| TABLE OF
CONTENTS
Introduction
The Hartman
Model
Construct
Validity Studies Of
The Hartman Value Profile
EEOC Studies
of The Hartman Value Profile
Instrument
Reliability Study of
The Hartman value Profile
Concurrent
Validity Studies of
The Hartman Value Profile
Criterion
Validity Studies
INTRODUCTION
Dr. Hartman constantly
reminded his students that "The proof of the pudding is in
the eating." In other words, regardless how logical,
rigorous or beautiful a system of thinking might be, if the
system does not provide usable and testable results, then it
does have the right to be called a science. There are four
keys to a science. A science is:
- based on objective
observations which are independent of any one
subject's perspective.
- based on a
mathematical measuring system.
- universally
applicable.
- subject to empirical
testing to confirm the observations.
As a result, the applications
of science must be valid and reliable.
The science of axiology,
founded by Dr. Hartman, meets all four of these conditions.
The science is based on formal value theory generated by
Hartman's value mathematics. The Hartman Value Profile is an
application of Dr. Hartman's axiology. As such, it is based
on value mathematics. Moreover, the norm for the profile is
generated prior to statistical evaluation of profile
responses and results from the logical relations of the
value system.
The primary task of the value
scientist is to establish the validity of the Hartman Norm
as a measure of reality and to substantiate the reliability
of the instrument. Value Resource Group has participated in
a variety of research projects to validate the Hartman value
Profile, substantiate the reliability of the test instrument
and demonstrate the usability of the results from the
Profile. In addition, other value scientists, axiologists,
have conducted studies to examine and confirm the validity
of the
Hartman Value Profile.
Standard statistical
validation consists of three types of validation studies:
construct validity, concurrent validity and criterion
validity. Construct validity examines the instrument itself
and determines that all of the items on the instrument are
relevant and have the prescribed effect on the area or
concept that they measure. In other words, does the profile
measure what it is suppose to measure. Concurrent validity
is obtained by correlating an instrument to other industry
accepted instruments. Criterion studies demonstrate that the
instrument can be used as a valid predictive measure within
a specific application or discipline.
Confirmation studies in all
three of the major areas of validation have been conducted
both internally by axiologists and psychologists at Value
Resource Group and externally by scientist and statisticians
outside Value Resource Group. In addition, internal and
external studies have been conducted to confirm that The
Hartman Value Profile and its results are reliable and do
not discriminate either by age, race or sex.
This report is a summary of
some of the key studies which have taken place to this date.
First, the report will examine The Hartman Model to provide
a brief introduction to the components of Hartman's Value
Mathematics as it applies to The HVP. Next, the report will
examine the results of a series of validity studies which
provide strong statistical confirmation that Hartman's model
is in fact a real and accurate interpretation of the way we
make value judgments.
THE
HARTMAN MODEL
According to Dr. Robert S.
Hartman's system for measuring "Value", value is a phenomena
of concept. According to his system, the value of anything
is determined by the extent to which it meets the intention
of its meaning. For example; a chair that all of the natural
properties contained within the definition of chair, is by
definition a "good chair". A chair that has only a few of
the natural properties contained within the definition of a
chair is by definition a "less than good" or "not as good"
chair, a fair chair, a poor chair, etc.
A "good chair", according to
Hartman, fulfills the intention of its definition. Given
that premise, Hartman's theories set forth a system of
mathematics to establish and prove the accuracy and utility
of his theories.
Hartman discovered that every
concept has three dimensions. Every concept has the
following types of value:
- The value of its
uniqueness,
- The value of its
function or role, and
- The value of its meaning
and purpose.
These three "Dimensions of
Value" are referred to as the following concepts:
- Intrinsic Value
Dimension,
- Extrinsic Value
Dimension,
- Systemic Value
Dimension.
Hartman's model objectively
measures the relative clarity and level of development of
each of these value dimensions. For Hartman, this measures
the structure and dynamics of a person's value system
relative to each unique concept being measured.
In addition to measuring
value and personal value systems, the Hartman system
measures a person's value judgments; their valuations.
According to Hartman, when a person makes a judgment about
any conceptual phenomena, they make judgments in terms of
value combinations within and between dimensional sets.
For example, there are three
Value Dimensions: Intrinsic, Extrinsic and Systemic. These
three Value Dimensions can be combined in eighteen (18)
different ways when one is making a value judgment.
Consider the concept WIFE.
The intrinsic value of WIFE
is a wife's value resulting from being
a unique, one of a kind, authentically original individual.
Wife has value because she is
unique.
This uniqueness constitutes her INTRINSIC Value.
However, what happens when we
make a value judgment about WIFE.
What are our options.
The value is: the uniqueness
of wife.
When making a judgment about
a wife's uniqueness, I can:
Value the Intrinsic Value
intrinsically = I love my wife's uniqueness.
Value the Intrinsic Value
extrinsically = I enjoy my wife's uniqueness.
Value the Intrinsic Value
systemically = I find my wife's uniqueness meaningful.
Disvalue the Intrinsic
Value intrinsically = I hate my wife's uniqueness
Disvalue the Intrinsic
value extrinsically = I dislike my wife's uniqueness
Disvalue the Intrinsic
Value systemically = I think my wife's uniqueness is
crazy
The Value Dimension of the
concept Wife that I am making value judgments about, is her
INTRINSIC VALUE.
