General - Baldus Study - Memos (Redacted)

    { 

                               



                                          

oho 

    

* fod 

      

ope 

          

1 

        

oes 

        

Joos, 

  

wy 3 ope 
of fe 

                            

    

      

> fond 
4 

              

    

   



      

Lo 

    

   

  

= al 2 

  

Fo) 

                

  

  

                

   

    

    

        

         

     



  

[7-25-86] 

.0005 level, for the race of victim variable after 

simultaneous adjustment for the 39 background variables. [GW 

4: Table 1, line 9 of "Midrange Diagnostic Model"]. A 

backward elimination multiple logistic regression analysis, 

which excludes variable which do not have a statistically 

significant relationship to the death sentencing outcome, 

also produced a race of victim coefficient of 1.45, which was 

significant at the .0003 level. [DB 31B] This means that 

after adjustment for the 39 most important background 

factors, the odds of receiving a death sentence on average 

are 4.3 times higher in white victim cases than they are in 

similarly situated black victim cases. [DB82 R ___ J]. For 

example, if among a group of cases of a similar level of 

culpability, the odds of receiving a death sentence were 1:1 

(or even) for black victim cases, the 4.3 odd multiplier 

indicates that the average defendant in this group with a 

white victim would face odds of 4.3 to 1 instead of the 1:1 

faced by the defendants with black victims. The .0005 level 

 



[7-25-86] 

  

of statistical significance associated with the race of 

victim coefficient indicates that over the long run a 

disparity of this magnitude would be produced by chance 

factors in a study of this size when in fact no active 

disparity existed only one out of 2,000 times. 

The comparable race of victim disparity estimated 

for the CSS core model with "least squares" (linear 

probability) regression procedures, is a coefficient of 

.08, significant at the .0005 level (z=3.8). [GW 4:Table 

1 (line 1 of Midrange Diagnostic Model)] This indicates 

that, on average, defendants with white victims will 

have an 8 percentage point higher risk of a death : 

sentence, than with the average defendant with a black 

victim. 

These findings indicate the overall or average 

effect of the victim's race over all the cases in the 

CSS. In fact, the race of victim effects are 

concentrated in the mid-range of close cases where the 

 



  

[7-25-86] 

facts do not clearly dictate a life or death sentence. 

Among those cases, the risk that a defendant will 

receive a death sentence is 20 percentage points higher 

if the victim is white than if the victim is black. [DB 

90 and GW 8]. Moreover, this is the range of cases in 

which McCleskey's case is found. [GW 8]. 

As indicated above, McCleskey's claim also rests on 

a second study (PRS) that was conducted independently of 

the CSS study just described. Its population of cases 

overlaps somewhat with the CSS and it differs in terms 

of the data sources used, the coding protocols applied, 

and the decision points in the capital sentencing 

process on which it focuses. Specifically, it examines 

the disposition of defendants convicted of murder at 

trial. This allows (a) a separate examination of the 

prosecutorial assiaion to seek a death sentence after a 

murder trial conviction is obtained and the jury 

decision to impose a death sentence, and (b) an analysis 

 



  

[7-25-86] 

of the combined impact of these two decision points. 

The core result of the PRS indicates that after 

adjustment for 31 statistically significant legitimate 

background factors culled from over 160 possible 

explanatory factors, the average defendant with a white 

victim has an 8 percentage point (significant at the .01 

level) higher risk of receiving a death sentence than 

similarly situated defendants with black victims. [DB 

98 line 6]. A logistic multiple regression analysis in 

the PRS which controlled for all statutory aggravating 

factors, estimates that defendants with white victims 

faced odds of receiving a death sentence that were 3 

times higher than similarly situated defendants with 

black victims. [DB 98 line 3]. Also, Professors 

Baldus, Woodworth and Pulaski have recently published a 

reanalysis of the PRS data which identified, with a 

logistic screening procedure, 20 statistically 

significant background variables from over 130 variables 

 



  

|7-25-8 >] 

and 300 interaction terms. This study reports a death 

sentence odds enhancement of 14.5, (significant at the 

.0001 level), for white victim cases that had mitigating 

circumstances, and an odds enhancement of 2.7, (which is 

not statistically significant) in cases not involving 

mitigating circumstances after adjustment for these 

background factors. Davis at 1385-87. The overall 

average effect is an odds enhancement of 4.3, 

significant at the .02 level, a result virtually 

identical to the result in the core model in the CSS. 

The race of victim disparity estimated in the PRS 

reanalysis with a procedure that is comparable to a 

least squares regression coefficient is 8.5 percentage 

points (significant at the .001 level). (Davis at 1401) 

This result is also almost identical to the comparable 

core statistic (8 points) from the CSS. The results of 

this reanalysis of the PRS data also reveal a 

concentration of race of victim effects in the mid-range 

 



  

of cases--again the average is about 20 percentage 

points. pila 1401, 1403. 

It is also of interest that Arnold Barnett of the 

MIT Sloan School of Management undertook a completely 

independent analysis of the PRS data for the National 

Center for State Courts. He read summaries of over 500 

of the cases and strictly on the bases of intuition and 

common sense developed an eighteen-level categorization 

system for identifying cases of similar culpability.. 

The results of this analysis showed an overall race of 

victim effect of 8 percentage points, significant at the 

.006 level. Davis at 1400. This analysis also showed 

race of victim effects of 20 percentage points 

concentrated in the mid-range of cases. 

Supplemental Analysis 

The similarity of the overall race of victim 

effects estimated in the CSS and PRS, and the similarity 

of results of the race victim analyses in three separate 

10 

 



  

[7-25-86] 

analyses of the PRS enhances ones confidence that these 

disparities are not a product of chance but reflect a 

real racial effect operating among a broad range of 

cases in Georgia's Capital charging and sentencing 

system. The design of the PRS + CSS also permit a 

camparison of the race of victim coefficients estimated 

in the two studies for the prosecutorial decision to 

seek a death sentence after a murder trial connection 

has been obtained and the jury penalty trial decision. 

As the following data indicate, there are also striking 

similarities in the results obtained in the two studies 

which focus on these individual decision points. 

| Prosecutorial decision to seek a death sentence 
  

following a murder conviction at trial. 
  

a. OLS Results - controlling simultaneously for 

from 130 to over 230 legitimate variables in 

the file 

(1) C&SS: b= .23 (p=.0001) (DB 95) 

11 

 



[7-25-86] 

  

(2) PRS: b= .16 (p.0004) (DB 95) 

4 Logistic results - controlling for all 

statistically significant legitimate variables 

in the file 

(1) C&SS: b= 1.24(p=.01) (DB96) 

(2) PRS: b= 1.2(p=.01) (DB96) 

Ze Jury Decisions to impose a death sentence 
  

a. Logistic results - controlling for 

statistically significant background factors 

(1) C&SS: b=1.6(p=.01) (DB 97) 

(2) PRS: b=.33(p=.53) (DB 97); the reanalysis 

of the PRS data which includes 

interaction terms shows an average race 

of victim effect of 14.5 percentage 

points, significant at the .005 level 

(Davis at 1403, note 3.) 

Professor Baldus and his colleagues also conducted, for both 

the PRS and CSS, a variety of other multiple regression analyses 

12 

 



  

[7-25-86] 

using different combinations of variables--all in effort to 

determine if the race of victim effects observed in their core 

analyses may be an artifact of the procedures or variables they 

employed. As the following tabulations (which present the 

background factors controlled for and the results obtained) 

indicate, strong and statistically significant race of victim 

effects persist in each analysis. 

1, C&SS 

a. All statutory aggravating circumstances (DB 78). 

(1) OIS: b= .07 (.p=.0014) 

(2) Logistic: b=1.02 (p=.003) 

b. All statutory aggravating circumstances and all 73 

mitigating factors for which data were available 

(DB 79). 

(1) OIS: b=.10 (p=.001) 

(2) Logistic: b=2.1 (p=.001) 

Cc. Background factors limited by various selection 

techniques. 

13 

 



  

[7-25-86] 

(1) Statistically significant factors selected in 

forward stepwise regressions 

(a) Fourteen variables selected in an OLS 

screening procedure. 

