1. Logistic Regression
• Logistic Regression - Dichotomous Response
variable and numeric and/or categorical
explanatory variable(s)
– Goal: Model the probability of a particular as a function
of the predictor variable(s)
– Problem: Probabilities are bounded between 0 and 1
• Distribution of Responses: Binomial
• Link Function:
1
log
)
(
g
2. Logistic Regression with 1 Predictor
• Response - Presence/Absence of characteristic
• Predictor - Numeric variable observed for each case
• Model - (x) Probability of presence at predictor level x
x
x
e
e
x
1
)
(
• = 0 P(Presence) is the same at each level of x
• > 0 P(Presence) increases as x increases
• < 0 P(Presence) decreases as x increases
3. Logistic Regression with 1 Predictor
are unknown parameters and must be
estimated using statistical software such as SPSS,
SAS, or STATA
· Primary interest in estimating and testing
hypotheses regarding
· Large-Sample test (Wald Test):
· H0: = 0 HA: 0
)
(
:
:
.
.
:
.
.
2
2
2
1
,
2
2
^
^
2
^
obs
obs
obs
X
P
val
P
X
R
R
X
S
T
5. Example - Rizatriptan for Migraine (SPSS)
Variables in the Equation
.165 .037 19.819 1 .000 1.180
-2.490 .285 76.456 1 .000 .083
DOSE
Constant
Step
1
a
B S.E. Wald df Sig. Exp(B)
Variable(s) entered on step 1: DOSE.
a.
x
x
e
e
x 165
.
0
490
.
2
165
.
0
490
.
2
^
1
)
(
000
.
:
84
.
3
:
819
.
19
037
.
0
165
.
0
:
.
.
0
:
0
:
2
1
,
05
.
2
2
2
0
val
P
X
RR
X
S
T
H
H
obs
obs
A
6. Odds Ratio
• Interpretation of Regression Coefficient ():
– In linear regression, the slope coefficient is the change
in the mean response as x increases by 1 unit
– In logistic regression, we can show that:
)
(
1
)
(
)
(
)
(
)
1
(
x
x
x
odds
e
x
odds
x
odds
• Thus e
represents the change in the odds of the outcome
(multiplicatively) by increasing x by 1 unit
• If = 0, the odds and probability are the same at all x levels (e
=1)
• If > 0 , the odds and probability increase as x increases (e
>1)
• If < 0 , the odds and probability decrease as x increases (e
<1)
7. 95% Confidence Interval for Odds Ratio
• Step 1: Construct a 95% CI for :
^
^
^
^
^
^
^
^
^
96
.
1
,
96
.
1
96
.
1
• Step 2: Raise e = 2.718 to the lower and upper bounds of the CI:
^
^
^
^
^
^
96
.
1
96
.
1
,
e
e
• If entire interval is above 1, conclude positive association
• If entire interval is below 1, conclude negative association
• If interval contains 1, cannot conclude there is an association
8. Example - Rizatriptan for Migraine
)
2375
.
0
,
0925
.
0
(
)
037
.
0
(
96
.
1
165
.
0
:
%
95
037
.
0
165
.
0 ^
^
^
CI
• 95% CI for :
• 95% CI for population odds ratio:
)
27
.
1
,
10
.
1
(
, 2375
.
0
0925
.
0
e
e
• Conclude positive association between dose and
probability of complete relief
9. Multiple Logistic Regression
• Extension to more than one predictor variable (either
numeric or dummy variables).
• With k predictors, the model is written:
k
k
k
k
x
x
x
x
e
e
1
1
1
1
1
• Adjusted Odds ratio for raising xi by 1 unit, holding
all other predictors constant:
i
e
ORi
• Many models have nominal/ordinal predictors, and
widely make use of dummy variables
10. Testing Regression Coefficients
• Testing the overall model:
)
(
.
.
))
log(
2
(
))
log(
2
(
.
.
0
all
Not
:
0
:
2
2
2
,
2
1
0
2
1
0
obs
k
obs
obs
i
A
k
X
P
P
X
R
R
L
L
X
S
T
H
H
• L0, L1 are values of the maximized likelihood function, computed by
statistical software packages. This logic can also be used to compare
full and reduced models based on subsets of predictors. Testing for
individual terms is done as in model with a single predictor.
