27Thousand
- 109
- 0
Astronuc said:Miss Silvy is somewhat correct - it's not quatitative. People are unpredictable. One cannot look at a person and know what they are thinking at any given moment. One simply has to experience (interact with) another person over time. And even then the other person might be unpredictable. Trying to fit the behavior of another person to a model is futile.
Something to consider, to put things into perspective it may not be quantitative science in the same way Physics or a test tube in chemistry are, but it's science in the same way that it uses the scientific method like the medical field does. I was looking at some of the follow-up studies, and if you have two groups of people who are exactly the same and you manipulate an independent variable and then measure a dependent variable, then there's a cause-effect relationship. Even if there's a confounding variable involved, something is still acting consistent here and can be further studied and put into use. The studies used actual statistical techniques to make their conclusions quantitative to see that there was statistical significance and effect sizes.
The problem is the effect size, etc, just stays in the peer-review journal articles. They don't usually try to find patterns across studies that deal with effect sizes/mathematical equations, but rather just look at patterns of the general principle (people say they care about the concept rather than details). I don't see equations in psychology textbooks (okay maybe just a few but not many), which makes it vague and doesn't say the other factors which come into play. My proposal is to do what Newton did and turn it more into a quantitative science, patterns across studies rather than just staying inside of the peer-review article itself (when they use null hypothesis testing). From what I've learned from those stats classes I took, from the actual Statistics Department rather than Social Science Department, whenever you have an actual correlation or effect size (even if they're weak) you can always create a mathematical equation to say a general probability to make predictions in a certain range (even if it's a weak equation). My plan is creating equations and then searching for patterns across studies, and then working from there to see how strong/weak these principles really are (by converting it into quantitative science format just like people will convert documents into PDF). It makes it more falsifiable, and thus better able to be refined/improved upon over time (like the natural sciences do). It's kind of like data mining, but there are some differences in my plans. Also in the the hard sciences they say the better something is at making predictions, the more likely it can be used for technology. They already use these principles to come up with predictions, but they don't usually come up with generalized mathematical equations across studies that they want to make testable. So maybe I could use that for my dating life! Also as I said in some of the other posts, I'm not planning on it only being an intellectual adventure but also combining a ton of practice/experience with it.