There are at least 6
different value judgments or evaluations I can make about
her INTRINSIC VALUE.
I can make the same six
judgments for the EXTRINSIC and SYSTEMIC Dimensions of Value
of her value.
By combining the INTRINSIC,
EXTRINSIC and SYSTEMIC judgment combinations, there are
eighteen different evaluations about the value of a wife
that can be made.
The fascinating discovery
that Hartman made was that the 18 valuational possibilities
are not made randomly because their relative value is not
relative in nature. The relative value of each judgment is
built into the structure and dynamics of the conceptual
system which generates decisions.
As a result of Hartman's
research, he discovered that the relative value of each one
of the 18 valuational possibilities is different and that
the hierarchical arrangement of their relative value is
constant across concepts. He demonstrated that the relative
value of the intrinsic valuation of the Intrinsic Value is
more valuable than any one of the other 17 valuational
possibilities, and the intrinsic disvaluation of the
Intrinsic Value is less valuable than any of the other 17
valuational possibilities. He discovered that regardless
whether one uses symbolic logic, the theory of types, set
theory or transfinite set theory, any quantification of the
18 valuational possibilities will always arrange the items
in the same hierarchical order.
Hartman discovered that there
is a value norm in the natural universe, in the phenomenal
world of concepts. This value norm enables us to objectively
measure and study the unique characteristics of the
structure and dynamic of any person's value and valuations
concerning any conceptual phenomena.
Value profiles based on the
Hartman Model measure the deviation of a subjects ranks,
given to each of the eighteen items in a profile, from the
ranks which the model stipulates to each item. The model and
its internal logic determine the correctness or
incorrectness of an individual's value judgments. The
profile measures a person's capacity for making value
judgments.
The Profile
Scores(axiological scores) are numerical. The lower the
number, the better the axiological score. The higher the
number, the worse the score.
Hartman developed a scoring
system for measuring the axiological scores which results
from the difference between an individual's ranking of the
items and the axiological norm.
The first scale consists of
four measures:
- The Differentiation
Score, for the capacity to differentiate values;
- The Dimension Score, for
the sense of proportion which results from the
equilibrium between value dimensions;
- The Integration Score,
for the capacity to solve problems and see the relevant
in the complex;
- The Dissimilarity Score,
for the capacity to distinguish between good and bad.
The second scale consists of
the "measure" of the value dimensions:
- Intrinsic Value, the
capacity for discerning values concerning uniqueness and
individuality.
- Extrinsic Value, the
capacity for discerning role, function and practicality.
- Systemic Value, the
capacity for discerning values concerning meaning,
purpose, order and system.
The third set of scales
concerns the measurement of the value dimensions. Each of
the three value dimension has a:
- Dimensional Score,
indicating the capacity for discerning the value
dimension in question.
- Integration Score,
indicating the capacity for problem solving in that
dimension.
Using Hartman's axiological
profile system and scoring system, we measure, analyze and
assess a person's value system and method for making value
judgments and translate these measurements into a variety of
descriptive and prognostic reports that both describe and
make various diagnostic predictions about a person.
CONSTRUCT VALIDITY STUDY OF THE HARTMAN VALUE PROFILE
Value Resource group has
participated and sponsored in several studies concerned with
the construct validity of an eighteen item inventory
designed to measure values placed by subjects within the
dimensions of value measured by Dr. Hartman's mathematics.
The focus of the studies is the measure of the congruency
between subject's rank ordering of the eighteen items and
the Hartman Norm which is based on his Value Mathematics.
Two different studies were performed. One study was an
internal study sponsored by Value Resource Group which
included 1777 subjects. The other study was sponsored by
Dollar General Corporation and was performed primarily by
Dr. Chuck McDonald (A forensic psychologist in Nashville)
and William Panak (a statistician). Thhe second study
included 6,35 individuals pulled at random from a data based
in the thousands.
Dr. Hartman's model orders
each of the possibilities of a value profile from 1 to 18.
Based on Dr. Hartman's model, a person's value structure
consists of three dimensions, intrinsic, extrinsic and
systemic. Each of these dimensions is valued from a
dimensional standpoint, i.e. the intrinsic valuation of
extrinsic value, the extrinsic valuation of a systemic
value, the systemic valuation of an intrinsic value, etc.
Each value dimension can be valued in three ways,
intrinsically, extrinsically and systemically, resulting in
a total of nine valuational combinations. In addition, each
value dimension can be either enhanced valuationally, this
is called a "composition", or diminished valuationally, this
is called a "transposition". As a result, each Value profile
is composed on nine "compositional" items and nine "transpositional"
items.
The key issue for construct
validity is do large groups of people, on the average, order
the items in the same way predicted by the model. Value
Research conducted studies on both Hartman Value Profile One
(A measure of one's capacities to make value judgments about
the world) and Hartman Value Profile Two (a measure of one's
capacities to make value judgments about oneself.
The focus of the studies was
on three areas:
- The construct validity
of the Profile as a whole.
- The construct validity
of the Valuational (compositions) and Disvaluational
(transpositions) items.
- The construct validity
of each of the 18 individuals items
of the Profile.
