INFORMATION TO USERS
This manuscript has been reproduced from the microfilm master. UMI
films the text directly from the original or copy submitted. Thus, some
thesis and dissertation copies are in typewriter face, while others may
be from any type of computer printer.
The quality of this reproduction is dependent upon the quality of the
copy submitted. Broken or indistinct print, colored or poor quality
illustrations and photographs, print bleedthrough, substandard margins,
and improper alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete
manuscript and there are missing pages, these will be noted. Also, if
unauthorized copyright material had to be removed, a note will indicate
the deletion.
Oversize m aterials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand corner and
continuing from left to right in equal sections with small overlaps. Each
original is also photographed in one exposure and is included in
reduced form at the back of the book.
Photographs included in the original manuscript have been reproduced
xerographically in this copy. Higher quality 6" x 9" black and white
photographic prints are available for any photographs or illustrations
appearing in this copy for an additional charge. Contact UMI directly
to order.

U niversity M icrofilm s International
A Beil & H owell Information C o m p a n y
3 0 0 North Z e e b R oad . Ann Arbor. Ml 4 8 1 0 6 -1 3 4 6 USA
3 1 3 /7 6 1 - 4 7 0 0
8 0 0 /5 2 1 - 0 6 0 0

Order N um ber 9134146

A n assessm ent of M ichigan police officers’ cognitive knowledge
retention o f information presented in M S U ’s Accident
Investigation One training program
Lee, Daniel G., Ph.D.
Michigan State University, 1991

UMI

300 N. Zeeb Rd.
Ann Aibor, MI 48106

An Assessment of Michigan Police Officers' Cognitive
Knowledge Retention of Information Presented
in MSU's Accident Investigation One
Training Program
By
Daniel G. Lee

A Thesis
Submitted to MSU
in partial fulfillment of the requirements
for the degree of

Doctor of Philosophy
Department of Educational Administration
1990

ABSTRACT
An Assessment of Michigan Police Officers' Cognitive
Knowledge Retention of Information Presented
in MSU's Accident Investigation One
Training Program
By
Daniel G. Lee

The purpose of this study was to investigate the degree
of police officers' cognitive knowledge retention of subject
matter presented

to them during their attendance of MSU's

Accident Investigation One training program.

The researcher

evaluated the relationship between the retention of cognitive
information and selected demographic and job related vari­
ables.
One hundred police officers who had previously attended
the AI-1 training course were asked to participate
evaluation.

in the

Each officer was asked to retake the post test

that was administered at the completion of their respective
AI-1 training program.

Part of this evaluation compared the

mean score of the original performance test with the mean
score

of

the performance

comparison were:
test,

retest.

The variables

used

for

(1) time elapsed since taking the original

(2) age of the officer,

(3) education level and (4)

Daniel G. Lee
number of accidents investigated in the 12 months preceding
their taking the performance retest.
The first analysis was a paired-samples t-test.

Paired

samples were used on the original performance test scores and
the performance retest scores for the entire sample.

This

same analysis was used for each of the groups formed using the
demographic variables.
For the second evaluation,

analysis of covariance was

used for each of the four demographic variables

(time, age,

education and experience) to determine their relationship with
the dependent variable (performance retest) and the covariate
(original performance test).
Results
1.

The 100 subjects, when treated as one group, showed

a significant loss of retention when comparing mean scores of
the original performance test and the retest.
2.

Dividing the sample into various time periods since

completing the training showed, that the longer the elapsed
time since taking the training, the lower the retest score.
3.

Analysis of officer education level and the perfor­

mance retest mean score, indicates there is not a significant
relationship between education level and mean scores.
4.

The

analysis,

based

upon

officer

age

and

its

relationship to the performance retest mean score, showed that
a significant relationship did not exist.

Daniel G. Lee
5.
during

The number of accidents investigated by officers
the

year preceding

the

performance

retest

was

not

significant in terms of its relationship to the performance
retest.

ACKNOWLEDGMENTS
The author is grateful to a number of individuals who,
during the course of this study, offered and provided assis­
tance which helped to make this endeavor possible.
Special appreciation is extended to Dr. Eldon Nonnamaker,
Professor
support,

in

Administration

understanding

and

and

Higher

availability

Education,

whose

as Chairman of my

Guidance Committee was invaluable.
Appreciation is extended to Dr. Louis Hekhuis, Professor
in

Administration

and

Higher

Education

and

Dr.

Robert

Gustafson, Professor in Education and Engineering, as members
of my guidance committee, their shared insight, direction and
encouragement

given

during

the

course

of

this

study were

deeply appreciated.
A special thank you is extended to Dr. Betty Giuliani,
Professor of LifeLong Education, who as a guidance committee
member provided a tremendous amount of assistance in research
methodology, statistical analyses and general encouragement.
Special

appreciation

is also extended to Mrs.

Eunice

Weber for her excellent typing skills and patience.
Finally, to my wife, Sandy, and my daughter Danielle, I
offer my deepest thanks for their encouragement, patience and
many sacrifices.
v

TABLE OF CONTENTS
Chapter
I.

THE PROBLEM
Introduction
..................................
1
Required Basic Police Training in
Michigan ..................................
4
Background of the Problem .....................
5
6
Problem Statement ..............................
Purpose of the S t u d y ............................ 7
Need for the S t u d y ...............................7
Research Questions
............................
9
Research Hypothesis ............................
9
Definition of T e r m s ..............................10
Organization of Study .......................... 11

II.

REVIEW OF LITERATURE
Historical Perspective of Retention
S t u d i e s .................................... 12
Retention Defined .............................. 14
Retention Measurements
........................ 14
Effects of Retention Interval ................. 15
Curve of R e t e n t i o n ..............................16
Retention Affected by Degree of
L e a r n i n g .................................... 17
Longitudinal Retention
........................ 22
Retention Affected by Strategies Used ......... 22
Age-Effect on Retention ........................ 24
S u m m a r y .......................................... 26
Discussion of Previous Research ............... 28

III. RESEARCH METHODOLOGY
The S a m p l e ...................................... 30
Testable Hypotheses ............................ 31
AI-1 Original Performance Requirements
. . . .33
Population Selection
.......................... 34
Performance Retest Procedure
................. 34
Test I n s t r u m e n t .................................. 35
Validity Concerns .............................. 36
Reliability Concerns
.......................... 37
Analysis Procedure - Treatment of Data
. . . .38
T-Test (Paired Samples) ........................ 38
vi

Analysis of Covariance
........................ 39
S u m m a r y .......................................... 39
IV.

ANALYSIS OF DATA
The F i n d i n g s .................................... 41
Introduction to D a t a ........................... 42
Performance Test Mean Scores by Variable
Subcategories
............................ 44
Differences between Mean Original
Performance Test and Performance
R e t e s t ...................................... 53
The Effect of Time on the Performance
Retest Mean S c o r e ......................... 55
The Effect of Education on the
Performance Retest Mean Score
........... 58
The Effect of Test Subjects' Age on the
Performance Retest Mean Score
........... 61
The Effect of the Number of Accidents
Investigated on the Performance
Retest Mean S c o r e ......................... 64
Review of the Results of the Analysis
Of Hypotheses 1 - 5 ......................... 67
The Combined Effect of Time and Number
of Accidents Investigated on the
Performance Retest Mean Score
........... 70
The Combined Effect of Time and Age
on the Performance Retest Mean
S c o r e ...................................... 72
S u m m a r y .......................................... 77
V.

SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
S u m m a r y .......................................... 78
P u r p o s e .................................... 78
The Research Methods and
P r o c e d u r e s ........................... 79
Major F i n d i n g s ............................. 80
C o n c l u s i o n s ...................................... 81
Recommendations ................................ 84
Recommendations for Future Research ........... 86
Implications
.................................. 87
Training and Recertification ............. 87
Education Beyond High School ............. 87
Other Traffic Related Training ........... 88
Police Officer A g e ......................... 89

vii

APPENDICES
..........................
A.
SUMMARY OF TEST DEVELOPMENT
B.
CURRICULUM OUTLINE
. . . .
C.
OFFICER INFORMATION SHEET .
D.
AI-1 TEST - PART 1
. . . .
E.
AI-1 TEST - PART 2
. . . .
BIBLIOGRAPHY

........................

viii

LIST OF TABLES

Table

Page

4.1

Mean, Standard Deviation, Minimum Range, and
Maximum Number of each of the Six Variables
. . . .45

4.2

Performance Retest Mean, Standard Deviation, and
Number of Cases for Each of the Three
Subpopulations of the Time Variable
. . . . . . . .

48

4.3

Performance Retest Mean, Standard Deviation and
Number of Subjects for each of the Three Subpopula­
tions of the Variable of E d u c a t i o n ................... 49

4.4

Performance Retest Mean Score, Standard Deviation
and Number of Subjects for each of the Four Subpop­
ulations of the Age V a r i a b l e ......................... 50

4.5

Performance Retest Mean Score, Standard Deviation
and Number of Subjects for each of the Five Subpop­
ulations ............................................... 52

4.6

Comparison of the Mean Test Scores of the 100
Subjects for the Original Performance Test and the
Performance Retest ..................................

54

4.7

Analysis of Covariance of the Performance Retest by
the Divisions of T i m e ................................ 56

4.8

The Effect of Time on the Grand Mean as Determined
by the Analysis of C o v a r i a n c e ....................... 57

4.9

Analysis of Covariance of the Performance Retest by
the Three Divisions of Education ...................

59

The Effect of Education on the Grand Mean as
Determined by the Analysis of Covariance ...........

60

4.10

4.11 Analysis of Covariance of the Performance
Retest by the Division of A g e ....................... 62
4.12 The Effect of Age on the Grand Mean as Determined
by Analysis of C o v a r i a n c e ........................... 63
ix

4.13 Analysis of Covariance of the Performance
Retest by the Division of Number of Accidents
Investigated 12 months prior to the Retest ........

65

4.14 The Effect of Number of Accidents Investigated
on the Grand Mean as Determined by Analysis of
C o v a r i a n c e ............................................ 66
4.15 Analysis of covariance of the Performance
Retest by All Time and All Education Levels

. . . .

71

4.16 Analysis of Covariance of the Performance
Retest by All Time and All Accidents Investi­
gated
................................................. 73
4.17 Analysis of Covariance of the Performance
Retest by All Times and All Age L e v e l s .............. 76

x

CHAPTER 1
THE PROBLEM
Introduction
There

were

1,704

persons

killed

and

155,713

persons

injured in 410,437 reported motor vehicle traffic accidents in
Michigan during 1988 (Michigan Traffic Accident Facts, 1987).
Michigan highway deaths have been at this level or higher for
several years. Within the United States during 1988 there were
over

48,000

people

killed

as

a

result

of

motor

vehicle

collisions (National Safety Council, 1989).
In the state of Michigan,

sworn police officers have a

statutory duty to investigate motor vehicle accidents.
police department

is required by state and

federal

Every
law to

report the details of traffic accidents to a central location
within their state.

The Michigan Department of State Police

is assigned the function of compiling accident records for
Michigan.
Information from highway traffic accident reports is the
basis for many decisions that are made by branches of federal,
state

and

local

governments.

In addition to governmental

units and various levels of courts, the insurance industry and
educational

institutions

base
1

decisions

upon

information

received from accident reports.

When a police officer files

a report, prosecutors make decisions about criminal charges
that can be made against a driver of a vehicle involved in a
serious personal injury or fatal collision.

That decision to

prosecute (or not) obviously can have a lasting effect on the
driver and families involved.
individuals,

units

corporations

are

of

government

based

upon

investigating police officer.
high.

Many civil suits directed at

the

and

private

report

and

public

submitted

The cost of civil

by
suits

the
is

Government units, insurance companies, individuals and

corporations in Michigan spend millions of dollars every year
defending themselves in civil suits and/or paying judgements
awarded by courts and arbitration panels.
Federal, state and county highway officials and engineers
use

the

compiled

decisions

on

data

from

police

highways

they

design,

reports

to

maintain

base
and

their

repair.

Decisions on speed limits, placement of signs or signals and
parking are only a few of the decisions that are often based
upon accident investigations made by police officers.
Many universities, including Michigan State University,
conduct research on specific aspects of highway transportation
and safety.

Grant dollars from state, federal, corporate and

private agencies are being used to perform evaluations which
are

used

Faculty

to
at

make

Michigan

recommendations
State

about

University

and

transportation.
University

Michigan are currently using accident reports prepared by

of

3
police officers as the primary or sole source of data

for

state funded research projects.
The

National

Highway

Traffic

Safety

Administration

(NHTSA), the Insurance Institute for Highway Safety and other
private and public groups use police reports as a basis for
recommendations for legislation and the development of federal
standards

that

control

certain aspects

of vehicle

design.

Classification of collisions and type of injuries received by
vehicle occupants originate with investigating police offi­
cers,

and

are

later

interpreted

determine vehicle safety features.

by

the

above

groups

to

This can obviously have an

effect on vehicle manufacture in terms of the cost to produce
vehicles that meet federal safety standards.
The

citizens

of

Michigan

who

own

and

operate

motor

vehicles are subject to the state's no-fault insurance law.
Whether the motoring public agrees with the insurance law or
not, in part the passage of that law was based upon accident
statistics compiled in police t r a f f i c accident reports.
Michigan

drivers

are

also

aware

of

the

insurance

required by their individual insurance company.

Most

premium

Insurance

premiums are in part based upon accident records which provide
the

information about the type of collision and resulting

injuries.
Despite the large number of motor vehicle collisions and
the

impact

the

results

of

the

investigation

can

have

individuals and the transportation system in general,
limited

mandatory

training

is

provided

in

the

on

only

specific

4
knowledge and skills the police officer needs to properly
investigate a traffic accident.
Required Basic Police Training in Michigan
Public Act 203 of 1965 was the legislation that required
statewide

control

of

employment and training standards

police officers in Michigan.

of

Since that time, most of the

training and education police officers received was governed
by

the Michigan

(MLEOTC).

Law Enforcement

Officers Training

Council

This council consists of three Sheriff's,

three

Chief of Police appointees, one member from the Detroit Police
Officers' Association, one member from Metropolitan Club, the
Attorney General and the Director of State Police.
MLEOTC determines the requirements for mandatory basic
police

training

programs.

and

advanced

in-service

of

training

In Michigan once an applicant meets all require­

ments to enter a police academy,
hours

police

mandatory

training.

he/she must complete 440

Only

18

hours

consists

of

instruction on how to complete the State's accident report
form; very little instruction is provided on how to actually
investigate a traffic accident.
The second area MLEOTC
classification

of

advanced

focuses upon
in-service

is categories

training

or

programs.

MLEOTC approval must be obtained by individuals or organiza­
tions prior to their presenting in-service programs.

5
Background of the Problem
There
address

are

the

agencies,

numerous

various

in-service

needs

of

training

police

work.

programs

that

Individuals,

colleges or universities who wish to present in-

service training to police officers must have MLEOTC approval.
MLEOTC establishes training standards, authorizes training and
is

also

responsible

for conducting

on-going

evaluation

of

training programs to assure a sustained level of quality.
MSU provides MLEOTC certified training to police offi­
cers.

One subject matter frequently presented by MSU over the

past decade is traffic accident investigation.

Other subjects

include radar, police alcohol enforcement, traffic engineer­
ing,

supervision,

management,

pursuit

driving

and

other

related police topics.
When
training
mastery

a

police

or
of

an
the

officer

in-service
subject

is

in

Michigan

completes

training program,
based

upon

or

the

academy

officer's

demonstrated

successful completion of a written post-test.

by

There is very

little research to date that addresses the need for continued
assessment of the police officer.
and

post-test

degree

of

results

retention

with

Emphasis has been in pre-

little

of knowledge

attention
and

skills

given
needed

to
to

the
do

specific tasks accurately and safely.
Michigan

State University

has

addressed the

need

for

additional and improved police officers' accident investiga­
tion knowledge by presenting training programs that deal with
accident investigation.

A five day (35 hour) training program

6
entitled

Accident

Investigation

One

(AI-1), developed

by

Michigan State University, Civil and Environmental Engineering
and the Highway Traffic Safety Programs has been certified by
MLEOTC.

The AI-1 training program has become the "standard"

for Michigan police officers to successfully complete once
they have been employed by a police agency and have accident
investigation duties.
To date, MSU has also relied on use of a post-test to
measure the officer's mastery of AI-1 training materials.
attempt has

been made to evaluate

No

the cognitive knowledge

retention of police officers who have received MSU and MLEOTC
certification after having completed the AI-1 training course.
Currently,

there

are

no

requirements

stating

that

police

officers must receive refresher training or be tested in later
years to determine if they have maintained their AI-1 knowl­
edge.
Problem Statement
Police

officers

are

required to

reports on motor vehicle accidents.

investigate

and make

The content of accident

reports can have a profound effect on many people,

either

directly or indirectly.

