Generic Mixture about R Flashcards
(15 cards)
If I need to check the skewness or kurtosis of a dataset in R, which command do I use?
describe(datasetname$variable)
How do I know when I have a right-skewed graph without plotting it?
By analysing the results provided by the command describe (datasetname&variable). If the skewness is positive then it is right skwed, if it is negative then it is left skewed.
How do I know when I have a positive or negative kurtosis?
By analysing the results provided by the command describe I look at kurtosis. If it is near 3 it is a positive kurtosis. If it is near -3, it is a negative kurtosis = more spread.
Which values can the Cronbach’s Alpha acquire? What does it mean?
Cronbach’s Alpha can go from 0 to 1. The value of 0.7 would be a good one but anything beyond 0.9 indicate too much redundancy.
What is the difference in output between describe(nameofthedataframe) and describe(dataframe$var1)?
describe(nameofthedataframe)= a list and metrics for each column name
describe(dataframe$var1)= creates just one line about this one column name with all metrics
How do I calculate in R mediation? What steps do I need to write down?
1) Effect of independent variable on mediator (with regression)
2) effect of independent variable on outcome (c’)
3) effect of mediator and independent variable on outcome (c)
How do I calculate in R moderation? What steps do I need to write down from a conceptual perspective?
1) mean center both predictor and moderator
2) use formula for regression analysis
so that the outcome is affected by the multiplication of predictor*independent variable.
What is the formula of a t-test?
difference between mean values/standard deviation from the mean (error)
What is the difference between a two-tailed hypothesis and a one-tailed or directed hypothesis?
two-tailed hypothesis: is there a difference between the two groups that I am looking at?
one-tailed hypothesis: what is the direction of the difference between two groups that I am observing?
When I say that I have a very large t-value, what does it mean truly?
It means that I have a bigger difference between two groups that I am looking at, compared to their standard error (how much they are supposed to vary anyways).
What does R^2 stand for conceptually?
R^2 explains how much variance in the outcome is directly coming by the predictors. It indicates how much the model fits what I am observing.
What are manifest variables?
Differently from latent variables, manifest variables can be directly observed and measured with one parameter.
How is Hierarchical regression calculated in R from a conceptual standpoint?
1) Regression analysis between predictor 1 and outcome
2) Regression analysis between predictor 1 + predictor 2 on the outcome
Why certain researchers indicate that Baron and Kenny’s formula evaluative method is not truly valid?
- Baron and Kenny take 0.05 as an arbitrary number that has to be followed precisely.
- Baron looks at non significance as proof of full MEDIATIOn while non significance you lack evidence but does not indicate mediation.
What does the Sobel Test measure? How does it measure it?
The Sobel test attempts to define if the effect of the mediator is STATISTICALLY SIGNIFICANT. It evaluates whether the product of from X to M and from M to Y is large enough to be significant.
This model works only when we have normally distributed values.
