Interpreting and Understanding Logits, Probits, and other Non-Linear Probability Models

Richard Breen; Kristian Bernt Karlson; Anders Holm

doi:10.1146/annurev-soc-073117-041429

Interpreting and Understanding Logits, Probits, and other Non-Linear Probability Models

Richard Breen, Kristian Bernt Karlson, Anders Holm

Department of Sociology

84 Citations (Scopus)

132 Downloads (Pure)

Abstract

Methods textbooks in sociology and other social sciences routinely recommend the use of the logit or probit model when an outcome variable is binary, an ordered logit or ordered probit when it is ordinal, and a multinomial logit when it has more than two categories. But these methodological guidelines take little or no account of a body of work that, over the past 30 years, has pointed to problematic aspects of these nonlinear probability models and, particularly, to difficulties in interpreting their parameters. In this review, we draw on that literature to explain the problems, show how they manifest themselves in research, discuss the strengths and weaknesses of alternatives that have been suggested, and point to lines of further analysis.

Original language	English
Journal	Annual Review of Sociology
Volume	44
Pages (from-to)	39-54
ISSN	0360-0572
DOIs	https://doi.org/10.1146/annurev-soc-073117-041429
Publication status	Published - 30 Jul 2018

Keywords

Faculty of Social Sciences
logit
probit
KHB method
Y-standardization
marginal effects
linear probability model
mediation

Access to Document

10.1146/annurev-soc-073117-041429

Cite this

@article{b2976fbb966f4cacaca6d6b8c109a6c0,

title = "Interpreting and Understanding Logits, Probits, and other Non-Linear Probability Models",

abstract = "Methods textbooks in sociology and other social sciences routinely recommend the use of the logit or probit model when an outcome variable is binary, an ordered logit or ordered probit when it is ordinal, and a multinomial logit when it has more than two categories. But these methodological guidelines take little or no account of a body of work that, over the past 30 years, has pointed to problematic aspects of these nonlinear probability models and, particularly, to difficulties in interpreting their parameters. In this review, we draw on that literature to explain the problems, show how they manifest themselves in research, discuss the strengths and weaknesses of alternatives that have been suggested, and point to lines of further analysis.",

keywords = "Faculty of Social Sciences, logit, probit, KHB method, Y-standardization, marginal effects, linear probability model, mediation",

author = "Richard Breen and Karlson, {Kristian Bernt} and Anders Holm",

year = "2018",

month = jul,

day = "30",

doi = "10.1146/annurev-soc-073117-041429",

language = "English",

volume = "44",

pages = "39--54",

journal = "Annual Review of Sociology",

issn = "0360-0572",

publisher = "Annual Reviews",

}

TY - JOUR

T1 - Interpreting and Understanding Logits, Probits, and other Non-Linear Probability Models

AU - Breen, Richard

AU - Karlson, Kristian Bernt

AU - Holm, Anders

PY - 2018/7/30

Y1 - 2018/7/30

N2 - Methods textbooks in sociology and other social sciences routinely recommend the use of the logit or probit model when an outcome variable is binary, an ordered logit or ordered probit when it is ordinal, and a multinomial logit when it has more than two categories. But these methodological guidelines take little or no account of a body of work that, over the past 30 years, has pointed to problematic aspects of these nonlinear probability models and, particularly, to difficulties in interpreting their parameters. In this review, we draw on that literature to explain the problems, show how they manifest themselves in research, discuss the strengths and weaknesses of alternatives that have been suggested, and point to lines of further analysis.

AB - Methods textbooks in sociology and other social sciences routinely recommend the use of the logit or probit model when an outcome variable is binary, an ordered logit or ordered probit when it is ordinal, and a multinomial logit when it has more than two categories. But these methodological guidelines take little or no account of a body of work that, over the past 30 years, has pointed to problematic aspects of these nonlinear probability models and, particularly, to difficulties in interpreting their parameters. In this review, we draw on that literature to explain the problems, show how they manifest themselves in research, discuss the strengths and weaknesses of alternatives that have been suggested, and point to lines of further analysis.

KW - Faculty of Social Sciences

KW - logit

KW - probit

KW - KHB method

KW - Y-standardization

KW - marginal effects

KW - linear probability model

KW - mediation

U2 - 10.1146/annurev-soc-073117-041429

DO - 10.1146/annurev-soc-073117-041429

M3 - Journal article

SN - 0360-0572

VL - 44

SP - 39

EP - 54

JO - Annual Review of Sociology

JF - Annual Review of Sociology

ER -

Interpreting and Understanding Logits, Probits, and other Non-Linear Probability Models

Abstract

Keywords

Access to Document

Fingerprint

Cite this