I think the answer is complex. Firstly, there is the universal and unavoidable human tendency to err. This seems to be largely random, which is another way of saying that its causes are unknown. However, there are factors that make random errors more likely, such as distraction, hurry, alcohol, sleepiness, emotions, stress or a full bladder.
Making errors also has a systematic component. If our understanding of the position is faulty or we do things the hard way, eg neglect development or block our own pieces, then errors will result. If my focus is too narrow, say only on pieces, without regard to pawn structure, or I ignore endgame considerations, then this too will create errors. Habitual but unjustified preferences, such as for Knight over Bishop, or for attack at all costs, are another source of error. It is essential to be adept at both attack and defence. A one-sided approach will make for patchy results. We all operate with various blindspots and biases in life, and chess is no different.
I avoid certain positions because of past failures, but in some cases my judgement may be invalid. Conversely, I may erroneously think that certain positions are favourable because of a past success, which was actually due to opponent error. Learning to play well means acquiring a large store of chess patterns, ie generalised positions, together with the knowledge of whether they are favourable or not. The trick is to recognise when one of these patterns appears in the current game, but the difficulty is that a single detail may invalidate the conclusion. The simplest example is Queen and King vs King. This is always a win - except if you stumble into stalemate.
Each of us builds up an understanding of what chess is about. This amounts to a simplified mental representation, one that can never capture the full complexity of the game. While this mental picture tends to improve with time and practice, we also acquire ideas that are invalid, or only partially valid, eg that being a piece up is always a win - bishop and king vs king is a draw. Once entrenched, these notions are very hard to unlearn. It is entirely possible to insist on playing the same invalid line in the Ruy Lopez, hoping that one day we will find the way out of the difficulties it saddles us with. A better resolution is to learn the book move and why it is superior. Nearly all of us have at some time had the feeling "I've gotten into a similar mess before, why again?!"
Furthermore, analysis is not a precise art. No matter how much we analyse a position, unless there is a fairly simple forced win, we cannot be sure of our conclusions. The exponential nature of chess possibilities means that we have to limit our investigation to moves that seem promising. The problem is our limited skill in knowing which moves need investigation.
Finally, for psychological reasons, it is difficult to think as actively and creatively on behalf of the opponent as for our own side. Mechanical analysis may work well for a computer, but it cannot do the trick for humans. Hence we miss some of our opponent's cunning moves.
In conclusion, careful and thorough analysis can only partially protect us from these and other sources of error. Here is my strategy to make fewer blunders:
a) Never hurry a move; always reconsider before hitting Submit. Analyse at leisure and repent in haste. Reconsider even if my move appears forced - there just might be another alternative.
b) Don't just make the move I planned two moves ago - do the analysis again. Being in an actual situation, rather than a hypothetical one, focuses the mind. Also, it never hurts to check my analysis again.
c) Don't make a move when I am sleepy, mentally tired or distracted. Chess requires close and undivided concentration.
d) Avoiding systematic errors is no easy feat. It requires flexibility and a willingness to learn new approaches to chess. Analysis of my lost games does help.
e) Recheck the main line, especially right at the beginning, in case I missed a better move for my opponent. This is crucial. Check for possible sacrifices!
f) Check what candidate moves I missed for myself, and redo the analysis if needed. Look wide, then look deep is a good prescription.
g) Do a last check for my blindspots - pins, sacrifices, bishop skewers, pawn forks, knight forks.
h) Apply a healthy dose of prophylaxis, ie plan actively on behalf of my opponent in response to my candidate move. What are they trying to do?
This checklist will not eliminate all blunders - nothing can, apart from illegal computer assistance. My pragmatic aim is to make fewer mistakes than my opponent. Since in nearly all games I have played victory resulted from a blunder, this approach seems to make sense.