I more or less agree. The observations do not change, but they may be incomplete, or we may be misinterpreting what those observations are actually telling us. This is the case with spurious correlations. A great example of a spurious correlation has to do with hand size and shoplifting. If you were to gather data on the hand sizes of convicted shoplifters in any region, and then compare that with the average hand size of the general population in that region, you would find that shoplifters have larger hands. This is actually true, but the connection between hand size and thievery is not direct. The real connection is that, overall, the majority of shoplifters are male. Since males have larger hands than females on average, this creates an indirect correlation between hand size and shoplifting. However, if you didn't already know that, you have pretty solid ground to formulate a theory suggesting that, the larger your hands, the more likely you'll engage in shoplifting.Fulmen wrote:Yes and no. Theories and estimates may change, but not the observations. Sure some observations or measurements will turn out to be wrong, but the bulk of them are correct and will not change over time. The interpretations however are always "in play".Munk wrote:Science is indeed a discipline based on provisional knowledge. Being as such, nothing is ever really outside of the possibility of being revised/refuted/etc.
Many terms have a slightly different meaning in science. A scientific theory is far more than "just a theory", it's a model that can accurately describe the way nature works. It's an hypothesis that have been verified by observations to the point where there is little doubt of it's correctness, although it may turn out to be just a special case of a more fundamental "law".
The term law is also misleading, as laws are something WE define and can break if we want to. The laws of nature are not the models we use to describe them but the nature of nature, the unchanging and unbreakable way of the universe. These are forever hidden from us, we can never know them for sure. All we can do is build better and better, yet fundamentally flawed models in an attempt to understand the world around us.
Any legitimate scientific theory is based off of objective and repeatable observations, however the causal link between observations is something people simply have to infer. Evidence is mute, so we have to speak for it the best we can, but sometimes we get it wrong -- either because we don't know the whole truth, or because we misinterpret the correlations. The more we can interconnect a theory to other theories, those of which are interconnected to yet other theories, the more likely that correlations are not spurious. But, the chance is never zero...and the more fundamental the phenomenon, the greater the impact a new discovery can make. Take Newtonian Physics for example. Almost everything we do (on a macro level) seems to support the validity of Newtonian physics...and that's A LOT of supporting evidence. Then we went and discovered quantum physics, which essentially turned Newtonian physics on its head (as well as everything that is built upon it). Newtonian physics is still true in a proximate manner, but its principles may not actually apply to certain phenomena as it "should", if it were completely true.
On a side note, I hate the "just a theory" argument people will try to use to discredit well-established scientific principles. I often see this being used with regard to evolution. While its true evolution is a theory, people have a gross misunderstanding of what that actually means in the scientific community. They erroneously equivocate "theory" with "guess". However, because evolution is just a theory ( ), I can only respond in one way... I say: "That is indeed true, yet so is gravity."
(most people think its a "law", but yet again, not really understand what that word means in a scientific context).