Which Factors in Verbal Communication Make Information-senders Irritated?

(1)

The Japanese Psychonomic Society

NII-Electronic Library Service

The JapanesePsychonomicSociety

The

_ILiPanese

_]burnal

_of

Rsychonetnic

Science

19se,

Vol,

9,

No.

1,51-53

WhichFactors

in

Verbal

Communication

Information-senders

Irritated?

Make

Machike

SANNOMIYA

IVbruto

U)tiversity

of

Edacation

The

_present

study examined three

facters

ln

communication which may make us

irritated

when we send verbal

information.

Those

are sending unit oi

information,

frequency

of

_the

receiver's misrecegnition of

information,

and waiting time

before

getting tne receiver's

response,

These

factors

were manipulated

in

a simulated voice recognition system.

Main

results were as

follows:

1)

The

above three

factors

influence

the

feeling

of

irritatlon

during

information-sending

tasks.

2)

Information-sending

tasks even with

the

smallest unit such as monosyllable

do

not

cause the

feeling

of

irritation

when misrecognition rarely occurs

(5%-

or

O%-frequency

per

unit) and waiting

time

is

very short

<O.025sec

per unit).

3)

Frequency

of misrecognition and waiting time can not compensate each other

to

reduce

the

feeling

of

irritation

during

information-sending.

Further

stuclies are necessary

in

order

to

see

if

there

is

a common mechanism

to

cause

irritation

of

inforrnation-senders

whether

their

receiver

is

a real

person

or not.

Key

Words:

irritation,

information-sending,

communication, sending unit, misrecognition,

waiting

time

We

often

feel

irritated

when

we can not

send

verbal

information

smoothly.

Hitherto

the

_problem

of emotion such as

irritation

during

information-sending

has

scarcely

been

discussed

in

the

fremework

of cognitive

psychology.

However,

as

Norman

(1980)

argued,

it

is

one of

the

important

issues

for

the

study of

human

cognition

_to

know

what

factors

make

information-senders

irritated

and

thereby

_prevent

smooth sending of

information.

Following

studies

on

man-machine cornmunication

_give

us some

sugges-tions

on

thls

_point.

Takagi

(1986)

reported

that

the

users

of

a

voice-activated

word-processor

had

negative

impressions

and

felt

tired

during

the

information-sending

task

when

the

accuracy of voice recognition was

poor.

Sannomiya

(1986)

examined

_the

_voice

recognition method

in

CAI

and

found

that

recognition accuracy and unit size of sending

information

had

some

psychological

effects.

She

also

iound

that

the

_{psychological}

load

in

sending

information

was

smaller

with

vocalization method

than

with manual

type-writing method when voice recognition was

performed

with

high

accuracy

(Sannomiya,

1987).

These

_findings

suggest

that

recogni-tion

accuracy of

information-receivers

and sending unit

information

infiuence

the

ieel-ing

of

iniormation-senders.

Also

the

speed

of

responses

of

information

receivers

seems

to

have

irritating

effect

when

senders

need

responses

from

thelr

receivers.

The

present

study examined

the

effect of

the

above

three

factors

about

the

feeling

of

irritation

on

the

_part

of

information-senders

in

order

to

test

the

follewing

hypotheses.

1)

Sending

unit:

Senders

are more

irri-tated

when

they

have

to

send

information

in

smaller unit.

2)

Frequency

of receivers' misrecognition

per

unit:

Senders

are rnore

irritated

when

the

receivers misrecognize

the

information

more

frequently

and

they

have

to

correct

it.

3)

Waiting

time

per

unit:

Senders

are

more

irritated

when

they

have

to

wait

for

a

longer

tirne

until

the

receivers recognize

the

information

and

_give

thern

a response.

Three

factors

were manipulated

in

a

(2)

The JapanesePsychonomic Society

52 The

_Japanese

_Journal

ofPsychonomic

Science

(1)

monosyllable

and

Bunsetsu

_as

_sending

unit.

(2)

20%,

5%,

and

O%

as

frequency

of

-

-v

-misrecognition

_per

_unit.

