Scribd
Upload
293 views

Log Explanation

This document provides guidance on interpreting regression coefficient estimates from four different model specifications: level-level, log-level, level-log, and log-log. It outlines the assumptions required to interpret results and describes how to interpret the coefficient for each model in terms of the expected change in the dependent variable given a one-unit change in the independent variable. Videos are also provided that walk through examples of running each type of regression in R and interpreting the results.

Uploaded by

MarkWeber
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
293 views

Log Explanation

This document provides guidance on interpreting regression coefficient estimates from four different model specifications: level-level, log-level, level-log, and log-log. It outlines the assumptions required to interpret results and describes how to interpret the coefficient for each model in terms of the expected change in the dependent variable given a one-unit change in the independent variable. Videos are also provided that walk through examples of running each type of regression in R and interpreting the results.

Uploaded by

MarkWeber
Copyright
© © All Rights Reserved
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 3

Teaching - Curtis Kephart > Econometrics Notes & R Code (UCSC Econ113) >

Interpret Regression Coefficient Estimates - {level-level, log-level, level-log & log-log regression}

Interpreting Beta: how to interpret your estimate of your


regression coefficients (given a level-level, log-level, level-log,
and log-log regression)?

Assumptions before we may interpret our results:


Running a Regression (Using R Statistics
 The Gauss–Markov assumptions* hold (in a lot of Software)
situations these assumptions may be relaxed - Step-by-step example of how to do a
particularly if you are only interested in an regression using R statistics software
approximation - but for now assume they strictly hold). (including the models below). I'll walk through
* If you're interested in more details, read the code for running a multivariate regression
the discussion here, or check out your textbook. - plus we'll run a number of slightly more
 Our coefficient estimates (our estimates of below) complicated examples to ensure it's all clear.
are statistically significant and practically significant. Video 16:30 -
 With a multivariate model, we assume that other www.youtube.com/watch?v=Ktks5K95uQM
independent variable(s) (x_2, x_3, ... x_n) are held
constant.

Model Dependent Independe Interpretation of β Video Review


or Response nt or
Variable Explanato Given a change in x, Given reader requests, I created
(y) ry how much do we expect y to short video explanations of how
Variable change by? to interpret regression estimates
(x)
Δy=β1Δx Interpreting Level-Level
Regression Coefficient
Level-level Regression y x “If you change x by one, Estimate Results
we’d expect y to change We run a level-level regression
by β1" and interpret the regression
coefficient estimate results.
Simple example of regression
analysis with a level-level model.

Video 5:00 -
www.youtube.com/watch?v=TJA
CbJspao0
%Δy=100⋅β1⋅Δx Log-Level Regression
Log-Level Regression “if we change x by 1 Coefficient Estimate
x (unit), we’d expect Interpretation
our y variable to change by We run a log-level regression
100⋅β1 percent” (using R) and interpret the
regression coefficient estimate
results. A nice simple example of
regression analysis with a log-
Technically, the level model.
interpretation is the following:
Video 6:40 -
www.youtube.com/watch?v=wXC
2kViEGz8

but the
quoted interpretation is appr
oximately true for values -0.1
< β1 < 0.1 (and it's much
easier to remember.)
Δy=(β1/100)%Δx Level-Log Regression
Level-Log Regression Coefficient Estimates
y "If we increase x by one We run a level-log regression and
percent, help understand the regression
we expect y to increase by coefficient estimates. A nice
(β1/100) units of y." simple example of regression
analysis.
Note, you cannot include
obs. for which x<=0 if x is Video 6:50
then logged. You either can't www.youtube.com/watch?v=L9ZL
calculate the regression 6_DB4fQ
coefficients, or may
introduce bias.
%Δy=β1%Δx Log-Log Regression
Log-Log Regression Coefficient Estimate Results
“if we change x by one We do a log-log regression and
percent, explain the regression coefficient
we’d expect y to change estimate results. Simple example
by β1 percent” of regression analysis with a log-
log model.
Note, you cannot include
obs. for which x<=0 if x is Video
logged. You either can't 5:30 www.youtube.com/watch?v=
calculate NZCSt9Wkpkl
the regression coefficients, o
r may introduce bias.

You might also like