This is cluster sampling because the company selects all users in one city as a complete group to test the new feature. Cluster sampling involves selecting entire groups (clusters) and including all members within the chosen cluster. The city represents a natural cluster, and all app users within that city receive the new feature. Random sampling (Choice A) would involve randomly selecting individual users regardless of their geographic location, not selecting an entire city as a unit.

This is cluster sampling because researchers randomly select entire hospitals (clusters) and then survey all doctors within those selected hospitals. Cluster sampling involves selecting complete groups and studying all members within the chosen clusters. The key characteristic is that once a hospital is selected, all doctors in that hospital are included in the study. Stratified sampling (Choice B) would involve dividing doctors into groups and randomly selecting some doctors from each group, rather than selecting entire hospitals.

This represents nonresponse bias because not all students who receive the email survey will respond, and those who do respond may differ systematically from those who don't. Nonresponse bias occurs when survey participants differ in important ways from non-participants, potentially skewing results. Students with strong opinions about online learning may be more likely to respond than those who are indifferent. Random sampling error (Choice B) refers to natural variation in samples, not systematic differences in response rates.

This is a stratified sample because the researcher divides students into grade-level strata and then randomly selects participants from within each stratum. Stratified sampling ensures representation from each important subgroup by sampling within predefined categories. By selecting 40 students from each grade level, the researcher guarantees proportional representation across all grades. Choice A incorrectly describes systematic sampling, which would involve selecting every nth student from a single ordered list.

This represents location bias because surveying people at a gym creates a sample that is not representative of the general population's exercise habits. Location bias occurs when the sampling location systematically excludes certain groups or overrepresents others. Gym-goers are likely to exercise more than the general population, making the sample unrepresentative for understanding overall exercise habits. Random sampling error (Choice B) refers to natural variation in samples, not systematic bias from sampling location.

This is an observational study because researchers monitor existing air quality conditions and correlate them with health records without manipulating any variables. Observational studies examine relationships between variables as they naturally occur without researcher intervention. The researchers observe environmental conditions and health outcomes without imposing treatments or controlling air pollution levels. An experiment (Choice A) would require researchers to manipulate air pollution levels, which would be unethical and impractical.

This is an observational study because researchers collect data on existing behaviors without manipulating or assigning commute methods. Observational studies involve measuring variables as they naturally occur without researcher intervention. The health department simply surveys people about their current commuting habits and stress levels without asking anyone to change their behavior. Choice A incorrectly suggests this is an experiment, but experiments require researchers to assign treatments.

This is an experiment because the researcher assigns runners to different routines (treatment groups). In an experiment, the researcher actively manipulates or assigns the treatment - here, randomly assigning runners to either do the new stretching routine or keep their usual routine. The random assignment of treatments is the hallmark of an experiment, allowing causal conclusions about the stretching routine's effect. Choice A incorrectly calls it observational because runners self-report soreness, but self-reporting the outcome doesn't make it observational - what matters is that treatments were assigned.

This is cluster sampling because researchers randomly select entire schools (clusters) and then survey all students within those selected schools. Cluster sampling involves selecting complete groups and studying all members within the chosen clusters. The key characteristic is that once a school is selected, all students in that school are included in the study. Stratified sampling (Choice B) would involve dividing students into groups and randomly selecting some students from each group, rather than selecting entire schools.

This is a stratified sample because the researcher divides the population into distinct subgroups (year levels) and then randomly samples from each stratum. Stratified sampling ensures representation from each important subgroup by sampling within predefined categories. By sampling 50 students from each year, the researcher guarantees that all class years are proportionally represented. Choice B incorrectly suggests convenience sampling, which would involve surveying only easily accessible students without systematic stratification.