BRITISH STANDARD BS 2846-2:

1981
ISO 2602-1980
CONFIRMED

Guide to

Statistical
Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI

interpretation of
data —
Part 2: Estimation of the mean:
confidence interval —

[ISO title: Statistical interpretation of

test results—Estimation of the
mean—Confidence interval]

UDC 31:519.222:312
 BS 2846-2:1981

Cooperating organizations

The Quality Management and Statistics Standards Committee, under whose

direction this British Standard was prepared, consists of representatives from
the following:

Confederation of British Industry

Consumers’ Association
Department of the Environment (Building Research Establishment)
Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI

Institute of Cost and Management Accountants

Institute of Quality Assurance*
Institute of Statisticians*
Institution of Electrical Engineers
Institution of Production Engineers*
Ministry of Defence*
National Council for Quality and Reliability
National Terotechnology Centre
Retail Consortium

The organizations marked with an asterisk in the above list, together with the
following, were directly represented on the Technical Committee entrusted
with the preparation of this British Standard:

Association of Public Analysts

Cement and Concrete Association
Chemical Industries Association
Department of Industry (Laboratory of the Government Chemist)
Department of Industry (National Engineering Laboratory)
Economist Intelligence Unit Limited
Institute of Petroleum
Ministry of Agriculture, Fisheries and Food
National Coal Board
Post Office
Royal Statistical Society

This British Standard, having

been prepared under the
direction of the Quality
Management and Statistics
Standards Committee, was
published under the authority of
the Executive Board and comes
into effect on
30 January 1981

© BSI 02-1999 Amendments issued since publication

First published June 1975
First revision January 1981 Amd. No. Date of issue Comments

The following BSI references

relate to the work on this
standard:
Committee reference QMS/1
Draft for comment 80/63802 DC

ISBN 0 580 11810 X

 BS 2846-2:1981

Contents

Page
Cooperating organizations Inside front cover
National foreword ii
0 Introduction 1
1 Scope 1
2 Field of application 1
Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI

3 References 2
4 Definitions and symbols 2
5 Estimation of the mean 2
6 Confidence interval for the mean 2
7 Presention of the results 3
Annex Confidence interval for the mean from the range 5
Table 1 — Values of t1 – a and of the ratio t 1 – a ⁄ n 4
Table 2 — Untitled 4
Table 3 — Untitled 5
Publications referred to Inside back cover

© BSI 02-1999 i
 BS 2846-2:1981

National foreword

This Part of this British Standard is identical with ISO 2602 “Statistical
interpretation of test results—Estimation of the mean—Confidence interval”
prepared by Subcommittee 2 of Technical Committee 69, Applications of
statistical methods, and published in 1980 by the International Organization for
Standardization (ISO).
The correct interpretation and presentation of test results has been assuming
increasing importance in the analysis of data obtained from manufacturing
Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI

processes based on sample determinations and prototype evaluations in industry,

commerce and educational institutions. It was for this reason that a series of
guides was prepared.
The standard consists of the following Parts:
— Part 1: Routine analysis of quantitative data;
— Part 2: Estimation of the mean: confidence interval — (Identical with
ISO 2602);
— Part 3: Determination of a statistical tolerance interval — (Identical with
ISO 3207);
— Part 4: Problems of estimation and tests relating to means and variances —
(Identical with ISO 2854);
— Part 5: Efficiency of tests relating to means and variances — (Identical with
ISO 3494);
— Part 6: Comparison of two means in the case of paired observations —
(Identical with ISO 3301).
Of all the ways of representing a group of observations usually the most
satisfactory single measure is the mean; it indicates their central position and as
such has extensive applications in the assessment of statistical data. This Part of
this British Standard specifies techniques for the estimation of the mean of a
normal population on the basis of a series of tests applied to a random sample of
individuals drawn from this population, but deals only with the case where the
variance of the population is unknown.
This standard supersedes BS 2846-2:1975, which is now withdrawn.
Terminology and conventions. The text of the international standard has
been approved as suitable for publication, without deviation, as a British
Standard. Some terminology and certain conventions are not identical with those
used in British Standards; attention is especially drawn to the following.
The comma has been used throughout as a decimal marker. In British Standards
it is current practice to use a full point on the baseline as the decimal marker.
Wherever the words “International Standard” appear, referring to this standard,
they should be read as “British Standard”.
Cross-references
International standard Corresponding British Standard
ISO 2854-1976 BS 2846 Guide to statistical interpretation of data
Part 4:1976 Techniques of estimation and tests
relating to means and variances
(Identical)
ISO 3534-1977 BS 5532:1978 Statistics—Vocabulary and symbols
(Identical)
NOTE Textual errors. When adopting the text of the international standard the printing errors
listed below were noticed. They have been marked in this British Standard and have been reported to
ISO in a proposal to amend the text of the international standard:
a) in 6.2.2, item b), line 8, “ν = n + 1” should be read as “ν = n – 1”;
b) in Table 1, column 10, line 14 “0,668” should be read as “0,678”;
c) in Table 1, column 10, line 28 “0,658” should be read as “0,458”.