The problem is that, up until now, no

attempt

to

has

been

made

measure

the

retention

of

AI-1

cognitive knowledge once the officer has been back in the
field

and

performing

automobile accidents.

the

assigned

duty

of

investigating

7
Purpose of the Study
The researcher’s purpose in this study was to investigate
the degree to which cognitive knowledge gained in the AI-1
training course may be retained over varying periods of time
while police officers are performing their assigned duties in
the field.
Need for the Study
The Michigan Office of Highway Safety Programs (MOHSP),
National

Highway

Traffic

Safety

Administration

(NHTSA),

Michigan State University, Civil and Environmental Engineering
(MSU CEE) and the MLEOTC indicated interest in a need for the
assessment of the retention of knowledge and skills mastered
in AI training programs.

The interest and need for this type

of study has been indicated from the above organizations by
the following.
1.

The

Michigan

Office

of

Highway

Safety

Planning

(MOHSP) which has funded MSU Highway Traffic Safety Programs
to develop and present accident investigation training has
advised MSU that funding for this type of training will end,
if "long term" benefit to the police officers and the state in
general cannot be demonstrated.
2.
Council

The

Michigan

(MLEOTC)

which

Law
is

Enforcement
responsible

Officers
for

Training

certifying

the

accident investigation training, needs documentation to help
it decide if annual refresher AI training should be mandatory.

8
3.

The National Highway Traffic Safety Administration

(NHTSA) is currently proposing national standards for police
traffic accident investigation.

The proposed standards will

address:
a.
b.

accident investigation training program content;
hours

of

accident

investigation

training

needed

before an officer will be recognized as a competent
investigator;
c.

standardized national testing; and,

d.

recertification

based

upon

additional

training

and/or testing, in the frequency of recertification
as a primary issue.
4.
tal

The Michigan State University Civil and Environmen­

Engineering,

HTSP)

which

is

Highway Traffic Safety
currently

developing

Programs

(MSU CEE

additional

accident

investigation training programs, while continuing to present
the Accident Investigation 1 (AI-1) training program, needs to
know how much cognitive knowledge from previous AI training
has

been retained.

departments

Administrators

frequently

request

from

Michigan

refresher classes

police
and/or

updated AI training be provided by Michigan State University.
A decision must be made to determine the need,

content and

frequency for this type of follow-up training.
5.

This

study

could

be

shared

with

other

training

and/or enforcement organizations which have accident investi­
gation training responsibilities.

9
Research Questions
This

study

was

research questions.

directed

at

answering

the

following

Hypotheses based upon these questions

follow.
1.

Is there any variation between the mean score of the

original performance tests administered to the test group at
completion of their respective AI-1 course and the performance
retest mean score of the test group as a whole.
2.

After varying intervals of time have elapsed since

completing the Accident Investigation 1 (AI-1) program, what
degree of retention will officers in one interval of time have
as compared to officers in other intervals of time?
3.

What relationships,

if any, exist between level of

education and the written performance retest scores?
4.

What

relationships,

if

any,

exist

between

the

performance retest scores and the number of vehicle accidents
an officer investigated within the one year period prior to
taking the performance retest?
Research Hypotheses
The following hypotheses were tested in this study:
Hypothesis 1 :

The total sample of officers when treated
as one group will not show significant
loss of retention as measured by the
original performance test scores compared
with the performance retest.

Hypothesis 2 :

When dividing the total sample of offi­
cers into groups based upon varying in­
tervals of time,
officers'
retention
levels will not vary significantly
between the groups as reflected by the
mean scores on the performance retest.

10
Hypothesis 3 : Officers' education level will not show a
significant relationship with performance
retest scores.
Hypothesis 4 : Officers' age will not show a significant
relationship
with
performance
retest
scores.
Hypothesis 5 : The number of vehicle traffic crashes
investigated by officers during the one
year period prior to their taking the
performance retest will not show a sig­
nificant relationship with performance
retest scores.
Definition of Terms
Accident: An unplanned event causing contact between two
vehicles or objects resulting in damage and/or injury.
AI-1:

The

Police

Traffic

Accident

Investigation

Training Program developed and presented by MSU HTSP.

1

The

course is a 35-hour program that provides instruction to inservice officers on how to conduct a comprehensive on-scene
investigation of a traffic accident.
Correlation:

A numerical index of the degree of rela­

tionship between two variables on the same population.
Elapsed Time: The amount of time between the completion
of AI-1 and taking the performance retest.
Investigation: A systematic examination and gathering of
all facts and related information connected with a situation
commonly called a highway vehicle collision.
Original Performance Test:

The AI-1 post test that was

administered to the police subjects in this study when they
originally took AI-1.

This is a written (paper and pencil)

11
test consisting of multiple choice,

true/false,

calculation

problems and picture identification.
Performance Re-Test:
post-test

required

to

be

Same as performance test.
administered

at

completion

AI-1
of

a

MLEOTC-certified training program.
Retention:

The ability to recall cognitive information

to correctly answer test questions.
Organization of Study
This research is presented in five chapters.
contains a review of the literature.
current

literature

related

to

aspects of cognitive retention.

Chapter 2

Included is a review of

research

concerning

various

The third chapter contains

a description of research methods used in this study along
with validity and reliability concerns and a summary.

Chapter

4 is a review of the implementation of this study and contains
the research findings.

The summary, conclusion, and recommen­

dations are presented in Chapter 5.

CHAPTER 2
REVIEW OF LITERATURE
Retention has been of great interest to psychologists,
educators,

researchers

of

human

development

concerned with the learning process in adults.

and

others

Investigations

of time-related retention under various conditions have been
reported in the literature.
concerned

with

knowledge

is

evaluating

retained

studies were reviewed.
retention

and

the

over

Since this study was primarily
the

degree

time,

to

two

which

types

cognitive

of

retention

First were studies of the nature of

process

by

which

retention

is

formed.

Second were studies of the relationship of retention to such
factors as using pre- and post-test to measure the retention.
The major purpose of this study was to investigate the
recenL.j.011

gj .

cogiixuxve

namely police officers.

knowledge

as

it applies to adults,

The studies cited in this review are

related to longitudinal retention of cognitive knowledge.
Historical Perspective of Retention Studies
Philosophers have analyzed the processes of the mind in
great detail.

This interest began in ancient Greece.

Between

the period of 1650 and 1850 the school of British Philosophy
wrote mostly about empiricism.

12

Its contributions were from

13
great philosophers such as John Locke, George Berkeley, David
Hume and James Mill.

Even though most of their philosophies

differed, they all agreed upon the basic doctrine of empiri­
cism, which states that all knowledge comes from experience or
that

the

contents of the mind

are

learned.

The European

rationalist philosophers such as Rene Descartes and Immanuel
Kant believed that peoples' ideas are innate and that knowl­
edge comes with heredity rather than experience.

According to

the rationalists, people experience events in terms of space,
time, and causality because this is the way the mind imposes
structure on one's experiences of the world and not necessari­
ly the way the world is.
Up to the 1930s, philosophers believed that people forgot
(did not retain information)

because of the Law of Disuse.

The notion was that the memory trace that was not exercised
faded away.

We still hear "use it or lose it" when talking

about knowledge and/or skills.

Thorndike (1914) stated that

"When a modifiable connection is not made between a situation
and a response during a length of time,
strength is decreased."

that connections's

McGeoch (1932) attacked the theory of

Law of Disuse from several points of view.

He said, "Time of

itself does nothing about the cause of forgetting."
role

time plays

operate.

is to provide the

cause

of

The only

forgetting to

He said that, to find the causes of forgetting, one

has to look at what happens during the retention interval.

He

stated "Extinction occurs when a habit is being exercised and
spontaneous recovery when it is not."

This is the opposite of

14
what the Law of Disuse predicts.

Retroactive interference

could result from everyday activity and during the retention
interval could play a part in lack of retention.
Retention Defined
According to the Thesaurus of Psychology:
(1988),

"Retention

is

the

persistence

of

Index Terms.

a

learned

act,

information or experience as measured by reproduction, recall,
recognition or relearning. It is applied to long term or short
term memory."

(p. 291)
Retention Measurements

Ebel (1972) stated that educators used tests to measure
students' achievement and evaluate their educational progress
within the classroom.

Law and Bronson

(1977)

stated that

"Criterion referenced tests have gained in popularity until
today they provide

an alternative to the more traditional

norm-referenced test."

Glazer and Cox (1981) carried on the

investigation to discover that "measuring of training success
can be

classified

referenced."

as

either norm-referenced

or

criterion-

This theory coincided with Law and Bronson's.

Norm-referenced means that test scores are compared to scores
made by some defined reference group.
relative scale of measurement.

They are compared on a

Criterion-referenced assess­

ments are made by assessing the effects of the training on
observable

job-related

critical area exists.

behavior.

Little

research

in this

15
Effects of Retention Interval
Ebbinghaus
effects

of

(1885),

retention

who

specifically

intervals,

investigated

discussed

the

properties

of

retention and characterized the rate of forgetting through
intervals which he called "classic retention function."

He

introduced a method whereby memory could be measured objec­
tively

during

a

situation

where

the

time

and

original learning could be carefully specified.
were

quantitative

as

well

as

objective.

degree

or

Observations

Ebbinghaus

was

obsessed with the ideas of precise measurements by control of
factors

such

as mental

learning took place.

attitude and the time

of day that

His publication in 1885 opened a new era

to scientific investigation.
The primary thrust of Ebbinghaus's contribution was the
theory of retention of associations over a period of time.
agreed with the British empiricism theory.
subjectively,

retention

depended

upon

He

To Ebbinghaus

two

factors:

the

initial depth of impression on memory and the amount of time
it took for the impression to fade.

He sought objective,

quantitative measurements that could be used to measure depth
of impression and time that the impression faded.

He replaced

depth of impression with counting the number of times the list
(of nonsense syllables) was repeated in the original learning.
To

determine

the

extent

the

impression

faded

(had

been

forgotten), he proposed that the list be relearned after an
interval of retention was given.

Forgetting could be measured

by comparing the number of trials

(time interval) needed to

16
relearn the list of words with the number of trials
needed to learn the list in the first place.

(time)

He called this

comparison "savings" score and established a percentage.
EXAMPLE OF FORMULA USED BY EBBINGHAUS

Percentage savings = 100 x

Trials to learn trials to relearn
Trials to learn

If it took 40 trials to learn the initial list, according to
the formula,

it would take 40 trials to relearn,

"savings"

equal 0 percent, no effort is saved as a result of previous
practice on the list.
the

time

of

So there is no evidence of retention at

relearning.

If

relearning

requires

no

effort

(practice) and can be recited perfectly at the first attempt,
"savings" equal 100 percent.
"savings" equal

50 percent,

If relearning takes 20 trials,
and so on.

The savings score

provides a very sensitive measure of retention.
retention was

about

32 percent

of the

Ebbinghaus1s

learned

list which

showed that the retention of learning in a 24-hour period was
considerable.

Repetition beyond the point of learned material

he called "overlearning".
Curve of Retention
Ebbinghaus (1885) is also credited with the idea of the
curve of retention.

"The curve of retention is a graphic

representation of what remains in the memory, displaying the
frequency of correct reproduction of learned information over
a

period

of

time

or

a

Dictionary of Education).

number

of

trials"

(International

Through the use of the curve of

17
retention,

Ebbinghaus

concluded

that

"if

the

quantity

of

matter to be learned is increased slightly, the time taken to
learn it increases considerably."
Ebbinghaus experimented in trying to plot the function of
retention

as

it

related

to

time.

He

memorized

lists

of

nonsense words with three letters such as bok, jiw, by taking
two consonants and placing a vowel between them so as not to
make a German word or sense.

He learned a list of these words

and later relearned another list of 1,228 words and made the
nonsense words by adding 13 syllables each.
dropped

off

rapidly

at

first.

After

19

His retention
minutes

he

had

retained about 59 percent and retained about 45 percent after
one hour.

He

found that the rate of

slower over longer periods of time.

forgetting was much
He retained about 28

percent of the information after two days while retaining 21
percent after 31 days.

The change in the rate of forgetting

over time intervals at first showed rapid decline and later
became slower at longer time frames. This he called "classic
retention

function".

mathematical model
reached

zero

as

to

the

He

explored

fit his
time

the

idea

of

retention curve.

interval

reached

making

a

Retention

infinity.

He

believed that his retention became lessened in the memory with
time.
Retention Affected by Degree of Learning
According to Ebbinghaus's theory,
line

relationship between

the

number

there is a straight
of original

learning

18
trials and the number of trials saved in relearning.

The more

trials of original practice, the less effort is necessary for
relearning.

For example:

if one spoke German as a child and

did not use this learning as an adult, one's memory of German
would come back with a little practice in German to the degree
that one knew it before.
Ebbinghaus's

theories

were

highly

criticized

as

to

methods he used but he made a great contribution at a time
when higher mental processes were the obsession of philosophi­
cal

psychology,

when

observations were

all

conditions of observation were uncontrolled.

subjective

and

He introduced a

method of measuring memory objectively where a given situa­
tion, time and degree of original learning, could be carefully
specified.

He also pioneered the retention research.

Kirkpatrick (1894) discovered that false recognition of
words semantically related to those in the list based on word
meaning of other words could cause the person to match the
wrong words to each other.
theory.

However,

Kirkpatrick refined Ebbinghaus's

Kirkpatrick's

theory

received

little

attention until Atkinson and Shiffrin (1971) made observations
using a two-process theory.

The sensory register (auditory or

visual)

information

registers

stimuli

store (STS) of memory.

last

about

working memory.

the

short-term

Data decay rapidly from visual sensory

information in about 1/2 second.
data

into

3 seconds.

In auditory information the

The

short-term

store

is

the

It is stored and copied into permanent memory

which is called long-term store (LTS).

Short-term store may

19
not last long enough for retention.
and prolongs the data in memory.

Rehearsal strengthens STS

LTS is like a storage place.

Memory can be retrieved and is guided by control processes.
Data come to the memory involuntarily and can be retrieved
without conscious effort.

(Example:

6x4=24 recalls without effort,
memory.)

a person who has learned

if the problem is stored in

This retention is related directly to LTS and by­

passes STS according to the authors.
Ausubel

(1968) tested the hypothesis that the "learning

and retention of unfamiliar but meaningful material can be
facilitated by advanced introduction of concepts."

The basis

of his theory was that cognitive structure was organized in a
hierarchy of concepts and subconcepts.
the

process

of

acquisition

and

In order to facilitate

retention

of

meaningful

material, Ausubel advocated the use of advanced organizers.
The most efficient way of facilitating retention is
to introduce appropriate subsumers and make them
part of the learning task. The introduced subsum­
ers thus become advance organizers' or anchoring
foci for the reception of new material. In effect,
they provide an introductory overview at the appro­
priate level of conceptualization, (pp. 8-9)
He

identified three important attributes which support the

need to incorporate them into the learning materials:

(1)

Advanced

for

organizers

that

are

specifically

constructed

specific learning materials are an efficient method used to
facilitate the retention of subject matter, (2) Organizers are
better used with factual materials than for abstract
cepts, principals or generalizations) materials,

(3) In

(con­

20
advance

learning

materials,

organizers

would

facilitate

integration and thereby retention would be more effective.
Blanton

(1972) clarified Ausubel's theory of Reception

Learning by stating "In theory of meaningful verbal learning,
a major concept is that new material which is to be learned
must be related to the learner's existing cognitive struc­
ture."
Tulving and Watkin (1973) gave an interesting twist to
Ebbinghaus's theory.

They had a list of five-letter words and

gave the subject two, three, four and five letters of the word
as cues and asked the subject to write the word from the list
they were to retain using the cues.
might be Pepsi.

(Examples:

They could be given Pe, Pep,

they were to recall the word Pepsi.)

the word

Peps, Pepsi,

The authors believed

that retrieval cues might be necessary for memory retrieval
and,

if

not

necessary,

would

at

least

be

very

helpful.

Critics have stated that the data from memory depend on the
memory cue and also on the surrounding nontext in which the
cue is given.

They have cited where guessing by using the cue

would have a 50 percent correct response.
Anderson and Biddle
between

increased

subsequent
concur

with

(1975)

application

retention.
Anderson

LaPorte
and

found a strong relationship
(practice)

of materials

and Voss

Biddle

in

(1975)

their

and

seemed

studies.

to

They

demonstrated that superior retention was the result of usage
of the information and performance feedback.

21
Singer (1975) stated that retention can be determined by
"measuring

the

difference

between

the

learned and the amount forgotten".

amount

originally

Acquisition of subject

matter to become functionally competent, lends complexity to
the students' ability to retain information by the referring
of concepts and content transfer to practical application of
knowledge.
style

Davey (1976) concluded that a student's cognitive

defined

as

stable

preferences

in

individuals

with

respect to conceptual categorization and perceptual organiza­
tion of the external

environment was a critical

factor in

maximizing retention and performance.
Sage

(1977)

stated

that

retention

is

the

proficiency after a period of non-practice.

savings

Mehrens

of

(1986)

concurred with Sage in his studies that "criterion referenced
tests have gained in popularity until today they provide an
alternative to the more traditional norm-referenced tests".
Glazer and Cox (1981) concurred with Mehrens by reporting that
"Measures of training success can be classified as either
norm-referenced or criterion-referenced." (pp. 352-359) Normreferenced means that test scores are compared to scores made
by

some defined

reference group.

relative scale of measurement.