At the profile level, the
studies examined the extent of isomorphism (correlation)
between Hartman's model of value structures and the rank
ordering of the value profile instruments. The question is
whether the applicants would rank the 18 items in order
consistent with Hartman's model, i. e., norm item one ranked
lowest on the average and norm item eighteen ranked highest,
with all other items ordered from lowest to highest norm
rank.
At the level of compositional
vs. transpositional items, Hartman's model claims that
individuals will rank the compositional items with a
numerical ranking of 1 to 9 and the transpositional items
with a numerical ranking of 10 to 18. In other words,
individuals, on the whole, will see "good" items
(compositional items) as "good" and "bad" items (transpositional
items) as ""bad".
At the level of each of the
individual 18 items, the test is whether each item response
matches the norm value predicted by the model.
The following summaries will
focus on the results of the (1) Internal Studies performed
within Value Resource Group and the (2) External Studies
performed by Dr. McDonald both For Hartman Profile One
(World profile) and Hartman Profile Two (Self Profile).
A Study of The Profile As
A Whole
The Hartman model predicts
that homogeneous groups of individuals will rank order the
value profile instrument in a predictable way. According to
Dr. Hartman's model, the 18 valuational items have a fixed
hierarchical order of value. Hartman claimed that if one
were to objectively assess the value system and value
judgments of large samples of humanity, the average obtained
rank of all items would match the theoretical order
predicted by his system perfectly.
Internal Studies:
The Hartman model predicts
that homogeneous groups of individuals will rank order items
in a predictable order. To test for this order, Page's "Z"
was calculated and compared to critical Z values for a
one-tailed test. To derive an index of concordance between
subjects within each group, Kendall's coefficient of
concordance "W" was calculated for the group.
HVP1(World Profile):
Results indicate that the isomorphism between the model and
the observed sample ranking is quite high (Page's Z =
149.51) and the average correlation between pairs of
respondents is excellent (Kendall's W = .81). The rank-order
correlation of expected and
obtained mean rankings for the 18 items is .96.
HVP2(Self Profile):
Results indicate that the isomorphism between the model and
the observed sample ranking is also quite high (Page's Z =
143.64) and the average correlation between pairs of
respondents is excellent (Kendall's W = .75). Also, the rank
order correlation expected and obtained mean rankings for
the 18 items is .95.
External Studies:
The Hartman model predicts
that homogeneous groups of individuals will rank order items
in a predictable order. Data was obtained from 6,354
subjects randomly pulled from a data based of thousands.
To test the predictability of
the test, the following procedures were used: Friedman's
Tow-way ANOVA by Rank, Page's Test for Ordered Alternatives,
Kendall's Coefficient of Concordance, and Spearman's Rank
Order Correlation.
HVP1(World Profile):
The results indicated the following:
- Friedman's Two-way ANOVA
by Rank:
Friedman's chi-squared = 82479.290 indicating that there
is a difference in the ranking of items at the 99.9%
confidence level.
- Page's Test for Ordered
Alternatives:
Page's L = 12877990.0 and Page's Z = 272.869 indicating
that there is a set of items ordered by respondents in a
way that supports the model at the 99.9% confidence
level.
- Kendall's Coefficient of
Concordance:
Kendall's W = 0.764 and Kendall's chi-squared =
82479.290 indicating that there is a significant
concordance between respondents at the 99.9% confidence
level.
- Spearman's Rank Order
Correlation:
Spearman r = .96 indicating that there is a very strong
and significant correlation between the rank order of
the model and the rank order of obtained rankings.
Hartman's Value Profile
One(World Profile) demonstrates strong support for Hartman's
model when overall items is examined.
HVP2(Self Profile):
The results indicated the
following:
- Friedman's Two-way ANOVA
by Rank:
Friedman's chi-squared = 87322.741 indicating that there
is a difference in the ranking of items at the 99.9%
confidence level.
- Page's Test for Ordered
Alternatives:
Page's L = 12964086.0 and Page's Z = 282.60 indicating
that there is a set of items ordered by respondents in a
way that supports the model at the 99.9% confidence
level.
- Kendall's Coefficient of
Concordance:
Kendall's W = 0.808 and Kendall's chi-squared =
87322.741 indicating that there is a significant
concordance between respondents at the 99.9% confidence
level.
- Spearman's Rank Order
Correlation:
Spearman r = .95 indicating that there is a very strong
and significant correlation between the rank order of
the model and the rank order of obtained rankings.
Hartman's Value Profile
Two(Self Profile) also demonstrates strong support for
Hartman's model when overall items is examined.
A Study of The
Compositional-Transpositional Items
The key to the compositional\transpositional
analysis is the question: Do individuals, on the average,
rank compositional items ("good items") as good and
transpositional items ("bad items") as bad. Compositional
items on a value profile have a norm ranking of 1 through 9.
Transpositional items have a norm ranking of 10 through 18.
Internal Studies:
To test the hypothesis
concerning the compositional\transpositional items, raw
ranks were recorded so that 0 represented no distortions
(e.g., compositional items were ranked as compositions and
transpositional items were ranked as transpositions) and a
1 represented a distortion (e.g., a compositional item
ranked as a transposition or a transpositional item ranked
as a composition. The statistical test used was a 2x2
chi-square test for independence of expected and obtained
ranks.
