91Ó°ÊÓ

When working with experiments like coin flips, the sample proportion is a critical concept to grasp. A sample proportion is a measure that represents the ratio of times an event occurs relative to the total number of trials. For example, in the context of flipping a coin 150 times, we are measuring how many of those flips result in heads.
In the example, we obtained heads 90 times out of 150 flips. This means our sample proportion, denoted as \( \hat{p} \), is calculated as follows:\[ \hat{p} = \frac{90}{150} = 0.6 \]This value is a way of summarizing our data and can be really helpful when analyzing whether our coin behaves as expected, or if perhaps it is biased. A sample proportion is a practical tool for summarizing data from a series of trials and is foundational in conducting hypothesis tests.

The concept of a null hypothesis is foundational in statistical testing. It sets a benchmark expectation under which the randomization distribution is centered. The null hypothesis is essentially a statement of 'no effect' or 'no difference.' It provides a point of comparison against which we measure the outcome of our experiment.
In the context of flipping a coin, the null hypothesis would be that the coin is fair. That means there is an equal probability of landing on heads or tails, mathematically expressed as a probability of 0.5 for heads.
When we establish this null hypothesis, we are assuming there is no bias in the coin. Thus, any deviation from this hypothesis in our observed sample proportion might suggest that the coin could be biased, but we need statistical tests to confirm it.

Statistical evidence is crucial when we want to determine whether the results of our experiments are significant or merely due to chance. It allows us to make objective conclusions based on the data collected.
To gather statistical evidence in our coin flip example, we generate a randomization distribution. This is done by simulating the experiment multiple times under the assumption that the null hypothesis is true. In this example, we assume the coin is fair, so the expected proportion of heads in our randomization distribution should be 0.5.
If the sample proportion from our actual experiment (0.6) is significantly different from this expected value, and falls outside the common range of simulated proportions, we may obtain enough statistical evidence to reject the null hypothesis. This would indicate that the observed data (90 heads out of 150 flips) is unlikely to have occurred by random chance alone and suggests bias in the coin.