((1)) OLS: b= .06 (p=.001) (DB 83:I4) 

((2)) Logistic b= 1.00) p=.01) (DB 

83:IIa) 

(b) Thirteen variables selected in a logistic 

screening procedure 

(1) OILS: b= .06 (p=.001) (DB 83:Ie) 

(2) Logistic: b= .85 (p=.01) (DB 83:IIb) 

Statutory aggravating circumstances and 126 

"factor" variables derived from a factor analyses 

of over 200 legitimate case characteristics 

(a) OILS: b= .07 (p=.01) (DB 83:Ib) 

All (230+) legitimate variables in the file 

controlled for simultaneously 

(1) OILS: b= .06 (p=.02) (DB 83:1a) 

14 

 



  

2 PRS. Six separate analyses involving from five to 160 

legitimate background factors. 

(a) OLS race of victim coefficients ranged from b= 

.08 to b= .10 (all significant at or beyond 

the .02 level) (DB 98) 

Professor Baldus and Woodworth also conducted a series of 

"worst case" analyses which addressed concerns raised by 

respondent about the coding of "unknowns" and the imputation of 

missing data on certain racial and procedural variables. 

Specifically these analyses either deleted cases with missing data 

or coded unknowns in a way that would tend to supress the 

magnitude of the coefficient estimates for the race of victim 

  

variable. (R___, : 73 The results of the worst case 

analysis conducted with the CSS core 39 variable model produced a 

least squares race of victim coefficient of .06 significant at the 

00° level (2=3.56) . [GW 4: Table 1 (line 3 of Midrange Diagnostic 

Model]. 

15 

 



  

[7-25-86] 

[Jack, we also did the very analyses which the 4th Circuit in 

Bazemore said should have been done i.e. we adjusted for "county 

to county differences in salary increases". We adjusted for 

Judicial circuit differences in death sentencing rates and the 

race of victim effects persisted. The results are referred to in 

the second full paragraph on p. 103 of our report dated July 29, 

1983. I'm looking to see if I testified about this adjustment]. 

Professors Baldus and Woodworth also conducted a series of 

regression analyses on subsets of cases deemed by District Judge 

Owen Forrester to be the most likely candidates for a death 

sentence. Judge Forrester defined the populations of cases in 

terms of the aggravating factors present, the absence of 

mitigating factors, and the strength of the evidence. The results 

partially presented below show persistent and substantial race of 

victim effects. 

1. OIS results 

a. Samples of 238 and 354 cases 

16 

 



[7-25-86] 

  

(1) Controls for the 39 conceptually and 

statistically most important background 

variables 

(a) b= 413+ .15 (p=.05) 

h
c
 
A
A
R
 

(2) Controls for 13/16 statistically significant 

background variables 

(a) b= .17 & .15 (p=.01) (David C. BRaldus 

Affidavit, 9/15/83 p.8) 

2. ILogistic results 

IN
 

T
R
S
 
R
A
A
T
 

1 

a. Samples of 354 cases 

(1) Controls for the 39 conceptually and 

statistically most important background 

e
y
 

| R
Y
 

[L
J 

se
me
 

variables, b=1.39 (p=.03) 

(David C. Baldus Affidavit 9/15/83 p. 9). 

Finally, Professor Baldus and his colleagues used a non- 

regression procedure to identify cases of similar culpability. 

I 
g
e
 

ja 
A
 

They read narrative summaries of the cases and classified them ’ 

according categories of factual comparability. The results of 

17 

 



  

[7-25-86] 

this analysis at DB 86 show large race of victim effects in cases 

with the B2 statutory aggravating circumstance present. Moreover, 

Professor Baldus and his colleagues conducted a subsequent more 

extensive reanalysis of the armed robbery cases in the PRS data 

set after classifying them with a five level a priori measure of 

case culpability. After adjustment for case culpability, the 

results showed an average race of victim effect of 25 percentage 

points, significant at the .01 level. Stetson L. Rev. 133, 178- 
  

80, 196 Table 7 n.Z2. 

Other recently published studies of Georgia's Capital 

Sentencing System suggest that the patterns observed in the PRS 

and the CSS have continued into the 1980's. In 1984 Samuel Gross 

and Robert Mauro used logistic regression procedures to estimate 

race of victim effects in death sentencing among all Georgia 

homicides reported to the FBI from 1/1/76 through 12/31/80, after 

adjustment for five legitimate case characteristics. They report 

a logistic race of victim coefficient of 1.97 (p=.001). Gross & 

Mauro at 78. This compares with the logistic regression 

18 

 



[7-25-86] 

  

coefficient of 1.45 in the core 39 variable CSS model. DB 82, & 

GW 4 (table 1). 

Baldus, Pulaski and Woodworth also report an unadjusted 28 

percentage point race of victim disparity among Georgia 

| contemporaneous felony cases reported to the FBI for the period 

1977-85, which is significant at the .001 level. Baldus, Pulaski   
and Woodworth, 15 Stetson IL. Rev. 134, 163 (line 10 Col. B). This   

figure compares with the 24 point and 12 point disparities 

respectively among CSS contemporaneous felony cases, involving 

black and white defendant's for the 1973-79 period. (DB | 

86:ICols.D&I.) 

Finally, Baldus, Pulaski and Woodworth conducted a state-by- 

state race of victim analyses among cases involving 

contemporaneous offenses during the 1977-85 period and estimated 

an average nationwide race of victim effect of 6.4 percentage 

points, significant at the .0001 level. Baldus, Pulaski and 

Woodworth, 15 Stetson L. Rev. 133 164 n 3 (1986).   

e
i
 

s
v
 

EY 

19 

 



  

[7-25-86] 

3. The plausability that the race of victim results were 

biased by omitted variables. 

The studies on which McCleskey relies adjust simultaneously 

and in various combinations for more background available than any 

previously conducted sentencing study. They include every 

plausible legitimate variable on which data were available in the 

records of the Supreme Court and the Georgia Board of Pardons and 

Paroles. Respondent and the 11th circuit suggest that omitted 

variables may have biased the results. Several of these omitted 

factors such as the defendants "looks" and "clothes" are legally 

questionable and generally ummeasurable. McCleskey at 899. 
  

(Bazemore implies that the omission of umeasurable case 

characteristics is acceptable). Also there is no basis in the 

record or in the 11th circuit's opinion for believing that these 

omitted variables are both important determinants of the death 

sentencing process and correlated with the race of the victim. 

Cc. Conclusion 

20 

CL
AR
E 

COR
EY 

E
C
E
 

4 
FO

RE
 
e
w
 

wr
 

oh 
RE

EL
 

| 
JY
RL
 
R
E
E
 

 



  

[7-25-86] 

The record includes on "impressive array" of evidence to 

indicate that Georgia engages in a pattern or practice of race of 

victim discrimination in its capital sentencing process. The race 

of victim effect is substantial in terms of the average impact it 

has on the likelihood that the average defendant with a white 

victim will receive a death sentence. The impact is very large in 

the mid range of cases. The high levels of statistical 

significance of virtually all of the race of victim disparities 

suggest that those estimates are not a product of chance; so also 

does the striking consistency of the results from the CSS and PRS 

and the results of the three different analyses of the PRS. 

Moreover, the consistency of results produced within each study in 

a wide variety of different analyses, indicates that the race of 

victim effects are not a function of the type of regression 

analyses used or the combinations of variables employed. Finally 

it is quite unlikely that any variable omitted from the study is 

both an important determinant of death sentencing outcomes in 

21 

 



[7-25-86] 

  

Georgia and correlated with the race of victim, conditions both of 

which must exist before an omitted variable could bias the 

estimated race of victim results. 

22 

 



  

[7-25-86] 

II: THE MAGNITUDE OF THE RACE OF VICTIM EFFECTS 

I. Looking at the average effects 

A. There are two ways to measure the disparities. One 

is the arithmetic difference between the death 

sentencing rates in the white and black victim 

cases; the other is with multipliers (a ratio of 

these rates (e.g. 2.0(.50/.25) or an odds 

multiplier of the type used in logistic regression 

analyses) 

We think you should draw attention away from 

the 6% disparity by focusing on the logistic 

measure derived from the 39 variable core 

regression. However, in dealing with the 6 point 

arithmetic disparity it can be compared, as you 

have done earlier, to the overall death sentencing 

rate of .05 among all cases; the white victim case 

disparity rate is 120% (.06/.05) higher than 

 



  

[7-25-86] 

average rate for all cases. (DB 59) It is ; 

probably more realistic to talk of the average rate 

among those convicted of murder, which is .12 

(115/1004). Still the race of victim disparity is 

50% (.06/.12) of the average rate. 