11. Example - ED in Older Dutch Men
• Response: Presence/Absence of ED (n=1688)
• Predictors: (p=12)
– Age stratum (50-54*
, 55-59, 60-64, 65-69, 70-78)
– Smoking status (Nonsmoker*
, Smoker)
– BMI stratum (<25*
, 25-30, >30)
– Lower urinary tract symptoms (None*
, Mild,
Moderate, Severe)
– Under treatment for cardiac symptoms (No*
, Yes)
– Under treatment for COPD (No*
, Yes)
*
Baseline group for dummy variables
12. Example - ED in Older Dutch Men
Predictor b sb Adjusted OR (95% CI)
Age 55-59 (vs 50-54) 0.83 0.42 2.3 (1.0 – 5.2)
Age 60-64 (vs 50-54) 1.53 0.40 4.6 (2.1 – 10.1)
Age 65-69 (vs 50-54) 2.19 0.40 8.9 (4.1 – 19.5)
Age 70-78 (vs 50-54) 2.66 0.41 14.3 (6.4 – 32.1)
Smoker (vs nonsmoker) 0.47 0.19 1.6 (1.1 – 2.3)
BMI 25-30 (vs <25) 0.41 0.21 1.5 (1.0 – 2.3)
BMI >30 (vs <25) 1.10 0.29 3.0 (1.7 – 5.4)
LUTS Mild (vs None) 0.59 0.41 1.8 (0.8 – 4.3)
LUTS Moderate (vs None) 1.22 0.45 3.4 (1.4 – 8.4)
LUTS Severe (vs None) 2.01 0.56 7.5 (2.5 – 22.5)
Cardiac symptoms (Yes vs No) 0.92 0.26 2.5 (1.5 – 4.3)
COPD (Yes vs No) 0.64 0.28 1.9 (1.1 – 3.6)
Interpretations: Risk of ED appears to be:
• Increasing with age, BMI, and LUTS strata
• Higher among smokers
• Higher among men being treated for cardiac or COPD
13. Loglinear Models with
Categorical Variables
• Logistic regression models when there is a clear
response variable (Y), and a set of predictor
variables (X1,...,Xk)
• In some situations, the variables are all responses,
and there are no clear dependent and independent
variables
• Loglinear models are to correlation analysis as
logistic regression is to ordinary linear regression
14. Loglinear Models
• Example: 3 variables (X,Y,Z) each with 2 levels
• Can be set up in a 2x2x2 contingency table
• Hierarchy of Models:
– All variables are conditionally independent
– Two of the pairs of variables are conditionally
independent
– One of the pairs are conditionally independent
– No pairs are conditionally independent, but each
association is constant across levels of third variable
(no interaction or homogeneous association)
– All pairs are associated, and associations differ
among levels of third variable
15. Loglinear Models
• To determine associations, must have a measure: the
odds ratio (OR)
• Odds Ratios take on the value 1 if there is no
association
• Loglinear models make use of regressions with
coefficients being exponents. Thus, tests of whether
odds ratios are 1, is equivalently to testing whether
regression coefficients are 0 (as in logistic regression)
• For a given partial table, OR=e
software packages
estimate and test whether =0
16. Example - Feminine Traits/Behavior
3 Variables, each at 2 levels (Table contains observed counts):
Feminine Personality Trait (Modern/Traditional)
Female Role Behavior (Modern/Traditional)
Class (Lower Classman/Upper Classman)
PRSNALTY * ROLEBHVR * CLASS1 Crosstabulation
Count
33 25 58
21 53 74
54 78 132
19 13 32
10 35 45
29 48 77
Modern
Traditional
PRSNALTY
Total
Modern
Traditional
PRSNALTY
Total
CLASS1
Lower Classman
Upper Classman
Modern Traditional
ROLEBHVR
Total
17. Example - Feminine Traits/Behavior
• Expected cell counts under model that allows for association
among all pairs of variables, but no interaction (association
between personality and role is same for each class, etc).