ii © BSI 02-1999
 BS 2846-2:1981

In order to clarify the meaning of the standard the following changes to the existing text of
ISO 2602-1980 have been recommended for amendment by the Quality Management and Statistics
Standards Committee:
a) in clause 2, paragraph 3 to be read as:
“The normality assumption is very widely satisfied: the distribution of the results obtained
under test conditions is frequently a normal or near normal distribution.”;
b) in 6.1.2, the final paragraph to be read as:
Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI

“In the case of grouped results, the calculated value of s should be corrected (“Shephard’s
correction”). As this correction is of secondary importance, it has not been given here.”.
A British Standard does not purport to include all the necessary provisions of a
contract. Users of British Standards are responsible for their correct application.
Compliance with a British Standard does not of itself confer immunity
from legal obligations.

Summary of pages
This document comprises a front cover, an inside front cover, pages i to iv,
pages 1 to 6, an inside back cover and a back cover.
This standard has been updated (see copyright date) and may have had
amendments incorporated. This will be indicated in the amendment table on
the inside front cover.

© BSI 02-1999 iii

 Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI

iv
blank

0 Introduction
The scope of this International Standard is limited
to a special question. It concerns only the estimation
of the mean of a normal population on the basis of a
series of tests applied to a random sample of
individuals drawn from this population, and deals
Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI
only with the case where the variance of the
population is unknown. It is not concerned with the
calculation of an interval containing, with a fixed
probability, at least a given fraction of the
population (statistical tolerance limits).
It is recalled that ISO 2854 relates to the following
collection of problems (including the problem
treated in this International Standard):
— estimation of a mean and of the difference
between two means (the variances being either
known or unknown);
— comparison of a mean with a given value and
of two means with one another (the variances
being either known or unknown, but equal);
— estimation of a variance and of the ratio of two
variances;
— comparison of a variance with a given value
and of two variances with one another.
The test methods generally provide for several
determinations which are carried out:
— on the same item (where the test is not
destructive);
— on distinct portions of a very homogeneous
product (a liquid, for example);
— on distinct items sampled from an aggregate
with a certain amount of variability.
In the first two cases, the deviations between the
results obtained depend only upon the repeatability
of the method. In the third case, they depend also on
the variability of the product itself.
The statistical treatment of the results allows the
calculation of an interval which contains, with a
given probability, the mean of the population of
results that would be obtained from a very large
number of determinations, carried out under the
same conditions. In the case of items with a
variability, this International Standard assumes
that the individuals on which the determinations
are carried out constitute a random sample from the
original population and may be considered as
independent.
1)
This subject is in preparation.
© BSI 02-1999
BS 2846-2:1981
The interval so calculated is called the confidence
interval for the mean. Associated with it is a
confidence level (sometimes termed a confidence
coefficient), which is the probability, usually
expressed as a percentage, that the interval does
contain the mean of the population. Only the 95 %
and 99 % levels are provided for in this
International Standard.
1 Scope
This International Standard specifies the statistical
treatment of test results needed to calculate a
confidence interval for the mean of a population.
2 Field of application
The test results are expressed by measurements of
a continuous character. This International
Standard does not cover tests of a qualitative
character (for example presence or absence of a
property, number of defectives, etc.).
The probability distribution taken as a
mathematical model for the total population is a
normal distribution for which parameters, mean m
and standard deviation s, are unknown.
The normality assumption is very widely satisfied:
the distribution of the results obtained under test
conditions is generally a normal or nearly normal
distribution.
It may, however, be useful to check the validity of
the assumption of normality by means of
appropriate methods1).
The calculations may be simplified by a change of
the origin or the unit of the test results but it is
