91Ó°ÊÓ

When conducting hypothesis testing, the first major task is to formulate the null hypothesis, denoted as \( H_0 \). This hypothesis represents the default or original assumption about a population parameter. In simpler terms, it's what we suppose is true before collecting any evidence. In the exercise provided, the null hypothesis states that the proportion of smartphone owners remains unchanged at 0.75. This means we are assuming nothing has altered, and the claim that 75% of people still own smartphones holds true.

Having a clear null hypothesis is essential because it acts as a starting benchmark. We use statistical tests to determine if collected data significantly contradicts this assumption. If there is strong evidence against the null hypothesis, we may reject it in favor of an alternative assumption.

The alternative hypothesis is the statement you want to test and potentially provide evidence for. It is denoted as \( H_1 \) or sometimes \( H_a \). Unlike the null hypothesis, the alternative hypothesis proposes that there is a significant effect or change in the population parameter. In our exercise, the alternative hypothesis suggests that the proportion of smartphone owners is different from 0.75. This opens up the possibility for both an increase or a decrease.

Formulating the alternative hypothesis is crucial as it guides the direction of our statistical test. By setting up an alternative hypothesis, we are essentially saying we are looking for evidence of a change or difference. Only if the evidence is sufficiently strong, can we reject the null in favor of our alternative assumption.

The Z-score is a pivotal concept in hypothesis testing. It measures how many standard deviations an element is from the mean. In essence, the Z-score quantifies the difference between the observed data point, in this case, our sample proportion, and the proposed population proportion under the null hypothesis.

In our exercise, the Z-score helps decipher whether the difference between the sample smartphone proportion (0.79) and the hypothesized proportion (0.75) is statistically significant after accounting for the variability of the sample. The Z-score formula is: \[ Z = \frac{\hat{p} - P_0}{\sqrt{\frac{P_0(1-P_0)}{n}}} \]
Calculating the Z-score gives us a basis to determine how likely or unlikely our observed data would be if the null hypothesis were true.

A large absolute Z-score suggests the data point is unusual under the null hypothesis, leading us potentially to reject it, especially when it exceeds the critical Z-value determined by the chosen significance level.

The significance level, denoted by \( \alpha \), is a threshold set by the researcher which dictates the probability of rejecting a true null hypothesis, known as Type I error. It essentially defines how much risk we are willing to take in wrongly rejecting the null hypothesis.

In our exercise, the significance level is set at 0.02, or 2%. This implies that there is a 2% chance of concluding there is a difference in smartphone ownership proportions when, in fact, there isn't. It acts as a cut-off point to determine whether the evidence from data is compelling enough to reject the null hypothesis.

By comparing the calculated Z-score against the critical values set by our significance level, we can decide whether our results are statistically significant. A lower significance level means more stringent criteria, requiring stronger evidence to claim a significant difference.