91Ó°ÊÓ

When conducting surveys or experiments, researchers often need to estimate a population parameter based on a smaller, randomly selected group, known as a sample. A common parameter of interest is the proportion of the population that has a specific characteristic, termed the 'sample proportion'.

Imagine a school where we want to find out how many students are interested in starting a coding club, but it's impractical to ask everyone. Instead, we survey a random group of 100 students, and 25 say 'yes'. Here, the sample proportion, often represented with the symbol \( p \), would be \( p = 25/100 = 0.25 \) or 25%. This gives us a statistical estimate of the true proportion of interested students in the entire school.

In our exercise, the bank determined the sample proportion of customers with both checking and savings accounts by dividing the number of individuals with both accounts (110) by the total sample size (250), which gives us \( p = 110/250 = 0.44 \). This means that 44% of the sampled customers have both types of accounts, and this proportion is used as an estimate of the true proportion for the entire customer base of the bank.

The 'standard error' (SE) reveals how much we expect our sample's statistic, like the sample proportion, to vary from one random sample to another. It's also a crucial element in gauging the precision of our estimate. Consider SE as an alarm system: the smaller it is, the more confident we can be that our sample proportion is close to the real deal.

A high school basketball coach trying to determine the average height of players on future teams may not have the exact heights of potential players. By taking random samples of players' heights over the years, the coach can use SE to predict how much the average height may fluctuate, helping to prepare for different types of players.

The formula for the standard error of a proportion is \( SE = \sqrt{ p(1-p) / n } \), where \( p \) is the sample proportion and \( n \) is the sample size. In our bank example, we calculated the SE to assess how much the sample proportion of customers with both accounts might vary if we took different samples from the population. It serves as the building block for creating a confidence interval around our sample proportion.

A 'Z-score' represents how many standard deviations a data point is from the mean of a distribution. It's a way of standardizing scores across different types of data, enabling us to compare apples with oranges. For example, if a student scores 90 on a math test with a mean of 80 and a standard deviation of 10, the Z-score would be 1 - this score is one standard deviation above the mean.

In the context of confidence intervals, the Z-score allows us to decide how far we'd wander from our sample proportion to capture the true population proportion with a certain level of confidence. It's like casting a fishing net of a specific size; the Z-score tells us how big that net is based on the confidence level we want.

For 95% confidence intervals, the Z-score is commonly 1.96, which means we're casting our net to capture the true proportion within 1.96 standard deviations from the sample proportion. In our exercise, employing a Z-score of 1.96 allowed us to calculate the range in which we are 95% confident the true proportion of the bank's customers with both checking and savings accounts lies.