91Ó°ÊÓ

Statistical hypothesis testing is like a trial in the justice system for numbers. It’s a structure we follow to decide if a certain belief about a population (like a country or a bunch of chocolate bars coming off a factory line) stands up to scrutiny. This process starts by making some initial assumption, which we call the null hypothesis.
Each test also has a alternative hypothesis, which is what we’d like to prove instead. Think of it as the 'challenge' to the status quo of the null hypothesis. When we 'run the numbers,' we're essentially seeing if there's strong enough evidence to favor this challenge. If there isn’t, then we'll stick with the original assumption. This kind of testing doesn’t exactly say our null hypothesis is true—more that there’s just no strong reason to believe otherwise.
Imagine you're telling your friends that your favorite soccer player scores more than anyone else. That’s your alternative hypothesis because that's what you want to prove. Your null hypothesis would be that they score the same as any other player—nothing special (ouch!). Trying to figure out if you're right is where hypothesis testing comes in, crunching the actual goal-scoring data to see if you can officially brag about your favorite player.

Mean comparison is a detective game with numbers. It involves comparing the average value—that is, the mean—of a group, to see if it's different from some other value we're interested in, like a standard or another group’s mean.
For example, when a teacher wants to know if the new style of teaching helped the class perform better on tests, they might compare the average (mean) score from this year to last year's. This is a comparison of means. It gets interesting in statistics because we can't just look at the numbers and say 'Hey, this year’s average is higher, so the teaching method's better!' We need to do a statistical test to check if the difference isn't just due to chance or a fluke.
Sometimes our comparisons are built to check for anything but equality. Like, say we’re trying to see if pups trained with a new treat will follow commands faster. If the alternative hypothesis is that they will indeed follow commands in less time, we’re on the lookout for an average time that’s less than the one we've set as a standard, just like checking if a mean is less than 50 in our given problem.

Research methodology is like the recipe for whipping up a science experiment or study. It's all about planning how you're going to find the answer to a question. Will you do experiments, send out surveys, or observe people or things in action?
The method must suit your question. For instance, if you want to know if people are happier on sunny days, a survey might be your best bet. But if you're looking to see if plants grow taller with a new fertilizer, you might go for an experiment with randomly assigned pots.
Good research methodology follows systematic steps to ensure that the study results are reliable and can be trusted to help make decisions or advance knowledge. It includes choosing the right hypothesis, like we've seen with null and alternative hypotheses, and setting up conditions for testing that hypothesis to lead the study to meaningful conclusions, whether that’s about a new medicine or the classroom success of a teaching strategy.