Question

Jared is interested in finding out which of two types of soda are preferred by ​students at his school.
​
​Explain how he can randomly choose fifty students to ​participate in a survey or an experiment.
​
​Then describe how both a survey ​and an experiment can be designed for the purpose of answering his ​question.

Answer 1

Randomly choosing 50 students
- Define the population and sampling frame: list all students at the school (e.g., from the registrar or homeroom lists).
- Use a simple random sample:
1. Assign each student a unique number 1…N.
2. Use a random number generator (online tool, calculator, Excel/R sample function) to pick 50 distinct numbers, or put all names in a hat and draw 50 names.
- Alternatives if you want guaranteed representation by grade or sex: do stratified sampling. Divide students into strata (grades), then randomly sample from each stratum in proportion to its size so the 50 reflect the school makeup.

Notes: get parental consent if required, and allow refusals (replace a refusal with the next randomly selected student).

Survey design (observational approach)
- Sample: use the 50 randomly selected students.
- Questionnaire: ask clear, neutral questions, for example:
- “Which of the two sodas do you prefer overall: Soda A or Soda B?” (or present both names)
- Optional: “How often do you drink soda?” and basic demographics.
- Avoid bias: don’t lead respondents, present choices neutrally, keep survey anonymous if possible to reduce social-desirability bias.
- Analysis: compute the proportion who prefer each soda and a confidence interval or perform a test for difference in proportions. This measures stated preference (may include brand/image influence).

Experiment design (causal / taste-test approach)
Goal: remove branding/packaging effects and measure actual taste preference.
Two good designs:

1) Within-subjects blind taste test (recommended)
- Each of the 50 students tastes both sodas, but the order is randomized for each student (half taste A then B, half B then A).
- Blind the samples: label them with codes so students don’t know which is which. Ideally the administrator handing out the samples is also blind to which code corresponds to which soda (double-blind), or at least the label key is held by an uninvolved person.
- Use palate cleansers (water, plain crackers) and space tastings a little to reduce carryover.
- For each student record which soda they prefer (or ask for a rating for each).
- Analysis: count preferences and test whether preference differs from 50/50 (binomial test or McNemar’s if paired ratings are used).

2) Between-subjects blind test
- Randomly assign the 50 students to two groups (25 each).
- Each student tastes only one soda (blind) and rates it for liking.
- Compare average liking scores between groups (t-test) or compare proportions saying “prefer” if you ask for binary preference.
- Simpler but less powerful than within-subjects because individual differences affect results.

Other experiment considerations
- Randomize order to control order effects.
- Ensure blinding to remove label/brand bias.
- Check for allergies or dietary restrictions and get proper consent.
- Record any contextual variables (time of day, hunger) that could affect taste.

Which to choose?
- A survey measures stated preference in real life (including brand influence); it’s easier to administer.
- An experiment (blind taste test) isolates taste and gives stronger evidence of which soda is preferred for flavor alone.
- Both are useful: you could do the survey first (brand-influenced preference) and then a blind experiment to see how much branding changes preferences.