Unit 3 Flashcards
What is the requirement for samples from a large population?
Samples must be random
This ensures that the sample accurately represents the population.
What is ‘sampling without replacement’?
Each item/person can only be chosen once
This changes the probability of selection for subsequent items.
What is ‘sampling with replacement’?
Every item has equal chance of being selected each time
This maintains the initial probability for each item throughout the sampling process.
What is a simple random sample?
Every group of size (n) has equal chance of being chosen
Can be achieved by methods like drawing names from a hat or using random number tables.
List two methods to choose a simple random sample.
- Name in hat
- Random number table or random number generator
What is a potential drawback of a simple random sample?
It might not represent a diverse population
This can lead to biases in the results.
What is a stratified random sample?
Divides population into groups (strata) and samples from each
Example: sampling from different categories of restaurants.
What is the advantage of a stratified random sample?
Represents well if the population is diverse
However, it is not as random and takes more time.
What is a cluster random sample?
Creates similar groups made of very mixed populations
Picks one group as a mini representation for the whole sample.
What is the advantage of using a cluster random sample?
Faster and easier to implement
Represents pretty well, especially for non-diverse populations.
What is a systematic random sample?
Every (n)th person/item
This method involves selecting individuals at regular intervals from a list.
What is a census?
Selects all items/individuals
A census is comprehensive but time-consuming, typically only used with small populations.
What is the major problem with sampling?
Bias
Bias can lead to certain responses being systematically favored over others.
What is selection bias?
Sampling wasn’t random
This occurs when the sample does not accurately represent the population.
What is survey bias?
Bias is in the survey
This type of bias affects the quality and reliability of survey results.
What is voluntary response bias?
People volunteer with strong opinions
This leads to a skewed representation of the population’s views.
What is convenience bias?
Bias from selecting individuals who are easiest to reach
This can result in a non-representative sample.
What is undercoverage bias?
Certain groups are left out of the sample
This results in an incomplete representation of the population.
What is non-response bias?
People are chosen but don’t respond
This can be mitigated by encouraging responses from selected individuals.
What is response bias?
Problems leading to untruthful/incorrect data
Examples include broken tools, self-reported biases, and confusing questions.
Who are the key participants in experimental design?
- Researchers
- Experimental units/subjects
- Evaluators
What is an explanatory variable?
Levels that are manipulated intentionally
What are treatments in the context of experimental design?
Combinations of levels of the explanatory variable
What is a response variable?
The outcome of the experiment
What is a confounding variable?
A variable related to both the explanatory and response variables that can affect the results
What are the four pillars of a good experimental design?
- Comparisons of two treatment groups
- Random assignment of treatments to experimental units
- Replication (more than one experimental unit per group)
- Control of confounding variables and differences
2C2R
What is a single-blind experiment?
Subjects do not know which treatment they are receiving
What is a double-blind experiment?
Both subjects and evaluators do not know which treatment is being administered
Why are control groups important in experiments?
They need placebos to assess the effect of the treatment
Fill in the blank: People have a natural motivation to get better, so it is important not to tell them they are getting a _______.
placebo treatment
What is Completely Randomized Design?
A design where each subject is assigned a number using random number generation.
What is Randomized Block Design?
A design used when a lurking variable cannot be eliminated by random assignment, where subjects are grouped into blocks based on that variable.
What is Matched Pair Design?
A design where subjects are paired based on similar variables, with one subject receiving treatment A and the other receiving treatment B randomly.
What is the purpose of using random assignment in experimental design?
To create equivalent treatment groups and reduce bias.
What does statistically significant mean?
It indicates that the results observed in the study are unlikely to have occurred by chance alone.
How can one find statistically significant results?
By conducting statistical tests and analyzing p-values.
Fill in the blank: In a Randomized Block Design, if a lurking variable cannot be removed by random assignment, it is accounted for by creating _______.
blocks of each.
What is the role of statistical inference in experiments?
It involves taking statistics from a sample to infer conclusions about a larger population.
What is the first step in conducting a perfect experiment?
Randomly selecting subjects.
What is the second step in conducting a perfect experiment?
Randomly assigning subjects to treatment groups.