Also its useful to look at the impact of the 

victim's race at the two most important steps in 

the process: the prosecutodal decision to seek a 

death sentence and the jury decision to impose one. : 

(DB 54) The average rate at which death is sought | 

in the C85 among those convicted of murder at trial 

is .31 (238/762). The average race of victim 

disparity in the prosecutial decisions to seek a 

penalty trial among these cases is between 11 

points (Davis L.R. at 1403) and 23 points (DB 95). 

Thus these race of victim disparities are from 35% 

(.11/31) to 75% (.23/.31) of the average rates at 

which prosecutors seek death after trial. 

 



[7-25-86] 

  

For juries, the average death sentencing rate 

is .53 (128/238) (DB 59), while the race of victim 

disparity estimated for the jury decisions ranges 

from .15 and .22 (DB 95) to 14.5 (Davis at 1403). 

This makes the race of victim disparity from 28% 

(.15/53) to 42% (.20/53) of the average death 

sentencing rate. 

You'll note that the race of victim disparity 

at these two decision points (prosecutor and jury) 

is about twice as large as it is when you examine 

the combined effects of both decision points i.e. 

8.5 points. This is a property of the "difference" 

measure — the overall disparity will generally be 

smaller than the disparities observed at each point 

in the process that contribute to the observed 

overall disparity. It is analogous to taking a 

fraction of a fraction e.g. 1/2 x 1/2 = 1/4. 

 



  

[7-25-86] 

You obtain the opposite effect when you use 

the ratio of rates as the measure of the effects of 

several decision points in a process. If the ratio 

measure at each of the stages is the same, the 

measure of their combined effects will be larger 

than the measure at a given stage in the process. 

Its analogous to compounding interest. For 

example, a 2:1 white victim/black victim ratio at 

both the prosecutorial and jury level will in 

theory result in a 4 to 1 (white victim/black 

victim) ratio when the combined effects of the two 

decision points are analyzed. The/odds multiplier 

has neither of these effects. That is) as a measure 

of overall combined effects of several decision 

points, it does not substantially enhance or 

diminish the effects observed at the separate 

stages in the process: 

 



  

[7-25-86] 

For this reason, we believe the logistic odds 

multiplier is the best measure of overall impact, 

on Thin 
supplemented by the measures based on the authentic 

' 

d¢fference and ratio measures at different 

culpability levels in tabulations like DB 90. This 

may help avoid the problem of the small sounding 

"six percent" or "six percentage point" measure of 

effect. 

II. Focusing on the mid range 

A. If you think the figure we used at trial (GW 8) is 

hard to understand, we recommend supplementing it 

with references to tables which show the large 

impacts in the mid range which are consistent with 

the "liberation" hypothesis. We have already made 

reference to these tables in our earlier 

discussion. 

In the record, DB 90 (table 42) shows the mid 

range effect in a way that is not hard to 

 



[7-25-86] 

  

understand. McCleskey is at level 5 in DB 90, 

although we don't think this is in the record. 

These effects are also quite apparent in Davis at 

1401 (overall effects in our reanalysis of PRS 

data)", Davis at 1403 (prosecutor and jury race or 

victim effects in our reanalysis of PRS); and Davis 

at 1400 (overall effects in Barnett's reanalysis of 

PRS). 

ITI. The magnitude of the race of victim regression 

coefficient compared with coefficients for other 

legitimate variables. 

A. It is appropriate to compare coefficients if they 

are coded the same way. In this case the race of 

victim variable is a "yes-No" variable as are all 

of the other variables in DB 82. From a 

statistical standpoint, in the CSS the race of 

victim has about the same importance, as an 

aggravating circumstance, as the variables listed 

 



  

[7-25-86] 

as items 6 through 11 in DB 82. For example, the 

presence of a white victim increases the 

culpability level of the case about the same degree 

as having two victims or having a stranger victim. 

A comparison of the coefficients estimated for 

the PRS reanalysis (Davis at 1385-87) shows the 

impact of the race of victim in the absence of 

mitigating circumstances (b= 2.66) to be as large 

or larger than almost all the other legitimate 

variables in the analysis. 

IV. The number of death STE but for race of victim 

drill 

A. These measures involves a comparison of the number 

of sentences actually imposed with the number that 

would have been imposed if (a) the white victim 

cases had been sentenced at the black victim rate, 

and (b) the white victim cases had been sentenced 

at the black victim rate. Simple calculations with 

 



  

[7-25-86] 

the data in DB 90 show that if the white victim 

cases had been sentenced at the same rate as the 

black victim cases, there would have been 

53%(62/116) fewer death sentences; also if the 

black victim cases had been sentenced at the same 

rate as the white victim cases there would have 

been 38% (44/116) more death sentences. 

Similar calculations in our reanalysis of the 

PRS data suggest that if the white victim cases had 

been treated like the black victim cases, there 

would have been a 55% fewer death sentences imposed 

in armed robbery cases (Stetson at 197) and among 

all cases there would have been 30% fewer a 

sentences imposed. (Stetson at 199). 

Another measure of the impact of the race of 

victim is the discussion in part III of this memo 

which identifies the categories of cases in which 

there is a greater than .50 probability that any 

 



[7-25-86] 

  

given death sentence is a product of race of victim 

discrimination. 

 



[7-25-86] 

  

III. The likelihood that McCleskey's sentence was in whole or 

in part the product of race of victim discrimination. 

The 11th Circuit said that our data were 

insufficient to show that McCleskey's "sentence was 

determined by the race of his victim, or even that the 

race of his victim contributes to the imposition of the 
  

penalty in his case" McCleskey v. Kemp at 898 (emphasis 

added). Since the hearing we have developed a 

statistical argument which addresses these questions 

with more precision than we presented to the court. The 

argument is presented in our Stetson article pp 206-09 

(attached) and we apply it here to the data in DB 90 

(table 42) and the data in the Davis article. 

In DB 90, McCleskey is in level 5. In a system 

that treated white victim cases like black victim cases 

you'd expect 4.42 death sentences among the white victim 

cases. Thus there are 4.58(9-4.42) sentences in the 

1 

 



  

[7-25-86] 

black defendant/white victim category at level five that 

are the likely result of race of victim discrimination. 

This means that there is a .51 probability (4.58/9) that 

McCleskey's sentence is among those 4.58 which were the 

product of race of victim discrimination. 

In fact applying this analysis to all of the cases 

in table 42, there is reason to believe that for any 

given death sentence in levels 3-5 of Col. C there is a 

greater than .51 liklihood that it is the product of 

race of victim discrimination. The rule of thumb is 

that if the ratio of the death sentence rate in the 

white victim cases exceeds the rate in the black victim 

cases by a factor of 2 or more, the probability that 

race caused a particular death sentence to be imposed 

will be greater than .50. Applying this method, the 

probability that race of victim was the determintive 

factor is greater than .51 for all of the death 

 



  

[7-25-86] 

sentences in the Black Def./White Vic category (Col B) 

at levels 3-5. 

When the same method is applied to our reanalysis 

of the PRS data we find McCleskey in level 5 in the 

overall analysis. Davis at 1401. Here the WV rate is   

1.0(18/18) and the BV rate is .56(5/9) meaning 8 of the 

white victim cases were probably a product of race of 

victim discrimination. This produces a probability of 

.44 (8/18) that McCleskey was one of the cases in which 

the race of victim was the determinative factor. If you 

look at all the white victim death cases in table 3 

(Davis at 1401), those at levels 3 + 4 had a risk close 

to McCleskey's while at levels 1 + 2 there is a risk 

greater than .50 that the death cases were a product of 

racial discrimination. 

For the prosecutor analysis (Davis at 1403), 

McCleskey is in level 4 (WV is 95(38/4) and B V is 

.71(12/17) . The probability that he advanced to a 

 



  

[7-25-86] 

penalty trial because of his victim's race is 

.28(14/50). Only in levels 1 & 2 does the risk appear 

to exceed .50. 

In the jury analysis, the white victim rate is 1.0 

(19/19) versus .83(5/6) for the black victim cases; 

McCleskey is in level 4 where the probability his 

sentence was imposed because of his victim's race is 

.16(3/19). In none of the levels does the risk that the 

victims race was determinative exceed .50. 

