfect.
The biggest issue in my opinion in fatigue. I'll get back to that in a bit, because the why is important, and the fatigue precluded sound judgement, objective analysis of a rapidly deteriorating situation at the Little Rock Airport.
Pilots suffer from a strong desire to complete the mission. It's part of our nature, particularly military-trained pilots. We WANT to complete the mission, we feel pressure, from ourselves, to get the flight to where it's supposed to go. Passengers, our company, everybody, expects us to get where we are supposed to be, on time.
I've seen this in myself and in others. When SWA ran off the runway in Burbank, we all looked at the numbers (crossing the threshold on a 5,800 foot runway at 70 knots above target speed with the flaps not in landing position) and thought, "how in the heck could those guys have thought this was OK?" but they were trying to complete the mission, despite the incremental deterioration of flight parameters and safety to the point of ludicrous exceedance of safe landing criteria.
I've not gotten to that extreme point, but there have been a couple of times when I landed and thought, "I made a mistake in the way we handled this, in the future, I will not do things that way, I will not allow my desire to "get there" let me exceed our standards for performance". Minor stuff in my case (e.g. I landed long, at 3,500' down the runway in IAH, on a very long runway. Our standard is to touchdown in the first 1/3 or 3,000', whichever is shorter, so, not a big deal, and we taxied safely to the gate, but in hindsight, I should have gone around and set up for another landing). That self-critique, and examination of decision-making, is critical to both safety and professional improvement. Minor stuff needs to be addressed and corrected, period, even if nothing bad happened.
The failure to examine minor deviations, and correct them, is a classic contributor to aviation mishaps. It's known as the "normalization of deviation". Look at the Space Shuttle. The o-rings shouldn't burn through. Period. The shouldn't. But after a mission in which they did, NASA kept flying the shuttle with o-rings that kept failing, until on one mission, Challenger, the o-ring burned though, caused an external fuel tank explosion, and killed everyone. Same thing happened a few years later with Columbia - ice chunks falling off the tank bracket. Kept happening. Engineers complained, but that deviation from safety parameters was "normalized" in NASA culture, until the ice chunks took out some tiles and the Columbia broke up on re-entry.
So, a proper safety culture requires, demands, that minor deviations be examined and corrected. On every flight.
Back to AA 1420.
They say you can boil a frog if you raise the temperature of the water in the pot one degree at a time, but the frog will immediately jump out of a pot of boiling water. Pilots are the same way, if the situation degrades one degree at a time, pilots will stick with that pot of water (situation) as it gets worse and worse by tiny increments. If we were presented with all of the factors at once (like being thrown into the boiling pot), we would jump out and abandon the mission, recognizing how stupid/risky/degraded things had become.
That's what happened here.
The AA Dispatcher told the crew that if they hurried up, they could arrive before the thunderstorms. Mission completion pressure applied, though not intentionally.
Then, the flight was routed around to the other side of the airport and the thunderstorms arrived before they did. But, they were already focused on that mission completion, so they pressed on. In the vicinity of thunderstorms is windshear. Always risky, often fatal. We take that seriously. But this crew was focused on getting into Little Rock.
The runway was wet in heavy rain. Not uncommon, but one more risk factor added, as braking on a wet runway is degraded.
The winds picked up. Another risk, and another degree in the pot temperature. By itself, no big deal, but added to the windshear potential and this situation was getting riskier by the minute.
The winds exceeded the AA flight manual limits when tower cleared them to land. That should have been a red flag - don't land above the wind limits, but the crew pressed on, one more degree in the pot, one more added risk. Or, perhaps they didn't actually hear what tower said, because they were focused on finding the runway in the driving rain. That's happened to lots of people, you hear what you expect, not what was said if what said was different than expectation.
In accomplishing the landing checklist, the spoilers weren't armed. That was the final error in the set that led to the crash. You simply have to have spoilers armed, so that when the plane touches down, the weight comes off the wing, and settles onto the wheels. Then the tires have friction on the runway (degraded by the wetness, but you still must have the tires pressed onto the runway, not skimming along with the weight of the airplane held up by the aerodynamic forces on the wings) so that the brakes can stop the airplane.
That was the final degree in the boiling pot. The moment that airplane touched down, without spoilers, on a wet runway, in gusty crosswinds that exceeded the limits, it was doomed. It was physically impossible to get it stopped.
Again - WHY?
The crew had been up for 16 hours. 16 hours is a long day to be at work in a highly technical, thinking job. They were tired, and said so on the voice recorder. When you're tired, you may not hear what tower said, you may not process the change in wind, or the missed spoilers, or any one of those mishap chain links that could've been broken to avoid the crash. They felt pressure to get there, imposed both externally (company) and internally (pilot desire to complete the mission), they failed to notice, or understand, the incremental degradation of the situation.
They were an experienced, trained crew from an airline with excellent safety and training programs and they just plain missed the clues that could've saved them. It's not easy to admit that you're too tired to be performing well enough to keep flying. It's extremely difficult to even recognize your own performance degradation from fatigue. People, especially pilots, just aren't' good at that. We have regulations that are supposed to provide limits to time spent on duty as a means to prevent fatigue-related cognitive performance degradation...but in 1999, those regulations allowed a crew of two pilots to be on duty for 16 hours without a break, if the crew felt they were OK. Again, the crew is probably the least-able to judge their own performance degradation.
I don't think we will ever know if this Captain had flown into thunderstorms before, or landed above crosswind limits before, and the "normalization of deviation" was a factor.
But I know for certainty that fatigue WAS a factor. That fatigue led to the degradation of his performance, cognitively, in the crew's failure to recognize the situation, and their failure to accomplish the most basic of checklists: the landing checklist, which includes arming spoilers.
Following this crash, the FAA embarked on a scientific study of sleep and the degradation in cognitive performance that comes with fatigue and long days in the cockpit. FAR 117 now regulates those parameters (duty day, fatigue, rest) much more rigorously.