home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
Internet Standards
/
CD2.mdf
/
ccitt
/
1992
/
e
/
e427.asc
< prev
next >
Wrap
Text File
|
1993-08-11
|
13KB
|
273 lines
Fascicle II.3 _ Rec. E.427 3
Recommendation E.427
COLLECTION AND STATISTICAL ANALYSIS OF SPECIAL QUALITY OF SERVICE
OBSERVATION DATA
FOR MEASUREMENTS OF CUSTOMER DIFFICULTIES IN THE
INTERNATIONAL AUTOMATIC SERVICE
This Recommendation is provided to permit the orderly
collection of data required for special studies to identify sources
of difficulty in customer use of the international automatic
telephone service.
When calls are made to points outside a customer's home
country, many different sets of ringing and busy tones are
encountered. In order to measure the effect of unusual sounding
ringing tones and busy tones on customer behaviour, it has been
decided to collect data on how long customers listen to such
foreign tones as well as to their national tones in order to
compare them.
The data are to be collected in the same manner as those
required for the completion of Table 1/E.422. These data are an
extension of those collected for Table 1/E.422, and, as an aid to
subsequent analysis, a copy of the current version of that table
should be used with the table of this Recommendation.
Table 1/E.427 contains questions numbered 1_9. Their
relationship to the questions of Table 1/E.422 is shown in
parentheses.
A preferred set of analyses for identifying the statistical
significance of differences between data collected from subscribers
when setting up national calls and the corresponding data collected
from subscribers when setting up international calls is given
below.
1 Determine the percentage change in any measure by use of the
formula:
Change (Ci) = x 100 j = A, B, C
i = 0_2, 2_5 . . ., > 30
where
fij is the observed frequency of calls of category i in the
country j,
Njis the total number of observations in the country j
sample,
fiH is the observed frequency of calls of category i in the
home country H, and
NHis the total number of observations in the home country
sample.
2 Compare the central location of the distributions by use of
the Kruskal_Wallis One_Way Analysis of Variance [1].
3 Compare the "forms" or "shapes" of the distribution by means
of the chi_square test [2].
4 Compare changes in single valued variables, e.g. percentage
incomplete_trunk_code, by use the chi_square test.
TABLEAU 1/E.427
(Supplement to Table 1/E.427)
Observations of international outgoing telephone calls for quality
of service
Additionaldetails regarding subscriber dialled calls
Outgoing international exchange
Group of circuit
Period from to
Number Percentage
Category Subt Tota Subt Tota
otal l otal l
Details of dialled calls a)b)c)
1. Calls with errors in the dialled
number d)
1.1 (6.1) Wrong number dialled . . 100
.
1.1.1 Wrong country code . . . .
. .
1.1.2 National trunk prefix (e.g. . . . .
"0") wrongly included . .
1.1.3 Wrong trunk code . . . .
. .
1.1.4 Wrong subscriber number . . . .
. .
1.2 (6.2) Incomplete number dialled . . 100
.
1.2.1 National (significant) number . . . .
not dialled or incomplete . .
1.2.2 Trunk code not dialled or . . . .
incomplete . .
1.2.3 Subscriber number not dialled . . . .
or incomplete . .
2. (5.3) Calls abandonned prematurely . . 100
before receipt of a tone or .
(6.3) announcement
Interval from end of dialling to
disconnecte):
0-5 s . . . .
. .
5-10 s . . . .
. .
10-20 s . . . .
. .
20-30 s . . . .
. .
30-50 s . . . .
. .
> 50 s . . . .
. .
3. Post dialling delay on all calls that . . 100
are maintened beyond the start of a tone .
or announcement
Interval from end of dialling to
tone or announcement:
0-5 s . . . .
. .
5-10 s . . . .
. .
10-20 s . . . .
. .
20-30 s . . . .
. .
30-60 s . . . .
. .
60-90 s . . . .
. .
> 90 s . . . .
.
Average excluded portionf)
4. Calls that encounter ringing tonesg)
4.1 (1) Completed calls . . 100
Interval from beginning of tone to
answer:
0-10 s . . . .
. .
10-20 s . . . .
. .
20-30 s . . . .
. .
30-50 s . . . .
. .
> 50 s . . . .
. .
4.2 (2.6.4) Incompleted calls . . 100
Interval from beginning of tone to
disconnect:
0-10 s . . . .
. .
10-20 s . . . .
. .
20-30 s . . . .
. .
30-50 s . . . .
. .
> 50 s . . . .
. .
Number Percentage
Category Subt Tota Subt Tota
otal l otal l
5. (3-2) . . 100
Calls that encounter busy/congestion .
tonesg)
Interval from beginning of tone to
disconnect:
0-2 s . . . .
. .
2-5 s . . . .
. .
5-20 s . . . .
. .
20-30 s . . . .
. .
> 30 s . . . .
.
6. (4.2) . . 100
Calls that encounter tones that the .
observer cannot identify
Interval from beginning of tone to
disconnect:
0-2 s . . . .
. .
2-5 s . . . .
. .
5-10 s . . . .
. .
10-30 s . . . .
. .
> 30 s . . . .
.
7. (3.3, 4.3)Calls encountering recorded . . 100
announcements .
Interval from beginning of
announcement to disconnect:
0-2 s . . . .
. .
2-5 s . . . .
. .
5-10 s . . . .
. .
10-30 s . . . .
. .
> 30 s . . . .
.
8. List types of errors in dialling and tone interpretation
which could not be categorized
p
9. List restrictions on subscriber sampleh)
a) The term "calls" throughout this table refers to circuit
seizures by outgoing traffic.
b) The data for each called country should be collected separately
and not combined with other countries.
c) The interpretation of these results cannot be made adequately
except by comparing them with similar results on national
calls.
d) The practicability of putting the observation in category 1 will
depend upon the observation access point and knowledge of
national numbering plan of the outgoing country and of the
destination country.
e) 0-5 s simplies 0 £ t £ 5.
5-10s simplies 5 < t < 10.
f) The "post-dialling delay"measurements may not represent the
actuel delay from the time the subscriber finishes dialling to
receipt of tone. To the extent that this measurement as
observed on the trunk excludes the time from completing of
dialling seizure of trunk, the average duration of this
excluded time should be reported.
g) Identification of tone categories should be made by service
observers who are trained to identify the tome categories
reliably.
h) If access to the trunks being observed is restricted to some
specified population of subscribers, e.g., heavy users, non-
coin users residents of large urban centres, such restrictions
should be noted and reported with the service observations.
References
[1] MARASCUILO (L. A.), McSWEENEY (M.): Non_Parametric and
Distribution_Free Methods for the Social Sciences, Wadsworth
Publishing Co., California, 1977.
[2] SIEGEL (S.): Non_Parametric Statistics for the Behavioural
Sciences, McGraw Hill, New York, 1956.