Finally when we turn to Fulton County police victim 

cases we see a similar pattern (DB 117). Five white 

victims and 2 black victims among the cases with a level 

of aggravation comparable to McCleskey; only one death 

sentence was imposed and it was in a white victim case. 

On the basis of all this evidence we would hardly 

assert that we have proven that McCleskey's sentence was 

the product of race of victim discrimination. We would 

feel confident in saying however, that there is a 

 



[7-25-86] 

  

reasonably strong probability that the victim race 

"contributed" to the imposition of the penalty in his 

 



   
      

. 274 

refer- 

licitly 

‘ovide 

Just 

e can 

Yick 

1064, 

iich— 

imate 

ot be 

made 

quite 

arans, 

ether 

based 

some 

veter- 

> Dis- 

wctors 

§ 23, 
Nash- 

2ights 

refer- 

© gen- 

ors a 

© poli- 

cerns 

refer- 

ies of 

evita- 

pellee 

with 

little 

ance, 

inato- 

stitu- 

refer- 

nder- 

‘avor- 

i into 

noply 

  

442 U.S. 278 PERSONNEL ADM'R OF MASSACHUSETTS v. FEENEY 

2) iC Fed / n, Bo 5 hs 

2295 
Cite as 99 S.Ct. 2282 (1979) 

federal laws that have prevented all but a 
handful of women from becoming veterans. 

There are two serious difficulties with this 

argument. First, it is wholly at odds with 

the District Court's central finding that 

Massachusetts has not offered a preference 
to veterans for the purpose of discrimina- 

ting against women. Second, it cannot be 

reconciled with the assumption made by 

both the appellee and the District Court 

that a more limiting hiring preference for 

veterans could be sustained. Taken togeth- 

er, these difficulties are fatal. 

To the extent that the status of veteran 

is one that few_|women have been enabled 

to achieve, every hiring preference for vet- 

erans, however, modest or extreme, is in- 

herently gender-biased. If Massachusetts 
by offering such a preference can be said 

intentionally to have incorporated into its 

state employment policies the historical 

gender-based federal military personnel 

practices, the degree of the preference 

would or should make no constitutional dif- 

ference. 'Invidious discrimination does not 
become less so because the discrimination 

accomplished is of a lesser magnitude? 

Discriminatory intent is simply not amena- 
ble to calibration. It either is a factor that 

has influenced the legislative choice or it is 

not. The District Court’s conclusion that 

the absolute veterans’ preference was not 
originally enacted or subsequently reaf- 

firmed for the purpose of giving an advan- 

tage to males as such necessarily compels 

the conclusion that the State is intended 

nothing more than to prefer “veterans.” 
Given this finding, simple logic suggests 

that an intent to exclude women from sig- 

nificant public jobs was not at work in this 
law. To reason that it was, by describing 

the preference as “inherently nonneutral” 

or “gender-biased,” is merely to restate the 

fact of impact, not to answer the question 

of intent. 

To be sure, this case is unusual in that it 

involves a law that by design is not neutral. 

23. This is not to say that the degree of impact 

is irrelevant to the question of intent. But it is 

to say that a more modest preference, while it 

might well lessen impact and, as the State 

The law overtly prefers veterans as such. 

As opposed to the written test at issue in 

Davis, it does not purport to define a job-re- 

lated characteristic. To the contrary, it 

confers upon a specifically described 

group—perceived to be particularly deserv- 

ing—a competitive headstart. But the Dis- 

trict Court found, and the appellee has not 

disputed, that this legislative choice was 

legitimate. The basic distinction between 

veterans and nonveterans, having been 

found not gender-based, and the goals of 

the_jpreference having been found worthy, 

ch. 31 must be analyzed as is any other 

neutral law that casts a greater burden 

upon women as a group than upon men as a 

group. The enlistment policies of the Arm- 

ed Services may well have discrimination on 

the basis of sex. See Frontiero v. Richard- 

son, 411 U.S. 677, 93 S.Ct. 1764, 36 L.Ed.2d 

583; cf. Schlesinger v. Ballard, 419 U.S. 498, 

95 S.Ct. 572, 42 L.Ed.2d 610. But the histo- 

ry of discrimination against women in the 

military is not on trial in this case. 

2 

The appellee’s ultimate argument rests 

upon the presumption, common to the crim- 

inal and civil law, that a person intends the 

natural and foreseeable consequences of his 

voluntary actions. Her position was well 

stated in the concurring opinion in the Dis- 

trict Court: 

“Conceding that the goal here 

was to benefit the veteran, there is no 

reason to absolve the legislature from 

awareness that the means chosen to 

achieve this goal would freeze women out 

of all those state jobs actively sought by 

men. To be sure, the legislature did not 

wish to harm women. But the cutting- 

off of women’s opportunities was an inev- 

itable concomitant of the chosen 

scheme—as inevitable as the proposition 

that if tails is up, heads must be down. 

Where a law's consequences are that in- 

argues, might lessen the effectiveness of the 

statute in helping veterans, would not be any 

more or less ‘neutral’ in the constitutional 

sense. 

(ZH Cor, 
2 Cnse i 

_J278 

  

    
  

  

  

  

  

  

  
  

  

    

     

   
    

   

     

   
   
   
   
   
    

    

   
    
    

  

    

  

     

   
   

    

    

  

   
    
   
   
   
    

   

  

      
  

       



F
H
 

B
I
 

A
 

H
R
 

AR
 

508 749 FEDERAL REPORTER, 2d SERIES 

Thomas C./BIBBS, Jr. Appellant, 

“vy, : 

John BLOCK, Secretary United States 

Department of Agriculture, Appellee. 

No. 83-1942. 

United States Court of Appeals, 

Submitted April 12, 1984. 

Decided Dec. 5, 1984. 

Rehearing En Banc Granted 

March 8, 1985. 

3. Civil Rights ¢=43 

When race is shown to have been a 
discernible factor in the employment deci- 
sion, the Mt. Healthy “same decision” test 
is inappropriate under the Burdine princi- 
ples applicable to a Title VII employment 
discrimination suits and once a plaintiff 
proves his claim of unlawful discrimination 
by showing that a discriminatory reason 
for promotion or selection more likely than 
not motivated the employer nothing more is 
required to establish Title VII liability and 
whether race factor was minor or major is 

irrelevant. Civil Rights Act of 1964, 

§8 701-718, as amended, 42 US.CA. 
Employee brought” Title VII “action... §§2000e to 2000e-17. 

against employer charging race and age 
SyserivimatiomSm promotions denial, i The 

United States District Court for the West- 
ern District of Missouri, Howard F. Sachs, 
J., denied recovery, and employee appealed. 
The Court of Appeals, Lay, Chief Judge, 
held that on finding that race was a discer- 
nible factor in the promotion decision the 
district court erred reversibly by applying 
an improper standard of causation to the 
mixed motive question when it required 
plaintiff to show that even eliminating the 
race factor he would have been promoted. 

Vacated and remanded with direction. 

1. Civil Rights 9.10 

Fact that work force in shop was ra- 

cially integrated did not insulate employer 

from liability for intentional discrimination 

on basis of race against individual employ- 

ee seeking promotion. Civil Rights Act of 

1964, 88 701-718, as amended, 42 U.S.C.A. 
§§ 2000e to 2000e-17. 

2. Civil Rights 43 

A plaintiff in a disparate treatment 

action under Title VII need not prove that 

the sole reason for the adverse employment 

decision was discrimination and plaintiff 

need only show that race was a “but for” 

reason. Civil Rights Act of 1964, §§ 701- 
718, as amended, 42 U.S.C.A. §§ 2000e to 
2000-117. 

* The HONORABLE THOMAS E. FAIRCHILD, 
Senior Circuit Judge for the United States Court 

4. Civil Rights ¢44(4) 

Federal Courts ¢=858 

Where race was found to be a discerni- 

ble factor in refusal to promote plaintiff 

black employee the plaintiff, who brought 

Title VII suit, was entitled to judgment 

under the Burdine test and it was reversi- 

ble error to apply the Mt. Healthy “same 

decision” standard of causation to the 

mixed motive question and require plaintiff 

to prove that he would have been selected 

if race had been disregarded. Civil Rights 

Act of 1964, 88 701-718, as amended, 42 

U.S.C.A. §§ 2000e to 2000e-17. 

Gwendolyn M. Wells, Kansas City, Mo. 

for appellant. 

Kenneth Josephson, Kansas City, Mo., 
for appellee. 

