SUMMARY SAMPLING IN THE AUDIT

S U M M A R Y

by P. C. Louwers

W ith reference to sampling in the audit the w riter establishes a certain

parallel in the developm ent since 1933 of the professional conceptions in

the N eth erlan d s an d in the U .S.A .

T h e discussion of the problem has been split up in three parts.

1. Perm issibility in principle of sam pling in the audit in accordance w ith

the present professional conceptions

2. C h aracter and technique of „m athem atical” (a select) sampling

3. Inquiry into the possible significance of „m athem atical” sam pling for

the audit.

A d (1) Permissibility in principle of sampling in accordance with the present

professional conceptions

T h e w riter enters into detail on the significance of the internal control

w ithin the scope of the conception „complete a u d it” developed by P rof.

Limperg and the developm ent of the auditing theory since 1926.

H e comes to the conclusion th a t uniform ity of opinion has now been

attained on sam pling in the N etherlands: w ithin the scope of a complete

audit it is the d u ty of the auditor to ascertain th a t deviations to such an

extent as to essentially influence the picture presented by the figures are non

existent (certainty judgm ent) and on the other hand to acquire the firm

conviction th a t the system and the functioning of the internal control during

the period under observation have been of such a n ature th a t even devations

of relatively small im portance w ould have been noticed (probability

judgm ent).

Some kind of sam pling can as far as the probability judgm ent is concerned

take an acceptable place in such a complete audit.

T h e conceptions expressed in the Am erican literature do not show essen­

tial differences w ith the N eth erlan d s professional conceptions.

A d (2) Character and technique of „mathematical" (a select) sampling

In this p art some principles of the statistical calculation of probabilities

are discussed. T h e so called sequential sam pling, the likelihood-ratio-

m ethod, the tables for norm al and tightened inspection, the control of X -

an d R -charts, the random -selection, the stratification-technique and the

difference betw een the „decision-m aking” and „estim ating” purposes which

w ith the various sam pling m ethods m ay be h ad in view. T h e w riter states

th at though sequential sam pling based on the likelihood-ratio-m ethod is

strongly propagated in the A m erican literature, only a thorough combined

investigation by auditors and statisticians m ay determ ine w hich m ethod (s)

m ay be considered most effective for the auditor.

A d (3) Inquiry into the possible significance of „mathematical" sampling

for the audit

A fte r first discussing the character, technique an d m erits of unscientific

sam pling and judgm ent sampling, the natu re of the sam pling, used in the

audit is gone into. It is concluded th a t conventional sam pling is a combi­

nation of unscientific and judgm ent sam pling subsequently the possible

significance of m athem atical sam pling is analysed. T h e auditor is said (also

in the U .S .A .) to have the feeling th at the personal judgm ent of the auditor,

which is rightly considered to be of vital im portance threatens to get lost

when m athem atical sampling is applied. T h e w riter talks about the relation

betw een sam pling, and the auditing theory and he comes to the conclusion

th at it is only the auditor and not the statistician w ho is to decide w hether

the technique of sam pling will in a certain case result in the judgm ent re­

quired. T h e w riter deals w ith the statistical requirem ent of hom ogeneity of

the mass and he points out th a t hom ogeneity is not a conception w ith a fixed

m eaning but th a t basical and proportional aspects of the m ass, taking into

account the object of the inquiry, play a part. T h e duties of the auditor are

for this purpose distinguished in:

.—■

the system atic choice of the relatively im portant moments

.— the critical choice of the sym ptomatic moments

-— the exam ination of w hat is left.

From the statem ents the conclusion is d raw n that:

1. A ccording to the present professional views sam pling can take a ju sti­

fiable place in the exam ination of d ata for which a probability judgm ent

suffices.

C onventional sampling how ever is not thorough enough and will have

to give w ay to ,,m athem atical" sam pling.

2. A pplication of „m athem atical” sam pling in investigating a great num ber

of d ata will in most cases result in a higher degree of certainty than the

complete observation.

3. A lso for the exam ination of d ata on which a certainty judgm ent must

be obtained, sam pling, organically combined w ith the other auditing

means, is an effective m eans of auditing.

4. A thorough combined research of auditors and statisticians is required

in order to arrive at a better understanding.