HVP1(World Profile):
The results indicated that, averaged across the entire
sample, all items were ranked at their appropriate
compositional- transpositional level (chi-square = 18.00, p
< .001).
HVP2(Self Profile):
Again, the results indicated that, averaged across the
entire sample, all of the items were ranked at their
appropriate compositional-transpositional level (chi-square
= 18, p < .001).
External Studies:
The key in these studies is
to identify deviant items, "bad" items ranked as "good" and
"good" items" ranked as "bad".
Items's 1 thru 9 = Range of
Median Ranks 2 - 8
Items's 10 thru 18 Range of
Median Ranks 11 - 18
The compositional and
transpositional items were split and Spearman's rank order
correlations was run on each half.
HVP1(World Profile):
The results for the compositional items (1 thru 9) were:
S
R = 0.84
The results for the
transpositional items (10 thru 18) were:
S
R = 0.84
These results indicate that
no items in Hartman's World Profile are grossly misconstrued
to the point of being seen with a valence opposite of that
which is expected.
HVP2(Self Profile):
The results of the compositional items (1 thru 9) were:
S
R = 0.83
The results of the
transpositional items (10 thru 18) were:
S
R = 0.98
Again, these results indicate
that no items in Hartman's Self Profile are grossly
misconstrued to the point of being seen with a valence
opposite of that which is expected under the Hartman model.
Summary Comment on Internal
and External Compositional-Transpositional Studies
The studies indicate that, on
the average, individuals do not see "bad" items as "good"
and "good" items as "bad. Individuals do, in fact, make
distortions and these distortions become important when
distinguishing them as individuals and when measuring their
capacity to make value judgments. The key is that, on the
whole, the compositional-transpositional analysis supports
Hartman's model that the differences are a measure of
reality.
A Study of Individual
Items
The two former classes of
construct validity studies, evaluation of the profile as a
whole and analysis of compositional- transpositional items
support Hartman's model. This analysis examines how close
the individual items are ranked against Hartman's
mathematical norm. In short, the studies look for individual
items which deviate significantly from their expected
rankings. The studies represented in this validity synopsis
are the external studies, those done by professionals
primarily outside of Value Resource Group, conducted By Dr.
Chuck McDonald.
Differences between the
obtained sum of ranks of all of the obtained ranks, and
differences between the obtained sum of ranks and the model
sum of ranks for each of the items were analyzed using
critical differences defined by Hollander and Wolf and
Daniels.
HVP1(World Profile):
Of all possible comparisons between items, at least 89.54%
were statistically predicted by the model, 7.84% were not
predicted by the model, and 2.61% were in a direction
opposite to that predicted by the model.
HVP2(Self Profile):
Of all possible comparisons between items, at least 89.54%
were statistically predicted by the model, 7.19% were not
predicted by the model, and 3.27% were in a direction
opposite to that predicted by the model.
In both HVP1 and HVP2, even
in the differences between predicted ranking sums and the
actual ranking sums, the order and distribution of the items
complies with the Hartman model.
Summary Conclusions
Regardless of whether one
studies the profile( both HVP1 and HVP2) as a whole, the
compositional or transpositional sets of items, or each of
the individual items, the obtained ranks strongly support
the validity of Hartman's constructs regarding human value.
The Construct Validity findings both in the internal and
external studies lend great confidence that the profile
instruments measure a person's basic value structure and the
dynamics of their value judgments.
EEOC STUDIES OF THE HARTMAN
VALUE PROFILE
The fundamental principle
underlying Federal Law and Regulation in the United States
pre employment and career assessment practices is as
follows:
"Employer policies
and practices which have an adverse impact on the
employment opportunities of any race, sex or ethnic
group are illegal."(EEOC, 1978)
The guidelines issues by the
Federal Government cover not only tests, but also any
selection procedure that is used as a basis for any
employment decision.
"Employment decisions
include, but are not limited to hiring, promotion,
demotion, membership, referral, retention, licensing and
certification."(EEOC, 1978)
The focus of the EEOC studies
for Value Resource Group is whether or not, (1) the rank
ordering of people using the Hartman Value Profile of
different sexes, from different racial groups and at
different ages is statistically similar or different; (2)
the axiological scores resulting from the Hartman Scoring
System of different sexes, from different racial groups and
at different ages is statistically similar or different;
and, (3) the clinical, interpreted scores of people of
different sexes, from different racial groups and at
different ages, is statistically similar or different.
EEOC studies were performed
internally by Value Resource Group focusing on the rank
ordering of the profile items. EEOC studies were also
performed externally for Value Resource Group by Dr. Chuck
McDonald and sponsored by Dollar General Corporation.
The external studies focused
on: (1) the rank ordering of the items, (2) the axiological
scores resulting from the Hartman Scoring System, and (3)
the clinical, interpreted scores.
Internal EEOC Studies:
These studies were conducted
to determine whether the Hartman Value Profile discriminated
between males and females, individuals of different age
groups and individuals of different races, specifically
"White"-"Black" differences. Two sample populations were
chose for (1) the male-female study, (2) the age study and
(3) the white-black study. These populations consisted of
200 cases in each population chosen at random from a
database of approximately 6,000 cases.
