Introduction to Statistics

Random Sample: each member of the population has the same chance of being selected Representative Sample: characteristics should represent those of the target population without bias Observational Study: no intervention by the investigator, no treatment imposed Experimental Study: Investigator has some control over the determinant Variables: Categorical – each observation falls into a feline number of groups Nominal: named variables with no implied order e. G. Personality type Ordinal: grouped variables with implied order e. G. Veil of education Continuous – measured variables Discrete: take discrete values e. G. Number of children Numerical: can assume any value within a certain range/elemental e. G. Height Types of Designs: True experiment: researcher has potential to randomly allocate observations to conditions Quasi-experiment: demonstrate a relationship between an IV/DVD researcher makes use of naturally occurring groups, can't make cause and effect statements Non- experiments (correlation design): question If there Is a relationship between variables, can't make cause & effect statementsBetween groups: two groups being compared on some outcome measure Within-subjects: participants experience each condition of an IV, with measurements of some outcome taken on each occasion Extraneous variables: variable present In an experiment, which might Interfere with the relationship between IV & DVD Confounding variables: mediating variable that can adversely affect the relation between IV/DVD Internal validity: extent to which a casual relationship can be assumed between IV & DVD.External validity: degree to which you can generalize the results of your study to mom underlying population T-test One sample t-test – A: data should arise from a normal population Paired t-test -A: must be independent, arise from a normal distribution & populations of same spreads Independent sample – A: normally distributed, homogeneity of variances, independen ce of the observations Correlation/Regression – A: the relation in the population is linear, the residuals in y have a constant standard deviation and the residuals arise from a normal distribution detests of good fit and test of independence – A: expected count has to be larger than five