Reply

• Author: Nelson C. Dometrius
• Published date: 01 June 1987
• Date: 01 June 1987
• DOI: http://doi.org/10.1177/106591298704000210
• Subject Matter: Articles

REPLY

NELSON

C.

DOMETRIUS

Texas

Tech

University

S SHOULD

be

obvious

from

the

citations

in

Mueller’s

original

arti-

cle

and

my

critique,

we

both

respect

and

repeatedly

utilize

each

X C

other’s

work.

On

the

contemporary

utility

of

Schlesinger’s

index,

however,

we

continue

to

disagree.

Mueller

has

included

a

substantial

amount

of

new

material

in

his

rejoinder,

material

which

deserves

examination.

1.

Appointment

Power.

Mueller

bypasses

my

argument

that

using

the

same,

or

similar,

16

agencies

across

time

is

an

inconsistent

measuring

in-

strument

due

to

changes

in

the

number

and

responsibilities

of

state

agen-

cies.

I

believe

this

to

be

a

real

measurement

problem

but

will

focus

my

comments

where

Mueller

does,

on

the

relative

versus

absolute

character

of

Schlesinger’s

appointment

power

measure.

Mueller’s

original

theoretical

question

was

one

of

absolute

change.

Using

a

relative

indicator

here,

particularly

one

divided

into

quintiles,

is

simply

inappropriate.

Observe

that

most

of the

changes

Mueller

discusses

in

his

rejoinder

are

based

on

the

raw

appointment

scores

or

the

1-50

ranking

of

the

states

drawn

from

the

raw

scores,

but

Mueller’s

original

analysis

was

based

on

Schlesinger’s

1-5

quintile

scores.

The

question

remains;

do

the

quintile

scores

reflect

absolute

changes?

I

think

not.

Mueller

tells

us

that

all

50

states

experienced

1960-82

changes

in

their

raw

appointment

scores,

but

note

how

Schlesinger’s

component

fails

to

re-

flect

this

-

the

scores

for

19

states

(38

percent

of

the

N)

remained

exactly

the

same

in

both

1960

and

1982,

ignoring

the

real

movement

that

took

place.

Another

18

states

moved

up

or

down

one

point

on

Schlesinger’s

5-point

scale;

equivalent

to

a

20

percent

increase

or

decrease

in

power.

Did

these

states

really

move

that

much,

or

did

a

slight

movement

become

mag-

nified

as

a

state

was

nudged

from

the

bottom

of

one

quintile

to

the

top

of

another?

For

states

showing

dramatic

movement,

the

scoring

system

may

be

more

reasonable,

but

it

still

contains

substantial

measurement

error.

Consider

Mueller’s

own

example:

Colorado

was

given

a

40

percent

increase

(from

1

to

3)

by

using

relative

quintiles,

when

its

actual

increase

was

only

19

percent

on

the

absolute

raw

score

(46

percent

versus

27

percent).

Why

unnecessarily

introduce

measurement

error

when

the

absolute

raw

appoint-

ment

scores

are

available?

2.Index

variation.

Considerable

time

is

spent

on

the

appointment

power

component

because

it

is

the

prime

source

of

variation

in

the

overall

index

(variation

contributed

by

the

other

three

components

is

virtually

nil)

and

is

thus

the

prime

determinant

of

the

statistical

results.

Mueller

seems

to

argue

that

if

states

have

added

new

formal

powers

to

their

governorships

this

will

merely

shift

the

mean

to

the

right

while

leaving

the

states

lumped

together

at

the

top.

Not

only

does

this

argument

make

the

unreasonable

assumption

that

the

amount

of

unmeasured

formal

power

is

constant

for

all

states,

it

is

inconsistent

with

the

one

attempt

to

incorporate

additional