Before LAY, Chief Judge, FAIRCHILD," 
Senior Circuit Judge, and McMILLIAN, 
Circuit Judge. 

LAY, Chief Judge. 

Thomas Bibbs brought this action under 
Title VII of the Civil Rights Act of 1964, as 
amended, 42 U.S.C. §8 2000e to 2000e-17 

of Appeals for the Seventh Circuit, sitting by 
designation. 

XX XX 
v  



  

5 v2 
BIBBS v. BLOCK 

511 
Cite as 749 F.2d 508 (1984) 

ason was not the true reason for defend- 
Re promotion decision. 

[2] Certain well-established principles 
govern our review of the district courts 

decision. First, under the teaching of Bur- 

dine, 450 U.S. at 253, 101 S.Ct. at 1093, the 
wyltimate burden of persuading the trier of 
fact that the defendant intentionally discri- 
minated against the plaintiff remains at all 
times with the plaintiff.” Second,-as the 
district court acknowledged, a plaintiff in a 
disparate treatment action under Title VII 

peed not prove the sole reason for the 
employment decision was the discrimina- 

- tion; plaintiff need only show race was a 
#put for” reason. McDonald v. Santa Fe 
Trail Transportation Co., 427 U.S. 273, 
282 n. 10, 96 S.Ct. 2574, 2580 n. 10, 49 

L.Ed.2d 493 (1976). The district court also 
cited Mi. Healthy City School District 
Board of Education v. Doyle, 429 U.S. 
274, 97 S.Ct. 568, 50 L.Ed.2d 471 (1977), 
which did not involve Title VII but present- 
ed the issue whether the exercise of a 
constitutionally-protected right was the ba- 
sis for a decision not to hire an untenured 
teacher. In Mt. Healthy, the court adopted 
a rule of causation, sometimes called the 
“same decision” ‘test; for cases involving 
employment decisions based on protected 
conduct. In such cases, "when the plaintiff 
has shown his protected activity was a 
“substantial” or “motivating” factor, un- 
like Burdine, the burden shifts and the 
employer must show by a preponderance of 
evidence “that it would have reached the 
same decision * * * even in the absence of 
the protected conduct.” 7d. at 287 ,96.8.Ct. 
at 2582. 

L In a mixed-motive context such as this, the 
challenged employment decision presumably 
was motivated by both pretextual (unlawful) 
and nonpretextual (lawful) reasons. Because 
“pretext” means that the true reason behind the 
decision is in fact an unlawful reason, not “part- 
ly lawful” and “partly unlawful” as in a dual 
motive case, some commentators have noted the 
difficulty of analyzing mixed-motive cases in 
pretext terms. See Brodin, The Standard of Causation in the Mixed-Motive Title VII Action: 
A Social Policy Perspective, 82 Colum.L.Rev. 292, 301 n. 40 (1982) (“the pretext approach is based On an assumption of single-motive decisionmak- ing”). 

  

  

The district court recited thesconfusion of 
our own cases and those of other circuits 2 
We also acknowledge that the Supreme 
Court has not addressed the question of 
the proper standard of causation in mixed- 
motive cases under Title VII." The district 
court resolved the dilemma by finding that 
“racial considerations probably did play a 
minor role in the selection process, through 
the influence of Tresnak, but that plaintiff 
would not have been selected for the posi- 
tion even if his race had been disregarded.” 
While finding that race was a “discernible 
factor,” the district court found “race was 
not a determining factor in the decision to 
promote Laube rather than plaintiff.” 

[3] At first blush, the “same decision” 
reasoning is attractive and would require 
us simply to decide from the record wheth- 
er the factual finding alone is clearly erro- 
neous. However, when race is shown to 
have been a discernible factor in the em- 
ployment decision, as the district court 
found below, we conclude the same-decision 
test is inappro riate under the principles of ; 
Burdine. Here, plaintiff has done more 
than put forth a prima facie case of dis- 
crimination, he has successfully proven 
thatwrace was a discriminatory factor in his 
employer’s refusal to promote him. ther 
than requiring proof that‘race was a “‘sub- 
stantial” or ‘determining’ factor in the 
decision, we find that a plaintiff proves his 
claim of unlawful discrimination by show- 
ing that “a discriminatory reason more like- 
ly [than not] motivated the employer.” 
Burdine, 450 U.S. at 256, 101 S.Ct. at 1095. 
Nothing more is required of a plaintiff to 

2. See Tribble v. Westinghouse Elec. Corp., 669 
F.2d 1193, 1197 (8th Cir.1982) (determining fac- 
tor); Nanty v. Barrows Co., 660 F.2d 1327, 1333 
(9th Cir.1981) (same-decision test at relief 
stage); Williams v. Boorstin, 663 F.2d 109, 117 
(D.C.Cir.1980) (same-decision test at liability 
stage); Satz v. ITT Fin. Corp., 619 F.2d 738, 746 
(8th Cir.1980) (a factor); Marshall v. Kirkland, 
602 F.2d 1282, 1289 (8th Cir.1979) (motivating 
factor); Cleverly v. Western Elec. Co., 594 F.2d, 
638, 641 (8th Cir.1979) (determining factor). 

{ 

    

    

  

   

    

   
   
   
   
   

    

   
   
   
   

      

   

    

   

    

   
   
   

  

   

    

   
   
    
   

  

   

              

   

   

          

    

    

  

    

   
   

   



  

  

   

  

    
    

   
    

  

    

  

   

  

   

512 

establish liability under Title VII. Once 
the trier of fact has found that race was a 
factor influencing the decision, we find it 
error to attempt to quantify race as a mi- 
nor factor. Under the factual record 
presented here, once race was found to be 
a “discernible factor at the time of the 
decision,” the additional conclusion that it 
was a minor factor is irrelevant to the Title 
VII analysis. As stated in Personnel Ad- 
ministrator v. Feeney, 442 U.S. 256, 2717, 
99 S.Ct. 2282, 2295, 60 L.Ed.2d 870 (1979): 
“Discriminatory intent is simply not amena- 
ble to calibration: It either is a factor that 
has influenced the legislative choice or it is 
not.” 3 

We find it inherently inconsistent to say 
that race was a discernible factor in the 
decision, but the same decision would have 
been made absent racial considerations. 
Thus, we think that once race is shown to 
be a causative factor in the employment 
decision, it is clearly erroneous to find that 
racial considerations did not affect the out- 
come of the decision. The analysis could 
be reversed to say that once it is shown 
that the same decision would have been 
made absent racial considerations, then it is 
clearly erroneous to find that race was a 
causative factor in the employment deci- 
sion. However, we affirm the district 
court’s finding that race influenced the em- 
ployment decision and do so for several 
reasons. First, a finding that race influ- 
enced the decision is required by Burdine, 
whereas the Supreme Court has never 
sanctioned use of the same-decision test in 
Title VII cases. Moreover, we find that 
the record supports a finding that race 
entered into the decision: the record re- 
veals that Tresnak, who was shown to be 
racially biased, was the key figure in the 
promotion decision; the district court 
found, and we agree, that the selection 
committee lacked credibility in their testi- 
mony regarding the selection process; and 
the selection procedure was suspect due to 
its subjective nature. Thus, we reject the 

3. See also Village of Arlington Heights v. Metro- 
politan Hous. Dev. Corp., 429 U.S. 252, 265, 97 
S.Ct. 555, 563, 50 L.Ed.2d 450 (1977) (“Rarely 
can it be said that a legislature or administrative 

  

749 FEDERAL REPORTER, 2d SERIES 

  

   

unnecessary finding by the district court 
that “plaintiff would not have been select- 
ed for the position even if race had been 
disregarded,” as irrelevant and clearly er- 
roneous. 

