ERP stuffie

Purpose of descriptive statistics - introduce

Introduce the players of the game (the variables that play a role in the analysis), inform the reader of the nature of the variable (discrete, continuous, own research etc.).

Measurement levels

- Nominal – only categories, no ranking - Ordinal – ordered categories, distances not clear - Interval – ordered/ranking, equal distances assumed - Ratio – ordered/ ranking, equal distances assumed, absolute zero, no negative numbers

Validity

extent to which measure correctly represents the concept of the study Internal validity – how well the study was done External validity – are the results generalized to other situations

Skewness

measure that indicates the symmetry of a distribution compared to a normal distribution

Kurtosis

tailedness of a distribution – how many observations are in the tails compared to normal distribution

R2

Proportion of the variation y that is explained by the linear combination of the x variables

Adjusted R2

adjust for the number of predictors in the model – how well the model fits the data, corrected by the degrees of freedom

F test

overall significance of the model

Dummy variable

categorical variable with only two values, 0 and 1. Value 1 satisfies a condition

Control variable

explanatory variables that are introduced in the regression model in order to assess or clarify the relationship between two or more variables (hypothesized relationship)

- Explanatory variables

if u estimate the models including those variables, you can solve the problem of multicollinearity

Multicollinearity

high correlation between at least two independent variables.

Heteroskedasticity

Uneven distribution of errors in the scatterplot, i.e., different variances for different observations (e.g., groups of observations have different variances, or the variance could depend on the size of the observation).

Reverse causation

the DV influences the IV Reverse causality occurs when the dependent and independent variables have been gathered at the same point of time. The time dimension is neglected and cause does not precede effect.

Endogeneity

correlation between the explanatory variables and error term

Exogenity assumption

iv in a regression model is not correlated with an error term

Moderation

relationship between X predictor and Y outcome depends on a third variable Z moderator

Mediation

a mediator explains the relationship between two variables

Logistic regression rather than OLS?

if the dependent variable is a categorical (usually, binary) variable

Exploratory and confirmatory factor analysis

EFA is used to identify and quantify factors or latent variables, whereas CFA is used for testing hypotheses about the structures of those latent variables

Factor loading

correlation between a factor and variable, can take values between -1 and 1 due to a correlation

Main uses of factor analysis

- To understand the structure of the set of variables in the analysis - To analyze a questionnaire to measure the underlying (latent) variable - To reduce the dataset to a manageable size while retaining as much of the original information as possible

Communality

- proportion of variance that a variable share with other variables in the analysis

eigenvalue

measure of how much of the common variance (communality) of the observed variables a factor explains.

Purpose of descriptive statistics - problems

- Signal possible problems and things to keep in mind, e.g. large standard deviations (why are those a problem? - large heterogeneity, might be two different subpopulations). Whether there are negative values or not, truncated distribution