Parameter
A numerical summary of population
Statistic
A numerical summary of sample;
A science of collecting, organizing and analyzing the data
Steps of Statistic; Identification
Population-A group of men or objects to be studied;
Individual-A singular man or object to be studied
Steps of Statistic; Collection
Sample-A representative subset of the population
Steps of Statistic; Organization
Descriptive Statistic-Arranged & collected data
Steps of Statistic; Conclusion
Inferential Statistic- Applying the result to the real world
* Due to its nature, one cannot be 100% sure of the validity of statistic application (AKA samping error)
Data
Info collected for conclusion;
An observed value(coule be both qualitative and quantitative)
Qualitative Variable
Attribute characteristics or classification; Adding or subtracting doesn't have meaning.
Quantitative Variable
Gives numerical measures, subtract and addition will have significant impact.
Discrete variable
Countable and finite. Concrete, part of quantitative variable
Continuous variable
Infinite # of possible values; cannot be counted and needs to be measured.
ie) Height, time
Variable
Characteristics of individual
Explanatory variable
The cause or treatment, independent. Explains the outcome of the sample
Response variable
The ourcome or result, dependent. Impacted by explanatory variable
Designed Experiment
When researchers deliberately control/manipulate the explanatory variable to determine its impact. Casuation achieved
Observational study
When reserachers simply observe. No messing with the treatment. Cannot establish casuation, only association is possible.
Lurking variable
Third variable that influecnes other variables and consequently the response variable. Often hidden.
Confounding
State where an experiment is affected by lurking variable; the outcome in most cases is uncredible.
Census
The complete list containing the individuals in the population. Takes lot of time and effort to make it, therefore done seldomly. (Some still argues there's a gaphole, as in reality with citizens government has no way of measuring the homeless population)
SRS
AKA simple random sampling. Need randomly selected n from N. Onces chose, men cannot be RE-SELECTED. If each n has equal chance of occruing, SRS is achieved.
Stratified Sampling
Separates the population into groups called strata where each belongers have sth in similarity. Then randomly select few representatives from each stratum. Gurantees the representation from one end to another.
Systematic sampling
Select every kth individual from the population by applying the formula k=N/n. Get a starting point between 1~k, and nail every person that comes up after kth turn. Does NOT NEED FRAME, therefore ideal when one only knows N.
Cluster sampling
Select some natrually-occuring groups within the cluster of population, and sample everyone in that selected group. Except then you can miss out the entire facade of the population, or worse, your intended target.
Convience sample
When one's lazy enough, she will go where sample is not collected based on randomness but from voluntary response, like when men decides to reply to the survey. Population tends to be biased.