HONG KONG STATISTICAL SOCIETY

SYLLABUSES
for
the Examinations of the Society

FOR EXAMINATIONS TO BE HELD IN 2008

and subsequently until further notice
(except for the traditional version of the Higher Certificate
which will be withdrawn after the examinations in 2008)

CONTENTS
Page
Summary of changes from previous version of syllabus

Introduction

Ordinary Certificate

Higher Certificate

Traditional version
Paper I: Statistical Theory
Paper II: Statistical Methods
Paper III: Statistical Applications and Practice
Modular version
Module 1: Data Collection and Interpretation
Module 2: Probability Models
Module 3: Basic Statistical Methods
Module 4: Linear Models
Module 5: Further Probability and inference
Module 6: Further Applications of Statistics
Module 7: Time Series and Index Numbers
Module 8: Survey Sampling and Estimation
Advised Academic Progression Through the Modules
Mathematics Background

11
13
15
17
19
21
22
23
24
25
26
27
28
29
30

Graduate Diploma

33
35
37
38
39
40
41
42
44
46
47
49

Statistical Theory and Methods Paper I

Statistical Theory and Methods Paper II
Applied Statistics Paper I
Applied Statistics Paper II
Option in Statistics for Economics
Option in Econometrics
Option in Operational Research
Option in Medical Statistics
Option in Biometry
Option in Statistics for Industry and Quality Improvement
Option in Social, Economic and Financial Statistics

SUMMARY OF CHANGES FROM PREVIOUS VERSION OF SYLLABUS

SUMMARY OF CHANGES FROM PREVIOUS VERSION OF SYLLABUS

1. Several important changes were made in the 2006 syllabus.

2. In 2007, a modular version of the Higher Certificate syllabus was introduced, to run
alongside the "traditional" version in 2007 and 2008, after which the "traditional" version
would be withdrawn.
3. There were a very few other syllabus changes in the 2007 syllabus.
4. There are no further changes in the 2008 syllabus.
For purposes of record, and to guide candidates who may be completing their examinations over
an extended period, the Note and list of Changes appended below show the changes made in the
2006 syllabus and, in italics, the further changes made in the 2007 syllabus, as mentioned in (1)
and (3) above.
Note
The previous [up to 2005] version was based on the syllabus first issued in 1996, which
remained unaltered in substance. New editions were issued in 1997 and 2000 with minor
editorial changes. A further new edition was issued in 2003 with clarification of the
mathematics background required for the Higher Certificate and Graduate Diploma.
Changes
Ordinary Certificate Paper I
Deletion of "Simple concepts of databases" at end of final section (rest of section unchanged).
Higher Certificate Paper I
Explicit inclusion of hypergeometric distribution in list in section 2.
New section "Inference" to clarify (not to increase) the content previously shown under
"Properties of distributions".
Section on "Association" rewritten:
(i)
to separate the discrete and continuous cases (for greater clarity, no change in
substance);
(ii)
to explicitly include product-moment correlation;
(iii) to reorganise the regression material vis--vis the new section in Higher Certificate
Paper II.
Higher Certificate Paper II
Removal (to Higher Certificate Paper III) of sub-section "Interpretation of published data".
Restriction of goodness-of-fit tests to the 2 test; Kolmogorov-Smirnov procedure deleted (it
remains in Graduate Diploma Statistical Theory and Methods Paper II).
Introduction of Wilcoxon rank sum test by name (in addition to Mann-Whitney U test).
Introduction of new section on use of regression and correlation (but note that use of the
Durbin-Watson statistic is deferred to Graduate Diploma Applied Statistics Paper I).
Higher Certificate Paper III
Minor changes to comments for the first item.
Additional section at end on "Interpretation of published data" (moved from Higher Certificate
3

Paper II).
Graduate Diploma Statistical Theory and Methods Paper I
Explicit inclusion of hypergeometric distribution in list in section 2 sub-section 3.
Graduate Diploma Statistical Theory and Methods Paper II
Introduction of Wilcoxon rank sum test by name (in addition to Mann-Whitney U test).
Graduate Diploma Applied Statistics Paper I
Insertion of the Durbin-Watson statistic in the section on analysis of variance and regression
analysis. Explicit inclusion of multicollinearity in this section.
Graduate Diploma Applied Statistics Paper II
Clarification of content of the demography section.
Graduate Diploma Option in Operational Research
Changes to the comments for unconstrained optimisation problems and for Monte Carlo
methods.
Graduate Diploma Option in Biometry
Explicit inclusion of multicollinearity in the section on regression modelling.

