Descriptive Data
gender gap, economic status, gpa, backgrounds, age etc..
Inferential Data
weather, sports
Population
all together
Census
gathering EVERYONE in the population
Parameter
(everyone in the COUNTRY) everyone everyone*
Sample population
smaller survey population-→ from these we can find the parameter easier
Quantitative
#’s you can do meaningful with… height, age, distance, gpa, body temp
Catagorical
gender, religion, area code, zip
Discrete Quantitaive Data
Anything that you can count
Continuous
anything that you can measure with the use of a tool
Undercoverage
missed by the method, (people who don’t check their mail, or it gets lost)
Non-Response
choosing not to respond to a survey- voluntarily
Variation
expected, differs everyday, the data will be changed
Observational Experiment
this study has no effect on the outcome
Experimental Experiment
adding treatment to the experiment, seeing if there is RELATIONSHIP
BIAS
what DIRECTION**
Specific Random Selection (SRS)
names in a hat; random number generator, -→ until the desired number is picked as the survey population
Specific Random Assighment
flip a coin 50%, use a dice for even and odd-→ anything to randomly assign the group of people divided into an experiment
Specific Measurable Results
COMPARE THE RESULTS OR OVERALL SCORE AVERAGE
Systematic
A survey method that lists out the survey population and then starts from a random number/name on the list and moves every nth of the way down until desired number of people are picked
Cluster
group the population by some means 2. pick all from some of the groups ex: there are 5 rows you choose all of the first 3 rows as your survey population
Stratified
group the population by some means 2. pick some from all of the groups ex: there are 5 rows and you pick 2 from each (fe) until you get your desired amount of participants
Generalization
applying a conclusion to the study that is over reaching the population -→ ex: all of Florida when only talking about Pinellas County
Loaded Questions
using very persuasive techniques to get a difficult answer form the person responding-→ do you still like dogs?
Voluntary Response Bias
most of the time a call in poll, or an optional survey where the STRONGEST OPINIONS ARE EMPHASIZED
Interview Bias
Where the person asking the survey could in some way change the outcome of the survey
Confounding
multiple factors and your not sure of the TRUE FACTOR causing the real results→ weight pills, tom brady
Sampling Error
When there is a lot of variation as well as undercoverage and voluntary response (the big red flags of an experiment)
Non-Sampling Error
lying, biased methods, human calculation errors, non-responsive
Blinding
just blinding one person in the experiment so there is a less likely chance of bias-- test giver or tested doesn’t know which treatment they are doing
Double Blinding
When the tester and people tested are both unaware of which treatment is given or taken
Placebo (effect)
the baseline treatment that is the control variable, placebo effect is when there are more positive results with the placebo treatment
Explanatory Variable
What is causing the reaction?
Response Variable
What may happen?
Treatment
What is being done? (verb)
Unit
to what population
Factor
What are we giving or doing? (most likely the treatment)
Level
What are the measurements (all #’s)
Blocking
making obvious groups to break down our study--→ gender blocking (VERY BIG)
Matched Pairs
where people in the experiment are matched with one another based on a key characteristic, and then are compared against other people at the end of the experiment to see a difference.