The language of Title VII itself recog- 
nizes the broad purpose of eliminating con- 
sideration of race from employment deci 
sions. “Title VII prohibits all discrimina- 
tion in employment based upon race, sex, 
and national origin. ‘The broad, overriding 
interest, shared by employer, employee, 
and consumer, is efficient and trustworthy | 

    

  

workmanship assured through fair and ..., 
neutral employment and personnel deci 
sions.” ” Id. at 259, 101 S.Ct. at 1096 
(emphasis added) (citing McDonnell Doug- 
las, 411 U.S. at 801, 93 S.Ct. at 1823). As 

Justice Marshall has stated: 

[I]t is important to bear in mind that 

Title VII is a remedial statute designed 
to eradicate certain invidious employ- 
ment practices. The evils against which 

it is aimed are defined broadly: “to fail 

... to hire or to discharge ... or other 
wise to discriminate ... with respect to 

.. compensation, terms, conditions, or 

privileges of employment,” and “to limit, 

segregate, or classify ... in any way 

which would deprive or tend to deprive 
any individual of employment opportuni- 

ties or otherwise adversely affect his 
status.” 42 U.S.C. § 2000e-2(a) (1970 

ed., Supp.V) (emphasis added). : 

International Brotherhood of Teamsters 
v. United States, 431 U.S. 324, 381, 97 
S.Ct. 1843, 1878, 52 L.Ed.2d 396 (1977) 
(Marshall, J., concurring and dissenting). 

[41 We find that the trial court erred in 
requiring plaintiff to prove not only that 
race was a factor in the employment deck 
sion, but to assume the additional burden 
of proving the hypothetical fact—that he 
would have been selected for the position if 
his race had been disregarded. Cf T oney 

v. Block, 705 F.2d 1364, 1370 n. 3 (D.C.Cir. 
1983) (Tamm, J., concurring). The practical 

body operating under a broad mandate made a 
decision motivated solely by a single concern. 

or even that a particular purpose was the ‘domi- 
nant’ or ‘primary’ one.”).    

  

  

 



  

    
CHRISTOFFERSEN v. UNITED STATES 513 

Cite as 749 F.2d 513 (1984) 

fect of placing the burden on plaintiff to 

isfy the Mt. Healthy same-decision stan- 

ird is to require plaintiff to disprove de- 

‘ondant’s allegations of its subjective in- 

‘ent. Such an onerous burden is unreason- 

‘ble in a case in which race has been shown 

‘0 be a factor in the decision and when the 

jofendant has superior access to proof of 

ts actual motivation.! Although proof of 

tual motivation would be within a defend- 

t's knowledge, we find it inappropriate, 

ander the principles of Burdine governing 

Title VII cases, to follow Mt. Healthy and 

shift the burden to the defendant to show 

that plaintiff would not have been promot- 

od even if his race had not been considered. 

See also Givhan v. Western Line Consoli- 

dated School District, 439 U.S. 410, 99 

3.Ct. 693, 58 L.Ed.2d 619 (1979); Wright 

Line, 251 N.L.R.B. 1083 (1980), enforced, 

662 F.2d 899 (1st Cir.1981), cert. denied, 

455 U.S. 989, 102 S.Ct. 1612, 71 L.Ed.2d 848 

(1982) (NLRB adopts the Mt. Healthy test 

to determine liability in § 8(a)(@3) cases). 

Mt. Healthy exacts a distinctly different 

standard relating to recovery and burden 

of proof in constitutionally-protected con- 

duct cases than does Burdine which ap- 

plied Title VII. This leads then to the 

Burdine conclusion that to establish liabili- 

ty in an individual disparate treatment case 

under Title VII, plaintiff need only show 

that racial reasons more likely than not 

influenced the employment decision. If 

plaintiff has established this ultimate fact, 

then he has satisfied the “but for” require- 

ment of McDonald. 

For the reasons stated above, the judg- 

ment for defendant is vacated and the case 

is remanded to the district court to enter a 

Judgment in favor of plaintiff and to con- 

sider the necessary remedy to make plain- 

tiff whole. Costs on appeal are awarded to 

Bibbs. Counsel for Bibbs is directed to 

submit a detailed and specific request for 

attorney’s fees for this appeal. 

  

  

4. Cf. Teamsters, 431 U.S. at 360 n. 45, 97 S.Ct. at 
1867 (“[T]he employer was in the best position 

to show why any individual employee was de- 
nied an employment opportunity. Insofar as 

the reasons related to available vacancies or the 

employer's evaluation of the applicants qualifi- 

Arthur L. CHRISTOFFERSEN and 

Theresa A. Christoffersen, 

Appellees, 

Vv. 

UNITED STATES of America, 

Appellant. 

No. 84-1420. 

United States Court of Appeals, 

Eighth Circuit. 

Submitted Oct. 10, 1984. 

Decided Dec. 5, 1984. 

Rehearing and Rehearing En Banc 

Denied Jan. 3, 1985. 

Taxpayers appealed from a judgment 

of the United States District Court for the 

Northern District of Iowa, Edward J. 

McManus, Chief Judge, 578 F.Supp. 398, 

denying their claim for refund. The Court 

of Appeals, Lay, Chief Judge, held that 

contract under which taxpayers were given 

option in the year 2021 to purchase annuity 

and which gave them the right to select the 

mutual fund in which to invest was not 

“annuity” qualifying for deferred tax treat- 

ment where, although life insurance compa- 

ny maintained shares in the mutual fund in 

its name, taxpayers maintained right to 

withdraw upon seven days’ notice and right 

to choose another fund in which to invest 

and thus remained beneficial owners of the 

funds. 

Reversed. 

1. Internal Revenue 3121 

Contract under which taxpayers were 

given option in the year 2021 to purchase 

annuity and which gave them the right to 

select the mutual fund in which to invest 

was not “annuity” qualifying for deferred 

cations, the company's records were the most 

relevant items of proof. If the refusal to hire 

was based on other factors, the employer and its 

agents knew best what those factors were and 

the extent to which they influenced the decision- 

making process.”). 

    

    

  

   
   

              

        

        

  

        
  

    

   

        

   

        

   
   

    

   

  

   

   

  

   

    

   

  

   
   

        

   

    

   
    

  

   

  

   

  

       

B
—
—
—
—
—
—
—
—
—
—
—
—
_
—
 

=
 

i 
“ 
a
 

cn
 

i. 
—-
— 

pi
 

          
              



NANTY v. BARROWS CO.      1327 
Cite as 660 F.2d 1327 (1981) 

the due process rights of administrative 
claimants. No showing has been made that 
the time limits set by the Second Circuit in 
Lee have proved unreasonably burdensome. 
We hold that in cases involving vehicles of 
transportation within a district director’s 
authority under 19 C.F.R. § 171.21 (1980), 
the director must ordinarily '® act on a peti- 
tion for remission or mitigation within 24 
hours of receipt. The claimant must be 
allowed the opportunity to make an oral 
appearance to argue his or her claim. Be- 
cause the administrative delay in processing 
von Neumann's petition violated his due 
process rights, we reverse and remand with 
instructions to enter judgment for von Neu- 
mann in the amount of $3,600 plus costs. 
See Lee, 538 F.2d at 33. 

CONCLUSION 

The district court finding that von Neu- 

mann violated 19 U.S.C. § 1497 by failing to 

declare his car is not clearly erroneous, 

therefore the seizure of his car was proper. 

The delay in processing von Neumann's pe- 

tition for remission or mitigation, however, 

violated his due process right to prompt 

consideration of his claim. Because we hold 

that Customs must ordinarily act on a peti- 

tion for remission within 24 hours of re- 

ceipt, we reverse and remand with instruc- 

tions to enter judgment for the plaintiff. 

AFFIRMED IN PART, REVERSED 

AND REMANDED IN PART. 

W 
o £ KEY NUMBER SYSTEM 

T 

15. The district director has authority to act in 

cases involving a total value of merchandise 

not exceeding $25,000. See note 5 supra. The 

due process requirements for cases involving 

greater value than $25,000, which are decided 

by the Commissioner of Customs, are not be- 

fore us. See 19 CFR. § 171.11(a) (1980). We 

note that the greater the value of the merchan- 

dise, the greater the potential for a burden on 

individual rights. On the other hand, the ad- 

ministrative machinery may need more time to 

Herbert NANTY, Plaintiff-Appellant, 
Vv. 

The BARROWS COMPANY, formerly 

Barrows Furniture Company, 

Defendant-Appellee. 

No. 80-5106. 

United States Court of Appeals, 

Ninth Circuit. 

Argued and Submitted Aug. 12, 1981. 

Decided Nov. 9, 1981. 

Rehearing and Rehearing En Banc 

Denied Dec. 28, 1981. 