Model:(PR,PC,RC)
– Evidence of personality/role association (see odds ratios)
Class=Lower Class=Lower Class=Upper Class=Upper
Role=M Role=T Role=M Role=T
Personality=M 34.1 23.9 17.9 14.1
Personality=T 19.9 54.1 11.1 33.9
88
.
3
)
1
.
11
(
1
.
14
)
9
.
33
(
9
.
17
:
Upper
Class
88
.
3
)
9
.
19
(
9
.
23
)
1
.
54
(
1
.
34
:
Lower
Class
OR
OR
Note that under the no
interaction model, the odds
ratios measuring the
personality/role association
is same for each class
19. Example - Feminine Traits/Behavior
• Intuitive Results:
– Controlling for class in school, there is an
association between personality trait and role
behavior (ORLower=ORUpper=3.88)
– Controlling for role behavior there is no association
between personality trait and class (ORModern=
ORTraditional=1.06)
– Controlling for personality trait, there is no
association between role behavior and class
(ORModern= ORTraditional=1.12)
20. SPSS Output
• Statistical software packages fit regression type models, where
the regression coefficients for each model term are the log of
the odds ratio for that term, so that the estimated odds ratio is e
raised to the power of the regression coefficient.
Note: e1.3554
= 3.88 e.0605
= 1.06 e.1166
= 1.12
Parameter Estimates
Asymptotic 95% CI
Parameter Estimate SE Z-value Lower Upper
Constant 3.5234 .1651 21.35 3.20 3.85
Class .4674 .2050 2.28 .07 .87
Personality -.8774 .2726 -3.22 -1.41 -.34
Role -1.1166 .2873 -3.89 -1.68 -.55
C*P .0605 .3064 .20 -.54 .66
C*R .1166 .3107 .38 -.49 .73
R*P 1.3554 .2987 4.54 .77 1.94
21. Interpreting Coefficients
• The regression coefficients for each variable
corresponds to the lowest level (in alphanumeric
ordering of symbols). Computer output will print a
“mapping” of coefficients to variable levels.
Cell (C,P,R) Class Prsnlty Role C*P C*R P*R Expected Count
L,M,M 0.4674 -0.8774 -1.1166 0.0605 0.1166 1.3554 34.1
L,M,T 0.4674 -0.8774 0 0.0605 0 0 23.9
L,T,M 0.4674 0 -1.1166 0 0.1166 0 19.9
L,T,T 0.4674 0 0 0 0 0 54.1
U,M,M 0 -0.8774 -1.1166 0 0 1.3554 17.9
U,M,T 0 -0.8774 0 0 0 0 14.1
U,T,M 0 0 -1.1166 0 0 0 11.1
U,T,T 0 0 0 0 0 0 33.9
To obtain the expected cell counts, add the constant (3.5234)
to each of the s
for that row, and raise e to the power of that
sum
22. Goodness of Fit Statistics
• For any logit or loglinear model, we will have
contingency tables of observed (fo) and expected
(fe) cell counts under the model being fit.
• Two statistics are used to test whether a model
is appropriate: the Pearson chi-square statistic
and the likelihood ratio (aka Deviance) statistic
e
o
o
e
e
o
f
f
f
G
f
f
f
log
2
:
Ratio
-
Likelihood
)
(
:
square
-
Chi
Pearson
2
2
2
23. Goodness of Fit Tests
• Null hypothesis: The current model is
appropriate
• Alternative hypothesis: Model is more complex
• Degrees of Freedom: Number of sample logits-
Number of parameters in model
• Distribution of Goodness of Fit statistics under
the null hypothesis is chi-square with degrees of
freedom given above
• Statistical software packages will print these
statistics and P-values.
25. Example - Feminine Traits/Behavior
Goodness of fit statistics/tests for all possible models:
Model G2 2
df P-value (G2
) P-value (2
)
(C,P,R) 22.21 22.46 4 .0002 .0002
(C,PR) 0.7199 0.7232 3 .8685 .8677
(P,CR) 21.99 22.24 3 .00007 .00006
(R,CP) 22.04 22.34 3 .00006 .00006
(CR,CP) 22.93 22.13 2 .00002 .00002
(CP,PR) 0.6024 0.6106 2 .7399 .7369
(CR,PR) 0.5047 0.5085 2 .7770 .7755
(CP,CR,PR) 0.4644 0.4695 1 .4946 .4932
The simplest model for which we fail to reject the null
hypothesis that the model is adequate is: (C,PR): Personality
and Role are the only associated pair.