The projects analyzed scores
based on the Hartman value profile and considered the
potential significance of differences in scores in the
sample populations. The statistical test selected was
two-sample parametric interval data t-test. Decision rules
on interpreting the t-test gave a value of 1.282 and above
for a significance of 0.20.
Fifty four data items
including axiological and clinical measures were analyzed to
determine whether any significance could be established
between sample populations. The following results were
determined:
(1) Male-Female
Discrimination:
No score item demonstrated any difference between the male
and female population.
(2) Age Discrimination:
No score item demonstrated any significance between age
ranges.
(3) Black-White
Discrimination:
No score item demonstrated any significance between the
black and white population.
External EEOC Studies:
The Dollar General Company
administered The Hartman profile-Part 1 (The World profile)
and Hartman Profile-Part 2(The Self Profile), as part of a
battery of tests, given to 1,075 persons who were either
employed by or who were seeking employment with the company.
The studies sought to establish whether or not the Hartman
Value Profile-Part 1 and Part 2 met the requirements of the
EEOC law regarding nondiscrimination in selection,
assessment and promotion situations based on (1) age and age
group, (2) sex and (3) race.
The studies were carried out
at three different levels: from the perspective of (1) the
individual rank items, (2) the axiological scores based on
Dr. Hartman's scoring system and (3) the clinical or
interpreted scores.
The obtained scores and
rankings were subject to a MEANS statistical analysis in
order to establish whether or not there was any
statistically significant difference between the scores for
the respective groups.
The one way analysis of
variance of the difference between the within groups and
between groups variances produced the following results:
1. F ratio: That is the ratio
of the within group and between group variances.
2. Level of significance: Assuming that both sample groups
are drawn from a population in which the mean of each group
is equal, with what frequency would we expect to get the
results that were achieved?
3. E ta : Measures the proportion of the total variability
in the dependent variable that can be accounted for by
knowing the values of the independent variable.
Age Discrimination Studies
The EEOC Age Discrimination
Studies were based on a sample size of 1,075 subjects who
were either employed by, or were seeking employment with
Dollar general during 1986-1987. The studies were based on
two groups of the subjects. One Grouping focused on the age
of the individual subjects beginning with "Below 18",
continuing at each age, "31, "41", etc. and ending with
"Over 70". The second grouping was based on the number of
subjects in each age group:
LO thru 29 = 421
30 thru 39 = 298
40 thru 49 = 200
50 thru HI = 156
Total = 1,075
There were three types of
studies, examining the rank items, the axiological scores
and the clinical, interpreted scores, examining the
possibility of discrimination in two ways:
- Is the pattern for
different age groups the same or different?
- If there is any
significant difference in the pattern for different age
groups, could such difference be construed to have a
detrimental impact on the employability or career
development of any particular group.
Based on The Individual
Ranking of The Items:
An analysis of the rankings which 1,075 persons of
different ages and different age groups assigned to the 18
items on the Hartman Value Profile 1 discovered the
following:
(1) For sixteen(16) of the
items by specific age and thirteen (13) of the items by age
group, there was no significant statistical difference
between the mean ranks of the different ages and age groups.
(2) Of the two (2) items by
specific age and five (5) items by age group where there as
a significant difference in the mean assigned to those
items, the Eta clearly indicated that less than 1% to 8% of
the difference was attributable to a persons age or age
group.
These findings support the
contention that the Hartman Value Profile- Part 1 rank
scores do not discriminate unfairly against persons of any
specific or general age group.
Based on The Axiological
Scores:
An analysis of one hundred
fifty one (151) axiological scores based on Dr. Hartman's
scoring system which were produced by the 1,075 persons of
different ages and age groups produced the following
results:
(1) One hundred thirty nine
(139) of the axiological variables by age and one hundred
thirty four (134) of the one hundred fifty one (151)
axiological variables by age group were the same.
(2) Of the twelve (12) by age variables that had different
mean scores, nine (9) of them did not have different means
when re-scored by age group. The remaining three (3), when
analyzed by Eta, were found to have less than 2% of their
variance due to age.
(3) Of the twelve (12)
variables by age group that had different means, less than
2% of the variance was due to age.
These findings lend strong
support to the contention that the 151 axiological variables
of the Hartman Profile - Part 1 do not unfairly discriminate
against persons of different ages or age groups.
The findings also lend
support to the contention that these axiological scores can
be used in hiring and promotion assessments without any
adverse effect on persons of different ages who belong to
different age groups.
Based on The Clinical,
Interpreted Scores:
An analysis of sixteen (16)
clinical, interpreted, scores based on Dr. Hartman's scoring
system which were produced by the 1,075 persons of different
ages and age groups produced the following results:
1) For eleven (11) of the
clinical scores by age and thirteen of the clinical scores
by age group, there was no significant statistical
difference between the mean scores for persons of different
ages and age groups.
(2) Five (5) of the clinical
scores by age and three (3) of the clinical scores by age
group ere significantly different; however, the Eta ratios
indicated that in no case was more than 9% of the variance
attributable to age or age group.
The findings support the
contention that the individual clinical scores, analyzed in
this study, do not discriminate against persons of different
ages or persons of different age groups.