(3)

_1sec

and

O.025

sec as waiting

tlme.

Monosyllable

is

the

smallest

unit

when

sending

Japanese

_sentences,

and

Bunsetsu

refers

to

the

smallest

_group

of words

under-standable

in

a

Japanese

sentence.

20%

of

rnisrecognition

is

_quite

usual

in

the

current

ability of voice recognition

_systems,

_5%

is

the

best

record of real

scores,

and

O%

is

an

ideal

state.

1

sec waiting

is

the

case of slow responseand

O.025sec

waiting

is

_the

case of

quite

rapid and almost real

time

response.

Method

Subjects

Twenty-four

Japanese

graduate

students

served as subjects.

They

consisted

of

14 males

and

10 females.

The

age was about

28.5 _years

eld

in

average.

Design

A2

×

3

×

2 factorial

design

was used.

The

first

factor

was

sending

unit

(monosyllable

'and

_Bunsetsu).

_The

_{second,tactor}

was

fre-quency

of

rnisrecognition

_per

unit

(20%,

5%,

and

O%).

_,

The

third

factor

was waiting

time

per

unit

(1

sec and

O.025 _sec>.

Only

_the

_first

factor

was

between

subjects.

Material,

Six

Japanese

texts

were used as

information

te

be

sent,

They

were

_printed

on

_papers.

Each

text

consisted

of

50 Bunsetsu

and

246.5

syllables

in

average.

Misrecognition

in

5%

and

20%

conditions

was manipulated

to

occur randomly according

to

_{pseudo-randorn}

numbers.

Misrecognized

letters

were

_prepared

by

replacing correct

Kana

letters

with

those

which were

Table

1. An

Example

6f

Patterns

of

2o%

Vol.

9,

No.

1 tically

similar

and

were often misheard

by

both

human

hearers

and

voice-recognition

systems.

Table1

shows

an

example

of

20%

mis-recognition

in

the

monosyllable

condition.

Equipment

A

personal

computer

(PC-9801:NEC)

was

used as a simulated receiver of

information

-

.

(a

voice

recognition

system)

and a microphone

(SM12A:SURE)

was used as an

information-mstrument.

Procedure

Subjects

were

instructed

to

send

the

_printed

six

texts

to

the

simulated receiver

(a

voice

recognition system)

_through

_a

_microphone

by

reading

out

them

in

each

information

unit.

They

were

told

that

the

voice

tion

system

could

change

human

yoi¢

e

into

Japanese

Kana

letters.

Then

the

way

of

sending

information

was explained

in

detail.

The

information-sending

unit

was

syllable or

Bunsetsu.

In

the

monosyllab}e

condition, a syllable appeared on

the

TV

display

when

subjects

uttered

the

syllable

and

_pressed

an appointed

key

(sending

key),

In

the

Bttnsetsu

condition, a

Bunsetsu

peared

on

the

display

when

they

uttered

the

Bunsetsu

and

_pressed

a sending

key.

In

both

conditions,

they

have

to

wait

lsec

(lsec

condition)

or

O.025sec

(O.025sec

condition)

to

get

a response

for

each unit

on

the

display,

that

is,

the

correctly

or

incorrectly

recognized

infermation

which

they

have

sent.

If

the

response

was

correct,

they

could

_go

to

the

unit.

When

it

was wrong,

they

had

to

repeat

the

same unit and

to

_press

a

key

for

correction.

An

accurate

response

was always

_given

once corrected.

After

each

of six

texts

₍₆

comblnations of

3 _types

_of

misrecognition

frequency

and

2 types

of

wait-Misrecognition

in

the

Monosyllable

Condition.

One

of the Qriginal sentence

to

be

vocalized

Misrecognized

responses

for

the

left

sentence vocalizecl

l:･

E

ig

vetc

omegeab:L

h o

k

v(S

iE)

Ru,tx

6 n

tsi

Ui

OVX,

.