dangerous to round off these results.
It is not permissible to discard any observations or
to apply any corrections to apparently doubtful
observations without a justification based on
experimental, technical or other evident grounds
which should be clearly stated.
The test method may be subject to systematic
errors, the determination of which is not taken into
consideration here. It should be noted, however,
that the existence of such errors may invalidate the
methods which follow. In particular, if there is an
unsuspected bias the increase of the sample size n
has no influence on bias. The methods that are
treated in ISO 2854 may be useful in certain cases
for identifying systematic errors.
1
 BS 2846-2:1981
3 References
ISO 2854, Statistical treatment of data — Problems
of estimation and tests of means and variances.
ISO 3534, Statistics — Vocabulary and symbols.
4 Definitions and symbols
Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI
The vocabulary and symbols used in this
International Standard are in conformity with
ISO 3534.
5 Estimation of the mean
5.1 Case of ungrouped results
After the discarding of any doubtful results, the
series comprises n measurements xi
(where i = 1, 2, 3, ..., n), some of which may have the
same value.
The mean m of the underlying normal
distribution is estimated by the arithmetic mean
x of the n results:
5.2 Case of results grouped in classes
When the number of results is sufficiently high
(above 50 for example), it may be advantageous to
group them into classes of the same width. In
certain cases, the results may also have been
directly obtained grouped into classes.
The frequency of the ith class, i.e. the number of
results in class i, is denoted by ni.
The number of classes being denoted by k, we
have:
The midpoint of class i is designated by yi. The
mean m is then estimated by the weighted mean
of all midpoints of classes:
6 Confidence interval for the mean
The confidence interval for the population mean is
calculated from the estimates of the mean and of the
standard deviation.
The alternative method of calculating the
confidence interval by use of the range is given in
the annex.
2 © BSI 02-1999
6.1 Estimation of the standard deviation
6.1.1 Case of ungrouped results
The estimate of the standard deviation σ,
calculated from the squares of the deviations
from the arithmetic mean, is given by the
formula:
where
xi is the value of the ith measurement
(i = 1, 2, 3, ..., n);
n is the total number of measurements;
x is the arithmetic mean of the n
measurements, calculated as in clause 5.1.
For ease of calculation, the use of the following
formula is recommended:
6.1.2 Case of grouped results
In the case of grouping by classes, the formula for
the estimate of the standard deviation is written:
For ease of calculation, the use of the following
formula is recommended:
where
yi is the mid-point of the ith class
(i = 1, 2, 3, ..., k);
k is the number of classes;
n is the total number of measurements;
y is the weighted mean of all mid-points of
classes calculated as in sub-clause 5.2.
In the case of grouped results, the calculated value
of s should be corrected (“Sheppard’s correction”). As
this correction is of secondary importance, it has not
been mentioned here.
 BS 2846-2:1981

6.2 Confidence interval for the mean
For a chosen confidence level (95 % or 99 %),
according to the specific case, a two-sided or a
one-sided confidence interval has to be determined.
6.2.1 Two-sided confidence interval
The two-sided confidence interval for the population
Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI
The values t0,975, t0,995, t0,95, t0,99 are those of
Student’s t distribution with ν = n + 1 degrees of
freedom.2)
These values are given in Table 1.
This table gives also the values of ratios

mean is defined by the following double inequality:

a) at the confidence level 95 %:
When values of n are greater than 60, it is
preferable to calculate the value of t by linear
interpolation from 120
---------- using Table 2.
Example: n
b) at the confidence level 99 %
n = 250
120
---------- = 0,48
n
6.2.2 One-sided confidence interval
t0,995 = 2,576 + 0,48 + (2,617 – 2,576)
The one-sided confidence interval for the population
mean is defined by one or other of the following = 2,596
inequalities:
a) at the confidence level 95 %: 7 Presentation of the results
7.1 Give the expression of the mean according to 5.1
or 5.2.
7.2 Express the confidence interval in the form of
or the double inequality of 6.2.1 or one of the
inequalities of 6.2.2, stating the confidence
level (95 % or 99 %). Indicate the number of results
discarded as being doubtful and the reasons for
discarding.
b) at the confidence level 99 %:

or

with x , if necessary, replaced by y , in the case of

results grouped in classes.

2)
See national foreword for details of printing errors in this clause.

