# COMM3710F11001frey

## COMM3710 - Fall 2011 - Section 001 - Reading Notes - Frey

### riosjosh

Inferential statistics (inductive statistics) is used to do two things: estimation, which is used to generalize the results from a sample to its parent group, and significance testing, which is the likelihood that the results are by chance. Researchers often estimate population characteristics (called parameters) on the basis of characteristics found in a sample (called statistics). Hence, estimation procedures are often referred to as parametric statistics. Estimates are based on two assumptions: that the variables are normally distributed in the population and a random sample has been taken.

Asymptotic – a curve that runs on forever without ever touching a line

In a normal distribution 99.72% of all scores are accounted for between ±3 SD or standard deviations. Scores from any research or survey can be converted to “z-scores” or standard scores. This way the scare of the evaluation can be eliminated. For example an IQ score of +2.75SD would be labeled as genius because we know that only 2.14% of the population lie between +2SD and +3SD. Data can also be manipulated to be normally distributed a common example of this is when a teacher curves a course or test by using the mean and standard deviation.

Kurtosis – refers to how pointed or flat the shape of a curve is

A normal curve is called mesokurtic because its highest peak is also the center point and has two tails coming off of it equally. A leptokurtic curve, or distribution, is tall and sharply pointed, with scores clustered around the middle. A flat curve or platykurtic curve has scores that are mostly evenly distributed.

Skewness refers to where the scores are located. A positively skewed curve, or distribution, is when most of the scores are on the left and the tail runs out to the right. A negatively skewed curve occurs when most of the scores are on the right and the tail goes off to the left.

Random sample means that everyone in the population has an equal chance at being selected. Thanks to Central Limits Theorem there are many ways we can use samples. When using smaller samples and non-random samples it is up to the researcher to check that a normal distribution has been obtained.