They are compared on

a

Criterion referenced assess­

ments are made by assessing the effects of the training on
observable

job-related

behavior.

Little

research

on this

critical area exists.
Gagne
different

(1978)
types

predicted
of

learning

that

students'

environments

exposure
would

to

affect

22
performance on tests of academic ability.

This study examined

performance in two different learning environments in order to
detect changes in retention of academic performance resulting
from participation in one of the two environments during the
learning-interval.

He stated that the traditional classroom

is not the only or the best learning environment; meaning that
the outdoors, home or job could be a learning environment.
Longitudinal Retention
The
defined

ultimate
by

results

Cornwell

(1980)

trainees after training."
es

of

learning

in

longitudinal

as

"on-the-job

retention

were

performance

of

He noted that "one of the weakness­

evaluations

designs

is

the

failure

to

do

longitudinal studies... and that the post-training application
of skills is difficult to measure without access to employees
back on the job"

(p. 99) .

Swierczek and Carmichael

(1985)

concurred with this theory.
Retention Affected bv the Strategies Used
The effectiveness of the strategies used to learn tasks
has been found to have a significant relationship to reten­
tion.
use

Waters (1982) studied the relationship between strategy

and

memory

performance

and

found

that

there

differences in recall as a function of strategy.
may have

resulted

from the

fact that most

were

no

This finding

students

adopt

strategies that are relative to the task that they are to
perform and that the study was limited to students who used

23
poor strategies or no strategies at all which would lead to
varying degrees of performance.
McLeod

(1983)

learning:

considered that there were two modes of

(1) One achieved by gradually mastering complex

cognitive skills by concentrating on sub-tasks to be mastered,
such as the correct phonetic sound for pronouncing a word.
This frees the learner to begin on another sub-task, namely,
the use of the word in a sentence or syntax.

The learner

often loses contact as to learning how these parts fit into a
whole sentence, paragraph and story.
learning was restructuring.

(2) The other mode of

It is the process of grouping

information into related units, rather than isolated bits.
Carrol
between

time

categories:

(1985)

stated

that

and

learning

and

(1) engaged time,

allocated time.
learning

there
he

was

broke

time

relationship
into

(2) time on a task,

and

(3)

will maximize the

Carrol cautioned that time spent in learning was

crucial but not always sufficient for achievement.
stated

three

He concluded that the extent of time spent in

(allocated and/or engaged time)

learning,

a

that

the

amount

of

smallest of three factors:

time

spent

on

a task

He also
was

the

(1) learner's perseverance (amount

of time student was willing to give actively in learning) , (2)
time allocated for learning, and (3) time needed for learning.
Pressley
strategies

and

(1985)

concurred

retention;

theory one step further.

with

however,

the
he

other

seemed

studies
to

take

on
the

He suggested that strategic knowl­

edge involved more than simple awareness that a particular

24
strategy was useful in cognitive learning.

He believed that

knowledge about the conditions for strategy use, the effort
required

and

the

situations may
particular

all

usefulness
contribute

strategy.

Such

of
to

the

strategy

the decision

strategy

in

previous

to

knowledge

adopt
would

a
be

expected to affect the strategic behavior adopted and in turn
to be affected by experience in task situations.
Cheng (1985) proposed that improvement in performance can
be due to a restructuring of the components of a task so that
they

are

units,

coordinated,

thereby

components

to

integrated

allowing
be

replaced

involving new components.
hands on a piano.

the
by

or

reorganized

procedure
a

more

involving
efficient

into

new

the

old

procedure

She gave an example of playing two

Cheng indicated that the difficulty does

not stem from any physiological limitation,

but rather from

the lack of s suitably structured skill.
Justice (1986) concluded that recall was better when the
text was structured and strategies such as a standard order of
presentation were used; for example:

introduction setting,

motivation of the protagonist and ending statement for a story
rather than the giving of unrelated verbal materials. Mixed
order of presentation resulted in the lowest recall proficien­
cy.
The Effect of Age on Retention
McGeoch (1932) attacked the law of disuse theory with the
explanation that the main reason for forgetting was because of

25
retroactive
Underwood
same

interference;

(1960)

subjects

this was

accepted

refuted McGeoch's theory,

had

been

used

over

and

conditions, which tainted the studies.

for

25 years.

stating that the

over

in

different

McGeoch's hypothesis

of retroactive interference was questioned.

Underwood argued

that the major cause of forgetting in most of his experiments
was not retroactive interference but proactive interference
which caused differences in recall.

Underwood discovered that

retention was closely related to the number of lists a subject
had learned previously in the same experience.
being tested was the first,
accurate.

and

retention was about 75 percent

When there were 20 or more previous lists, reten­

tion was below 20 percent.
lists

When the list

retention

was

The relationship between previous
almost

identical

to

Ebbinghaus's

curve.
Postman (1961) claimed;

"Interference theory occupies an

unchallenged position as the major significant analysis of the
process of forgetting."
when

scores

were

Gladis and Braun (1958)

corrected

for

differences

found that

in

original

learning, there were no age differences in retention.

Mehrens

(1986) concurred with this finding.
Scovel (1982) designed an experiment to investigate the
second

language

learning

process

in

the

setting

in which

language learning is most likely to occur in the United States
of America.

Flavell

(1982)

did a similar experiment which

entailed the initial stages of language learning. He assumed
that

since

the

experiment

entailed

the

initial

stages

of

26
language learning, there would be considerably more "consis­
tency and homogeneity" in the subjects' responses relative to
their age groups than in their "subsequent cognitive manage­
ment"

of

difference

this

learning

event.

among age-related

Results

learning

revealed

in terms

of

little
overall

test-score results.
Summary
Ebbinghaus's

(1913)

findings

seemed

to

indicate

that

there is a positive relationship between the time that data
are given and the retention of these data.

However,

in his

studies he did not consider the content of data or procedures
for the data to be given which seems important in trying to
prove the extent of such relationship

(time intervals

and

retention of data).
Ebbinghaus's study showed evidence of a positive rela­
tionship of properties of retention and the rate of forget­
ting.

Kirkpatrick's and Atkinson's investigations revealed

similar trends about this relationship.

Retention seems to

appear in stages or in time frames and is manifested by (1)
the time that data were presented,

(2) the learning of the

data (3) the amount of time needed to recall the data, and (4)
how much of the data were retained.
In summary, the following conclusions seem to be support­
ed by evidence from previous studies:
1.

From grade level to grade level there seems to be an

increased ability for retention.

27
2.

Adults

retain data,

if they are useful

to their

needs (job) .
3.

Attempts have been made to determine the relation­

ship of memory and retention but no firm conclusions have been
reached.
4.

During

a

retention

interval, when

forgetting

is

correlated with time, there is sometimes retroactive interfer­
ence.
5.

Some

research

has

suggested

that

the

causes

of

forgetting over short- and long-retention intervals are not
different,

while

other

research

shows

that

forgetting

is

greater immediately after learning and the forgetting process
slows as time increases.
6.

Most of the forgetting studied in memory experiments

can be classified as either proactive or retroactive interfer­
ence.
7.

By restructuring components of tasks so that they

are organised,

integrated and reorgani zed,

performance and

retention are increased.
8.

There

are

no

age

differences

in

retention

when

scores are corrected for differences in original learning.
This

review

of

the

literature

led

the

researcher

to

conclude that little, if any, specific research has been done
to determine whether police officers retain cognitive informa­
tion over varying periods of time.

However,

studies have

dealt with longitudinal assessment of knowledge retention.

It

seems that the latter authors/researchers have based their

28
theories upon Ebbinghaus's ideas and have expanded the idea to
conclude

the

restructuring

of

learning

into

related

units

rather than isolated bits, as has been promoted by learning
theorists up to the present time.
Discussion of Previous Research
This research investigated the possible relationship that
time, age, educational level and/or the number of accidents
investigated may have

to the police

officers

longitudinal

retention of cognitive knowledge gained in the AI-1 training
program.

The testable hypotheses stated in chapter 3 are

based upon information the researcher gained from the review
of the literature.

Hypotheses can be stated in the null form,

which does not necessarily reflect the researcher's expecta­
tions.

The

null

form states

that

no

relationship

exists

between the variables.
In
expect

cases
a

direction,

where

significant

the

researcher has

difference

to

strong

occur

in

reasons
a

to

specific

hypotheses can be written to reflect direction.

Borg and Gall (1974) stated that "hypotheses that state (sic)
a specific expected direction for the finding, should only be
used when there is little or no possibility that the findings
will yield a difference in the opposite direction."
Information provided by Ebbinghaus, Kirkpatrick, Anderson
and Biddle indicated that retention can be affected by time
and use of the cognitive knowledge.

A search and review of

the literature did not provide adequate information to enable

29
this researcher to make predictions about the relationship
that the variables of age and educational level may have to
retention.

CHAPTER 3
RESEARCH METHODOLOGY
The Sample
The researcher's purpose in this study was to determine
the degree to which cognitive knowledge mastered by police
officers who

attended the Accident

Investigation

1

(AI-1)

training program has been retained over varying periods of
time.

The police officers who participated in this longitudi­

nal study were:
1.

AI-1

trained

officers

from the

Eaton

County

from the

Ingham

County

Sheriff's Department;
2.

AI-1

trained

officers

Sheriff's Department;
3.

AI-1 trained officers who attended three MSU ad­
vanced level AI training programs during September
of 1989.

The
separate

participating
groups

based

completion of AI-1.

officers
upon

were divided

into

three

the total elapsed

time

since

The time divisions were,

months through 12 months,

(1) more than 4

(2) more than 12 months through 36

months, and (3) more than 36 months.

30

In this study the focus

31
was

on

the

retention

of

cognitive

knowledge

over

varying

periods of time, as measured by scores on written tests.
Testable Hypotheses
The hypotheses
form.

for this

study are stated in the null

The null hypothesis states that there is no relation­

ship among the variables identified for this research, which
are mean scores for the original performance test, mean scores
of the performance retest, time, age, educational level and
the number of traffic accidents investigated during the 12
months prior to taking the performance retest.

The statisti­

cal tools selected to test the hypotheses and the levels of
confidence are identified in other sections of this chapter.
Two research hypotheses were tested to examine AI-1 cognitive
knowledge retention over time.

Three additional hypotheses

were tested to determine whether age, education level and the
number

of

traffic

accidents

investigated

were

related

to

retention over time.
The first evaluation was to determine whether there was
a significant difference between the computed mean score of
the original performance test taken by the police officers
included in this study at completion of their respective AI-1
program as

compared

to the mean

score

of

the performance

retest taken by the same group of police officers.
Hypothesis 1 :

The total sample of officers when treated as
one group will show significant loss of reten­
tion as calculated by the performance retest.

32
The second evaluation was to determine whether there was
a significant difference in mean scores based upon the amount
of elapsed time since completion of AI-1 training.

The mean

scores of the performance retest were used to compare the
three groups, which were divided by time segments as follows:
Group 1

=

more than 4 months to 12 months

Group 2

=

more than 12 months to 36 months

Group 3

=

more than 36 months

Hypothesis 2 :

When dividing the total sample of officers
into three groups based upon interval of time
since completing the AI-1 program, officers'
retention levels will vary significantly as
reflected by the group mean scores on the
performance retest.

The third evaluation was to determine whether a relation­
ship existed between education level and retention over time.
The three subgroups for education were defined as follows.
Group 1

=

High

School

Diploma

but

less

than

an

Associate Degree
Group 2

=

2-year Associate Degree but less than a
4-year degree

Group 3
Hypothesis 3 :

=

4-year degree or more

Officers' education level will
significant
relationship
with
retest scores.

not show a
performance

The fourth evaluation was to determine whether a rela­
tionship existed between age and retention over time.

The

four subgroups for education level were defined as follows:
Group 1

=

24 to 34 years of age

Group 2

=

More than 34 to 37 years of age

33
Group

3

=

More than 37 to 42 years of age

Group

4

=

More than 42 years of age

Hypothesis

4:

Officers' age will not show a significant
relationship with performance retest scores.

The fifth evaluation was to determine whether there was
arelationship
officers
retest.

between

investigated

the number
and

mean

of

scores

traffic
on

the

accidents
performance

The sample was divided into five subgroups based on

the number of personal injury and/or fatal accident investiga­
tions conducted during the 12-month period before they took
the performance retest.

The five subgroups were defined as

follows.
Group 1

=

No accidents investigated

Group 2

=

1-10 accidents investigated

Group 3

=

11-35 accidents investigated

Group 4

=

36-65 accidents investigated

Group 5

=

66 or more accidents investigated

Hypothesis 5 :

The number of personal injury and/or fatal
vehicle traffic crashes investigated by offi­
cers during the one year period prior to
taking the performance retest will show a
significant positive relationship with the
mean scores on the performance retest.

AI-1 Original Performance Requirements
To complete the AI-1 training course successfully, MLEOTC
required the officer to achieve a score of 70 percent or above
on the original performance test.

The performance test has a

total of 62 questions, which equals a total of 100 points.

34
Population Selection
The

most

practical

procedure

to

obtain

a

sample

of

Michigan police officers was to administer the performance
retest at the beginning of other training programs that MSU
HTSP presents at central training facilities.

During Septem­

ber 1989, MSU scheduled three advanced Accident Investigation
Training programs.

One requirement of attending the advanced

program was that officers have successfully completed the AI-1
training.

One advanced program was conducted at each of the

following

three

sites:

Macomb

County

Criminal

Justice

Training Center, Detroit Metropolitan Police Academy in Wayne
County and the Southfield Police Department in Oakland County.
Over the past several years a large percentage of the officers
who attended AI-1 received their training at these same three
locations.
Additional officers included in the study were the AI-1
trained officers from the Ingham and Eaton County Sheriffs'
Departments = The officers from these two departments received
their AI-1 training through Lansing Community College Justice
and

Law

Center.

Approximately

80

percent

of

the

police

officers in Michigan who have attended AI-1 received their
training at one of these four training facilities.
Performance Retest Procedure
Officers attending the three advanced training programs
in Wayne, Macomb and Oakland Counties were asked at the start
of the class session if they would be willing to take the

35
performance retest.

All officers in the class were willing to

be involved in the retest.
The officers from the Ingham and Eaton County Sheriffs'
Departments were assigned to a classroom within their police
facility.

The

performance

beginning of duty.

the

individual

was

scheduled

at

the

Test subjects did not receive advanced

notice of the retest.
of

retest

This procedure reduced the possibility

officers

preparing

for

the

performance

retest.
In police departments, it is not uncommon for administra­
tion

to

order

personnel

to

do

certain

things.

In

this

performance retest situation, only those who were willing to
participate were given the test.

Several testing sessions at

each sheriff's department were necessary to obtain the maximum
number of AI-1 graduates from each department.
Test Instrument
The measurements taken were test scores from a validated
test (Nerbonne, 1980), which was also certified by the MLEOTC
(see Appendix D & E) .

The written paper and pencil perfor­

mance retest instrument was the same one used at completion of
the AI-1 training program that each subject attended.

Scoring

of the performance retest was completed in the same fashion as
the post-test administered at completion of each AI-l course.
The same scoring key and assessment were used.
designed

to

knowledge.

measure

the

degree

of

AI-1

The test was

content

cognitive

36
Validity Concerns
The same two primary instructors have taught all the AI-1
courses

that

the

participants

in

this

study

completed.

Training course material outlines were revised and/or updated
but

the

basic

content

has

remained

the

procedure of presentation and material

same.

This

set

content reduced the

possible variations that could arise if many instructors using
various materials were employed.
The total number of subjects in the study was 100.
upon

the

number

of

subjects,

individual groups were small.
ered

a threat

research,

if

to validity
any

cell

for

some

of

the

Small cell size can be consid­
(Borg

individual

sizes

Based

cell

& Gall,
size

1971).

created

concern, a comparison of that cell was not made.

In this
a validity
A minimum

number of five was the smallest acceptable/testable cell size
for this study.
The

use

of

identical

original

performance

tests

and

performance retests was not a threat to internal validity for
the following reasons:
1.

The time period that elapsed between the testing

periods was between four months and 120 months.
2.

At completion of the AI-1 post-test, the students

received their scores but were not provided with copies of the
test instrument or the correct answers.
3.

The test

instrument was the same; therefore,

the

test content could not be more difficult or easier as a result
of test change.

37
In longitudinal surveys of this nature, maturation can be
a variable that affects internal validity.

In this situation

biological or psychological processes could have an effect on
remembering over a period of time.
periods

Other events over extended

of time may also affect retention.

