Table Of ContentTHE SIMPLIFICATION  OF  COMBAT  AIR  CREW 
CRITERIA WITH SPECIAL REFERENCE 
TO RANK-ORDER  RATINGS
DISSERTATION
Presented  in P artial Fulfillm ent  of  the  Requirements 
for  the Degree  Doctor  of  Philosophy  in the 
Graduate  School  of  The  Ohio  State 
University
tt?  By
a*
Mr FRANCES  ESTEP,  B.  S.,  M.  A. 
The  Ohio "state  University
1951
Approved by:
Adviser
PREFACE
The  Personnel  Research  Center  of  Wayne University  is  one  of  the 
several  university groups  currently performing  contract  research for 
m ilitary agencies.  Dr.  Roger M.  Bellows,  D irector  of the  Center, 
has  facilitated  the  development  of  th is  research  study through making 
data available  from the  Center  and  especially through his  personal, 
considerate,  and patient  attention to  the  demands  of  the  project.
Dr.  Carroll  L.  Shartle  of  The  Ohio  State University has been 
adviser  to the project,  contributing  his  knowledge  of  governmental 
and m ilitary research to  the  effective  guidance  of  this  study.
Dr.  Robert  J.  Wherry  and  Dr.  John  R.  Kinzer were  consulted  on 
certain phases  of  the  research,  and both have  criticized  an  early 
draft  of  the  dissertation,  offering  suggestions  for  its  improve
ment.  Mrs.  Katherine Dockeray performed  numerous  kindnesses  for 
the w riter while  the project  was being  carried  on.
Appreciation  is  expressed  to Rutledge  Jay and  to  John  Weaver 
of  the  Personnel  Research Center for  assistance  on  statistic al 
d etails.
-  x  -
86250,3
TABLE OF  CONTENTS
Page
PREFACE      i
CHAPTER  I  CRITERIA  OF  COMBAT  PROFICIENCY.............................  1
CHAPTER  II  DESCRIPTION  OF THE DATA...........................................  lfc
The  Crew  Ratings.............................................................  17
Fam iliarity with  the  Crews  Ratings...................  17
Individual  Ratings........................................................  19
CHAPTER  III  STATISTICAL ANALYSIS  OF THE RELIABILITY  OF
THE RATERS'  RATINGS...................................................  21
CHAPTER  IV  FACTORIAL  COMPOSITION OF THE VARIABLES.  .  .  50
CHAPTER V  ANALYSIS  OF  OTHER  STATISTICAL RELATIONSHIPS  57
CHAPTER VI  CONCLUSIONS  WITH SUGGESTIONS  FOR  FUTURE
RESEARCH...............................................................................  78
BIBLIOGRAPHY  .........................................    8l
AUTOBIOGRAPHY  .......................................................................................................  85
-  ii  -
LIST  OF  FIGURE AMD  TABLES
Page
SAMPLE  3"  X  5"  CARD USED  FOR RANKINGS............................ 18
CORRELATIONS BETWEEN EACH RATER'S  RATINGS  ON ALL 
VARIABLES............................................................................................. 2h
CORRELATIONS BETWEEN PAIRS  OF RATERS  ARRANGED  IN 
ORDER  OF MAGNITUDE  OF THE  CORRELATION  COEFFICIENTS 25
MATRICES  OF  INTERCORRELATIONS BETWEEN RATERS FOR 
EACH  OF  NINE VARIABLES,  VARIABLE 1 ................................. 27
MATRICES  OF  INTERCORRELATIONS BETWEEN RATERS  FOR 
EACH  OF  NINE VARIABLES,  VARIABLE 2 ................................. 29
MATRICES  OF  INTERCORRELATIONS BETWEEN RATERS  FOR 
EACH  OF NINE VARIABLES,  VARIABLE  3 ................................. 30
MATRICES  OF  INTERCORRELATIONS  BETWEEN RATERS  FOR 
EACH  OF  NINE VARIABLES,  VARIABLE  k ................................. 31
MATRICES  OF  INTERCORRELATIONS BETWEEN RATERS  FOR 
EACH  OF NINE VARIABLES,  VARIABLE  5 ................................. 32
MATRICES  OF  INTERCORRELATIONS BETWEEN  RATERS  FOR 
EACH  OF  NINE VARIABLES,  VARIABLE 6 ................................. 3^
MATRICES  OF  INTERCORRELATIONS  BETWEEN RATERS  FOR 
EACH  OF NINE VARIABLES,  VARIABLE  7 ................................. 35
MATRICES  OF INTERCORRELATIONS BETWEEN RATERS  FOR 
EACH  OF NINE VARIABLES,  VARIABLE  8 ................................. 36