Sex Discrimination Studies
The EEOC Sex Discrimination
Studies were based on a sample size of 1,075 subjects who
were either employed by, or were seeking employment with
Dollar General during 1986-1987.
There were three types of
studies, examining the rank items, the axiological scores
and the clinical, interpreted scores, examining the
possibility of discrimination in two ways:
- Is the pattern for men
and women the same or different.
- If there is any
significant difference in the pattern for men and women,
could such difference be construed to have a detrimental
impact on the employability or career development of any
particular group.
Based on The Individual
Ranking of the Items an analysis of the rankings which 1,075
male and female subjects assigned to the 18 items of the
Hartman Value Profile - Part 1, provided the following
results:
(1) For fourteen (14) of the
eighteen (18) items of the Hartman Profile - Part 1, there
was no significant statistical difference between the mean
ranks of the male and female subjects.
(2) Of the four (4) items
where the mean ranks of the male and female subjects were
significantly different, the Eta clearly indicated that less
than 1% of the difference in both groups was due to sexual
gender.
These findings support the
contention that The Hartman Value Profile - Part 1 rank
scores do not discriminate unfairly against either men or
women.
Based on The Axiological
Scores:
An analysis of one hundred
fifty one (151) axiological scores based on Dr. Hartman's
scoring system which were produced by the 1,075 male and
female subjects produced the following results:
(1) One hundred fifteen (115)
of the one hundred fifty one (151) axiological variables
used in the study for the group of 1,075 men and women had
mean scores that were equal.
(2) Of the thirty six (36)
axiological scores that had mean scores that were
statistically different, the Eta analysis revealed that only
from 1% to 2% of the variance between the two groups was due
to sexual gender.
These findings provide strong
support to the contention that the use of the axiological
scores, used in this study, do not unfairly discriminate
against either men or women. The findings also support the
contention that the axiological scores can be used in hiring
and promotion assessments without any adverse effect on
either men or women.
Based on The Clinical,
Interpreted Scores:
An analysis of sixteen (16)
clinical, interpreted, scores based on Dr. Hartman's scoring
system which were produced by the 1,075 male and female
subjects produced the following results:
For all sixteen (16) of the
clinical scores, used in this study, there was no
significant statistical difference between the mean ranks of
men and women.
These findings strongly
support the contention that the clinical scores, used in
this study, do not discriminate unfairly against either men
or women and that the use of the clinical scores will have
no effect on the hiring or promotion of either men or women.
Race Discrimination
Studies
The EEOC Race Discrimination
Studies were based on a sample size of 1,075 subjects who
were either employed by, or were seeking employment with
Dollar General during 1986-1987. The Group of 1,075 subjects
included the following:
Asian = 6 subjects
Black = 69 subjects
Hispanic = 6 subjects
Indian = 1 subject
Caucasian & Other = 992 subjects
Total = 1075 subjects
Based on The Individual
Ranking of The Items:
An analysis of the rankings which 1,075 persons of different
racial origins assigned to the 18 items of the Hartman Value
profile - Part 1 provided the following results:
(1) For fourteen (14) of the
eighteen (18) items of the Hartman Profile - Part 1, there
was no significant difference between the mean ranks of
persons of different racial origins.
(2) Of the four (4) items
where the mean ranks of persons of different origins were
significantly different, the Eta clearly indicated that less
than 3% of the difference for any variable was due to their
racial origin.
These findings strongly
support the contention that the Hartman Value Profile - Part
1 rank scores do not discriminate unfairly against persons
of different racial origins.
Based on The Axiological
Scores:
An analysis of one hundred
fifty one (151) axiological scores based on Dr. Hartman's
scoring system which were produced by the 1,075 subjects of
different racial origins produced the following results:
(1) One hundred twenty four
(124) of the one hundred fifty one (151) axiological
variables used in this study had mean scores that were
equal.
(2) Of the twenty seven (27)
axiological scores that had mean scores that were
statistically different, the Eta revealed that only from 2%
to 3% of the variance between the two groups was due to
racial origin.
These findings strongly
support the contention that the use of the axiological
scores, used in this study, do not unfairly discriminate
against persons of different racial origins. The findings
also support the contention that these axiological scores
can be used in hiring and promotion assessments without any
adverse effect on persons of different racial origins.
Based on The Clinical,
Interpreted Scores:
An analysis of sixteen (16) clinical, interpreted, scores
based on Dr. Hartman's scoring system which were produced by
groups of different racial origins produced the following
results:
For all sixteen (16) of the
clinical scores used in this study, there were no
significant statistical difference between the mean ranks
for persons of different racial origins.
These findings strongly
support the contention that the clinical scores, used in
this study, do not discriminate unfairly against persons of
different racial origins and these clinical scores will not
have any adverse effect on the hiring or promotion of any
person because of their racial origin.
Summary Conclusions for
The External EEOC Studies
An analysis of 1,075 subjects
broken down by age, sex and racial origin and evaluated from
the perspective of the individual profile rankings, the
axiological scores and the clinical, interpreted, scores
indicate that:
When the mean scores by rank,
axological scores and clinical scores for persons of
different age, race or sex are statistically different, the
differences in mean scores are due to other factors than
sex, age or race.
These findings support the
utilization of the Hartman Value Profile, the Hartman
Scoring System and the Clinical, Interpreted Scores based on
the scoring System for use in making hiring or promotional
decisions without adverse effects due to age, sex or race.