'ali;k

1]h:

i

b}AJ

ix

f#epD-titL

N

-(F

as

ip

6nr,

fik

O

maVik

D

t4

e

8L

IM

e

bl)(DKvah:g

･L,o'(T,

_{tavipmtwtiDfofo.to"(!ettkL,rti65hts.}

a J5

_ig

_t)

_tiL

sDOvhet ,L

hl

vst)

ts

'(i

is

Mtti

6

℃

tsLNObgts<Oihth:"D"fu6tLtf6

oeLh-etstoshgJltkogvkeoigltot}t

t)in<otrLw,Lhststuss-cfDxLiolfgs

(3)

The JapanesePsychonomic Society

M.Sannomiya:

Whichfactors

in

ing

time),

subjects

rated

on

a

5-point

scale

their

feeling

of

irritation

which

they

felt

during

the

information-sending

task.

The

order

of

six

texts

and

the

combination of

misrecognition

frequency

and waiting

time

were counterbalanced among subjects.

Resultg

Table

2

shows

the

mean score of rating

of

irritation

in

each

condition.

Table

2. Mean

Score

of

Rated

Irritation,

Waiting

time verbal communicatien?

Sending

unit

Frequency

ofmlsreco-gnition

1

sec

O,025

sec

MSDMSD

Monosyllable

Bunsetsu

20%5%o%20%5%o%

g'r6L7''-o.6i'-217s

1.ol

2.58

1.26

1.92

O.95 2,67

1,18

1.33

O.85

2.00 1,29

1.58

O.95 I,58

O.76

1.58

O.64

1.75 e.83

1,17

O.30 Note,

The

range of score

is

from

1(never)

to5

<very

much).

A

2

×

3x2

(sending

unit ×

frequency

of mis-recognitionxwaiting

time)

analysis ef

vari-ance

yielded

significant

main

effects

of

send-ing

unit

(F(1,

22)=:9.74,

P<O.Ol),

frequency

of

misrecognition

(F(2,44)==14.56,

P<O.Ol),

and waiting

time

(F(1,

22)=26.67,

P<O.Ol).

And

the

interactions

between

unit

and

fte-quency

of

misrecognition

(F(2,

44)=5.04,

P<

O.05)

and

between

_.unit

and waiting

time

(F(1,22)=6.08,

P<O.05)

were significant.

Subtests

showed

that

the

difference

between

monosyllable

_and

Bunsetsu

was significant

when

frequency

of

misrecognition was

20%

(t(22)==3.85,

P<O.Ol)

or when waiting

time

was

1

sec

(t(22)z3.47,

_P<O.Ol),

but

not

signi-ficant

when accuracy was

100%

and

response

time

was

O.025

sec

(t(22)==O.57)

or when

ac-curacy was

95%

and response

_time

was

O.025

sec

(t(22)=:]O.98).

Diseussion

The

main

fipdings

in

the

_present

study

were

as

follows:

1)

Three

factors

in

communication-send-53

ing

unit of

information,

frequency

of

the

receivers' misrecognition, and waiting

time

before

the

receivers' responses-respectively

have

a significant effect

to

make

information-senders

irritated.

2)

Even

if

sending unit

is

very small

(monosyllable),

information-senders

do

not

feel

more

irritated

than

in

the

case of

larger

unit

(Bunsetsu)

when

the

frequency

of

mis-recognition

is

very

low

(5%

or

O%

_per

unit) and waiting

time

is

very short

(O.025

sec

_per

unit).

3)

Those

who send

information

in

very

small unit

(monosyllable)

feel

irritated

when misrecognition occurs

frequently

(20%

per

unit) even

if

they

can

_get

a response almost

in

real

time

(O.025sec

per

unit).

And

they

feel

irritated

when

they

have

to

wait

lsec

per

unit

to

_get

a response

even

if

recognition

is

_perfect.

In

this

sense

the

two

factors-frequency

of misrecognition

and

waiting

time-can

not compensate each other

to

re-duce

senders'

irritation.

The

communication

which was examined

in

this

study was

between

a

human

inform-ation-sender and a simulated receiver whose recognition was realized on a