MATRICES  OF  INTERCORRELATIONS BETWEEN RATERS FOR 
EACH  OF  NINE VARIABLES,  VARIABLE  9 ................................. 38
MATRIX  OF  INTERCORRELATIONS BETWEEN VARIABLES 
FOR  EACH RATER,  Rater  1  —  Group  Commanding 
O fficer................................................................................................. kl
MATRIX  OF  INTERCORRELATIONS  BETWEEN VARIABLES 
FOR  EACH RATER,  Rater 2  —  Squadron Commander  (a) k2
MATRIX  OF  INTERCORRELATIONS BETWEEN VARIABLES 
FOR  EACH RATER,  Rater  3  —  Squadron  Commander  (b) ^3
MATRIX  OF  INTERCORRELATIONS  BETWEEN VARIABLES 
FOR  EACH RATER,  Rater  H  —  Squadron  Commander  (c) kk
-  iii  -
Page
TABLE XVI  MATRIX  OF  INTERCORRELATIONS BETWEEN VARIABLES
FOR EACH RATER,  Rater  5  —  Group  Executive 
O fficer.........................................................................................  li-5
TABLE XVII  MATRIX  OF  INTERCORRELATIONS BETWEEN VARIABLES
FOR  EACH RATER,  Rater  6  --  Group  Operations 
O fficer.....................................................• .  . ........................  1+6
TABLE XVIII  MATRIX  OF  INTERCORRELATIONS BETWEEN VARIABLES 
FOR  EACH RATER,  Rater  7  —  Group Bombardier 
O fficer.........................................................................................  1+7
TABLE XIX  AVERAGE RELIABILITY COEFFICIENTS  ARRANGED  IN
ORDER OF MAGNITUDE ON ALL VARIABLES FOR  EACH 
RATER.............................................................................................  1+9
TABLE XX  CORRELATIONS BETWEEN NINE  CRITERION VARIABLES  .  51
TABLE XXI  AVERAGE  RELIABILITY  OF THE NINE CRITERION
VARIABLES  ARRANGED  IN ORDER  OF MAGNITUDE  OF
THE RELIABILITY COEFFICIENT..........................................  53
TABLE XXII  CORRELATIONS BETWEEN RATINGS  ON  EACH VARIABLE
AND  THE TOTAL RATINGS  WITH THE  INDIVIDUAL 
VARIABLE  TAKEN  OUT.................................................................  55
TABLE XXIII  JUDGED  FAMILIARITY  OF  EACH RATER  WITH  CREWS
ON EACH VARIABLE,  Variable  1..........................................  58
TABLE XXIV  JUDGED FAMILIARITY  OF  EACH RATER WITH  CREWS
ON  EACH VARIABLE,  Variable  2 ..........................................  59
TABLE XXV  JUDGED  FAMILIARITY  OF  EACH RATER  WITH  CREWS
ON EACH VARIABLE,  Variable  3..........................................  59
TABLE XXVI  JUDGED  FAMILIARITY  OF  EACH RATER  WITH  CREWS
ON EACH VARIABLE,  Variable  k..........................................  60
TABLE XXVIII  JUDGED FAMILIARITY  OF  EACH RATER  WITH  CREWS
ON  EACH VARIABLE,  Variable  6..........................................  60
TABLE XXIX  JUDGED  FAMILIARITY  OF  EACH RATER WITH  CREWS
ON EACH VARIABLE,  Variable  7..........................................  61
TABLE XXX  JUDGED FAMILIARITY OF  EACH RATER  WITH  CREWS
ON EACH VARIABLE,  Variable  8..........................................  61
TABLE XXXI  RELATION BETWEEN FAMILIARITY WITH A VARIABLE
AND  GOODNESS  OF RATINGS  ON THAT VARIABLE.  . . .   63
-  iv  -
Page
TABLE XXXII  COMPARISON OF RATERS  ON  SEVERAL  INDEXES  OF
GOODNESS  OF RATING.............................................................  65
TABLE XXXIII  AVERAGE RATER RELIABILITY COEFFICIENTS  FOR
EACH VARIABLE FOR  INDIVIDUAL RATINGS  OF  FIVE 
MEMBERS  OF  THE BOMB  TEAM...............................................  67
TABLE XXXIV  COMPARISON BETWEEN RATINGS  GIVEN TO FIVE 
MEMBERS  OF  THE BOMB  TEAM AND THE RATINGS 
GIVEN  TO THE  CREW  OF  WHICH THEY WERE A 
MEMBER........................................................................................  69
TABLE XXXV  RANK  OF  EIGHT  LEAD  CREWS  AND  THEIR AIRCRAFT
COMMANDERS...............................................................................  71
TABLE XXXVI  RATER RELIABILITIES  ON  OVER-ALL  CREW RATINGS
FOR  EACH VARIABLE.................................................................  72
TABLE XXXVII  COMPARISON BETWEEN RELIABILITY  OF THE  OVER-ALL