INSTRUMENT RELIABILITY STUDY
OF THE HARTMAN VALUE PROFILE
The instrument reliability
study of the Hartman Value profile (including HVP1 and HVP2)
is an internal study performed by Value Resource Group. The
reliability of an instrument measured the probability that
the results of the assessments generated by the instrument
ar not a result of chance. This study analyzes basic Hartman
Axiological scores as well as Interpreted scores.
Two samples used in this
analysis were selected from an applicant database generated
for Dollar General Corporation over a three year test
project. Two hundred (200) retest situations were used for
the reliability analysis. The 200 profiles used were
selected in no particular order from a larger population of
applicants re-tested at different times.
The first analysis examined
each of the sample files and the score items generated by
computer scoring of the Hartman Value profile.
The instruments were
evaluated according to Hartman's scoring scheme and the
Interpreted Factors based on the axiological combinations.
This analysis consisted of calculating the arithmetic mean
and standard deviation for each of the data items in each of
the data files analyzed.
The second analysis processed
the data files concurrently and produced the Spearman Rank
order Correlation analysis. This analysis is designed to
compare the differences of each item of each subject. To
achieve a significance level of .001, the analysis needed a
Rank order Coefficient of 0.549. The group Rank Order
Coefficient for the test samples as 0.974. The strength of
the coefficient provides an extremely high level of
significance and confidence in the reliability of the
Hartman Value Profile.
CONCURRENT VALIDITY STUDIES
The most significant
statistical evaluation for the purposes of concurrent
validation of the Hartman Value profile centers around the
correlation of the HVP scores with comparable elements from
industry accepted psychological testing instruments. This
correlation evaluates the probability that the measure of
correlation or association obtained was due to chance. Any
probability less than p < .05 points to a statistically
significant association.
To this date, by far the most
significant work correlating the HVP to industry accepted
models is centered in the work of Dr. Leon Pomeroy. Dr.
Pomeroy is now retired from his position of Chief of a
biofeedback and stress management unit at the Veterans
Administration Medical center Outpatient Clinic in Brooklyn.
He holds advanced degrees in the fields of both psychology
and biology from the University of Texas at Austin and
University of Massachusetts at Amherst. Dr Pomeroy is a
respected clinical and researcher in the field which he
discovered, Behavioral Axiology. he has published many
papers on the subject of stress, lectured on the stress
correlations between axiology and psychology, conducted
years of research on the Cross Cultural Correlations of The
Hartman Value Profile and has served as editor or associate
editor for psychological journals. Currently, Dr. Pomeroy is
in private practice, consulting, writing and serving as
President of The Hartman Institute.
Dr. Pomeroy and Dr. John
Davis published findings of their concurrent validity study
of The Hartman Value profile as it relates to the Minnesota
Multiphasic Personality Inventory (MMPI), Cattell 16PF,
Personal Belief Inventory(PBI), Cornell Medical Index (CMI),
and Auto Lethality Index (ALI).
Two sets of data were used in
this study, one group collected in 1981 including a
population of 68 patients, 180 students and 125 doctors. The
second group collected in 1982 had a population of 72. Both
groups were largely male. The MMPI instruments were scored
by a VA computer facility, the IPAT data was scored by IPAT,
the HVP profiles were scored at the University of Tennessee
by Dr. Davis, and the PBI, CMI, and ALI were scored by Dr.
Pomeroy.
All comparisons listed in the
report had a probability of p < .05.
In the correlation of the HVP
with the MMPI, one hundred and twenty eight (128) HVP/MMPI
correlations of less than p < .05 were located. Of these
128, more than seventy five (75) had a correlation of better
than p < .001.
The PBI/HVP correlation
indicated eighteen (18) scores with a correlation of less
than p < .05.
The CMI/HVP correlation
indicated six (6) scores with a correlation of less than p <
.05.
The AL/HVP correlation
indicated four(4) scores with a correlation of less than p <
.05.
These are highly significant
correlations pointing to the high validity for the HVP in
clinical applications. The correlations with the MMPI are
highly significant and indicate that the measure of one's
value system and capacity for making value judgments is a
significant measure which has valid and useful applications.
The data from Dr. Pomeroy's
studies establish a concurrent validity argument for the
Hartman value profile in the measurement of self defeating
behavioral states commonly referred to as psychology and
commonly measured by such instruments as the MMPI and CAQ.
Moreover, the empirical results of the study provide
credibility for the Hartman Value Profile as well as the
science of axiology itself.
Dr. Pomeroy has continued to
collect and analyze data which he has presented as updates
to his research at the annual meetings of the Hartman
Institute. He is also currently writing a book on the
results of his studies.
Additional correlation
studies have been conducted by Dr. John Austin, a long time
member and current Chairman of The Hartman Institute. Dr.
Austin has conducted many studies on the value of the
Hartman Value Profile for education, investigating the use
of the HVP to identify highly gifted students. He conducted
a series of studies on the measure of moral value
correlating variables from the Hartman Value Profile with
variables from the Rokeach value Survey, Kohlberg's Theory
of Moral Development and Allport-Vernon- Lndzey Study of
Values. The correlations from his study of HVP and
Kohlberg's Moral variables and Rokeach variables show a
higher significance (where the Rho or measure of statistical
correlation was largely above .900) than the variables from
the study of specific values from the Allport et. al. study.
