Multilevel Models

Multilevel models cannot be used to analyze data with ...

• ... students nested in schools

• ... countries nested within persons

• ... observations nested within participants

• ... patients nested within clinics

• the expected value of y of an average person

• the random slope coefficients

• the distribution of the person i's y values

• the residual variance

• the fixed effect for variable u

• the random slope coefficients

• the expected deviation of person i's values from the intercept

• the residual variance

Which software cannot be used to estimate multilevel models?

• SPSS

• R

• MPlus

• all of those

In multilevel models, the intraclass correlation coefficient represents ...

• how many participants we need to get enough statistical power

• the proportion of unexplained variance that is due to between-person differences

• the proportion of unexplained variance that is due to within-person differences

• how much the random intercept correlates with the random slope

BLUP stands for

• best linear unbiased predictor

• bayesian limited upper parameter

• biased linear upper parameter

• bayesian linear universal predictor

How do we need to transform the data so that we can test a within-person effect of X on Y?

• person-mean centering of Y

• person-mean centering of X

• take the mean of X as a predictor

• take the square root of X as a predictor

In multilevel models, cross-level interactions can be tested by ...

• jumping into lake Como

• interacting two level 1 variables

• taking the mean of the level 1 variable and interacting it with the level 2 variable

• interacting a level 1 and a level 2 variable

In R, how do you access the third row and the 2nd column of a data.frame object

• dat[2,3]

• dat$2:3

• dat[3,2]

• dat$2$3

what does this command do to the data.frame dat?

• it describes the functional form of the table in the data

• it applies the dat function to each element of the table

• it helps you to find the best function for your table

• it counts how many observations are missing and not missing per variable