This research

included the types of variables associated with maturation;
they were considered

only in the age,

education

level and

experience comparisons which are addressed in Hypotheses 3, 4
and 5.
Reliability Concerns
Some
reliability

factors
of

can

change

adversely
scores,

affect

especially

measure of pre-and post-test results.

the

meaning

those

used

and
as

a

This researcher did not

deal with change scores in the pre-test, post-test fashion.
The mean scores of the original performance test were analyzed
against the mean scores of the performance retest,

and the

retest means were also compared to the mean scores of the
groups based upon time, education and experience.

Generally,

change scores will be more reliable when experimental data are
not being analyzed.

The reliability of the test used in this

type of study was defined as the consistency of the test in
making estimates of the student's level of mastery of the
test's domain (Borg & Gall, 1981).
The performance retest instrument was identical to the
original performance test given to each subject.
valid test and it does assess retention.

It is a

The AI-1 curriculum

38
and original performance test were developed

in accordance

with M S U 1s Learning System Design professors Davis, Alexander
and

Yelon.

The

AI-1

performance

test

was

validated

by

Nerbonne (1980) in a dissertation entitled "An Evaluation of
the Time Formats Used in Teaching the MSU HTSP Introductory
Traffic Accident Investigation Course."

(See Appendix A for

validation summary of Nerbonne.)
Analysis Procedure— Treatment of Data
The Statistical Package for the Social Sciences
was used to analyze the data.
tive

statistics

(group

(SPSS)

In the first analysis, descrip­

mean

and

standard

deviation)

were

computed for each comparison group in the study.
T-Test ^Paired Samples)
This

study,

like many

educational

dealt with small numbers of subjects.

research

projects,

In this situation, the

t-test is an appropriate statistical tool to determine whether
the

means

differ

significantly

from

one

another.

The

t

distribution takes into account the nonnormal distribution of
standard errors when the sample size is small
1971).

(Borg & Gall,

Because the performance retest scores in this research

could be greater or less than the original performance scores,
the differences in retention had to be examined through the
use of the two tailed t-test of significance.

Paired samples

were

score

used

on the

original

performance

test

performance retest scores for the entire sample.

and

the

The same

39
analysis was used for each of the groups formed using demo­
graphic variables.

The .05 level of significance was used.
Analysis of Covariance

Analysis
matching

of covariance provides

groups

experience.

It

on
is

variables
also

such

designed

a post-hoc method

as
to

age,

education

determine

of
and

whether

a

difference between two groups on a particular variable can be
explained by another difference that exists between the two
groups (Borg & Gall, 1971).

For this study, the performance

retest was the dependent variable,
mance test was the covariate.

and the original perfor­

Analysis of covariance makes

selected groups equal with respect to one or more control
variables.
Analysis of covariance was used

for each of the

four

demographic variables (time, age, education level and experi­
ence)

to

variable

determine

(performance

performance
measured

their

the

test).
effect

relationship

retest)
An

and

additional

that

the

the

with

the

dependent

covariate

(original

analysis

demographic

of

covariance

variables

(age,

education level and experience) had on the variable of time.
For acceptance or rejection of the hypotheses a .05 level of
significance was used.
Summary
The researcher's purpose in this study was to investigate
the relationship between the retention of cognitive informa­
tion,

originally learned in the AI-1 training program,

and

40
selected demographic and job-related variables, including age,
education

level

and

accident

investigation

experience.

Retention was measured by comparing group mean scores on the
original AI-1 performance test with group mean scores on the
same test taken in mid-1989

(the performance retest).

In this chapter the research hypotheses were presented
and the basis of study population selection was explained.
Data-collection procedures were explained,

and the methods

that were used to conduct the analysis were described.

CHAPTER 4

ANALYSIS OF DATA
The Findings
This

study was

designed to

investigate the degree to

which cognitive knowledge gained in the Accident Investigation
1 (AI-1) training course was retained over varying periods of
time while the police officer continued to perform his/her
assigned duties.

The methods and procedures for the investi­

gation were presented
chapter

are

the

in the preceding

findings

analysis of the data.
the difference

resulting

chapter.

from

the

In this

statistical

Data analyses in this chapter include

in retention as measured by comparisons

of

original performance test means and performance retest means
against correlations of time, age, education level, and number
of traffic accidents investigated in the 12 months preceding
the performance retest.
The

sections

of this

following headings:
Test

Mean

Scores

chapter

are presented under the

Introduction to the Data,
by

Variable

Subcategories,

Performance
Differences

between Mean Original Performance Test and Performance Retest,
The Effect of Time on the Performance Retest Mean Score, The
Effect of Education on the Performance Retest Mean Score, The
41

42
Effect of Test Subject's Age on the Performance Retest Mean
Score, The Effect of the Number of Accidents Investigated on
the Performance Retest Mean Score, Review of the Results of
the Analysis of Hypotheses 1-5, The Combined Effect of Time
and

Education

on

the

Performance

Retest

Mean

Score,

The

Combined Effect of Time and Number of Accidents Investigated
on the Performance Retest Mean Score, and The Combined Effect
of Time and Age on the Performance Retest Mean Score.
Introduction to the Data
The total number of subjects in the final statistical
analysis was 100:
ment,

23 from the Eaton County Sheriff Depart­

17 from the Ingham County Sheriff Department,

and 60

from the advanced level training programs conducted in Wayne,
Macomb,

and Oakland Counties.

In Eaton County,

out of the

total number of deputies who previously took AI-1, only two
deputies did not take the performance retest; one deputy was
on sick leave and one was on vacation.

One qualified deputy

from Ingham County did not participate

in the survey;

his

current assignment to the regional drug unit precluded his
participation.

All officers attending the advanced training

programs volunteered to participate in this research.
A total of 112 officers completed the questionnaire and
took the performance retest.
to be

eliminated

reasons.

from

the

Twelve of the participants had
total

number

for the

following

43
1.

Two

officers'

records

were

not

listed

in

the

computer where all AI-1 records were kept; therefore, neither
the dates that they took AI-1 nor their original performance
scores could be located.
2.

Three officers had attended similar programs through

the Northwestern University Traffic Institute at Evanston,
Illinois, but did not attend MSU's AI-1 program.

(They had to

be eliminated from the final analysis.)
3.

Seven officers who volunteered to be part of the

research

attended

AI-1

programs

hat

were

presented

on

a

segmented basis.
MSU Highway Traffic Safety Programs presented the fiveday

AI-1

training programs

in two formats.

It was

most

frequently presented in five consecutive days (Monday through
Friday).

The second format for the five days of training was

one day a week for five consecutive weeks (i.e., every Tuesday
for

five

weeks).

The

seven who attended

the segmented

programs were not used in the final analysis to eliminate the
possibility

that

differences

in

cognitive

retention

might

exist as a result of training program format.
Approximately 3,500

Michigan

attended MSU's AI-1 training.

police officers

Only about 400 of the 3,500

attended the one-day-a-week segmented format.
3,100

attended

AI-1

training

have

programs

consecutive day (Monday through Friday)

The remaining

presented
format.

in

five

44
Performance Test Mean Scores
by Variable Subcateaories
The
included

data

collected

performance

from

retest

the

scores,

participating
the

number

subjects
of

months

between original performance test and performance retest, age,
education level, and the number of traffic accidents (personal
injury and fatal) investigated during the 12 months preceding
the performance retest.
Table 4.1 lists the variables used in this evaluation:
original

performance

test,

performance

retest,

time,

education, and number of accidents investigated.

age,

Table 4.1

shows that:
1.

The test subjects' mean score by percentage on the

original performance test was 89.71, standard deviation 10.39,
minimum score 72, and a maximum score of 100.
2.

The test

subjects'

mean score by percent on the

performance retest was 75.51, standard deviation 6.55, minimum
score 55, and a maximum score of 100.
3.

The test subjects' mean time in months since taking

the original performance test was 47.85,
40.61;

the minimum time was 4 months,

standard deviation

the maximum was

130

months.
4.

The test subjects' mean age in years when taking the

performance retest was

37.55,

standard deviation 6.02;

the

minimum age was 24, the maximum age was 49.
5.
years

of

The test subjects' mean education level in number of
education

beyond

high

school

was

2.32,

standard

45
deviation 1.23; minimum number of years was 1, the maximum was
6.
6.

The test subjects' mean number of accidents investi­

gated in the 12 months preceding the performance retest was
62.77, standard deviation 103.09, minimum number investigated
was 0, and maximum was 486.

TABLE 4.1.

Mean, Standard Deviation, Minimum Range, and
Maximum Number for each of the Six Variables
RANGE
MAXIMUM
(N)

MEAN

Post test score
(0/0)
Retest score
(0/0)
Time in months
Age in years
Educ. in years
No. of Acc. Inv.

89.71

6.55

72

100

75.51

10.39

55

100

47.85
37.55
2.32
62.77

40.61
6.02
1.23
103.09

4
24
1
0

130
49
6
486

Summary and Comments;

S.D.

MINIMUM
(N)

VARIABLE

Table 4.1 Information

There was a difference of 14.2 percentage points between
the post test (original test)
test being the higher.

and retest, with the original

The standard deviation for the post

test was lower; i.e., the test scores for the post test were
more closely grouped than those of the retest resulting in a
standard deviation of 10.39,
test.

or 3.84 higher than the post

On the post test two of the 100 officers in the study

sample scored 100 percent; however, it should be noted that of
the approximate 3,500 officers who have taken the post test

46
over the past 12 years, only eight have scored 100 percent on
the test.

Only one officer scored 100 percent on the perfor­

mance retest.
The variable of time for this evaluation was based upon
the number of months that elapsed since the officer completed
the

original

performance

test.

standard deviation of time,

Based

upon

the

mean

and

it can be seen that there was a

wide spread from the shortest and longest period of time.
Although the longest period of time was 130 months, there was
only one officer with 90 or more elapsed months.
The only variable that had less deviation than age was
education.

Age measured in years had a standard deviation of

6.02 with a mean of 37.55; there was at least one subject in
every age group between 24 and 49 years of age.
The variable of education had a range from one year of
college to six years of college.

Every officer in the test

group had a high school education.
retest

subjects

had

two

years

of

The largest number of
college.

Three

retest

subjects had more than four years of college; one officer had
five years of college and only two officers had six years of
education beyond high school.

There were no test subjects who

had more than six years of college.
The last variable in Table 4.1 is the number of accidents
(personal

injury

and/or

the

test

subjects in the year preceding the performance retest.

The

range was

103

from 0 to

fatal)

investigated

by

486 with a standard deviation of

accidents investigated.

Most of the test subjects reporting

47
large numbers of accidents investigated were from the state's
largest

metropolitan

departments

where

there

were

large

numbers of traffic accidents and only a small traffic unit
assigned to perform investigations.
population

of

100,

19

officers

Out of the total research

investigated more

than

88

accidents and only five officers investigated 400 or more.
There were 15 officers who did not investigate any accidents
during the year preceding the performance retest.
Table 4.2 is a listing of the three subpopulations of the
time variable.

Information included is the mean score, the

standard deviation of the performance retest for each subpopu­
lation, and the number of cases in each subpopulation of the
time variable.
1.
than

Table 4.2 shows that:

The test subjects who have an elapsed time of more

4 months

to

12

months

since

completing the

original

performance test had a mean score of 80.69, standard deviation
of 10.32; a total of 26 subjects.
2.

Test subjects who had an elapsed time of more than

12 months to 36 months

had a mean score of 76.09,

standard

deviation of 10.05; a total of 31 subjects.
3.

Test subjects who had an elapsed time of more than

36 months had a mean score of 72.02,
9.26; a total of 42 subjects.

standard deviation of

48
TABLE 4.2 Performance Retest Mean, Standard Deviation, and
Number
of
Cases
for
Each
of
the
Three
Subpopulations of the Time Variable
TIME/PERFORMANCE RETEST

MEAN

S.D.

1

More than 4 months
to 12 months

80.69

10.32

26

2

More than 12 months
to 36 months

76.09

10.05

31

3

More than 36 months

72.02

9.26

42

Summary and Comments:

CASES

Table 4.2

The mean score of the performance retest is listed for
each time period.

The shortest time period, 4-12 months, had

the highest mean score;

12-36 months had the next highest

score, and 36 months or more had the lowest mean score.

From

the shortest time period to the longest, each category dropped
approximately 4 percent; the regression of scores is consis­
tent from time period to time period.

The standard deviation

only varied a few percentage points among all time categories.
Table 4.3 lists the three subpopulations of the education
variable.

Information included is the mean score,

standard

deviation of the performance retest for each subpopulation,
and the number of cases in each subpopulation.
Table 4.3 shows that:
1.

The test subjects who had a high school diploma but

less than an associate degree had a mean score of 72.66 on the
performance

retest,

standard

deviation

subjects were in this subpopulation.

of

10.99;

33

test

49
2.

Test subjects who had an associate degree but less

than afour-year degree
deviation
3.

had a mean score of

77.19, standard

9.99; 42 test subjects were in the subpopulation.
Test subjects who had a four-year

degree or more

than four years of college had a mean score of 76.44, standard
deviation of 9.8; 25 test subjects were in the subpopulation.

TABLE 4.3 Performance Retest Mean, Standard Deviation and
Number of Subjects for each of the Three Subpopula­
tions of the Variable of Education
EDUCATION/PERFORMANCE
RETEST

MEAN

S.D.

1

High school diploma
but less than
associate degree

72.66

10.99

33

2

Associate degree but
less than four-year
degree

77.19

9.99

42

3

Four-year degree or
more

76.44

9.8

25

Summary and Comments:
Performance

CASES

Table 4.3

retest

scores

were

the

highest

for

the

education level group that had an associate degree but less
than a four-year degree; 42 of the total test group fell into
this category.
group

did

not

The four-year degree or more education level
do

as

well

as

the

associate

degree

group;

however, there was only .75 percent difference between the two
levels.

Both of the degree-level subjects scored approximate­

ly 4.5 percent higher than the group that had high school

50
diplomas.

Standard deviation varied only 1.19 between the

three education levels.
Table
variable.
retest

4.4

the

four

subpopulations

of

the

age

Information in the table includes the performance
mean,

subpopulations,
subpopulation.
1.

lists

standard
and

the

deviation
number

for
of

each

subjects

of

the

in

each

Table 4.4 shows that:

Test subjects from 24 to 33 years of age had a mean

score of 79.79, a standard deviation of 11.08 on the perfor­
mance retest; a total of 24 subjects were included.
2.

Test subjects 34 to 37 years old had a mean score of

73.07, standard deviation 8.25; 26 subjects were included.
3.

Test subjects 38 to 42 years old had a mean perfor­

mance retest score of 76.53, standard deviation of 11.22; 26
subjects were in this subpopulation.
4.

Test subjects who were older than 42 years had a

mean score of 72.75, standard deviation of 9.8; 24 subjects
were in the subpopulation.

TABLE 4.4 Performance Retest Mean Score, Standard Deviation
and Number of Subjects for each of the Four Subpop­
ulations of the Age Variable
AGE/PERFORMANCE RETEST

MEAN

S.D.

CAS!

1

24 to 33 years

79.79

11.08

24

2

34 to 37 years

73.07

8.25

26

3

38 to 42 years

76.53

11.22

26

4

43 years or older

72.75

9.8

24

51
Summary and Comments:

Table 4.4

The effect of age on the performance retest shows some
inconsistent variations among the four divisions.

There were

one or more officers in each year for the range of 24 to 49
and as table 4.4 shows, the study sample of 100 was almost
evenly distributed among the four age divisions.

The youngest

group and the next-to-oldest group had the highest performance
retest scores and these same two groups also had the highest
standard deviation.

Ages 34 to 37 years of age (next to the

youngest group) and the oldest group of 43 years or more had
the lowest retest scores but a smaller standard deviation.
Table 4.5 lists the five subpopulations of the number of
accidents investigated in the year preceding the performance
retest.

Information in the table includes the performance

retest mean score, standard deviation, and number of subjects
in each subpopulation.
1.

Table 4.5 shows that:

Test subjects who investigated no accidents in the

preceding 12 months had a performance retest mean score of
69.84,

standard

deviation

8.78;

13

subjects

were

in this

subpopulation.
2.

Test subjects who investigated 1-10 accidents had a

mean score of 77.80, standard deviation of 11.30; 21 subjects
were in this subpopulation.
3.
a

mean

Test subjects who investigated 11-30 accidents had
score

of

75.37,

standard

subjects were in this subpopulation.

deviation

of

10.65;

24

52
4.

Test subjects who investigated 37-65 accidents had

a mean score of 81.55, standard deviation of 9.45? 20 subjects
were in this subpopulation.
5.

Test subjects who investigated 66 or more accidents

had a mean score of 71.31,

standard deviation of 7.76;

22

subjects were in this subpopulation.

TABLE 4.5 Performance Retest Mean Score, Standard Deviation
and Number of Subjects for each of the Five Subpop­
ulations
ACCIDENTS INVESTIGATED/
PERFORMANCE RETEST

MEAN

1

No accidents
investigated

69.84

8.78

13

2

1-10 accidents
investigated

77.80

11.30

21

3

11-30 accidents
investigated

75.37

10.65

24

4

31-65 accidents
investigated

81.55

9.45

20

5

66 or more accidents
investigated

71.31

7.76

22

Summary and Comments:

S.D

CASES

Table 4.5

Zero to 486 accidents were investigated by individual
officers during the year preceding the performance retest.
Test subjects who investigated zero accidents had the lowest
mean score and also the next-to-lowest standard deviation.