These results are to be expected since the Hartman Value
Profile measures one's capacity to make value judgments
rather than the measure of one's specific values.
CRITERION VALIDITY STUDIES
The construct validity
studies establishes that the HVP measures what it is
designed to measure, the capacity to make value judgments.
The test - retest reliability study demonstrates that the
profile results can be counted on to be a reliable indicator
over time of an individual's capacity to make value
judgments. The EEOC studies gives confidence that the HVP
analyses can be used without discriminating by age, race or
sexual gender. The concurrent validity studies lend weight
and credibility to the validity and reliability of the HVP
based analyses as accurate measurements of the ability to
make value judgments. The accumulated weight of these
studies gives confidence that the HVP based analyses are
accurate, valid and reliable diagnostic predictors.
The internal and external
studies, as well as the concurrent studies, confirm that HVP
based analyses are an accurate measure of a person's
specific capacity to make value judgments, such as the
ability to pay attention to practical values and concrete
detail. The information generated can be validly and
accurately utilized to identify value patterns in
individuals, groups and to identify resources for enabling
an individual to learn how to develop an understanding of
their value capacities and chart a development course.
Criterion validity measures
the capacity of the HVP based analyses to be utilized as
valid predictors. In other words, how well do the HVP and
the report instruments function as an indicator of
performance in specific areas such as sales, management and
customer relations. Can the HVP analyses be used to decide
ahead of time which individuals have a better chance of
succeeding in a specific environment?
Value Resource Group has
collected data from a variety of organizations, data within
different geographical areas of organizations, within
specific performance areas and between performers who were
demonstrating either ability or lack of ability. One
consistent factor which VRG has continuously found and
pointed out to clients is that Value Patterns which indicate
"Success" and thus which can be validated as predictors vary
between companies, within companies, and within performance
areas. Unless the diagnostic patterns measured by the HVP
are empirically correlated to those factors which measure
success for a specific performance function in an individual
company and in an individual geographical location for the
company, the information cannot be reliably utilized to
decide ahead of time who will and who will not succeed.
CRITERION STUDY OF THE SALES PERFORMANCE MODEL
One particular statistical,
criterion study conducted for Value Resource Group analyzed
the Sales Performance Model. The study was conducted by Tim
Garton Associates in Lemont, Ill.
The Sales performance Model
was initially based on the findings of a Yale University
study, published in the Journal of Personal Psychology, that
identified and evaluated 5,000 successful sales people over
a five year period. The findings of the study identified the
major factors of sales success as "empathy" and "ego drive".
Value Resource Group expanded this study to develop a
measure of a person's capacity to relate, handle rejection,
be a self starter, have versatility of motivation, be able
to think and see what needs to be done, manage stress and be
able to organize and plan. The Sales model included 49 basic
factors which are based on an integrated analysis of scores
from the Hartman Scoring System.
The criterion study was
conducted as an empirical validation of the sales
interpreted functions of the Sales Performance Model. The
population used for this study consisted of one hundred
(100) non- sales persons, eighty seven (87) moderately
successful sales persons (50,000 to 100,000 in yearly
commissions) and fifty (50) successful sales individuals
(100,000 to 500,000 in yearly commissions). The sales
population was selected primarily from insurance and estate
planning individuals with a proven three year performance
record. The non-sales population was selected at random from
the database of Dollar General job applicants.
Methodology Of The Sales
Criterion Study
The initial scoring process
for The Hartman Value profile takes the ordinal rankings and
converts them into interval scales or scores. The scoring
process weights the Hartman Rank data according to the
Hartman Scoring System. Value Resource Group further weights
the interval data and generates new interval data. This new
interval data is used to determine 49 sales factors which
are used to measure sales strengths and development areas.
This statistical analysis
analyzes the final interval data and variables and considers
the significance of the differences in the values of
distributions between a sample of non-sales persons, mid-
level sales persons and top-sales persons.
With interval scales data in
3 sample populations, an analysis of variance test resulting
in a F ratio was chosen. Decision rules on interpreting the
F ratio were values of 2.00 and above for 0.05 level
significance and 4.00 and above for 0.01 levels of
significance.
For variables not deemed
significant in the ANOVA test, a new test of the two extreme
populations, non-sales and top-sales was constructed. Given
the smaller size of the top-sales population and the
dominance in size of the non-sales population, the Kruskal-
Wallis test was applied.
SUMMARY OF
RESULTS
This study focused on three
sample populations:
1. Non-sales - 100 cases
2. Mid-sales - 87 cases
3. Top-sales - 50 cases
The results of comparative
tests of all three populations indicate that of the 49
variables analyzed:
1. 26 were significant at the
.01 level
2. 14 were significant at the 0.05 level
3. 2 were significant at the 0.10 level
4. 7 were not significant
Subsequent studies of the two
extreme populations, non-sales and top-sales, indicated
significance for two of the seven not deemed significant.
Overall, these results
indicate a very powerful statistical model with a high level
of statistical significance relating to samples drawn on the
key characteristic which the model is intended to predict --
sales ability.
-- End -- |
