Multiple Linear Regression Results Table
(from Hofferth, S. L. (2006) Residential father family type and child well-being:
investment versus selection. Demography, 43(1), pages 53-77)
Take Home Message
 If a coefficient is negative, then more of that covariate
leads to less of the outcome variable (holding the values of
the other covariates constant)
 If a coefficient is positive, then more of that covariate leads
to more of the outcome variable (holding the values of the
other covariates constant)
 The p-value gives the probability of getting that coefficient
value in your sample ASSUMING that the coefficient in
the population is 0.
 The coefficient in the population is (almost) never 0.
 Do NOT use p-values to decide on the importance of a
covariate (i.e. the size of its effect). The p-value does not
necessarily reflect the importance of a covariate. It reflects
measurement error and sample size.
 The importance of a covariate (e.g., whether it has a small,
medium, or large effect) is a judgment made by the
researcher.
 Consider using standardized regression coefficients
Logistic Regression Results Table
(from Eloundou-Enyegue, P. M. & Williams, L. B. (2006) Family size and
schooling in sub-Saharan African settings: A reexamination. Demography, 43(1),
pages 25-52)
Note that Exp(B) is the odds ratio.
Take Home Message
 If a coefficient is negative  the odds ratio is less than 1,
then more of that covariate makes the outcome less likely
(holding the values of the other covariates constant).
 If a coefficient is positive  the odds ratio is more than 1,
then more of that covariate makes the outcome more likely
(holding the values of the other covariates constant).
 The p-value gives the probability of getting that coefficient
value (and corresponding odds ratio) in your sample
ASSUMING that the coefficient in the population is 0.
 The coefficient in the population is (almost) never 0.
 Do NOT use p-values to decide on the importance of a
covariate (i.e. the size of its effect). The p-value does not
necessarily reflect the importance of a covariate. It reflects
measurement error and sample size.
 The importance of a covariate (e.g., whether it has a small,
medium, or large effect) is a judgment made by the
researcher.
 The purpose of this class is to convince you to display your
results in terms of predicted probabilities and to show you
how to do this.