HONG KONG STATISTICAL SOCIETY

Introduction to the Examinations
The professional examinations conducted by the Hong Kong Statistical Society provide
a route into the statistical profession at introductory, intermediate and advanced levels.
They will be suitable for many people who wish to enter into or advance within he
profession.
This may be particularly helpful for those who cannot pursue an
undergraduate degree course or who wish to convert from another discipline after some
years' work.
The examinations consist of papers at three levels, namely Ordinary Certificate, Higher
Certificate and Graduate Diploma. Taken in sequence, these levels provide a
comprehensive syllabus of study in applied statistics.
The Ordinary Certificate consists of two compulsory papers.
In 2008, the Higher Certificate is offered in two versions. The traditional version
consists of three compulsory papers - Statistical Theory, Statistical Methods, Statistical
Applications and Practice. This version is available for the last time in 2008. The
modular version consists of eight modules which may be taken singly or in any
combination. Each module is separately certificated. The Higher Certificate as a whole
is awarded on successful completion of six modules (subject to some restrictions on
choice of modules); the shelf life of modules leading to this Certificate is unlimited.
Full details of both versions of the Higher Certificate are contained in this document.
Candidates who expect to complete the entire Certificate in 2008 may prefer to take the
traditional version. Candidates who prefer to spread their studies over a longer period
should take the modular version. The overall standards of the two versions are identical.
The Graduate Diploma consists of five compulsory papers. These are Statistical Theory
and Methods I and II, Applied Statistics I and II, and an Options Paper. The subjects in
the Options Paper take the form of half-syllabuses in the following subject areas:
Statistics for Economics, Econometrics, Operational Research, Medical Statistics,
Biometry, Statistics for Industry and Quality Improvement; candidates are required to
answer questions on two of these half-syllabuses. It is not necessary to take all five
papers of the Graduate Diploma at the same sitting; the regulations for splitting papers
over more than one sitting are set out in the Graduate Diploma section of this document.
The Graduate Diploma is widely recognised and respected, locally and internationally,
by employers in the public and private sectors. This qualification is also widely
recognised by universities as being sufficient for admission to an MSc course in
Statistics.
Individuals entering for examinations must be Full Members/Student Members of the
Society or Examination Associates. Full details of the procedures for entering the
examinations are available from the Society. Copies of past papers are also available
from the Society, as are suggested reading lists for the examinations.
5

The examinations are held in May each year. The Society provides statistical tables in
each examination. Copies of these tables may be obtained from the Society so that
candidates may familiarise themselves with the layout. Candidates are expected to make
appropriate use of calculators in the examinations. Regulations concerning calculators
are available from the society.
Candidates should read this syllabus jointly with the Guide to Examinations and
Examination Associate Membership, which contains full details of the procedures for
entering the examinations and is downloaded from the Society website.
The address for all examinations correspondence is the Hong Kong Statistical Society
Examination Office, c/o HKU School of Professional and Continuing Education, Rm
313, 3/F, Admiralty Centre, 18 Harcourt Rd, Hong Kong. Email may be sent to
exam@hkss.org.hk . Examinations materials are also available on the Society website
http://www.hkss.org.hk .

All examinations materials is available on the Society website at

http://www.hkss.org.hk/Exam/Exam.htm

ORDINARY CERTIFICATE IN STATISTICS

The object of this first level certificate is to provide a sound grounding in the principles
and practice of statistics with emphasis on practical data collection, presentation and
interpretation. It is intended both as a first qualification, an end in itself, and as a basis
for further work in probability and statistics, as for example in the more advanced
syllabuses of the Society. The normal study time required is estimated as one year.
The background level assumed for this examination is a good Hong Kong Certificate of
Education Examination (HKCEE) in mathematics, or an equivalent qualification or
relevant work experience.
The examination consists of two three-hour papers. There are no restrictions on the
number of questions that may be answered on each paper. Questions are of different
lengths. The marks available for each question are printed on the examination papers.
Questions on either paper may come from any part or combination of parts of the
syllabus, and the syllabus is to be seen as a unified whole. Both papers must be taken at
a single sitting. It is not permitted to sit the papers separately in different years.

SYLLABUS FOR ORDINARY CERTIFICATE

The Ordinary Certificate examination consists of two papers. Each paper contains from 7 to 9 (inclusive)
questions. There are no restrictions on the number of questions that may be attempted.
The Ordinary Certificate syllabus covers the essential ideas of statistics in practice. Anyone familiar with
the syllabus should be able to carry out supervised statistical work of a routine kind, or be able to apply
statistical methods, at an elementary level, within work of a more general kind. Paper I, Collection and
Compilation of Data, includes the key topics of data collection in the field: what data should be collected,
who from, and how will the data be captured? It also includes work on data processing, including a brief
introduction to statistical computing. Paper II, Analysis and Presentation of Data, encompasses the basic
techniques of descriptive statistics. Two ideas are of paramount importance: what is the most appropriate
way to analyse the data, and how can the message within the data most clearly be communicated?