A

mean score for the officers who were in the division of 66 or
more accidents investigated (or an actual range of 66 to 486)

53
had the smallest standard deviation of the five divisions and
also had the next-to-lowest mean score.

The mean percentage

score of the two groups varied by only 1.47 percent.

The

lower mid-range groups of 1-10 and 11-30 accidents investigat­
ed had mean scores that varied by 2.43 percent and a variance
in standard deviation of only .65 percent.

The division of

officers who investi-gated 31-65 accidents had the highest
mean score, 11.71 percent more than the lowest mean score.
Differences between Mean Original Performance Test
and Performance Retest
The

following

directed hypothesis

was

tested

for the

combined mean scores of the 100 test subjects:
Hypothesis 1 :

The total sample of officers when treated
as one group will show significant loss
of retention as calculated by the perfor­
mance retest.

At the end of each training program each student was
given

an

original

beginning of
retest

each

test

follow-up testing

(post

test) .

session

At

the

a performance

(same as post test) was given to 100 previously Ai-1

trained officers.
100.

performance

The maximum possible score on each test was

Table 4.6 provides a statistical analysis of the two

tests.
To determine whether a significant loss in the mean score
of the original performance test occurred, as compared to the
mean score of the performance retest, a paired sample t-test
analysis was used.

The .05 level of significance was selected

as the basis for accepting or rejecting the hypothesis.

The

54
critical value of the t-statistic in this test was the .05
significance level and the calculated t-statistic was

.00.

Table 4.6 shows that:
1.

The 100 test subjects had a mean score of 89.71 on

the original performance test and a mean score of 75.51 on the
performance retest.
6.55 and 10.39.
2.

The respective standard deviations were

The difference of the means was 15.07.

The t-value was 14.20 with 99 degrees of freedom,

the critical t value was 1.661.
3.

The tail

significance

set

probability was

at

.05.

This

.00,

means

with the

that

the

level of

directional

hypothesis was not rejected.

TABLE 4.6 Comparison of the Mean Test Scores of the 100
Subjects for the Original Performance Test and the
Performance Retest
N
Original 100

MEAN

SD

DM

89.71

6.55
15.07

Retest

100

t-VALUE

df

15.45

99

CF

1.661

2-TAIL
PROBABILITY

.00

75.51 10.39

N
MEAN
SD
DM
t-VALUE
df
CF
TAIL PROBABILITY

=
=
=
=
=
=
=
=

Number of subjects instudy
Mean scores of the two tests
Standard Deviation
Difference between the two means
Computed t value
Degrees of freedom
Critical t value
Level of significance

55
Summary and Comments;

Table 4.6

Hypothesis 1 stated that the mean score of the original
performance

test

would

be

performance retest mean.

significantly

higher

than

the

Not only was the original test 14.2

percent higher, the standard deviation was 3.84 points lower
than

the

retest

standard deviation.

This means

that

the

retest scores were both lower and less consistent than the
original performance test scores.
The Effect of time on the Performance
Retest Mean Score
The

following

varying periods

hypothesis

was

tested

to

determine

if

of elapsed time since taking the original

performance test had an affect on the performance retest:
Hypothesis 2 :

When dividing the total sample of tests
of the officers into three groups based
upon interval of time since completing
the AI-1 training program,
officers'
retention levels will vary significantly
as reflected by the group mean score on
the performance retest.

To determine whether the mean of the performance retest
was influenced by the main effect of time,

the scores and

three time periods were examined through the use of analysis
of covariance.
the

possible

With the analysis of covariance controlling
effect

of

the

original

performance

score,

positive or negative results of the analysis can be attributed
to the effect of time.
The 100 subjects'
time segments:

tests were divided by the following

56
Group 1 = Four to 12 months since tests were taken
Group 2 = More than 12 months to 36 months since tests
were taken
Group 3 = More than 36 months since tests were taken
Table 4.7 provides the results of the analysis of the perfor­
mance retest by time with the results of the original perfor­
mance test being controlled through use of covariance.

Table

4.7 shows that:
1.

There were 2 degrees of freedom;

the sum of the

squares was 1235.96 with a mean square of 617.981.
2.

The computed value of F is 9.031; its critical value

at a .05 significance level was 3.103
3.

The probability area is .00.

The hypothesis was not

rejected.

TABLE 4.7 Analysis of Covariance of the Performance Retest by
the Divisions of Time

2

617.981

1235.96

COMPUTED
VALUE OF F

9.031

CRITICAL
VALUE OF F
at .05

COMPUTED
SIGNIFICANCE
OF F

3.103

•

MEAN
SQUARE

o
o

SUM OF
df SQUARES

Table 4.8 provides additional detail of the three periods
of time and provides the basis for not rejecting the hypothe­
sis.

Table 4.8 shows that:

57
1.

The grand mean of the three groups was 75.65.

2.

The unadjusted by analysis of covariance plus-or-

minus value for the grand mean were Group 1 = 5.32, Group 2 =
.45, and Group 3 = -3.62.
3.

The adjusted by analysis of covariance plus-or-minus

values for the grand mean were Group 1 = 4.86, Group 2 = 1.01,
and Group 3 = -3.75.

TABLE 4.8 The Effect of Time on the Grand Mean as Determined
by the Analysis of Covariance
TIME
GROUP

GRAND
MEAN

UNADJUSTED
AMOUNT

ADJUSTED
AMOUNT

1

75.65

5.32

4.86

80.97

80.51

2

75.65

.45

1.01

76.1

76.66

3

75.65

-3.62

-3.75

72.03

71.9

NEW GRAND
MEANS
U
A

GROUP 1 = Four months to 12 months since tests were taken
GROUP 2 = More than 12 months to 36 months since tests were
taken
GROUP 3 = More than 36 months since tests were taken
Grand Mean
Unadjusted Amount

=
=

Adjusted Amount

=

U & A

=

Mean of N100
Difference + or - to grand mean with
no control for possible effect of
original performance test
Difference + or - to grand mean when
controlling for possible effect of
original performance test
New means per group when adding or
subtracting the unadjusted (U) and
adjusted (A) amounts

58
Summary and Comments:

Tables 4.7 and 4.8

Data in Tables 4.7 and 4.8 address the directed hypothe­
sis that the variable of time had an effect on the performance
retest

and

reflected

that

by

retention

the group

levels

mean

varied

score.

significantly

Table

4.7

lists

as
the

results of the use of analysis of covariance which controls
for the possible effect that the original performance test may
have

on

the

outcome

of

the

analysis.

As

a

result,

the

differences in the mean scores are the result of time, not the
original post test scores.

The .00 significance level listed

in Table 4.7 indicates that there were significant differences
among the three time periods.
The data in Table 4.8 provides additional results of the
analysis by showing the grand mean for the total N=100.

The

analysis of covariance adjusted grand mean ranges from 4.86
percent for group 1 to -3.75 for group 3 or a difference of
8.61 percent.

If the analysis of covariance was not used the

ranges would be 5.32 percent to 3.62 percent or a difference
of 8.94 percent.
The Effect of Education on the Performance
Retest Mean Score
The

following

hypothesis

was

tested

to

determine

if

varying levels of education affect the performance retest:
Hypothesis 3 :

To determine
performance

Officers' education level will not show a
significant relationship with the perfor­
mance retest scores.
the possible

effect

on the mean

retest with the main effect

of the

of education,

the

59
scores and education levels were tested by the analysis of
covariance.

The

100

test

subjects

were

divided

by

the

following education levels.
1.

Less than two years of college

2.

Associate degree but less than four

3.

Four-year-degree or more

years of college

Table 4.9 contains the sum of the squares, mean square,
degrees of freedom, the F values computed on the means of the
performance

retest,

and

effects of education.
1.

There were

significance

level

with

the

main

Table 4.9 shows that:
2 degrees of

freedom,

the sum of the

squares was 135.59 with a mean square of 67.79.
2.

The computed value of F was .813; its critical value

at a .05 significance level was 3.10.
3.

The probability area was .447 which is larger than

the .05 level.

The hypothesis was not rejected.

TABLE 4.9 Analysis of Covariance of the Performance Retest by
the Three Divisions of Education
SUM OF
df SQUARES
2

135.59

Table

MEAN
SQUARE

7.79

4.10

COMPUTED
VALUE OF F

.813

provides

CRITICAL
VALUE OF F
at .05

COMPUTED
SIGNIFICANCE
OF F

3.10

.447

additional

details

for the

three

divisions of education and provides the basis for not reject­
ing the hypothesis.

Table 4.10 shows that:

60
1.

The grand mean of the three groups by education was

75.51.
2.

The unadjusted by analysis of covariance plus-or-

minus values for the grand mean were Group 1 = -2.84, Group 2
= 1.68, and Group 3 = .93.
3.

The adjusted by analysis of covariance plus-or-minus

values for the grand mean were Group 1 = -1.68, Group 2 = .83,
and Group 3 = .83.

TABLE 4.10

EDUCATION
DIVISION

The Effect of Education on the Grand Mean as
Determined by the Analysis of Covariance
GRAND
MEAN

UNADJUSTED
AMOUNT

ADJUSTED
AMOUNT

NEW GRAND
MEANS
U
A

1

75.51

-2.84

-1.68

72.67

73.83

2

75.51

1.68

.83

77.19

76.34

3

75.51

.93

.83

76.44

76.34

GROUP 1 = Less than 2 years of college
GROUP 2 = Associate degree but less than a four year degree
GROUP 3 = Four-year--degree or more
Grand Mean
Unadjusted Amount

=
=

Adjusted Amount

=

U & A

=

Mean of N100
Difference + or - to grand mean with
no control for possible effect of
original performance test
Difference + or - to grand mean with
controlling for possible effect of
original performance test
New means per group when adding or
subtracting the unadjusted (U) and
adjusted (A) amounts

61
Summary and Comments:

Tables 4.9 and 4.10

Table 4.9 data indicate that the directed hypothesis of
officers*

education level will not show a significant rela­

tionship with the performance retest scores; the hypothesis
was not rejected.

A .05 level of significance was chosen and

the analysis of covariance determined the significance level
to be .447.
Table 4.10, as does Table 4.8, lists the percentage that
the mean scores of N=100 must be adjusted when the total is
divided into three groups and the analysis of covariance is
applied.

The adjustment amounts of -1.68 for Group 1, .83 for

Group 2, and .83 for Group 3 show a trend but not as strong as
the trend in Table 4.8.
show

a

stronger

analysis

of

education

trend;

covariance

did

The unadjusted amounts in Table 4.10

not

this

indicates

which

the

determines

significantly

affect

value

that
the

of

amount

mean

of

the
of
the

performance retest scores.
The Effect of Test Subjects* Age on
the Performance Retest Mean Score
The following hypothesis was tested to determine whether
varying ages of the test subjects affected the performance
retest:
Hypothesis 4 :

Officers' age will not show a significant
relationship
with
performance
retest
scores.

To determine the possible influence on the mean of the
performance retest with the main effect of age groupings, the
scores

and

age

divisions

were

tested

using

analysis

of

62
covariance.

The

100

test

subjects

were

grouped

by

the

following age levels.
Group 1 = 24 to 34 years of age
Group 2 = More than 34 to 37 years

of age

Group 3 = More than 37 to 42 years

of age

Group 4 = More

than 42 years of age

Table 4.11 contains the sum of the squares, mean squares,
degrees of freedom, the F values computed on the mean of the
performance

retest,

effect of age.
1.

and

significance

level

with

the

mean

Table 4.11 shows that:

There were

3 degrees of freedom;

the sum of the

squares is 3029.86 with a mean square of 757.46.
2.

The computed value

of F was

1.972;

its critical

value at a .05 significance level was 2.70.
3.

The probability area is .123 which is greater than

the significance level of .05.
not make

a

difference;

The age of the officer does

therefore,

the

hypothesis

was

not

rejected.

TABLE 4.11

Analysis of Covariance of the
Retest by the Division of Age

SUM OF
df SQUARES

MEAN
SQUARE

3

159.16

444.49

COMPUTED
VALUE OF F
1.972

Performance

CRITICAL
VALUE OF F
at .05

COMPUTED
SIGNIFICANCE
OF F

2.70

.123

63
Table 4.12 provides additional details for the four age
divisions

and

hypothesis.

provides

the

basis

for

not

rejecting

the

Table 4.12 shows that:

1.

The grand mean of the four groups by age was 75.51.

2.

The

unadjusted

amount

by

analysis

of

covariance

plus-or-minus values for the grand mean were Group 1 = 4.28,
Group 2 = -2.43, Group 3 = 1.03, and Group 4 = -2.76.
3.

The

adjusted amount by analysis of covariance plus-

or-minus values for the grand mean were Group 1 = 3.16, Groups
2 = -1.86, Group 3 = 1.07, and Group 4 = -2.30.

TABLE 4.12

The Effect of Age on the Grand Mean as Deter­
mined by Analysis of Covariance

AGE
DIVISION

GRAND
MEAN

UNADJUSTED
AMOUNT

1

75.51

4.28

3.16

79.79

78.67

2

75.51

-2.43

-1.86

73.08

73.65

3

75.51

1.03

1.07

76.54

75.58

4

75.51

-2.76

-2.30

72.75

73.21

Summary and Comments:

ADJUSTED
AMOUNT

NEW GRAND
MEANS
U
A

Tables 4.11 and 4.12

Table 4.11 shows the result of the analysis of covariance
of

the performance

retest

by

age divisions.

placed the significance level at .123.

effect of an officer's

results

The F of .123 exceeded

the selected significance level of .05.
the

The

The hypothesis for

age was written

in a directed

64
fashion

as

a

result

of

the

Chapter

2

literature

review;

therefore, the hypothesis was not rejected.
Table

4.12

is

additional

documentation

of

why

an

officer's age did not bear a significant relationship to the
performance
control

retest.

the

Analysis

effect

of

the

of

covariance

original

was

used

to

performance

test

to

determine the effect of age on the performance retest.

The

primary results of this type of analysis were listed under the
column

"adjusted amount"

1.07,

and

-2.30

are

in table 4.12.

the

percentage

The 3.16,

points

used

-1.86,
for

the

respective age divisions of 1 to 4 to adjust the grand mean of
75.51

percent.

The

adjustment

amounts

show

no

trend

to

support the regression of scores based upon age; therefore,
age was

not

a

significant

variable

on

performance

retest

scores.
The Effect of the Number of Accidents
Investigated on the Performance Retest
Mean Score
The following hypothesis was tested to determine whether
the performance retest was affected by the number of accidents
investigated

by

the

test

subjects

during

the

previous

12

months:
Hypothesis 5 :

The number of personal injury and/or
fatal vehicle traffic crashes investigat­
ed by officers during the one year period
prior to taking the performance retest
will show a significant positive rela­
tionship with the mean scores on the
performance retest.

65
To determine the possible influence on the mean of the
performance

retest with

the main

effect

of the

number

of

traffic crashes investigated, individual scores and the five
subgroups were tested by the analysis of covariance.

The test

subjects were divided into the following five groups.
Group 1 = No accidents investigated
Group 2 = 1-10 accidents investigated
Group 3 = 11-30 accidents investigated
Group 4 = 31-65 accidents investigated
Group 5 = 66 or more accidents investigated

TABLE 4.13

Analysis of Covariance of the Performance
Retest by the Division of Number of Accidents
Investigated 12 months prior to the Retest

SUM OF
df SQUARES

MEAN
SQUARE

4

177.759

711.037

COMPUTED
VALUE OF F

CRITICAL
VALUE OF F
at .05

2.248

2.50

COMPUTED
SIGNIFICANCE
OF F
.070

Table 4.13 contains the sum of the squares, mean squares,
degrees of freedom, the F values computed on the mean of the
performance retest, and significant level with the main effect
of number of accidents investigated.
1.

There were

Table 4.13 shows that:

4 degrees of freedom;

the

sum of the

squares was 711.037 with a mean square of 177.759.
2.

The computed value of F is 2.248; its critical value

at a .05 significance level was 2.50.

66
3.

The probability area was .070 which was greater than

the significance level of .05, which means that the number of
accidents

investigated during the

performance

retest

was

not

12 months

significant;

preceding

the

therefore,

the

hypothesis was rejected.
Table

4.14

provides

additional

details

for

the

five

divisions of number of accidents investigated and provides the
basis for the nonsignificant finding.
1.

The grand mean

of the

Table 4.14 shows that:

five

groups

by

number

of

accidents investigated was 75.51.
2.

The unadjusted by analysis of covariance plus or

minus values for the grand mean were Group 1 = -5.66, Group 2
= 2.30, Group 3 = -.13, Group 4 = 6.04, and Group 5 = -4.19.
3.