26. Adjusted Residuals
• Standardized differences between actual and
expected counts (fo-fe, divided by its standard
error).
• Large adjusted residuals (bigger than 3 in
absolute value, is a conservative rule of
thumb) are cells that show lack of fit of
current model
• Software packages will print these for logit
and loglinear models
27. Example - Feminine Traits/Behavior
• Adjusted residuals for (C,P,R) model of all
pairs being conditionally independent:
Adj.
Factor Value Resid. Resid.
PRSNALTY Modern
ROLEBHVR Modern
CLASS1 Lower Classman 10.43 3.04**
CLASS1 Upper Classman 5.83 1.99
ROLEBHVR Traditional
CLASS1 Lower Classman -9.27 -2.46
CLASS1 Upper Classman -6.99 -2.11
PRSNALTY Traditional
ROLEBHVR Modern
CLASS1 Lower Classman -8.85 -2.42
CLASS1 Upper Classman -7.41 -2.32
ROLEBHVR Traditional
CLASS1 Lower Classman 7.69 1.93
28. Comparing Models with G2
Statistic
• Comparing a series of models that increase in
complexity.
• Take the difference in the deviance (G2
) for the
models (less complex model minus more
complex model)
• Take the difference in degrees of freedom for
the models
• Under hypothesis that less complex (reduced)
model is adequate, difference follows chi-square
distribution
29. Example - Feminine Traits/Behavior
• Comparing a model where only Personality
and Role are associated (Reduced Model)
with the model where all pairs are associated
with no interaction (Full Model).
• Reduced Model (C,PR): G2
=.7232, df=3
• Full Model (CP,CR,PR): G2
=.4695, df=1
• Difference: .7232-.4695=.2537, df=3-1=2
• Critical value (=0.05): 5.99
• Conclude Reduced Model is adequate
30. Logit Models for Ordinal Responses
• Response variable is ordinal (categorical
with natural ordering)
• Predictor variable(s) can be numeric or
qualitative (dummy variables)
• Labeling the ordinal categories from 1
(lowest level) to c (highest), can obtain the
cumulative probabilities:
c
j
j
Y
P
Y
P
j
Y
P ,
,
1
)
(
)
1
(
)
(
31. Logistic Regression for Ordinal Response
• The odds of falling in category j or below:
1
)
(
1
,
,
1
)
(
)
(
c
Y
P
c
j
j
Y
P
j
Y
P
• Logit (log odds) of cumulative probabilities are modeled
as linear functions of predictor variable(s):
1
,
,
1
)
(
)
(
log
)
(
logit
c
j
X
j
Y
P
j
Y
P
j
Y
P j
This is called the proportional odds model, and assumes the
effect of X is the same for each cumulative probability
33. SPSS Output
• Note that SPSS fits the model in the
following form:
1
,
,
1
)
(
)
(
log
)
(
logit
c
j
X
j
Y
P
j
Y
P
j
Y
P j
Parameter Estimates
-1.027 .102 101.993 1 .000 -1.227 -.828
.165 .094 3.070 1 .080 -.020 .351
-.001 .133 .000 1 .993 -.263 .260
0a . . 0 . . .
[ATTITUDE = 1]
[ATTITUDE = 2]
Threshold
[RACE=0]
[RACE=1]
Location
Estimate Std. Error Wald df Sig. Lower Bound Upper Bound
95% Confidence Interval
Link function: Logit.
This parameter is set to zero because it is redundant.
a.
Note that the race variable is not significant (or even close).
34. Fitted Equation
• The fitted equation for each group/category:
165
.
0
0
165
.
0
)
Nonwhite
|
2
(
)
Nonwhite
|
2
(
logit
:
te
Mod/Nonwhi
or
Neg
166
.
0
)
001
.
0
(
165
.
0
)
White
|
2
(
)
White
|
2
(
logit
:
Mod/White
or
Neg
027
.
1
)
0
(
027
.
1
)
Nonwhite
|
1
(
)
Nonwhite
|
1
(
logit
:
onwhite
Negative/N
026
.