Paper I: Collection and Compilation of Data

The origin, use and interpretation of published or
administrative data.

Overview of official statistics in a country of the

candidate's choice. What statistical series are
produced, how are the data collected and to what
uses are the data put?

Elementary ideas of sampling methods.

Definitions of population and sampling frame.
Methods of selecting samples (including practical
problems) and implications of sample size:
simple random sampling, systematic sampling,
cluster sampling, quota sampling, stratified
random sampling and multi-stage sampling.

No formulae are required.

Advantages and disadvantages of each method,
including considerations of cost and accuracy.

Pilot surveys, censuses, sample surveys, personal

interviews, self-completion questionnaires, postal
and telephone enquiries.
Serial surveys longitudinal or cross-sectional. Problems arising
in the collection of data, late returns, 'freak' values
and their treatment.

No formulae are required.

Non-sampling errors.
Identification and
interpretation of bias error (e.g. from nonresponse, errors in defining the population,
enumerator distortion, etc).

No formulae are required. Candidates may be

asked to provide examples of how different types
of bias error may arise.

Design of simple questionnaires and forms for

collection of data. Formulation, classification and
coding of questions, including verification.
Making questionnaires suitable for data
processing and analysis; use of missing value
codes.

Candidates should be able to produce their own

simple questionnaire and data form.

Distinction
between
experimental studies.

observational

and

Use of computers for data storage and retrieval.

Access to a computer when preparing for the

examinations is not necessary.

Paper II: Analysis and Presentation of Data

Use of rough checks for order of magnitude and
leading digits in results.

Candidates are encouraged to check that their

answer 'looks about right'.

Approximation, limits of accuracy, rounding and

accuracy of recording. Percentages, ratios, rates
and linear interpolation. Distinction between
discrete and continuous data.

Error in the result of a simple expression when

the consistent parts are all rounded.

Construction and uses of frequency tables for one

or more variables and contingency tables. Tables
for presenting collections of results together with
summary tables of frequencies, relevant averages,
standard deviations, etc.

Credit will be given for neatness and clarity of

presentation.

Graphs and diagrams, their use in analysis and

presentation. Construction, uses and limitations of
scatter diagrams, time charts, stem and leaf
diagrams, histograms, bar charts, pie diagrams,
frequency and cumulative frequency curves and
boxplots (box and whisker plots).

Credit will be given for neatness and clarity of

presentation.
Candidates should use graph paper, a ruler and
sharp pencil to draw diagrams.

Sample measures of location and dispersion.

Arithmetic mean, median, mode, percentiles,
range, inter-quartile range, variance, coefficient of
variation; their uses and limitations as measures;
their calculation from frequency tables and raw
data; graphical methods of estimation. Distinction
between inter- and intra-subject variation.

Formulae should be known, except that the

definition of 'hinges' is not required.

Probability as a measure of uncertainty. Link

between probability and relative frequency.
Allocation of probabilities in 'equally likely' cases.
Mutually exclusive events. Independent events.
Addition and multiplication of probabilities with
simple applications. Use of Venn diagrams and
tree diagrams.

The long-run concept of probability, e.g. from

tossing a coin repeatedly, should be understood.

Calculation of least squares regression line and its

interpretation. Correlation as a measure of linear
association between two variables. Productmoment correlation coefficient. Spearman's rank
correlation coefficient.

Derivation of the least squares estimates (by

calculus) is not required.
Candidates should know when it is appropriate to
use each method.

Simple moving averages for detecting trends and

for smoothing time series. Seasonal data.

Candidates should know when it is appropriate to

use additive or multiplicative models for
seasonality, and how each is calculated.

Knowledge of weighted forms of moving average.

No calculation is required.

Simple and weighted averages of price relatives.

Construction of aggregate (Paasche, Laspeyres
and Fisher) averages. Simple chain-based indices.
Limitation and use of index numbers e.g. in
assessment of productivity and prices.

Knowledge of how and why index members are

used in real life is assumed.

Candidates should be able to estimate percentiles

from a cumulative frequency curve (ogive).

Permutations and combinations are not required.

continued on next page

Interpretation. Translation of written statements

into tabular forms; simple fallacies, typical
misleading distortion in popular published graphs.
Answers to questions about tables and charts.

Candidates may be asked to explain what the

data presentation tells the reader.

Writing of clear and concise reports on numerical

data in different contexts.

Candidates may be tested on their spelling and

grammar as well as their logic.