CREW RATINGS  ON EACH VARIABLE  IN TWO  LOCATIONS  7^
TABLE XXXVIII  INTERCORRELATIONS BETWEEN NINE VARIABLES  USING
SUM OF RANKS BY FIVE RATERS..........................................  75
TABLE XXXIX  RELATION BETWEEN RATINGS  GIVEN TO THE  CREW  ON 
VARIABLE  9 AND THE RATINGS  GIVEN TO  INDIVIDUAL 
CREW MEMBERS  ON NINE VARIABLES  . . . .....................  76
-  v  -
THE SIMPLIFICATION  OF  COMBAT  AIR  CREW CRITERIA WITH 
SPECIAL REFERENCE TO RANK-ORDER  RATINGS
CHAPTER  I 
CRITERIA OF  COMBAT  PROFICIENCY 
Among the  several  areas  of  contribution by psychologists  to  the 
m ilitary program are  the  selection,  classification,  placement,  and 
training  of  personnel.  Effective testing  of  research  in  these  areas 
depends  upon  identification  of  the  successful  performance behavior  to 
be  duplicated by future  groups  of  personnel.  Selection  and  placement 
techniques,  such  as  psychological  tests  and  other measurements,  must 
demonstrate  an  acceptable  level  of validity to be  recommended  for use.
The  end  evaluation  of  such techniques  and  procedures  is  th eir  relation
ship  to the  successful  accomplishment  of  the m ilitary goal.  This  goal 
usually  implies  success  in  combat  situations,  although there  are  selected 
instances which  involve  only non-combat  activity.
During World  War  I  psychologists  applied  the meager  techniques  then 
available to  the  special problems  of  m ilitary  classification  and place
ment.  The  inadequacy  of  the  instruments  seemed  rooted  in th eir  failure 
to predict  successful  and  unsuccessful  performers  in the  designated 
tasks.  In the  absence  of  a  definition  of  the  kind  of  job  proficiency 
sought,  validation  of  the predictive devices was  not  possible.  C riteria 
had  not  yet been the  object  of  a  sim ilar volume  of  research  then  going 
forward  on the  construction  of  psychological  te sts.  Jenkins  has  de
scribed  the history  of psychologists'  experience with validation thus:
The  events  of World War  I  taught  American psychologists 
the necessity  of  validation.  The  experience  of  the  next 
two  decades  taught  them much  about  the  technique  of 
validation.  It  remained  for  World  War  II  to  drive  home 
to the psychologists  at  large  the necessity for  devoting 
-  1  -
2
-    -
much time  and  thought  to the basis  for  validation.!
During World War  II  and  the  years  since  then,  evaluation  of  the  criterion 
itse lf  has  been  considered prerequisite  to  later phases  of  validation 
research.  If  criterion data  of  a  known  and  acceptable  level  of 
dependability are  not  available  for  the particular  sample  of  subjects 
to be  studied,  the  research  project  founders before  it begins.
In practice,  the  adequacy  of  criteria  varies.  There  were  some 
attempts  during  World  War  II  to  obtain measures  of  proficiency under 
combat  conditions but  the worth  of  the  criteria  obtained was  found  to be 
negated by  circumstances  beyond  the  control  of  the  research workers.
There  was  no  assurance  that  criteria  gathered under  combat  conditions 
would be  collected  system atically because  of  the  pressure  of  other 
more  urgent  activity.  C riteria  labelled  as  combat  criteria may have 
been made  for  seme  other purpose,  such  as  adm inistrative  action,  and 
were  not  necessarily appropriate  for psychological  research.  If  the 
criteria were  to be  used  to  evaluate  training procedures,  there was 
sometimes  a  considerable  delay before  combat  criteria  could be  obtained 
for use  in validation.  In  the meantime,  both training  requirements  and 
definition  of  combat  proficiency may have  changed.  Such  an  example  arose 
during the  la tte r part  of  World War  II when the United  States  shifted 
from defensive  to  offensive warfare.
