Logic and Facts

To establish the truth, there are 3 essential steps. First- is the argument logically valid? This can be broken down into the questions of- Are the premises required contradictory?
(note that a contradictory statement is not invalid- it is meaningless. If (contradiction) then Y is ALWAYS true. It is, as such, incapable of providing input. IE, a test that always comes up positive is no better than simply declaring the statement true without any measurement. IE in any designation of C->Y is equivalent to 0=0. This principle ignores scale- it's like multiplying a final product by 0, all information is lost if any link is invalid. It should be noted, that autorepair routines are a valuable asset, in that YOU can yourself, fix a target argument such that it becomes a valid argument. Even the best scientists often make mistakes especially in long and convoluted math. Sometimes these mistakes are easily fixed, and the argument does not need to be sent back to its source for maintenance.)

Are all statement transformations done such that they lead to equivalent results?
(IE, take the statements of X->Y and ~X (not X)->~Y. These are NOT equivalent. IE, assume X is true if the tested variable is a hawk, and Y is true if the variable is a bird. In all cases wherein X is true, so is Y. However, a dove is not a hawk, but it is a bird. Any argument wherein we start with the premise of X->Y and derive a statement such as ~X->~Y is logically invalid. This is, once again, due to the fact that given any derivative from which the original truth table is changed, I can obtain any result from any premise. As such, the truth of the premises has no relation to the truth of the final statement.)

Once these routines are complete the second step is- Is the statement mathematically valid?
This is when outside information is first factored into the issue. Each premise is measured against additional known information. The validity of the premises is based on how much outside information would have to be rendered false for the the premise to be declared true, vs how much outside information would have to be rendered false for the premise to be declared false. A strength level is assigned to each premise. Note the difference between a premise and a variable. IE "If cattle exist the so does the sun god." is a premise. The existence of cattle and the existence of the sun god are both variables. The RELATION is the premise. Since there are many scenarios that include cattle but not a sun god, I can call this premise into question. It should be noted, that it is impossible to render such a premise impossible (any such design would already be logically invalid.) but it is possible to assign probabilities to each premise. As some premises drop below a certain threshold I can furthermore declare the effort involved with further exploration to be not worth the benefits derived thereof, thus setting a virtual 0 to the premises strength.

Finally, the issue turns to the variables themselves. It should be noted that in many arguments, rendering one or more variables false will not destroy the argument. For instance- X or Y->Z. Even if Y is false Z (the variable I care most about) is still true, given that X is true. That said, it becomes necessary to, with a similar approach as before, judge the strength of each variable. It can be stated that the probability that Z is proven (be careful not to dismiss Z just because a false proof is given. IE "given that oil continues to be used to generate electricity, then we will be nuked. Oil will continue to be used to generate electricity, so we must prepare for a nuclear attack with a nuclear defense shield." Note that the starting premise is weak (that using oil will necessarily lead to getting nuked) but the conclusion is not necessarily a bad idea. At least, many better arguments could be given for that same conclusion.) , given a statement such as X or Y->Z is equal to- The probability of the premises being accurate multiplied by the probability that any set of variables that leads to Z is accurate.

So, this leads to the issue, of what tools to use to asses the probability of the premises and variables. The first, and most obvious is direct input. For instance, one does not need expert advice to know that the sky often displays a blue color. Any variable that requires the sky to lack its color in order to function requires the complete dismissal of all input throughout your entire life, and thus becomes abstractly improbable. This method is, however, limited. Especially when dealing with variables. Oftentimes, events can avoid contradicting personal experience even if you know a good deal about a subject. IE, say there was a statistic that said 2% of females aged 16-28 in town X get raped every year. As a resident female who lived her entire life in town X and hasn't been raped, you must realize that according to the statistic provided, this scenario has a nearly 77% chance to occur. As such, the personal experience of living in town X and not getting raped is nearly worthless in terms of judging the issue.
This is when you turn to aggregate data. So long as the collection method is valid and the source is trustworthy, statistics are the best and only way to answer questions upon which you lack data. Any amount of individual data points can be averted by declaring them a coincidence. However, by aggregating them, they become much stronger- facts are like sticks. Breaking one is always easy, but breaking them all at once is astronomically harder.

Sometimes, though, you will inevitably run across issues which fall into one of two categories- the statistics behind the premises are so astronomically convoluted that you can't understand what they mean or the data is so lacking that it's impossible to use a amalgamated mass. This is when the opinions of outside analysts become important. IE, a biologist can declare that his data indicates a coming epidemic, that requires immediate action to avert. The reasons for his declaration can easily be absurdly arcane, but if the other biologists around him agree that his data means what he thinks it means, and if he is a person who has successfully predicted 3 similar events, his data is probably valid. Alternatively, say a murder occurred and the murderer is confirmed to have destroyed all video evidence and left no dna evidence. However, he didn't know that his target's child was hiding in the closet and saw the entire thing, leaving a good enough testimony to identify him from amongst all the people who could physically have achieved the act. One must judge if the source would want to mislead you, or have some sort of vendetta against the presumed murderer, or if it is more like that the child is simply telling the truth. Oftentimes, these questions are actually very easy to make, but sometimes a very confusing situation arises- one in which 2 valid sources come to opposite conclusions. Since this method is a step within the process of judging argument validity, presuming that any given argument has not already been judged true or false, it WILL inevitably be reached. (assuming data exists. Even so, when faced with a unsolvable problem, people tend to pull in their friends, coworkers, etc. to help them come to a conclusion.) As such all problems are supported by one group and opposed by another, and the constituent strength of the groups supporting and opposing the argument must also be factored in. Note that group size is of minimal importance compared to other factors. IE, 10 doctors telling me that my symptoms are caused by disease X is more meaningful than 10000 non-doctors saying the same thing. Although popular opinion can be used as an indicator, it is often caused by such things as "going with the flow" or by the conclusion simply being the most obvious, even though anyone who really knows what he's talking about knows 10 proofs that make the conclusion false. (Often, 1 way hash arguments, wherein the true argument is thoroughly proven, but the proof is so long/difficult, that the average person cannot follow it, either due to willpower+available time or simple intellect failure. Note that lack of will to follow an argument is not necessarily a bad thing, and is in fact a mathematical necessity. By definition, listening to one argument requires that you are not listening to another. If you just start listening to every available argument, you will soon find that there are STILL far more that you haven't listened to.)