© BSI 02-1999 3
 BS 2846-2:1981

Table 1 — Values of t1 – α and of the ratio t 1 – a / n

Confidence level Confidence level Confidence level Confidence level

Two-sided case One-sided case Two-sided case One-sided case

95 % 99 % 95 % 99 % 95 % 99 % 95 % 99 %

t 0,975 t 0,995 t 0,95 t 0, 99

n t0,975 t0,995 t0,95 t0,99 n --------------- --------------- ------------ ------------
Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI

n n n n

2 12,71 63,66 6,314 31,82 2 8,985 45,013 4,465 22,501

3 4,303 9,925 2,920 6,965 3 2,484 5,730 1,686 4,021
4 3,182 5,841 2,353 4,541 4 1,591 2,920 1,177 2,270
5 2,776 4,604 2,132 3,747 5 1,242 2,059 0,953 1,676
6 2,571 4,032 2,015 3,365 6 1,049 1,646 0,823 1,374
7 2,447 3,707 1,943 3,143 7 0,925 1,401 0,734 1,188
8 2,365 3,499 1,895 2,998 8 0,836 1,237 0,670 1,060
9 2,306 3,355 1,860 2,896 9 0,769 1,118 0,620 0,966
10 2,262 3,250 1,833 2,821 10 0,715 1,028 0,580 0,892
11 2,228 3,169 1,812 2,764 11 0,672 0,956 0,546 0,833
12 2,201 3,106 1,796 2,718 12 0,635 0,897 0,518 0,785
13 2,179 3,055 1,782 2,681 13 0,604 0,847 0,494 0,744
14 2,160 3,012 1,771 2,650 14 0,577 0,805 0,473 0,708
15 2,145 2,977 1,761 2,624 15 0,554 0,769 0.455 0,668
16 2,131 2,947 1,753 2,602 16 0,533 0,737 0,438 0,651
17 2,120 2,921 1,746 2,583 17 0,514 0,708 0,423 0,627
18 2,110 2,898 1,740 2,567 18 0,497 0,683 0,410 0,605
19 2,101 2,878 1,734 2,552 19 0,482 0,660 0,398 0,586
20 2,093 2,861 1,729 2,539 20 0,468 0,640 0,387 0,568
21 2,086 2,845 1,725 2,528 21 0,455 0,621 0,376 0,552
22 2,080 2,831 1,721 2,518 22 0,443 0,604 0,367 0,537
23 2,074 2,819 1,717 2,508 23 0,432 0,588 0,358 0,523
24 2,069 2,807 1,714 2,500 24 0,422 0,573 0,350 0,510
25 2,064 2,797 1,711 2,492 25 0,413 0,559 0,342 0,498
26 2,060 2,787 1,708 2,485 26 0,404 0,547 0,335 0,487
27 2,056 2,779 1,706 2,479 27 0,396 0,535 0,328 0,477
28 2,052 2,771 1,703 2,473 28 0,388 0,524 0,322 0,467
29 2,048 2,763 1,701 2,467 29 0,380 0,513 0,316 0,658
30 2,045 2,756 1,699 2,462 30 0,373 0,503 0,310 0,449
40 2,024 2,707 1,682 2,430 40 0,320 0,428 0,266 0,384
50 2,008 2,680 1,676 2,404 50 0,284 0,379 0,237 0,340
60 2,000 2,664 1,673 2,393 60 0,258 0,344 0,216 0,309
Table 2
120
n ---------- t0,975 t0,995 t0,95 t0,99
n
60 2 2,000 2,664 1,673 2,393
120 1 1,980 2,617 1,658 2,358
` 0 1,960 2,576 1,645 2,326

4 © BSI 02-1999
 BS 2846-2:1981

Annex Confidence interval for the Table 3

mean from the range Confidence level Confidence level
Two-sided case One-sided case
If the measurements are arranged in ascending
order of magnitude, so that x1 <x2 <... < xn, then 95 % 99 % 95 % 99 %
w = xn – x1 is defined as the sample range. Still n q0,975 q0,995 q0,95 q0,99
assuming that the population is normally
Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI

