Notes on Statistics

Hypothesis Testing

• experimental design
• control/randomization
• quasi-experimental design
• cross-sectional analysis
• longitudinal analysis
• panel studies
• null hypothesis
• variables
  • independent (X)
  • dependent (Y)
  • intervening
• relationships
  • linear
  • curvilinear
  • bivariate
  • multivariate
• spuriousness
• operational definition
• validity
  • internal
  • external
• reliability

Measurement

• nominal
• ordinal
• interval
• ratio

Descriptive Statistics

• parameter vs. statistic
• measures of central tendency
  • mean
  • median
  • mode
  • standard deviation
• variance
• kurtosis (platykurtic, leptokurtic)
• skew

Significance Testing

• significance level
• critical region
• binomial distribution
• Type I & Type II error
• one-tail/two-tail test
• sample size
• standard error of mean
• central limit theorem
• t-test of difference of means
• significance vs. measures of association

Nonparametric Statistics

• chi-square
• rank-order correlation (Spearman's r)

Parametric Statistics

• normal curve
• z-scores
• PRE measure
• general linear model (regression line)
• homoscedasticity
• independence
• scattergram
• a
• b
• e (or u)
• correlation
• coefficient of determination
• multiple regression
• R-square
• partial correlation coefficient
• beta weight (standardized partial correlation coefficient)
• analysis of variance
• factor analysis
• Logit/analysis of covariance
• data transformations

Choosing Inductive Statistics

• if Y=scalar and X=categorical, use ANOVA
• if Y=scalar and X=scalar, use r or R
• if Y=categorical and X=scalar, use Logit/ANCOVA
• if Y=categorical and X=categorical, use Chi-square

MSU