Question

question 11
1 pts
ch 18 before conducting a test of significance, which of the following should you do? check all that apply.
set an appropriate significance level.
inspect the data you have collected for outliers or other issues that might challenge the assumptions of the particular statistical test you wish to use.
determine what sample size you need for your inference to be successful.
question 12
1 pts
ch 18 inference is most reliable when the data come from a random sample or a randomized comparative experiment.
true
false

Explanation:

• Question 11:
• Setting a significance level ($\alpha$) is fundamental as it defines the threshold for rejecting the null hypothesis. For example, a common $\alpha = 0.05$ means a 5% chance of Type - I error (rejecting a true null).
• Inspecting data for outliers (e.g., using box - plots) is crucial. Outliers can violate assumptions of many statistical tests (e.g., normality assumption in t - tests).
• Determining sample size (using power analysis) ensures the test has enough power (probability of detecting a true effect) to make reliable inferences.
• Question 12:

Random sampling (selecting units randomly from a population) and randomized comparative experiments (randomly assigning treatments) reduce bias. In random sampling, each member of the population has an equal chance of being selected, and in randomized experiments, treatment assignment is random. This randomness allows for valid statistical inference (e.g., using probability theory to generalize results or establish causal relationships).

Answer:

• Question 11: All the checked options (Set an appropriate significance level, Inspect the data you have collected for outliers or other issues that might challenge the assumptions of the particular statistical test you wish to use, Determine what sample size you need for your inference to be successful) are correct.
• Question 12: A. True.