Soc110 study guide

Reference group

What is the group being compared to in a dummy variable

Significance of a variable

P-value tells us what

Spurious relationship

Disappears after controlling for a 3rd variable

Controls for the other predictors in the model

Partial slope differs from a bivariate slope because:

Evidence is consistent with casuakl claim, but alternative explanations can still exist

What cautious conclusions about social media use causing depression is true

2

Race=white, black, and other, how many dummy variables?

Yes

educ=p.02, should we reject it?

No, only the reference group

Do dummy variables mean compared to all?

Interaction

Education predicts income among low-wealth but not high-wealth, whats it called?

Each additional year of education means a .110 decrease in children, controlling for sibs

educ=-.110, predicting number of children, controlling for siblings, means what?

Penalizing for additional variables

Adjusted R2 compared to r2

educ because it has a higher absolute value

Educ beta = -.191, sibs =- .127, which is stronger?

Little or no association because its is close to 0

Phi coefficient near .01 suggests what about a 2x2 table

Interaction

If the x-y association differs across the levels of z, it means what pattern?

Model fit improves, education explains part of the original race difference

R2 adjusted changes from .009 to .294, when adding educ and reduces the race coefficient, which is the interpretation

First, before independent

Stata regression syntax, dependent var goes where?

Time order

Casual criterion means that cause must occur before the outcome

chain/mediating relationship

X affects M, affects Y

Y-intercept, when all values are 0, alpha = the predicted value of y

Multiple regression, the intercept is:

Standardized beta coefficient is useful because;

Comparing predictors measured in different metrics

There is almost always alternative explanations

Why can observational research rarely prove causality definitively?

Variance in y explained by x in a bivariate = how much x and y relate

R2 tells us what about regression

6

Trained survived at 42, trained died at 14, untrained survived at 24, untrained died at 48, what is the odds ratio

Equal group size is not a casual criterion

Equal group size are required for causality, why is it wrong?

Association (as x changes, so does y) time order (happens before what we are predictions) and eliminating all other reasonable expectations

3 criteria for causality

Black students are -1.23 lower at a starting point when white =0

White as a reference, Black = -1.23, what does it mean?

Complete explanation/spurious

Zero-order relationship disappears after controlling for friend group, what does it mean?

Educ explains some of the racial differences in occupational prestige

Model with race, educ, spousal edu, the black coefficient shrinks after educ is added. What does that mean?

44.26

If the intercept is 46.49 and the black coefficient is -1.23, what is a black student at?