getting your bearings

Now that you know where you're going, do you know where you are? South Newfane.

empirical, normative

concept

variable, constant

conceptual definition

measure, measurement (operational definition, operationalization)

unit of analysis--individual, aggregate

ecological fallacy

measurement error--random, systematic

(measure) validity

(measure) reliability

validity & reliability in terms of targeting

panel surveys, as opposed to cross-sectional surveys

level of measurement--nominal, ordinal, interval

index scores (aka indices)

feeling thermometer

hypothesis

variables, dependent & independent

measures of central tendency--mean, median, mode

distribution & measures of dispersion--range, skewness, kurtosis, variance, standard deviation

univariate analysis & bivariate, multivariate analysis

crosstabulations (continency analysis)

difference of means (comparison of means) tests

requirements for a causal statement--covariation, temporal priority, elimination of rival hypotheses

internal validity

external validity

implications for X-->Y hypotheses when controlling for Z variables:  spurious relationship; additive effects; interaction effects

zero-order v partial relationships

research designs--experimental, quasiexperimental, nonexperimental

specification error

theory

population parameter(s), census

sample statistic(s), sample

statistical inference

sampling frame

simple random sampling

representative sample

random sampling error

standard deviation, variance

normal distribution, normally distributed variable(s)

standard error of a sample mean, of a sample proportion

confidence level

confidence boundary, upper & lower

tests of statistical significance: Χ2 (Pearson's chi square)

    one-tailed (-sided), two-tailed (-sided) tests of statistical significance

degrees of freedom

null hypothesis, alternative hypothesis

measures of association:  Cramer's V, Lambda, Somers' d

Pearson's correlation coefficient, r; adjusted r2

simple linear regression (with one independent variable); multiple regression (with two or more independent variables): 

unstandardized regression coefficient(s) and associated values for assessing statistical significance---standard error(s), t-value(s), p value(s)

in multiple regression, standardized regression coefficients (aka Betas)

constant

correlation coefficient (r in simple and R in multiple regression) and adjusted r2  or adjusted R2 for proportion of variance explained

F test for the statistical significance of the equation

 

Asked to explain the difference between a positive relationship and a negative relationship, one student said, "One describes the first month and the latter describes six months into it."