The adjusted by analysis of covariance plus-or-minus

values for the grand mean were Group 1 = -2.50,

Group 2 =

1.50, Group 3 = .20, Group 4 = 3.97, and Group 5 = -3.81.

TABLE 4.14

The Effect of Number of Accidents Investigated
on the Grand Mean as Determined by Analysis of
Covariance

NO. OF
ACCIDENTS
INVESTIGATED
DIVISION

GRAND
MEAN

UNADJUSTED
AMOUNT

1

75.51

—5.66

-2.50

69.85

73.01

2

75.51

2.30

1.50

77.81

77.01

3

75.51

-.13

.20

75.38

75.71

4

75.51

6.04

3.97

81.55

79.48

5

75.51

-4.19

-3.81

71.32

71.7

ADJUSTED
AMOUNT

NEW GRAND
MEANS
U
A

67
Summary and Comments;
Table

4.13

Tables 4.13 and 4.14

shows

that

the

significance

level

of

the

effect of the number of accidents investigated on the perfor­
mance retest was at a .07 level.

A significance level of .05

was selected for this study; if a .10 level had been chosen
then the .07 would have been significant.
Table 4.14 provides additional detail about the relation­
ship of the number of accidents investigated on the perfor­
mance retest scores.

Looking at the percent adjustment to the

grand mean, the adjusted amount (from 1 to 5) was -2.50, 1.50,
.20, 3.97, and -3.81.

Division 1 is zero accidents investi­

gated and Division 5 is 66 or more accidents investigated.
Both Divisions 1 and 5 show a negative amount, which does not
show normal regression or related consistency.

Division 2 (1-

10 accidents), Division 3 (11-30 accidents), and Division 4
(31-65

accidents)

provide

some

indication

that

as

more

accidents are investigated there is an increase in performance
retest

scores.

The

analysis

procedure

chosen

for

this

research evaluated the effects of all divisions of accidents
investigated and the results of that analysis indicate that
the number of accidents investigated does not have a signifi­
cant effect (.05 level) on the mean score of the performance
retest.
Review of the Results of the
Analysis of Hypotheses 1-5
The preceding pages included the analysis of the five
hypotheses

proposed

for

this

study.

Variables

of

time,

68
education,
mance

test

accidents investigated,
were

compared to

age and original perfor­

the performance

retest.

The

results of the five tests were:
Hypothesis 1 :

The total sample of officers when treated
as one group will show a significant loss
of retention as calculated by the perfor­
mance retest.

This directional hypothesis was not rejected; there was
a significant loss of retention based upon a comparison of the
original performance test and the performance retest.
Hypothesis 2 :

When dividing the total sample of offi­
cers into three groups based upon inter­
val of time since completing the AI-1
training program,
officers'
retention
levels will vary significantly as re­
flected by the group mean score on the
performance retest.

This directional hypothesis was not rejected; there was
a significant loss of retention based upon the three separate
time periods as compared to the performance retest.
Hypothesis 3 :

Officers' education level will not show a
significant relationship with the perfor­
mance retest scores.

This directional hypothesis was not rejected; there was
not a significant relationship between the officers' education
level and the performance retest scores.
Hypothesis 4 :

Officers' age will not show a significant
relationship
with
performance
retest
scores.

This directional hypothesis was not rejected; age did not
have a significant relationship with the performance retest
mean score.

69
Hypothesis 5 :

The number of personal injury and/or
fatal vehicle traffic crashes investigat­
ed by officers during the one-year period
prior to taking the performance retest
will show a significant positive rela­
tionship with the mean score on the per­
formance retest.

This directional hypothesis was rejected; the number of
accidents investigated by an officer during the preceding year
prior to taking the performance retest was not significant in
terms of the relationship to the performance retest.
Each

hypothesis

in this

research was

directed.

The

direction was based upon information obtained in the litera­
ture review in Chapter 2 of this research.
hypotheses were directed at

(1) test/pretest,

education, and (4) age were not rejected.

The first four
(2) time,

(3)

The fifth hypothe­

sis dealing with the number of investigations was rejected.
The five research hypotheses, which were the basis for
this research, were discussed in detail in the preceding pages
of this chapter.

The remainder of Chapter 4 includes three

areas of additional analyses of the same variables of perfor­
mance retest scores, time in months since completing the AI-1
training,

officer

education

level,

age,

and the number of

accidents investigated during the year preceding the perfor­
mance retest.

Analysis of covariance and paired sample t-

tests were also used for the additional analysis.
Analysis for the five research hypothesis evaluated the
possible effect of the single variables of time,

education,

age, and number of accidents investigated on the performance
retest scores.

Various combinations of the variable were used

70
in the additional research to determine whether two variables
together showed a significant relationship with the perfor­
mance retest.

The following combinations were tested:

Main Effects
1.

2.
3.

Covariate

A
B
A
B

Time since completing AI-1
Education level
Time since completing AI-1
Age

Performance Retest
Score
Performance Retest
Score

A
B

Time since completing AI-1
Number of accidents
investigated in year
preceding performance
retest

Performance Retest
Score

Other combinations

of main effects such as

"education

and

age," "number of accidents investigated," and "age plus other
combinations" were also analyzed.
The primary problem encountered in this analysis was that
the total

sample size was too small.

additional division of the N=100,
analyses dropped below five.

As a result of the

cell size in many of the

The reliability of the analysis

is suspect when cell size becomes this small.
discussed

in

Chapter

3,

and

it was

Cell size was

determined

that

data

generated by a cell size of less than five would not be used
in this research.
The Combined Effect of Time and Education
on the Performance Retest Mean Score
To determine whether the mean score of the performance
retest was influenced by the combined main effects of time and
education,

all times and all education levels were examined

71
through the use of analysis of covariance.

With the analysis

of covariance controlling the possible effect of the original
performance

score,

positive

or

negative

results

of

the

analysis can be attributed to the combined effects of time and
education.

TABLE 4.15

df

Analysis of covariance of the Performance
Retest by All Time and All Education Levels
SUM OF
SQUARE

MEAN
SQUARE

CRITICAL
COMPUTED VALUE OF
VALUE OF
F at
F
.05

COMPUTED
SIGNIFICANCE
OF F

Main
Effects

4

1297.75

324.43

4.712

2.49

.002

Time

2

1227.78

613.89

8.916

3.11

.000

Educa­
tion

2

61.79

30.89

.449

3.11

.640

Table 4.15 shows the results of the analysis described
above:
1.

There were 4 degrees of freedom, 2 each for time and

education;

the

sum

of

the

squares

was

1297.75 for

main

effects,

1227.78 for time and 61.79 for education with main

effects,

1227.78 for time and 61.79 for education with mean

squares of 324.43 for main effects, 613.89 for time, and 30.89
for education.
2.

The computed value of F was 4.712 for main effects,

8.916 for time and .449 for education; the critical value of

72
F at a .05 significance level for main effects is 2.49, time
3.11, and education 3.11.
3.

The probability area for main effects was .002, time

.000 and education
significant,

time

.640, which means that main effects are
was

significant,

but

education

was

not

significant.
Summary and Comments:

Table 4.15

Data in Table 4.15 show the combined effect of all time
and

all

education

on

the

performance

retest.

The

table

contains the levels of significance and a .002 was significant
for the combined effects of time and education.
individual levels of significance,

Looking at

it can be seen that time

was significant but education was not.

The results of the

analysis show that the effect of time causes the main effect
to

be

between

significant.
time

and

There

education.

was

not

a

two-way

Hypothesis

interaction

3 dealt with

the

variable of education, and the analysis indicated that there
was not a significant relationship between education and the
performance retest.

Hypothesis 2 dealt with the variable of

time, and the analysis of that variable indicated that there
was

a

significant

loss

of

retention based upon

the

three

separate time periods as compared to the performance retest.
The Combined Effect of Time and Number of Accidents
Investigated on the Performance Retest Mean Score
To determine whether the mean score of the performance
retest was affected by the combined main effects of time and

73
number of accidents investigated, all time and all divisions
of accidents investigated were examined through the use of
analysis

of

covariance.

With

the

analysis

of

covariance

controlling the possible effect of the original performance
score.

Positive or negative results of the analysis can be

attributed to the combined effects of time and education.

TABLE 4.16

Analysis of Covariance of the Performance
Retest by All Time and All Accidents Investi­
gated
SUM OF
SQUARE

df

MEAN
SQUARE

CRITICAL
COMPUTED VALUE OF
VALUE OF
F at
F
.05

COMPUTED
SIGNIFICANCE
OF F

Main
Effects 6

1739.217

289.869

4.384

2.33

.001

Time

1189.693

594.864

8.997

3.11

.000

503.256

125.814

1.903

2.72

.118

2

Number
of Acc. 4
Inv.

Table 4.16 contains the results of the analysis described
above:
1.

There were 6 degrees of freedom,

for the number of accidents investigated.

2 for time and 4

The sum of the

squares was 1739.21 for the main effects, 1189.69 for time and
503.25 for the number of accidents investigated; mean squares
of 289.86 for the main effects,

594.86 for the variable of

time and 125.81 for the variable of the number of accidents
investigated.

74
2.
4.384,

The computed value of F for the main effects was
variable of time was 8.997,

investigated is 1.903.

and number of accidents

The critical value at a .05 signifi­

cance level for the main effects was 2.33, time was 3.11, and
number of accidents investigated was 2.72.
3.

The probability area for the main effects was .001,

time was

.000,

and the probability of number of accidents

investigated was

.118.

This means that the main effect of

time was significant, but the number of accidents investigated
in the year preceding the performance retest was not signifi­
cant.
Summary and Comments:

Table 4.16

Data in Table 4.16 address the combined effect of all
time and all numbers of accidents investigated during the year
preceding the performance retest.

The main effects of the

combination of time and accidents investigated was significant
at

a

level

of

.001.

Looking

at

the

significance

level

individually, time was significant but the number of accidents
investigated was not.
Time

was

more

The main

significant

than

effects were
the

number

significant.
of

accidents

investigated, but combining the two variables made the main
effects significant.
cant at a

Although the main effects were signifi­

.001 level,

there was not a two-way

interaction

between the variable of time and number of accidents investi­
gated.
analysis

Hypothesis 2 addresses the variable of time, and the
of

the

variable

indicated

that

there

was

a

75
significant loss of retention based upon the three divisions
of time.

Hypothesis 5 addressed the variable of number of

accidents investigated during the year preceding the perfor­
mance retest.

That hypothesis was rejected;

the number of

accidents investigated was not significant in terms of its
effect on the performance retest.
The Combined Effect of Time and Age on the
Performance Retest Mean Score
To determine whether the mean score of the performance
retest was affected by the combined main effects of time and
age, all times and all ages were evaluated through the use of
analysis

of

covariance.

With

the

analysis

of

covariance

controlling for the possible effect of the original perfor­
mance

score,

any

effect

that

the

analysis

shows

can

be

attributed to the combined effects of all times and all age
groups.
Table 4.17 shows the results of the analysis described
above ?
1.

There were five degrees of freedom, two for time and

three for age;

the sum of the squares was

1343.8 for main

effects, 941.317 for the variable of time and 107.839 for the
variable of age.

The mean square for the main effects was

268.76, time had a mean square of 470.658, and age was 35.946.
2.

The

computed value

of

F was

effects, 7.646 for time and .584 for age.

4.366

for the main

The critical value

at a .05 significance level for main effects was 2.21, time
3.11, and age 2.49.

76
3.

The probability area for main effects was .001, time

was .001 and age was .627.
are

significant,

time

was

This means that the main effects
significant,

and

age

was

not

significant.

TABLE 4.17

Analysis of Covariance of the Performance
Retest by All Times and All Age Levels
SUM OF
SQUARE

df
Main
Effects 5

MEAN
SQUARE

1343.8

CRITICAL
COMPUTED VALUE OF
F at
VALUE OF
F
.05

COMPUTED
SIGNIFICANCE
OF F

268.760

4.366

2.21

.001

Time

2

941.317

470.658

7.646

3.11

.001

Age

3

107.839

35.946

.584

2.49

.627

Summarv and Comments:

Table 4.17

Data in Table 4.17 show the combined effect of all time
periods and all age levels on the means of the performance
retest snores.

The table lists the levels of significance, a

.001 is significant for the combined effects of time and age.
Looking at the individual levels of significance,
seen

that

time

was

significant

significant at a .627 level.

at

.001

and

it can be

age

was

not

The main effects were signifi­

cant as a result of the stronger level of significance of the
variables

of

time.

There

between time and age.

was

not

a

two-way

interaction

Hypothesis 2 addressed the variable of

time, and the analysis of that variable indicated that there
was

a significant

loss

of

retention based upon

the

three

77
separate time periods on the performance retest.

Hypothesis

4 addressed the variable of age, and it was determined that
age did

not have

a

significant

effect

on the

performance

retest score.
The paired samples of time and the variables of age,
education, and number of accidents investigated were signifi­
cant

due

to

the

strong

significance

level

of

time.

The

variable of time continued to be a significant factor in all
tests, even though sample size was smaller when comparing two
variables

instead

of

only

one,

as

was

done

in

the

five

hypotheses.
Summary
The findings of the investigation were presented in this
chapter.

These findings showed the relationship that time,

education, age, and number of accidents investigated analyzed
singly had with the mean of the performance retest scores.
Also

presented

were

the

combined

effects

that

time

and

education, time and age, and time and the number of accidents
investigated had on the mean of the performance retest scores.
The following chapter contains the summary, conclusions' and
recommendations.

Chapter 5
SUMMARY, CONCLUSIONS AND RECOMMENDATIONS
The findings of this investigation were presented in the
preceding chapter.

This chapter contains:

a summary of the

investigation, conclusions based upon the findings, recommen­
dations,

recommendations

for further research and implica­

tions.
Summary
Purpose
The writer's purpose in this study was to investigate the
degree to which cognitive knowledge gained by Michigan police
officers in the Michigan State University Accident Investiga­
tion

1

(AI-1)

training program was

periods of time.
had been made

retained

over

varying

The problem was that, until now, no attempt

to measure

the

retention of AI-1

cognitive

knowledge once the officer was back in the field and perform­
ing his/her assigned duties in the field.

This investigation

was designed to answer the following questions:
1.

Is

there

a

significant

difference

between

computed mean score of the original performance test taken by
the police officers included in this study at completion of

78

the

79
their respective AI-1 program as compared to the mean score of
the performance retest taken by the same group of officers?
2.

Is there a significant difference in mean scores

based upon the amount of elapsed time since completion of AI-1
training.
3.

Is there a significant difference in mean scores of

the original performance test and the performance retest based
upon the education level of individual officers.
4.

Is there a significant difference between officers'

age and retention based upon the means of the performance
retest and original performance test.
5.

Is

there

a

significant

difference

between

the

officers retention level and test mean scores based upon the
number of accidents the officers investigated during the year
preceding the performance retest.
The Research Methods and Procedures
During spring and summer 1989,

100 police officers who

completed the AI-1 training program volunteered to retake the
AI-1 post test.
test

were

The officers who volunteered to retake the

attending

Michigan

counties

officers

who

of

other

MSU

Oakland,

volunteered

to

training
Wayne

take

the

and

programs
Macomb.

retest

were

in

the

Other
police

officers from the Eaton and Ingham County Sheriff Departments.
Each volunteer completed a questionnaire that requested
specific information about age, education level, elapsed time
since completing AI-1,

and number of accidents investigated

80
during the year preceding their participation in this evalua­
tion.

Once the test and questionnaire had been completed, the

data were tabulated and placed on a summary sheet.

Other

information placed on the summary sheet was the date that each
volunteer completed AI-1

and also his/her post test

score

obtained at completion of the AI-1 training program attended.
Summary data were divided by age categories, educational
levels,

number

of

accidents

investigated,

elapsed

months

between the two tests and original performance score
test) and performance retest scores.

(post

The groups were compared

by using paired sample t-tests and the analysis of covariance
to

determine

whether

there

was

a

significant

difference

between the mean scores based upon the variables of time,
education, age and number of accidents investigated.
Major Findings
The major findings of this investigation were:
1.

The total sample of 100 subjects, when treated as

one group with no sub-divisions, showed a significant loss of
retention based upon a comparison of the mean scores of the
original performance test and the performance retest.
2.
periods

When the total sample was divided into various time
since

completing

the

AI-1

training

program,

the

analysis showed that, the longer the elapsed time since taking
the training, the lower the retest score.

81
3.
and

Based upon the analysis of officer education level

the

performance

retest

mean

score,

there

was

not

a

significant relation between education level and mean score.
4.

The

analysis,

based

upon

officer

age

and

its

relationship to the performance retest mean score, showed that
a significant relation did not exist.
5.
during

The number of accidents
the

year preceding

the

investigated by officers

performance

retest

was

not

significant in terms of its relationship to the performance
retest.
Conclusions
The

following conclusions

are based

upon

the

general

findings of this investigation.
1.

Of the variables included in the investigation, the

overall most significant one was the amount of elapsed time
for the officer between taking the original performance test
and the performance retest.