In a’suit under the equal employment 

opportunity provisions of the Civil Rights 

Act of 11964) the United States District 

Court for the District of Arizona, at Phoe- 

nix, Wesley E. Brown, J., 483 F.Supp. 101, 

after short trial entered judgment for de- 

fendant employer. The Court of Appeals, 

Reinhardt, Circuit Judge, held that: (1) evi- 

dence showing that plaintiff was qualified 

applicant for job as furniture delivery truck 

driver but was rejected summarily, without 

inquiry or even opportunity to file applica- 

tion, and was told that there was no job to 

be filled, while after his rejection the posi- 

tion remained open and two Caucasians 

were hired for the job made prima facie 

case of racialidiserimination in employment, 

and District Court’s finding to contrary was 

clearly erroneous; (2) in view of fact that 

employer knew nothing about minority job 

applicant at time of rejection and offered 

no explanation for action in rejecting such 

applicant, there was failure by employer to 

articulate legitimate, nondiscriminatory 

reason for rejection and to meet burden of 

overcoming the prima facie case; (3) plain- 

tiff was entitled at least to injunctive re- 

operate when the district director is not autho- 

rized to act. 

16. Unlike the court in Lee, we have set forth 

times to be followed under ordinary circum- 

stances. We do not preclude the possibility of 

some reasonable extension being permissible if 

unusual circumstances, such as holidays or un- 

availability of needed personnel, require some 

brief delay. A claimant might also waive de- 

lays in the administrative process.     

   

  

   

        

   
   

  

    

  

   
    

   

        

   
       

   
   

   
   
   

  

    

  

    

   

    

   

  

   
    

    

        

   
   

    

     

      

  

   



= 
; 

£ 

£ 
| 
8 
: 

} 
: 

3 
= 

= 
3 

FT
E 
T
H
E
R
 
O
N
 

A
V
E
S
 

OR
 

GU
N 
E
R
I
E
 
|
 

MO
DE

TO
RA

 
TO
 

WD
 C

NM
I 

AC
 
VT

E 
Wh

 
T
O
R
Y
 

he
 

| 

    

1328 660 FEDERAL REPORTER, 2d SERIES 

lief; and (4) District Court not having con- 

sidered issue of what relief was required or 

effect of defendant’s claim regarding supe- 

rior qualifications of others, remand for 

further hearing was necessary. 

Case reversed and remanded with in- 

structions. 

1. Civil Rights e=44(1) 

Analytic process to be used for evaluat- 

ing evidence of discrimination in equal em- 

ployment opportunity cases was as set forth 

in opinion. Civil Rights Act of 1964, § 701 

et seq. as amended 42 U.S.C.A. § 2000e et 

seq. 

2. Civil Rights e=44(1) 

To establish prima facie case of race 

discrimination under equal employment op- 

portunity provisions of Civil Rights Act of 

1964, complainant must show that he be- 

longs to racial minority, that he applied and 

was qualified for job for which employer 

was seeking applicants, that despite his 

qualifications he was rejected and that, af- 

ter his rejection, position remained open 

and employer continued to seek applications 

from persons of complainant’s qualifica- 

tions. Civil Rights Act of 1964, § 701 et 

seq. as amended 42 U.S.C.A. § 2000e et seq. 

3. Civil Rights <=44(1) 

Plaintiff may meet his initial burden, 

in proving race discrimination under equal 

employment opportunity provisions of Civil 

Rights Act of 1964, simply by offering evi- 

dence adequate to create inference that em- 

ployment decision was based on discrimina- 

tory criterion illegal under the Act, i. e., 

evidence indicating that it was more likely 

than not that employer's actions were based 

on unlawful considerations. Civil Rights 

Act of 1964, § 701 et seq. as amended 42 

U.S.C.A. § 2000e et seq. 

4. Civil Rights e=44(1) 

Federal Courts &=858 

Evidence showing that plaintiff was 

qualified applicant for job as furniture de- 

livery truck driver but that he was rejected 

summarily, without inquiry or even oppor- 

tunity to file application, and was told that 

there was no job to be filled, while after his 

rejection the position remained open and 

two Caucasians were hired for the job made 

prima facie case of racial discrimination in 

employment, and district court’s finding to 

contrary was clearly erroneous. Civil 

Rights Act of 1964, § 701 et seq. as amend- 

ed 42 U.S.C.A. § 2000e et seq. 

5. Civil Rights e=43 

In view of fact that employer knew 

nothing about minority job applicant at 

time of rejection and offered no explana- 

tion for action in rejecting such applicant, 

there was failure by employer to articulate 

legitimate, nondiscriminatory reason for re- 

jection and to meet burden of overcoming 

prima facie case of unlawful racial discrimi- 

nation in employment. Civil Rights Act of 

1964, § 701 et seq. as amended 42 U.S.C.A. 

§ 2000e et seq. 

6. Civil Rights e=46 

When legitimate candidate for job has 

demonstrated that he has been subject of 

unlawful discrimination in employment 

process, he is entitled to injunction against 

future or continued discrimination. Civil 

Rights Act of 1964, §§ 701 et seq., 706(g) as 

amended 42 U.S.C.A. §§ 2000e et seq. 

2000e—-5(g). 

7. Civil Rights ==46 

When legitimate candidate for employ- 

ment is summarily rejected because of un- 

lawful discrimination, court in determining 

right to job and monetary relief must still 

determine whether, absent such discrimina- 

tion, plaintiff would have been hired. Civil 

Rights Act of 1964, § 701 et seq. as amend- 

ed 42 U.S.C.A. § 2000e et seq. 

8. Civil Rights ¢=46 

When plaintiff had made case of dis- 

crimination under equal employment oppor- 

tunity provisions of Civil Rights Act of 

1964, court at relief stage imposed heavy 

burden of proof on defendant employer, and 

job applicant having proved unlawful dis- 

crimination in employment process was to 

be awarded full relief, i. e., position retroac- 

tively, unless defendant showed by clear 

and convincing evidence that, even in ab- 

\ \ / \ \ 
\ - 

~~ ss 
/\ F\ 

/ \ 

; 

  

  

F
E
 
P
I
E
R
 
S
R
E
 

  

sence of di 

would not § 

position. C 

et seq., 70 

§§ 2000e et 

9. Civil Rig 

Use of 

many insta 

ties but also 

discriminato 

in lower ech 

criteria sho 

cism. Civil 

seq., 706(g 

§§ 2000e et 

10. Civil Ri 

Danger: 

ria that ha 

impact case 

ployment 

treatment ( 

mand was t 

its consider 

had met b 

Civil Right{ 

706(g) as a 

seq., 2000e 

11. Civil Ri 

By pro} 

tion, and p 

tiff prevai 

whether or 

ey damages 

tled to attq 

employmen 

Rights Act 

1964, § 70 

§ 2000e-5( 

See pu 
for othe 
definitio 

Alice L. 

Monson, Ph 

lant. 

* Honorable 

United Staf] 

nation.  



  

Se \ XxX \ 

'NANTY v. BARROWS CO. 1333 
Cite as 660 F.2d 1327 (1981) 

BEYOND McDONNELL DOUGLAS 

Nanty requests, inter alia, an injunction 
prohibiting future, or continued, discrimina- 
tion by Barrows against him, an injunction 
ordering Barrows to hire him as a furniture 
delivery truck driver, and monetary relief 
for lost wages and benefits. 

[6] When a legitimate candidate for a 
job has demonstrated that he has been the 
subject of unlawful discrimination in the 
employment process, he is entitled to an 
injunction against future, or continued, dis- 
crimination.’ The purpose of such an order 
is to ensure that, at the very least, the 
applicant will receive full and fair consider- 
ation from the employer if he seeks similar 
employment in the future. In view of our 
holding that Nanty proved his claim of un- 
lawful discrimination, he is entitled to an 
injunction against future, or continued, dis- 

crimination. 

[7,8] Frequently, the finding of unlaw- 
ful discrimination necessarily determines 
the plaintiff’s right to the job and monetary 
relief. However, that is not the case here. 
When a legitimate candidate for employ- 
ment is summarily rejected because of un- 
lawful discrimination, ‘we still must deter- 
mine whether, absent that discrimination, 
he:would have been hired!" Before consider- 
ing whether Nanty is entitled to an order 
directing Barrows to hire him, we set forth 
the burden of proof that is applicable to the 
determination of this issue. The McDonnell 
Douglas rule that the ultimate burden of 
proof remains on the plaintiff is not appli- 
cable after unlawful discrimination has 
been proven. League, etc. v. City of Sali- 
nas Fire Dept., 654 F.2d 557, 559 (9th Cir. 
1981). At the relief stage, we apply the 

better qualified than Nanty does not affect our 
conclusion. It is obvious from what we have 
said earlier that the allegedly superior qualifi- 
cations of the two individuals subsequently 
hired was not the reason that Barrows unlaw- 
fully refused to interview Nanty or permit him 
to file an application. Nor, in fact, does Bar- 
rows make any such claim. 

