chapotraphouse

In science you use a null hypothesis to determine like, is the thing I'm proposing even worth doing? Is my hypothesis completely wrong?

To give an example because my plain-english attempts to explain stuff are usually cruddy: I see two doors, 🚪A and 🚪B. One door, 🚪A is of average width and height and the other, 🚪B is about 2/3rd's the height, same width.

People walk through the 🚪doors, I look and think I see a pattern. What I think is that people who are tall, will not go through 🚪B, why? Yeah maybe because tall people really really don't like it. Maybe they are entitled since they are generally viewed favourable in terms of dating on average at least. Maybe tall people just don't see the 🚪B, they only see things at a certain height.

Really I know it's physically possible, I just need to start somewhere, so I put people into two bins, tall and not-tall. I count 100 people walking through the 🚪doors and my null hypothesis, is something I can be confident in saying is not true. Like the opposite of what I want. My null hypothesis then is that no tall people go through 🚪B

Ok let's look at the data to see if we can reject my null hypothesis or not. Excuse the table formatting

             🚪A  🚪B
        Tall: 46   3
    Not-Tall: 37   14

Oh wow look at that! Ok now I can move on to my real question, which is, do people who are tall like to go through 🚪doors which are of lower height? This is much harder. Maybe I'll say if there is a preference of like 90%, so 90% of tall people go through regular height 🚪doors that means that they don't like going through shorter height 🚪doors. I also need to know how many people generally like to go through 🚪doors of shorter height. Because if tall people prefer regular 🚪doors as much as non-tall people, then height of the individual, tall or not, would not play a role in the difference.

Since I was able to eliminate the idea that tall people do not go through shorter 🚪doors, I can move on to other questions which depend on the initial null hypothesis to be false. Because, why waste time if our hunches or guesses (i.e our _hypotheses) are not on the right track?

Ok, does this relate. I'm skipping stuff for brevity, and I think you can still get it bcuz of scientific socialism & all that goodness. These 'failures' can tell us what not to do. And more importantly, these 'failures' failed in a specific way! Why that specific way? Because the overall 'failure' is the result of a ton of smaller 'failures' and 'successes' and we will say stuff that didn't really affect anything. Did the weather during a specific guerrilla attack lead to the end failure? Did a comrade in a mid-level position decrease the fighting power of the fighters? Are guerrilla tactics different in the 80s (I think it was the 80s?) and because tactics from the civil war in China were used, that was what ultimately led to failure?

The thing is there is likely some useful information. You're not wrong when you say it is a failure and thus not worth studying or that there is no point considering it. And others with different knowledge may find something useful. The trick is–you cannot know until you try. Because a presumption or prediction is a model, or something in your head. And as dialectical materialists we must accept that what is in our head is not the same as the external world, i.e. the stuff _outside of our head.

Again it's not that you are wrong; it's only, that is a difficult, perhaps impossible thing to know without actually putting in the work and effort.

One other way, think of these as experiments, not all experiments succeed, and why they fail helps you figure out how to run the experiment next time.