Three time divisions were used.

The shortest division included the time period of 4 months to
12 months.
score.

This group had the highest performance retest mean

Th'.s finding was

supported by the

literature review in Chapter 2.
retest

scores

for

each

results

of the

The mean of the performance

successively

longer

time

period

regressed at a consistent rate, with the longest elapsed time
group achieving the lowest performance retest score.

Standard

deviation scores for the three time periods were very consis­
tent and all were within 1.06 percentage points of each other.

82
Time is a significant factor and the results of this research
show that cognitive knowledge gained during the AI-1 training
program is not fully maintained over varying periods of time.
The longer the elapsed time period since completing the AI-1
training equals less retention of cognitive knowledge.
2.

The

variable

of

officer

education

level

significant effect on the mean of the performance

had

retest.

There was no correlation between the three levels of educa­
tion.
years

Officers with a high school diploma and less than two
of

score.

college

had

the

lowest

performance

retests

mean

Officers with an Associate Degree but less than a

four-year degree had a mean score higher than officers with
less than two years of college and also higher than officers
with a four-year degree or more.

As a result, there was not

a significant correlation between the three levels of educa­
tion;

therefore,

a significant relationship does not exist

between education and the mean of the performance retest.
3-

An age range of 24 to 43 years of age existed for

the one hundred test subjects in this investigation.

It was

determined that age of the individual officers did not have a
significant relationship to the mean score on the performance
retest. The youngest age group had the highest mean score and
the oldest group had the lowest mean score.

These two age

groups had what could be expected in terms of the possible
effect

of

age

on

retention,

however,

the

two

age

groups

between the youngest and oldest offered no resemblance of a
significant

correlation

in conjunction with

the

other age

no

83
divisions.
retest

Therefore, the variation between the performance

mean

scores

for

the

various

age

levels

were

not

significant.
4.

Overall,

the relationship between number of acci­

dents investigated during the year preceding the performance
retest and the performance retest mean score did not reach
significance.
mean

scores

However, there was some correlation between the
of the

investigated.

five

divisions

of

number

of

accidents

Division 1 (no accidents investigated during

the year preceding the performance retest)
adjusted mean score.

had the

lowest

Based upon the literature review,

a

lower score could be expected of those who do not utilize
their knowledge.
largest

number

However,
of

Division 5, the group with the

accidents

investigated,

performance retest mean score.

had

the

lowest

This result cannot be ex­

plained based upon the investigation or information contained
in the literature review.

It is known that the officers in

the division with the largest number of accidents investigated
were

from

Michigan's

largest

police

departments,

where

officers have many other assignments and their jurisdictions
have high traffic volumes and high accident ratios.

It would

be useful to investigate the possible relationship between
size of department, number of accidents investigated and AI-1
training cognitive knowledge retained over time.
5.

Analysis to show the results of combining variables

to determine their relationship to the mean of the performance

84
retest was made.
tion.

Combined variables of (1) time and educa­

time and number of accidents, and (3) time and age

(2 )

resulted in time continuing to be significant in combination
with education and number of accidents investigated but not in
combination with a g e . As a result of the strong significance
level of time, the main effects in all three tests were also
significant.
6.

The hypotheses investigated in this research were

formulated on the general findings of the literature review.
Chapter 2.

Each hypothesis was directed or based upon the

expected

outcome

as

Previous

researchers

determined
have

by

the

suggested

literature

that

review.

time would

be

a

significant factor in the retention of cognitive knowledge,
and it proved to be in this study, as well.
education

level

and

number of accidents

Officer age and
investigated were

found not to be significant, based upon the directed hypothe­
ses concerning these variables.
Recommendations
Based upon this research,
experience

in

accident

as well as the researcher's

investigation,

such

as

training

approximately 4,000 Michigan police officers in AI-1, review­
ing reports of AI-1 graduates and assisting them in their
investigations, the following recommendations are made:
1.
knowledge

Time
retention.

is

a

critical

Refresher

variable

AI-1

in

training

AI-1

cognitive

should

be

required of all police officers who have the responsibility of

85
investigating accidents.

An officer should receive refresher

training every two years.
2.

Refresher training should include the primary areas

of roadway evaluation, measuring, sketching and photography.
3.

Refresher AI-1 training should be included in the

form of role call training. A 15-minute video tape on each of
the

four

subjects

identified

in

Recommendation

2

would

"sharpen" the officers' skills in these four critical areas of
recording information.
4.

Investigation of large numbers of accidents does not

improve

cognitive

knowledge

receive

refresher

training

retention.
regardless

Officers
of

the

should

number

of

accidents they investigate in any given time period.
5.

"Use it or lose it" is a viable concept.

An officer

who has not had AI-1 training for three or more years and has
not been assigned to accident investigation duty should not be
permitted to perform full-scale investigations.
this

category np^ds more than

roll-call

should be required to have a one-day

An officer in

training.

He/she

(eight hours)

review

and/or update program on AI-1 subjects.
6.

Age is not a significant factor in police officer

cognitive knowledge retention of AI-1 training information.
Provided that he/she has completed AI-1 training,

has been

continually assigned to investigate accidents and receives the
recommended AI-1 role call training, a police officer should
not be removed from the duty of investigating accidents solely
because of age.

86
7.

Advanced college degree requirements of more than an

associate's degree should not be mandatory for an officer to
gain entry to a police departments traffic unit.

However,

once officers are in the traffic unit they should be encour­
aged to continue their education.
Recommendations for Future Research
1.

This investigation should be repeated, taking into

consideration:
a.

using test subjects who have all completed the same
level of additional AI training programs, i.e., all
officers who have completed AI-1 through AI-8.

b.

using two sets of test subjects, one group consist­
ing of officers who have general police responsi­
bility
the

in addition to accident

second

assigned

group consisting

the

specific

task

of
of

investigation and
officers
accident

who

are

investi­

gation and traffic enforcement.
c.

using test

subjects

similar to the

ones used

in

this research with the additional variables of size
of department and accident ratios for the individu­
al departments.
2.
completed

Follow-up
in

the

research
form

of

question in the post test.

to

an

this

item

evaluation

analysis

of

should
each

be

test

This would help identify other

cognitive knowledge areas in which an officer becomes weaker
through time.

87
Implications
The research, analysis, and findings based on the study
have implications in the following areas.
Training and Recertification
The variables reviewed in this study and the findings can
benefit the traffic services related function in law enforce­
ment.

For various reasons,

Al-trained police officers are

temporarily transferred to other assignments.
concern is:
role,

The question of

When the officer is returned to a traffic-related

is he/she still qualified to perform traffic-related

duties?

The

Council,

which

Michigan

Law

sets training

Enforcement

Officers

standards,

is

Training

faced with the

problem of deciding how often an officer should be sent to
recertification school.
variable

that

should be

This writer suggests that time is a
considered when

organizations

individuals have to answer this type of question.

and

How long it

has been since the officer received training and how many
accidents the officer has investigated during the past year
can have a significant effect on the officer's ability to
perform his assigned task effectively.

Additional study is

needed to address specific time periods and the effects of the
number of accidents

investigated

in those time periods

on

retention.
Education Beyond High School
This researcher addressed education and its effect on
retention of cognitive knowledge gained in the AI-1 training

88
program.

College education in the form of an associate degree

obtained from a two-year institution was positively related to
retention of AI-1 cognitive knowledge,
participated in this evaluation.

of the officers who

Officers who had a four-year

degree or more had lower scores than those with associate
degree's.

Several two-year institutions in Michigan offer an

associate degree in law enforcement.

Some of the two-year

institutions offer college classes that are oriented toward
the skills and knowledge needed to perform traffic related
duties effectively.

Three Michigan colleges offer a two-year

law enforcement associate degree in conjunction with the basic
police academy.
the

police

It would be interesting to determine whether

officers

who

scored

higher

on

the

performance

retest had such a degree.
Other Traffic Related Training
Other traffic-related training programs are available to
Michigan police officers,
legal updates,

such as radar,

court testimony,

alcohol detection,

selective enforcement,

other levels of accident investigation.

and

Other training might

have improved the scores of the officers who took the perfor­
mance

retest.

Future

researchers

could determine whether

additional training affects an officer's ability to investi­
gate traffic accidents.

Evaluations of this type could be

employed to identify the variables that are positively related
to

retention

of

cognitive

knowledge

training programs listed above.

gained

in

the

other

89
Police Officer Age
This research included officers in the age range of 24 to
43 years.

Information from the literature review suggested

that age does not become a factor in cognitive retention until
later years.

Future researchers could address the issue of

longevity and its effect on an accident investigator's knowl­
edge.

The issue of how long an individual has been performing

a specific skill and its relationship to retention has not
been addressed.

An evaluation addressing this issue could

help determine whether police officers should specialize in
certain functions such as accident investigation.

APPENDICES

APPENDIX A
SUMMARY OF TEST DEVELOPMENT

92

Summary of Test Development
The purpose of developing the Traffic Accident Investiga­
tion tests was to evaluate the effectiveness of the Michigan
State University Highway Traffic Safety Center's AI-1 course
formats.

Emphasis was placed on developing and using test

items that were highly relevant to
tion.

the objectives of instruc­

The validity of item content was ascertained by the

judges.

Although such matters as the shape of score distribu­

tions, the indices of difficulty and discrimination, and the
reliability coefficients were of secondary importance, these
data were presented on each content area for both the pre-test
and the post-test.
The following observations can be made:
1.

The pre-test

scores

for the total

test and each

content area were normally distributed.
2.

The post-test distributions of four content areas

deviated
course

from normaility,

content.

distributed.

One

showing the students'

content

area

of

five

mastery of

was

normally

However, the post-test total score distribution

retained normality.
3.

The reliability coefficients on the post test were

higher than the corresponding coefficients on the pre-test.
The pre-test scores included a large guessing factor.
actual

reliability

coefficients

The

of the post-test might be

higher than those reported, since mastery of items leads to an
underestimate of test reliability.

Ebel stated that if a test

93

includes many items on whcih the average score is near 100%,
the underestimate of reliability could be quite large. 59
The data presented support the adequacy of the tests for
the purpose they served.

59Ibid., p. 415

APPENDIX B
CURRICULUM OUTLINE

95

Unit 1.1
Highway T r a f f i c S a f e t y Program
C i v i l & Environmental E n g i n ee r i ng
Michigan S t a t e U n i v e r s i t y

Curriculum Outline
1.

Introduction/Overview
1.1 Introduction & Overview
1.2 Student Assessment (Pre-Test)
1.3 Expected Student Competency Levels

2.

Measuring and Recording
2.1
2.2
2.3
2.4

3.

Field Sketching
Field Measurements
Accident Photography
Field Exercise #1

Elements of Traffic Accidents
3.1 Multiple Causation Theory
3.2 Elements of Traffic Accidents

4.

Speed Determination
4.1
4.2
4.3
4.4
4.5
4.6

5.

Roadway Evaluation
5.1
5.2
5.3
5.4
5.5
5.6
5.7

6.

Symbols/Abbreviations, Speed 4 Velocity
Determining Drag Factors
Determining Speed from Skidmarks
Determining Yaw/Sideslip and Critical-Curve Speeds
Fall, Flip-Vault and Combined Speeds
Field Exercise #2

Roadway Evaluation Introduction
Final Positions
Tiremarks
Metal Scars
Debris
Fixed Objects
Field Exercise #3

Vehicle Evaluation
6.1
6.2
6.3
6.4
6.5

Types of Vehicle Damage
Thrust and Collapse
Ground Contact
Recording Damage to Vehicles
Field Exercise #4

96

2

7.

Legal
7.1 Duties Required by Statute
7.2 Authority to Gather Accident Information
7.3 Enforcement Authority at Accident Scenes

8.

Course Review
8.1 Review of Course Content

9.

Course and Student Evaluation
9.1 Student Evaluation (Post-Test)
9.2 Course Evaluation by Students

APPENDIX C
OFFICER INFORMATION SHEET

98

Name

Social S ecu rit y N o . ____________________

(L a s t )
(Firs11
(M .I . I
Birthdate ______________________________

Age

.

AI-1 Course L o c a t i o n ______________________ AI — 1 Course D a t e s __________________
AI-1 Course Score -- P R E T E S T __________________ POST T E S T ________________________
Employing A g e n c y ________________________________Phone

(

)_____________________

Agency A d d r e s s _____________________________________________________________________
(street)
(cityl
(zip)
Total number of Police personnel (sworn o n l y ) _________________________________
R a n k _______________

Ed ucation

HIGH SCHOOL

1 2

3

4

C OLL E G E

1 2

3

4

Years of Police S e r v i c e
Years a s s i g n e d s p e c i f i c a l l y to t r a f f i c ______
Approximate number of traffic acc idents investigated by you in last months:
Fatal
Personal Injury
Pro per ty Damage
_________________
6 m o n t h s _________ ____________________________
12 m o n t h s _________ ____________________________ _____________________________
Accident

Investigation cou rse s c omp let e beyo nd AI — 1 - cir cle

have comple ted

AI

2

3

4

Other traffic related training:
Content
Whe re

A ssigned

to:

_Patrol Division
Traffic Division
Training Division
o t h e r : ____________

5

6

7

8

9

the ones you

10

W h e r e . Wh e n . Le n g t h . and Basic Content
When
Length

Func t i o n :

Put i e s :

___General Patrol___________________ A dm ini s t r a t i v e
___ Traffic E n for cem ent_________ ___ S u per vis ory
Tr aining______________________ ___ "Line" operations
___ o t h e r : _______________ _____________o t h e r : _____________

Please answer the following que stion on the 1 to 5 scale (1 = di sag r e e and
a = agree).
Since c o m p l e t i o n of AI-1.
Circle one number for each question.
1.

I still ma int ain the knowle dge and skills

2.

My d ep art men t encour age s and allows me
AI-1.
1 2
3
4
5

.!.

The qu ality of my acc ide nt

4.

I frequently refer back to my AI-1
the investigation.
1 2
3
4
5

gained

to use

investigation has

the

in AI-1.

1 2

3

4

information ob tai ned

improved.

training material

1

2

in order

3

4

5
in

5

to complete

Cse the following space (including the back) to describe: 1) How AI-1 can be
improved 2) How or if AI-1 training has ben ef i t e d you and your depa rtm ent and
3) If you c urr ent ly do not have a good grasp on AI-1 content and skill - why
not.

APPENDIX D
-1 TEST - PART

100

DO NOT WRITE ON THIS TEST BOOKLET
Booklet # _ _ _

AI-I
Test
Part 1

Questions 1 thru 15 will consist of you watching slides projected on the
screen. The instructor will ask you a question about each of the slides as
they are projected. Choose your answer to the question from the answers given
below. Mark your answer on the answer sheet provided. There Is only one
correct answer to each question. Questions 1 thru 15 are worth 3 points each.
1.

Answers:
A.
B.
C.

Acceleration
Braking
Not Sure

2 . Answers:

A.
B.
C.
D.
3.

Answers:
A.
B.
C.
D.

4.

Skidmark
Scuffmark
Tireprint
Not Sure

Answers:
A.
B.
C.
D.

7.

Rub off
Imprint
Superimposed contact damage
Not Snrp

Answers:
A.
B.
C.
D.

6.

Gap skid
Skip skid
Tire mark
Not Sure

Answers:
A.
B.
C.
D.

5.

Triangulation
Coordinate
Offset
Not Sure

Coordinate
Offset
Triangulation
Not Sure

Answers:
A.
B.
C.
D.

Yawmarks
Tireprints
Skidmarks
Not Sure

101

Page 2
8.

Answers:
A.
B.
C.
D.

9.

Acceleration
Braking
Not Sure

Answers:
A.
B.
C.
D.

11.

Tire grinding
Pavement grinding
Scratch
Not Sure

Answers:
A.
B.
C.

10.

DO NOT WRITE ON THIS TEST BOOKLET

Furrow
Rut
Tireprint
Not Sure

Answers:
A. Contact damage only
B. Induced damage only
C. Both contact and induced damage
D. Not Sure

12.

Answers:
A. Contact damage only
B. Induced damage only
C. Both co n t a c t and induced d a m a g e
D. Not Sure

13.

Answers:
A.
B.
C.

14.

Controlled
Uncontrolled
Not Sure

Answers:
A. Straight
B. Curved
C. Overlapping
0. Not Sure

15.

Answers:
A.
B.
C.
D.

Rotated to its left (clockwise)
Rotated to its right (counter-clockwise)
No rotation
Not Sure

102

Page 3

(Questions 16 thru 60 are worth 1 point each)

16. When one vehicle crosses over onto the wrong side of the road and
occupies the path assigned to another vehicle, this is referred to as:
A.
B.
C.
D.
E.
17.

Encroachment
Ma ximum engagement
First harmful event
Disengagement
Not Sure

Which of the following correctly lists the five types of metal scars that
can be found on the roadway at accident scenes?
A.
B.
C.
D.
E.

