Following a discussion with the mods at CogSci, we've migrated the question. They feel that it's appropriate for them. Given that the asker is pretty clear about wanting the psychological underpinnings of code readability, I agree with that assessment. Although psychology and other cognitive sciences and social sciences touch software development, we can't provide expert answers here in those areas. If the question was approached from a software engineering, software quality, or software metrics perspective (the kinds of things that would be published by the IEEE, ACM, or BCS), it would have stayed here.
I'd say that it is mostly on-topic and isn't a reference request. In this particular case, a good answer should cite reputable studies over experiences. However, I would suspect that it would be a synthesis of academic research and their findings rather than a collection of links.
Code readability falls under "software engineering", which is on-topic here, and determining methods for determining how readable code is seems to be a valid research topic. The question is if we have people who can adequately answer the question, especially from a psychological background. I can answer it from a software metrics perspective - there is some academic research that has tried to find things similar to the Flesch-Kincaid Grade Level or Gunning-Fog Index for software. There's also work to examine the layout of code (indentation, line-length, etc.) and determine automated methods of computing a readability metric that corresponds to human surveys of the same code.
However, I'm concerned about the desire for "psychological factors". We aren't psychologists here. I'm comfortable talking about the research in software quality and metrics, but not the psychological underpinnings of why these are what they are. That may be better suited to CogSci.