6. See, 42 U.S.C. § 2000e-5(g); cf. James v. 
Stockham Valves & Fittings Co., 559 F.2d 310, 
356 (5th Cir.), cert. denied 434 U.S. 1034, 98 
S.Ct. 767, 54 L.Ed.2d 781 (1977); E.E.O.C. v. 

opposite rule and impose a heavy burden on 
the defendant. Id. ‘Where a jobrapplicant 
has proved unlawful discrimination in the 
employment process, he must be awarded 
full relief, i. e., the position retroactively, 
unless the “defendant shows ‘by “clear and 
convincing evidence” that even in the ab- 
sence of discrimination the rejected appli- 
cant would not have been selected for the 
open positions’ w Marotta v. Usury, 629 F.2d 
615, 618 (9th Cir. 1980) [citing Day v. Math- 
ews, 530 F.2d 1083 (D.C.Cir.1976)].” Id. at 

558. This court, in League, explained the 

reason for this rule: 

The burden of showing that proven dis- 

crimination did not cause a plaintiff’s re- 

jection is properly placed on the defend- 

ant-employer because its unlawful acts 

have made it difficult to determine what 

would have transpired if all parties had 

acted properly. See Day, 530 F.2d at 

1086. 

Id. at 559.7 

Barrows contends that Nanty was not as 

qualified as the Caucasians who were hired, 

and that as a result Nanty would not have 

been hired even in the absence of discrimi- 

nation. The district court did not consider 

this contention. It simply made a finding, 
which we have held to be erroneous, that 
Nanty was not qualified for the job. For 

reasons we have explained earlier, Barrows’ 

contention regarding the superior qualifica- 

tions of those actually hired is not relevant 

to our determination of unlawful discrimi- 

nation in the employment process. It is 

relevant, however, to the question whether 

Nanty would have been hired, absent such 

discrimination, and thus to the question of 

Nanty’s right to the job. Since the district 

New York Times Broadcasting Service, Inc., 

542 F.2d 356, 361 (6th Cir. 1976). 

7. The burden we have placed on the defendant 

at the relief stage of the case is consistent with 

the Supreme Court’s cbservation in East Texas 

Motor Freight System, Inc. v. Rodriquez, 431 

U.S. 395, 97 S.Ct. 1891, 52 L.Ed.2d 453 (1977), 

that when unlawful discrimination has been 

established, the defendant is entitled to prove 

that the plaintiff would not have been hired. 

Id. at 404 n.9, 97 S.Ct. at 1897 n.9. 

    

    

   

      

   
   

    

   

  

   
   
    

  

   

    

    
   

   
    

  

   
    
   

  

    
   
   
   

   

     

     

   
    

     

     

   
      

            

  



RE
T 

AL
AR
 

T
T
 

M
t
 

LE
TT

E 
U
T
R
 
B
T
 

ARIAL 
88 

wo
 

i 
A
R
H
 i

k 
E
r
e
 

[11 
I
i
 

U
T
 

I 

  

1334 660 FEDERAL REPORTER, 2d SERIES 

court has not previously considered the is- 

sue of what relief is required or the effect 

of Barrows’ claim regarding the superior 

qualifications of others, we remand for a 

further hearing, limited to the question of 

the proper relief. At the hearing, the dis- 

trict court should afford Barrows the oppor= 

tunity to prove by “clear and convincing” 

evidence that Nanty was less qualified than 

those who were hired and that, in the ab- 

sence of discrimination, he would not have 

been hired.? 

[9,10] On remand, the district court will 

also be faced with Barrows’ contention that 

Nanty would not have been hired because 

he did not possess the three personal attrib- 

utes which Barrows contends are a prereq- 

uisite for employment as a furniture deliv- 

ery truck driver, i. e., that an applicant be 

neat, articulate and personable. We there- 

fore offer some guidance to the district 

court on this matter. Subjective job crite- 

ria present potential for serious abuse and 

should be viewed with much skepticism. 

Use of subjective job criteria not only has, 

in many instances, a disparate impact on 

minorities,’ but also provides a convenient 

pretext for discriminatory practices. Sub- 

jective criteria may easily be asserted as the 

reason for an adverse employment decision 

when, in fact, the reason was discriminato- 

ry. Moreover, where the job in question 

involves skills which are primarily physical 

8. We note that the answer to this question 

depends in part on Barrows’ hiring practices at 

the time of Nanty’s rejection. If Barrows filled 

job vacancies with the first qualified applicant, 

then the fact that the two Caucasians who 

applied later were better qualified is plainly 

irrelevant to the issue whether Nanty would 

have been hired, absent discrimination. As we 

have noted, the burden, with respect to all 

aspects of the relief issue, is on Barrows. 

9. See Note, Title VII and Employment Discrimi- 

nation in “Upper Level” Jobs, 73 Colum.L.Rev. 

1614, 1630 (1973). We do not suggest that the 

principles developed in disparate impact cases 

are necessarily applicable in the context of a 

disparate treatment case. However, the dan- 

gers in the use of subjective job criteria that 

have been identified in disparate impact cases 

are also present in disparate treatment cases. 

As we explain in the text following this note, 

the district court, on remand, must be sensitive 

to these dangers in its consideration of whether 

/ 

or mechanical, or are tangible or objective 

in nature, as does the job of furniture deliv- 

ery truck driver, it is more likely that sub- 

jective criteria can be used as an excuse for 

discrimination. In a case involving higher 

echelon employment, the skills for which 

are necessarily measured in more subjective 
terms, the same potential for abuse exists, 

but the use of subjective criteria is less 

inherently suspect. 

[11] Upon remand, exacting scrutiny of 

Barrows’ claim is required. The district 

court should carefully examine the rele- 

vance of any subjective criteria used by 

Barrows to the job performance of a furni- 

ture delivery truck driver, the extent to 

which Barrows has made use of such ecrite- 

ria for legitimate reasons in the past, and 

whether the criteria are being used as a 

cover for discrimination in this instance.l® 

DISPOSITION 

The case is reversed and remanded with 

instructions to enter an injunction prohibit- 

ing Barrows from future, or continued, dis- 

crimination against Nanty and for further 

proceedings consistent with this opinion. 

Ww 
o £ KEYNUMBERSYSTEM 

5 

Barrows has met its burden in the relief stage 
of the case. 

10. Nanty also requests, and is entitled to, rea- 

sonable attorney's fees. Section 706(k) of Title 

VII, 42 U.S.C. § 2000e-5(k), provides that “the 

court ... may allow the prevailing party ... a 

reasonable attorney’s fee as part of costs ...” 

(emphasis added). A plaintiff is a “prevailing 

party” if he (1) prevails on a significant issue in 

the litigation which (2) achieves a benefit 

sought in the lawsuit. Manhart v. City of Los 

Angeles, 652 F.2d 904, 907 (9th Cir. 1981). By 

proving unlawful discrimination, Nanty pre- 

vailed on a significant issue. Id. ‘‘By securing 

an injunction’ against future, or continued, dis- 

crimination, Nanty achieved a benefit sought in 

his lawsuit. Id. He need not be awarded mon- 

ey damages to achieve a benefit sought in his 

lawsuit. Accordingly, Nanty is a ‘‘prevailing 

party,” and is entitled to attorney’s fees. On 

remand, the district court is to determine what 

constitutes a reasonable fee.   

BILL JOH 

NATION! 

BO 

No 

United § 

Argued anc 

De 

As Modified 

Rehearin 

Restaura 

Labor Relatiq 

cross-petitiond 

Court of Ap 

held that: (1 

nation that si 

for union actj 

finding that 

labor practic 

waitresses w 

discharged w 

plying that a 

tion would b 

practice; (4) { 

tion that reg 

enjoin picketi 

against picke 

in fact and w 

waitress for | 

picketers for { 

and (5) Boa 

stop to the Ia 

to withdraw 

Enforce 

1. Labor Rel; 

Test of 

ion employee 

whether unig 

cause behind 

bor Relations 

US.C.A. § 15 

2. Labor Rel 

Determir 

discharge of