Scratch,
Scratch,
Scratch,
Scratch,
Not Sure

scrape,
scrape,
scrape,
imprint,

chip, imprint, furrow
chip, chop, groove
furrow, rip, imprint
gouge, rip, furrow

18. How should a measurement of 5 feet and 6 inches be written
sketch, as recommended in the text?
A.
B.
C.
D.
E.
19.

Date of accident
Scale of sketch
Direction north lies by the compass
Name of person making sketch
Not Sure

Compute the minimum initial speed of a vehicle that laid down 90 feet
of locked wheel skidmarks on a pavement surface with a coefficient of
friction of .72
A.
B.
C.
D.
E.

21.

5'6"
5.5'
566"
Not Sure

All but one of the following should appear on every field sketch.
Identify the exception.
A.
B.
C.
D.
E.

20.

on a field

37 mph
40 mph
44 mph
54 mph
Not Sure

Measurement(s) to be taken first at the scene of an accident should be:
A.
B.
C.
D.
E.

Marks or residues of a temporary nature
Permanent gouges on the pavement
The longest distances that have to be measured
The width of the streets involved
Not Sure

103

Page 4
22.

DO NOT WRITE ON THIS TEST BOOKLET

When calculating the speed of a vehicle that slid on a road with a
6 % downgrade:
A. .06 is added to the coefficient of friction
B. The 6 % is not considered, as that variable is already compensated
for in the basic speed formula
C. .06 is subtracted from the measured skid distance
D. The 6 % is not considered if, to determine the drag factor, the test
skid is made down the same grade the accident vehicle slid on
E. Not Sure

23.

The "shadow" of a skidmark is:
A. The part of a skidmark in which a locked wheel loses contact
with
the ground when it bounces or skips
B. The indistinct part of a skidmark left before a tire becomes hot
enough to smear
C. The distance through which brakes are slowing the vehicle before
they ire applied hard enough to lock the wheels
D. The superimposing of one skidmark on another
E. Not Sure

24.

When skidmarks left by a car are curved:
A. Each should be measured in a straight line from one end to another
B. Each should be measured along the curve
C. The distance should be measured from the center point of the car
where it began to slide, to the center point where it stopped sliding
D. The longest skidmark should be measured along the curve
E. Not Sure

25.

In estimating speed from skidmarks, it is important to remember that you
are determining the:
A. Exact speed of the vehicle prior to the collision
B. Ma x i m u m speed of the vehicle prior to the collision
C. Mi ni mum speed the vehicle would have to be travelling to lay down
the skidmarks
D. Actual crash speed
E. Not Sure

26.

The coefficient of friction is:
A. The ratio of force necessary to slide an object at uniform speed on a
surface, to the pressure of the object against that surface
B. The amount of friction generated between the brake shoes and the
brake drums
C. The amount of grade, plus or minus
D. The amount of buckling that occurs when objects collide
E. Not Sure

104

Page 5
27.

DO NOT WRITE ON THIS TEST BOOKLET

"Superelevation" is:
A. A raised stretch of road over a railroad track
B. Number of feet a road rises for each 100 level
C. A measure of the sharpness of a curve
D. Slope measured across the road on a curve
E. Not Sure

feet along the road

28. Photographs are admissible in evidence only when:
A. The photographer is first called to testify
B. They are not gruesome or bloody
C. They are material and relevant to the issues in the case and a
proper foundation is laid for their introduction
D. No one objects at trial to their introduction into evidence
E. Not Sure
29.

That event in the accident which stabilizes the accident situation is:
A. First harmful event
B. Initial contact
C. Disengagement
D. Stopping
E. Not Sure

30. During your investigation of an accident, you learn that an unidentified
vehicle forced Vehicle #1 off the road. There was no collision between
the unidentified vehicle and Vehicle #1. The unidentified vehicle is
properly referred to as:
A.
B.
C.
D.
E.

A hit and run vehicle
A disengaged traffic unit
An evasive action unit
A non-contact unit
Not Sure

31. Skidmarks at the scene of an accident:
A.
B.
C.
D.
E.
32.

Can only be used if the vehicle that slid is found at rest on those
skidmarks
Can be used to show the physical condition of the driver
Are useless unless there are four identifiable marks
Can be useful in determining initial positions of vehicles
Not Sure

A vehicle with a speed of 45 mph is travelling at a speed of _ _ _
per second (f ps ).
A.
B.
C.
D.
E.

38 fps
45 fps
54 fps
66 fps
Not Sure

feet

105

Page 6
33.

An accident-involved driver's decision to operate his vehicle while
under the influence of alcoholic beverages, is a _ _ _ _ _ _ _ _ _ _ _ _ con­
tributing factor to that collision.
A.
B.
C.
D.
E.

34.

Traction instability
Grade and/or slope
Crucial event speed
Critical speed
Not Sure

A dent pressed into vehicle body parts by some stronger object which
clearly shows it shape, is called:
A.
B.
C.
D.
E.

36.

Condition
Operational
Physical
Mental
Not Sure

A velocity above which a particular highway curve cannot be negotiated
by a motor vehicle without yawing, is called:
A.
B.
C.
D.
E.

35.

DO NOT WRITE ON THIS TEST BOOKLET

Obscured contact motion
Collapse
Imprint
Rub-off
Not Sure

When the accident scene is blanketed with heavy snow, the measuring
method you are most likely to use is t h e _ _ _ _ _ _ _ _ _ _ _ method.
A.
3.
C.
D.
E.

Triangulation
Coordinate
Offset
Engineering
Not Sure

For the following TRUE-FALSE QUESTIONS, mark answer "A" on the answer sheet
for "True", answer "B" for "False" and answer "C" for "Not Sure."
37.

The purpose of making urgent measurements is simply to locate temporary
and short-lived positions with respect to landmarks which are permanent.

38.

When using the triangulation method, you should measure two triangles
for every spot you want to locate.

39.

When using triangulation to locate points on an accident sketch, the
investigator should select permanent points for two corners and a
temporary object for the third corner.

40.

In measuring the total length of skidmarks, for determining minimum
initial speed, you should include any and all gaps as part of the
overall skidmarks.

106

Page 7

DO NOT WRITE ON T H IS T E ST BOOKLET

41.

The drag factor/coefficient of friction of a road surface can best be
established by consulting the table of drag factors in J.S. Baker's
text, "Traffic Accident Investigation Manual."

42.

The essential difference between tireprlnts and skidmarks is one of
rolling as compared to sliding.

43.

The speed computed from accident skidmarks does not represent all of the
actual speed of the vehicle just before the brakes were applied.

44. Unless there are four distinct marks, skidmarks are of no value as evidence.
45. A vehicle travelling at 65 mph on a surface with a drag factor of .55 will
have a braking distance of about 210 feet.
46.

A car and driver together weigh 3,000 lbs and skids to a halt in 30 feet
from a known speed. If six additional passengers weighing a total of
1,000 lbs are added to the car and all other conditions remain the same,
the test skid would be approximately 40 feet long.

47. Using a drag factor of .75 and a speed of 30
of a vehicle would be approximately 58 feet.

mph, the skidding distance

48. Grass pinched between a tire and wheel rim would indicate that the
vehicle moved violently sideways.
49.

Collapsed parts of vehicles involved in accidents can indicate the direction
and, to a degree, the amount of force which did the damage.

50. Reconstruction of an accident is nearly always based upon measurements
made at the accident scene during the initial investigation.
51.

Contact damage usually makes spider-web appearing or circular cracks
in windshields.

52.

The accident investigator must, know the reaction time of the driver
involved, to accurately estimate speed from skidmarks.

53. Superelevation is the rise or fall across.a roadway on a curve.
54. Skidmarks begin just as soon as the brakes on a car are applied.
55.

Physical facts at the scene cannot lie, but their significance may be
frequently overlooked or misinterpreted.

56.

In determining the radius of a curve, the entire length of the curve
must be measured.

57. An accident begins to happen at the instant of impact or upset.
58. If a vehicle had the right front wheel lock up while the others remained
free-rolling, the vehicle would turn clockwise.
59.

If the two rear wheels locked while the two front wheels remained freerolling, the vehicle would slide straight ahead.

60.

A field sketch is drawn to scale w h i l e a ma p is not to scale.

APPENDIX E
AI-1 TEST - PART 2

108

Highway Traffic Safety Programs
Civil & Environmental Engineering
Michigan State University
AI-I_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
Test
PRINT Your Name Clearly
Part 2
Score _ _ _ _
i

SPEED CALCULATION PROBLEMS (5 points for each question)
Do your work on this page to answer these two problems. Use the reverse
side for your calculations 1f necessary. Show your work as completely as
possible. Partial credit may be given even if you do not arrive at the
correct answer. Round your answer down to the nearest whole mph.
Problem A
The following skidmarks were straight and made on a level asphalt
surface.
Accident Vehicle Skidmarks:

Test Skid Vehicle Skidmarks:

LF «

196-

RF «

196-

LF =

39!i

RF «. 3 9 l£

LR *

196—

RR «

197—

LR =

4o£

RR =

4oi

Test Skid Vehicle Speed * 30
Question: What was the vehicle's minimum Initial speed? _ _ _ _ _ mph
Not Sure _ _ _ _ _
(

Problem B
A vehicle failed to negotiate a curve, leaving yawmarks on the road.
You determine that the yawmarks have a middle ordinate of 11 inches when
you use a 60 ft. chord. You also note that the curve superelevation is
Test skids you conducted at 30 mph on a level stretch of roadway as
close to the accident site as possible resulted in the following:
Left Front Rioht Front Left Rear Riaht Rear Averaae
Test #1

46-

4?0

43-

ol

Test #2

Ol
in

44^

41-

42-

Question: What was the vehicle's sideslip (yaw) speed? _ _ _ _ _
Not Sure _ _ _ _ _ _

I

R2

mph

BIBLIOGRAPHY

110

BIBLIOGRAPHY
Anderson, R. C . , and Biddle, W.B.
"On Asking People Ques­
tions About What They Are Reading." Psychology of
Learning and Motivation. Edited by G. Bower, New York:
Academic Press, 1975.
Andre, T . , and Thieman, A.
"Learning Concepts by Reading."
Contemporary Educational Psychology 13 (1988): 296-307.
Atkinson, R. C. and Shiffron, R. M.
"Human Memory: A
Proposed System and Its Control Process." In K. W.
Spence and J. T. Spence (Eds.) The Psychology of
Learning and Motivation (Vol. 2), New York: Academic
Press, (1968): 314-315.
Ausubel, D. P.
"Relationships Between Didactic and
Experimental Learning." Paper prepared for the
National Center for Research in Vocational Education
Symposium on Learning-In-Work, Washington, D.C.,
October, 1980.
Ausubel, D. P.
New York:

Educational Psychology:
A Cognitive V i e w .
Holt, Rinehart and Winston, 1968.

Bandura, A. Social Learning Theories.
Learning Press, 1971.

New York:

General

Belbin, Eunice, and Downs, Sylvia.
"Interference Effects
From New Learning: Their Relevance to Design of Adult
Training Programs.: Journal of Gerontology. 20,
(1965): 154-159.
Bilodeay, E. A. and Levy, C. M.
"Long-Term Memory as a
Function of Retention Time and Other Conditions of
Training and Recall." Psychology Review. 71, (1964):
24-41.
Blanton, B. Elgit.
"Reception Learning and Advanced
Organizers:
Implications for Reading".
In National
Reading Conference. Edited by Frank Greene.
Milwaukee:
The National Reading Conference, (1972):
4-5.
Borg, Walter R. and Gall, Meredith D. Educational Research:
An Introduction. New York:
David McKay Company, Inc.,
1974.

Ill

Bruner, J. S.
"The coursed in Cognitive Growth."
Psychology. 10, (1964):
1-15.

American

Carrol, J. B.
"The Model of School Learning:
Process of an
Idea." In C. W. Fisher & D.C.
Berliner (Eds.)
Perspective on Instructional Time. New York:
Longman,
(1985):
29-58.
Cheng, P.W.
"Restructuring versus Automaticity:
Alternative Accounts of Skill Acquisition.
Psychology
Review. 92, (1985):
214-223.
Cornwell, J. B.
"Justify Your Training Efforts with on-theJob Performance".
Training. Vol. 17., No. 8 (August,
1980): 34-36.
Davey, B.
"Cognitive Styles and Reading Achievement."
Journal of Reading. Vol. 20, (1976):
113-120.
Ebbinghaus, H. Uber Das Gedachtnis. Leipzig:
Dunker &
Humbolt, 1885.
Translated by H. Ruyer and C. E.
Bussenius.
New York:
Columbia Teachers College, 1913.
Ebel, R. L. Essentials of Educational Measurement.
Englewood Cliffs, N.J., 1972.
Flavel, L. J.
"On Cognitive Development".
Development. 53, (1982):
1-10.

Child

Frankel, M. T. and Rollins, H. A.
"Associative and
Categorical Hypotheses of Organization in the Free
Recall of Adults and Children." Journal of
Experimental Child Psychology. 40, (1985):
304-318.
Gagne, E. D.
"Long-Term Retention of Information Following
Learning From Prose".
Review of Educational Research.
48, (1978):
629-665.
Glasss, Gene V. and Hopkins, Kenneth D. Statistical Methods
in Education and Psychology. Englewood Cliffs, NJ:
Prentice Hall, 1984.
Glazer, Waldman and Cox, Hilda.
"The Use of Similarity
Judgments to Assess the Effectiveness of Instruction".
Education. Vol 100, No. 4, (1980):
352-359.
Justice, Elaine M. and Weaver-McDougal, Rebecca G.
"Adults'
Knowledge About Memory: Awareness and Use of Memory
Strategies Across Tasks". Journal of Educational
Psychology. 81(2), (1986):
214-229.
Kirkpatrick, E. A.
"An Experimental Study of Memory".
Psychological Review. Vol. 1, (1894):
602-609.

112

Kirkpatrick, D. L.
"Techniques for Evaluating the
Effectiveness of Technical Training." Training and
Development Journal. (June, 1971):
86.
LaPorte, R. E. and Voss, I. F. "Retention of Prose
Materials as a Function of Post-Acquisition Testing".
Journal of Educational Psychology. 67, (1975):
259266.
Law, Alexander, and Bronson, William H.
"The Evaluation
Improvement Program". Evaluation Trainer's Guide.
Princeton, N J : Educational Testing Service, 1977.
McGeoch, J. A.
"Forgetting and the Law of Disuse".
Psychological Review. 39, (1932):
352-370.
McGeoch, J. A. and Irion A. L. The Psychology of Human
Learning. New York:
David McKay, 1952.
McLeod, Beverly, and McLaughlin, Barry.
"Restructuring or
Automaticity? Reading a Second Language".
Language
Learning, 33, 1 (March 1983):
135-158.
Mehrens, W. and Lehmann, I. Measurement and Evaluation in
Education and Psychology. New York: Rinehart and
Winston, Inc., 1986.
Michigan Department of State Police, Michigan Traffic
Accident Facts. Col. R. T. Davis, Director, Lansing,
Michigan, 1988, 5.
National Safety Council.
1989, 5.

Accident Facts. Chicago, Illinois,

Page, Terry G. and Thomas, J. B . ; with Marshall, R.
International Dictionary of Education. London:
Kegan,
Paul: New York: Nicholas Publishing Co., (1977):
381.
Pressley, D. L . , MacKinnon, G. E. and Waller, T. G. (Eds.).
Metacognition. Cognition and Human Performance. New
York: Academic Press, (1985):
111-153.
Sage, G. H.
Introduction to Motor Behavior:
A
Neurospsychological Approach. 2nd ed. Menlo Park,
California: Addison Wesley, 1977.
Scovel, T.
"Questions Concerning the Application of
Neurolinguistic Research to Second Language
Learning/Teaching".
TESOL Quarterly. 16, (1982):
332.

323-

Singer, Robert N. Motor Learning and Human Performance. 3rd
ed. New York: Macmillann, 1981.

113

Swierczek, Fredricks and Carmichael, Lynne.
"The Quantity
and Quality of Evaluating Training".
Training and
Development Quarterly. (January, 1985):
95-99.
Thesaurus of Psychology:
Index Terms. 5th ed., American
Psychological Association, 1988, Vol. 3.
Thorndike, E. L. The Psychology of Learning.
Teachers College, 1914.
Tolman, E. C.
"Psychology:
Koch (Ed.). New York:

New York:

A study of A Science".
McGraw-Hill, 1959.

In S.

Tulving, E. and Watkin, J. J.
"Continuity Between Recall
and Recognition".
American Journal of Psychology. 86,
(1973):
739-748.
Underwood, B. J. and Postman, L.
"Extraexperimental Sources
Of Interference In Forgetting." Psychological Review.
67, (1960):
73-95.
Waters, H. S.
"Memory Development in Adolescence:
Relationship Between Metamemory, Strategy Use and
Performance." Journal of Experimental Child
Psychology. 33, (1982):
183-195.
Weisss, Thomas Michael; Hover, K. H . ; and Mills, D. F.
Psychological Foundations of Education. Dubuque, Iowa:
Wm. C. Brown, 1963.