distributed, the confidence interval for the 2 6,353 31,828 3,157 15,910
population mean can be determined from the 3 1,304 3,008 0,885 2,111
sample range when the number of measurements is 4 0,717 1,316 0,529 1,023
small, say 12 or less. The practical convenience of
5 0,507 0,843 0,388 0,685
this calculation is that it is faster; its disadvantage
is that it leads to a confidence interval which is 6 0,399 0,628 0,312 0,523
generally wider and which is more sensitive to 7 0,333 0,507 0,263 0,429
departures from the assumed normal form of the 8 0,288 0,429 0,230 0,366
observations.
9 0,255 0,374 0,205 0,322
Two-sided confidence interval
10 0,230 0,333 0,186 0,288
The two-sided confidence interval for the population
11 0,210 0,302 0,170 0,262
mean is defined by the following double inequality:
12 0,194 0,277 0,158 0,241
a) at the confidence level 95 %:
x – q0,975 w < m < x + q0,975 w From: E. Lord. The use of range in place of the
standard deviation in t-test (Biometrika,
b) at the confidence level 99 %: Vol. 34, 1947, pp.41-67), with entry q0,95 (n = 2)
x – q0,995 w < m < x + q0,995 w corrected.
One-sided confidence interval
The one-sided confidence interval for the population
mean is defined by one or other of the following
inequalities:
a) at the confidence level 95 %:
m < x + q0,95 w
or
m > x – q0,95 w
b) at the confidence level 99 %:
m < x – q0,99 w
or
m > x – q0,99 w
The coefficients q0,975, q0,995, q0,95, q0,99, are given
in Table 3.

© BSI 02-1999 5
 Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI

6
blank
 BS 2846-2:1981

Publications referred to

See national foreword.

Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI

© BSI 02-1999
 |
|
|
|
|
|
|
|
|
BSI Ð British Standards Institution
|
|
|
|
|
|
| BSI is the independent national body responsible for preparing British Standards. It
|
| presents the UK view on standards in Europe and at the international level. It is
| incorporated by Royal Charter.
|
|
| Revisions
|
|
| British Standards are updated by amendment or revision. Users of British Standards
|
| should make sure that they possess the latest amendments or editions.
Licensed Copy: :FULLNAME, : DATE, Uncontrolled Copy, (c) BSI

|
|
| It is the constant aim of BSI to improve the quality of our products and services. We
|
| would be grateful if anyone finding an inaccuracy or ambiguity while using this
| British Standard would inform the Secretary of the technical committee responsible,
|
| the identity of which can be found on the inside front cover. Tel: 020 8996 9000.
|
| Fax: 020 8996 7400.
|
|
| BSI offers members an individual updating service called PLUS which ensures that
|
| subscribers automatically receive the latest editions of standards.
|
|
| Buying standards
|
| Orders for all BSI, international and foreign standards publications should be
|
| addressed to Customer Services. Tel: 020 8996 9001. Fax: 020 8996 7001.
|
|
| In response to orders for international standards, it is BSI policy to supply the BSI
|
| implementation of those that have been published as British Standards, unless
|
| otherwise requested.
|
|
| Information on standards
|
| BSI provides a wide range of information on national, European and international
|
| standards through its Library and its Technical Help to Exporters Service. Various
|
| BSI electronic information services are also available which give details on all its
|
| products and services. Contact the Information Centre. Tel: 020 8996 7111.
|
| Fax: 020 8996 7048.
|
|
| Subscribing members of BSI are kept up to date with standards developments and
| receive substantial discounts on the purchase price of standards. For details of
|
| these and other benefits contact Membership Administration. Tel: 020 8996 7002.
|
| Fax: 020 8996 7001.
|
|
| Copyright
|
|
| Copyright subsists in all BSI publications. BSI also holds the copyright, in the UK, of
|
| the publications of the international standardization bodies. Except as permitted
| under the Copyright, Designs and Patents Act 1988 no extract may be reproduced,
|
| stored in a retrieval system or transmitted in any form or by any means ± electronic,
|
| photocopying, recording or otherwise ± without prior written permission from BSI.
|
|
| This does not preclude the free use, in the course of implementing the standard, of
|
| necessary details such as symbols, and size, type or grade designations. If these
|
| details are to be used for any other purpose than implementation then the prior
| written permission of BSI must be obtained.
|
|
| If permission is granted, the terms may include royalty payments or a licensing
|
| agreement. Details and advice can be obtained from the Copyright Manager.
|
| Tel: 020 8996 7070.
|
|
|
|
|
|
|
|
|
BSI |
|
389 Chiswick High Road |
|
London |
|
W4 4AL |
|
|
|
|
|
|