91Ó°ÊÓ

When dealing with a binomial distribution that has a large sample size, it becomes feasible and often simpler to use the normal approximation instead of calculating exact probabilities for each possible outcome. This technique leverages the central limit theorem, which states that the distribution of the sample mean will approximate a normal distribution as the sample size becomes large, given certain conditions. In the context of hypothesis testing, this allows us to apply more straightforward techniques, such as using Z-scores to find probabilities.

To use normal approximation here, you'll need two key parameters from the binomial distribution:

• The mean (\( \mu \))
• The standard deviation (\( \sigma \))

It helps to remember:

• \( \mu = n \times p \)
• \( \sigma = \sqrt{n \times p \times (1-p)} \)

where \( n \) is the number of trials and \( p \) is the probability of success in a single trial. These allow you to transform the binomial distribution into a normally distributed approximation. This makes hypothesis testing more manageable when checking for rare events or infrequent outcomes in larger populations.

The binomial distribution is a discrete probability distribution of the number of successes in a sequence of independent experiments. Each experiment is a Bernoulli trial, meaning it has only two possible outcomes: success or failure. The likelihood of a success on any given trial is denoted by \( p \).

In our exercise, the trials represent customers deciding whether to prefer soap brand A. Here, the probability that any one customer prefers brand A (a success) is 20%, i.e., \( p = 0.2 \). The number of trials, \( n \), is the sample size, which is 400 potential customers.

• Number of trials, \( n = 400 \)
• Probability of preference (success), \( p = 0.2 \)

The importance of the binomial distribution in hypothesis testing lies in its ability to describe the expected variability of outcomes (such as number of customers preferring a product) in a population. In practice, it is often paired with normal approximation techniques when the conditions (i.e., sufficiently large sample size) for the latter are met.

Once you've decided to use the normal approximation, you'll often need to standardize your findings using a Z-score, which is a measure that tells you how many standard deviations an element is from the mean of the distribution. This is crucial in hypothesis testing, as it helps convert binomial outcomes into a normal distribution framework, making it easier to determine probabilities.

The Z-score is calculated using the formula:\[ Z = \frac{X - \mu}{\sigma} \]where:

• \( X \) is the observed number of successes
• \( \mu \) is the mean of the distribution (\( \mu = 80 \) in our exercise)
• \( \sigma \) is the standard deviation of the distribution (\( \sigma = 8 \) in our solution)

For example, if 92 customers preferred brand A, the Z-score would be:\[ Z = \frac{92 - 80}{8} = 1.5 \]This Z-score allows us to reference standard normal distribution tables to find the probability of observing such an outcome or more extreme, under the null hypothesis. It essentially translates a specific instance into a standardized result that can be universally read through these tables.