In lieu  of  combat  criteria,  more  available measures  were used 
without  knowledge  of  th eir  relationship -to  the  ultim ate  goals  of  the 
m ilitary  establishment.  C riteria which were utilized within the  United
-'-John  G.  Jenkins,  V alidity for what?  Journal  of  Consulting 
Psychology,  19^6,  10,  93.
3
-    -
States bore  an unknown  relationship  to measures  of  success  of men who 
were  under  enemy fire   or who were  living under  stress  conditions.
In the  search fo r a  criterion to use  for  the  evaluation  of 
psychological  techniques  for  combat  air  crews  the  greatest  problem 
seems  to be  to  find  the  one  which  is  the best  measure  of  combat 
effectiveness.  Sisson  has  pointed  out  the  general  unsuitability  of 
common  industrial  criteria for m ilitary personnel  research.1  Combat 
efficiency  is not  only hard to  define but  it  is  also  d ifficu lt  to 
reduce  to  quantitative  indexes.  He  cites  five  industrial  criteria 
commonly used,  which have  few  counterparts  in m ilitary work,  (l)  An 
index  of  output  per  unit  of  time  is  conceivably possible  to  tabulate, 
but  it  is  doubtful whether  the  number  of  shots  fired  or  the  number  of 
bombs  dropped  is  related  to  the  job  of m ilitary personnel;  in  any event, 
it  seems  that  there  are most  likely a rtific ia l  lim its  to  the  amount  of 
"production"  an  individual  could  turn  out.  (2)  An  index  of  the  quality 
of  production:  Sisson  commented,  "In  the  job  of  destroying  the  enemy's
capacity to wage war,  the  absurdity  of  speaking  in  such  terms  as  number 
of  rejected  products,  cost  of  spoiled work,  or  customer  complaints  is 
sufficiently  evident  to  need  no  further  elaboration."2  (3)  An  index  of 
employee  turnover  cannot be based  on voluntary  quits  as  would be  possible 
in the  civilian  commercial  establishment.  (4)  An  index  of  employee  satis
faction,  or  rather  dissatisfaction,  is  available  in  great  quantities,  but 
without much u tility   for  criteria.  (5)  An  index  of  train ab ility   is  a 
possible  criterion for m ilitary personnel  research,  and was  commonly used
1E.  D.  Sisson.  The  criterion  in Army personnel  research.  In 
George  A.  KeHy  (Ed.),  Hew Methods  in Applied Psychology.  College  park, 
Md.:  University  of Maryland,  19^7•  Pp.  17-18.
2Ibid.,  17.
-  h  -
in  early phases  of Army  personnel  research.
There  are,  according to Flanagan,  four main  categories  of  combat 
criteria:  (l)  objective  measures  of  combat  proficiency,  (2)  records
of  adm inistrative  actions  in  combat  units,  (3)  ratings based  on  direct 
and  systematic  observation  of  combat  effectiveness,  and  (4)  ratings  based 
on  general  impressions,  reports,  and  incidental  observations.1
In the  firs t  category  of  objective  records  of  combat  proficiency, 
the most  readily accessible  criteria,  the  AAF  has  used  records  of bomb
ing  accuracy as  obtained  from  strike  photographs,  records  of  airplane 
accidents  related  to  combat  operations,  records  of  numbers  of  enemy 
planes  shot  down,  and  records  of  combat  casualties.
In  the  second  category,  records  of  adm inistrative  actions  in  combat 
units,  several  criteria  are widely used because  of  availability  in  spite 
of being based  upon the  subjective  judgments  of  superior  officers.  The 
adm inistrative  actions  include promotions,  special  awards  and  decorations, 
reclassifications  due  to  anxiety reactions  or  lack  of  combat  proficiency, 
special  duty assignments,  such  as  to  lead  crews  or  for  important  combat 
missions,  official  ratings  of  efficiency while  assigned  to  a  combat  unit.
The  third  category  of  criteria  includes  ratings based  on direct  and 
systematic  observation  of  combat  effectiveness.  The  evaluations  include 
ratings  of navigational  proficiency based  on  a  study  of  the  logs  prepared 
by the  navigator  on various  missions,  ratings  regarding the  success  or 
failure  of bombing missions based  on  direct  observation  of  the  results  or
1 John' C.  Flanagan  (Ed.).  The  Aviation Psychology Program  in 
the  Army Air Forces,  AAF Aviation Psychology Program Research Report  No.  1. 
Washington,  D.  C.:  U.  S.  Government  Printing  Office,  19V8.  Pp.  276ff.