1
)
001
.
0
(
027
.
1
)
White
|
1
(
)
White
|
1
(
logit
:
hite
Negative/W
Y
P
Y
P
Y
P
Y
P
Y
P
Y
P
Y
P
Y
P
For each group, the fitted probability of falling in that set of categories
is eL
/(1+eL
) where L is the logit value (0.264,0.264,0.541,0.541)
35. Inference for Regression Coefficients
• If = 0, the response (Y) is independent of X
• Z-test can be conducted to test this (estimate
divided by its standard error)
• Most software will conduct the Wald test, with the
statistic being the z-statistic squared, which has a
chi-squared distribution with 1 degree of freedom
under the null hypothesis
• Odds ratio of increasing X by 1 unit and its
confidence interval are obtained by raising e to the
power of the regression coefficient and its upper
and lower bounds
36. Example - Urban Renewal Attitudes
• Z-statistic for testing for race differences:
Z=0.001/0.133 = 0.0075 (recall model estimates -)
• Wald statistic: .000 (P-value=.993)
• Estimated odds ratio: e.001
= 1.001
• 95% Confidence Interval: (e-.260
,e.263
)=(0.771,1.301)
• Interval contains 1, odds of being in a given category or
below is same for whites as nonwhites
Parameter Estimates
-1.027 .102 101.993 1 .000 -1.227 -.828
.165 .094 3.070 1 .080 -.020 .351
-.001 .133 .000 1 .993 -.263 .260
0a . . 0 . . .
[ATTITUDE = 1]
[ATTITUDE = 2]
Threshold
[RACE=0]
[RACE=1]
Location
Estimate Std. Error Wald df Sig. Lower Bound Upper Bound
95% Confidence Interval
Link function: Logit.
This parameter is set to zero because it is redundant.
a.
37. Ordinal Predictors
• Creating dummy variables for ordinal
categories treats them as if nominal
• To make an ordinal variable, create a new
variable X that models the levels of the
ordinal variable
• Setting depends on assignment of levels
(simplest form is to let X=1,...,c for the
categories which treats levels as being
equally spaced)
![SPSS Output
• Note that SPSS fits the model in the
following form:
1
,
,
1
)
(
)
(
log
)
(
logit
c
j
X
j
Y
P
j
Y
P
j
Y
P j
Parameter Estimates
-1.027 .102 101.993 1 .000 -1.227 -.828
.165 .094 3.070 1 .080 -.020 .351
-.001 .133 .000 1 .993 -.263 .260
0a . . 0 . . .
[ATTITUDE = 1]
[ATTITUDE = 2]
Threshold
[RACE=0]
[RACE=1]
Location
Estimate Std. Error Wald df Sig. Lower Bound Upper Bound
95% Confidence Interval
Link function: Logit.
This parameter is set to zero because it is redundant.
a.
Note that the race variable is not significant (or even close).](https://image.slidesharecdn.com/chapter15c-250321120326-e6fd94fc/85/LogisticRegressionDichotomousResponse-ppt-33-320.jpg)
![Example - Urban Renewal Attitudes
• Z-statistic for testing for race differences:
Z=0.001/0.133 = 0.0075 (recall model estimates -)
• Wald statistic: .000 (P-value=.993)
• Estimated odds ratio: e.001
= 1.001
• 95% Confidence Interval: (e-.260
,e.263
)=(0.771,1.301)
• Interval contains 1, odds of being in a given category or
below is same for whites as nonwhites
Parameter Estimates
-1.027 .102 101.993 1 .000 -1.227 -.828
.165 .094 3.070 1 .080 -.020 .351
-.001 .133 .000 1 .993 -.263 .260
0a . . 0 . . .
[ATTITUDE = 1]
[ATTITUDE = 2]
Threshold
[RACE=0]
[RACE=1]
Location
Estimate Std. Error Wald df Sig. Lower Bound Upper Bound
95% Confidence Interval
Link function: Logit.
This parameter is set to zero because it is redundant.
a.](https://image.slidesharecdn.com/chapter15c-250321120326-e6fd94fc/85/LogisticRegressionDichotomousResponse-ppt-36-320.jpg)
