Ethics in Experiments/Questioning Mistakes
This is not a principle of experimental design, because, for example, it would be unethical to assign certain people to an experimental group of smokers, and others to a control group of non-smokers.
What is randomization?
A form of studying a sample by assigning different treatments to different groups, the comparing the results at the end.
What is an experimental study?
Choosing subjects for a study with no regard for their personal characteristics, and every individual in the population has an equal chance of being selected.
What is a simple random sample?
An example of *this* would be asking your followers to rank Donald Trump's performance on a scale of 1-10, and recording those who respond.
What is a Self-selected sample?
This occurs when a subject in the control group unknowingly changes their behavior, believing they are receiving a treatment when they are in fact not.
What is a placebo effect?
This question type assumes that a piece of information is true (maybe incorrectly) as part of the question. It does not necessarily push the respondent in a certain direction, though. For example: Has Trump's tough stance on foreign powers increased American power in the world?
What is a loaded question?
In contrast with random sampling, which allows you to make an inference about the population, you can make an inference about this if you randomly assign your sample to different experimental groups.
What is cause and effect?
For this sampling method, you could break a population into groups based on their personal characteristics, like race, career, number of children, etc, then take a simple random sample from each group to ensure your sample has a mix of different sorts of people.
What is a stratified random sample?
An example of *this* would be surveying only the students that take classes with you.
What is a convenience sample?
What is a control group?
A question whose answer options are ambiguous; multiple answer options might apply to a single respondent; the answer options are not mutually exclusive.
Unlike a statistic, which describes the sample, *this* describes the population.
What is a parameter?
What is a Cluster Sample?
An example of this would be asking people to state whether or not they have ever been convicted of a crime.
What is a response bias?
This is when you first divide experimental subjects into two groups based on a characteristic, then break each of those groups into all treatment categories (i.e., each group has a placebo sub-group, treatment 1 sub-group, treatment 2 sub-group, etc...)
What is blocking, in experimental design?
A question that, when answered, actually implies answers to two separate topics; it is ambiguous which topic - if not both - is being answered.
What is a double-barrel question?
An example of *this* would be an experiment where the subjects do not know which brand of citrus soda they are drinking, but the experimenters do.
What is a single-blind experiment?
It is important to do this because otherwise, there will not be enough subjects to avoid the influence of random chance. For example, if you only have five test subjects, and just 2 of them happen to be serial killers, it will appear as though 40% of the population are serial killers.
Why is it important to have a large enough sample size?
A survey gathering data on people's opinion of meat alternatives at fast food restaurants, where respondents must return the survey by mail is an example of this.
The number of subjects in each group of an experiment with a sample of 1,998 subjects, studying the relationship between two brands of melatonin and perceived quality of sleep (don't forget to include a control group!!!).
This type of question is likely to push a respondent toward a certain answer, more so than the respondent would have if the question was worded fairly. It does not necessarily assume anything incorrect. For example, "Since he was convicted of grand theft auto, shouldn't the defend receive the death penalty?"
What is a leading question?
An example of *this* would be a study where you do not know which brand of cigarettes people are smoking, but you record whether people who smoke get cancer or not, and whether people do smoke get cancer or not, and compare those results.
What is an observational study?
In this method, you could maybe randomly select just a few of the states in the U.S., then randomly select just a few of the counties, then randomly select 1,000 people from each group: annual household income below 50k, between 50k-100k, and above 100k.
What is multi-stage sampling?
When conducting a survey of customers' opinions of the food quality at Hoodbachi, the researcher only surveyed people between the ages of 15-25.
What is coverage bias?
The number of subjects in each group of an experiment studying the relationship between cavities and soda, where the total sample size is 120, subjects are blocked into 'male' and 'female,' and there are two experimental groups for 'dark' and 'clear' sodas.
What is 20?