10

HIGHER CERTIFICATE IN STATISTICS

TRADITIONAL VERSION
The Higher Certificate in Statistics is intended both as a qualification in statistics at a
more advanced level than that of the Ordinary Certificate, an end in itself, and as a basis
for further work in statistics up to the highest undergraduate level, as for example in the
Society's Graduate Diploma in Statistics. The normal study time for the Higher
Certificate is estimated as two years.
Entry to the Higher Certificate assumes knowledge of the material in the Ordinary
Certificate in Statistics, or equivalent. Candidates should therefore ensure that they are
familiar with this material before sitting for the Higher Certificate.
The examination consists of three three-hour papers. Each paper contains eight questions
of which candidates are asked to answer five. All three papers must be taken at the same
sitting, except in the case of candidates who are resitting a single "referred" paper from a
previous sitting. Candidates doing this may also sit some examinations for the Graduate
Diploma at the same time (subject to some restrictions).
Candidates who fail marginally in any one paper in the 2008 sitting may, at the
discretion of the examiners and provided their overall level of performance reaches an
approved level, be directed to one or more modules in the modular version of the Higher
Certificate so that they may complete the Certificate in a subsequent year.
Candidates are reminded (see the Introduction section above) that this version of the
Higher Certificate is offered for the last time in 2008.

11

SYLLABUSES FOR HIGHER CERTIFICATE

TRADTIONAL VERSION
The Higher Certificate examination consists of three papers, namely Statistical Theory, Statistical
Methods, and Statistical Applications and Practice. Each paper contains eight questions, of which five
are to be answered.
Candidates will be expected to be able to interpret computer output from statistical packages. Detailed
knowledge of specific packages is not required.
Knowledge of the mathematical topics listed below is required. The ability to apply this mathematics will
be examined within the statistical contexts defined by the syllabuses for the three papers.

Algebra
Permutations and combinations. Simple partial fractions. Solution of linear and quadratic equations.
Manipulations and solution of simple inequalities. Trigonometric functions. Summation of series with
notation. Limits of sequences and functions. Geometric series. Exponential and logarithmic functions,
including their expansions in series and the results

lim(1 + n 1 ) n = e and lim(1 + x ) n = e x

n
n
n

Positive numbers raised to any real power. Use of the binomial theorem with any real index.
Solution of simple sets of linear equations having unique solutions.

Differential calculus
Differentiation. Derivatives of polynomial, trigonometric, logarithmic and exponential functions, and of
sums, products, quotients or functions of these functions. Maxima and minima; simple examples of
asymptotes; points of inflexion (defined as those points at which a curve crosses its tangent). Graphical
representation of functions and simple examples of curve-tracing. Some examples of partial
differentiation.

Integral calculus
Integration. Indefinite and definite integrals, including those with infinite limits. Integrals of algebraic,
exponential, logarithmic and trigonometric functions. Simple examples of integration by substitution and
by parts, including reduction formulae.

12

Paper I: Statistical Theory

The aim of this syllabus is to develop the theoretical framework of probability and probability distributions
to a level where it can be used in standard statistical applications.
Probability
Definition of probability. Addition and multiplication of probabilities, conditional probability,
statistical independence. Bayes' theorem.
Distributions
Random variables.
Discrete and continuous
probability distributions. Probability function,
probability
density
function,
cumulative
distribution function.
Simple theory of
elementary probability distributions, including
binomial, Poisson, geometric, negative binomial,
hypergeometric, Normal, exponential, gamma and
uniform.
Properties of distributions
Expectation and variance; their general properties
and values for standard distributions.

Distributions, means and variances of sums of

independent and identically distributed random
variables and simple functions, such as aX + b.
Linear combinations of independent Normally
distributed variables. Use of probability and
moment generating functions to establish the
distribution of sums of random variables.
Inference
The likelihood function.

In this syllabus, a frequentist view of probability

is adopted.

Derivation of distribution function

probability density function and vice versa.

from

Candidates should be able to show that the

probability function (probability density function)
of the named distributions sums (integrates) to
unity. Some knowledge of the derivation and
application of these distributions is expected.
Derivation of the expected value and variance of
random variables with the distributions listed
above. Questions may be set for other simple
distributions.
Results for distributions of sums of Normal and
exponential random variables should be known.
Distribution of

a X
i

when X i has N i , i

distribution.

Plots of likelihood against the parameter.

Estimation of a single parameter of a distribution

using the method of moments and the method of
maximum likelihood.

Maximisation of likelihood using graphical

methods and calculus. Questions may be set
involving simple distributions other than those
listed above.

Calculation of approximate variance of a

maximum likelihood estimator using second
derivative of log likelihood

Use in constructing a confidence

Regularity conditions may be assumed.

interval

continued on next page

13

Association
Simple bivariate discrete distributions. Joint,
conditional and marginal distributions.

Distributions with probabilities presented in a

two-way table.

Joint probability density function for bivariate

continuous distributions.
Conditional and
marginal distributions.

Simple cases only. Bivariate integration is not

required.

The bivariate Normal distribution.

Familiarity
distribution.
required.

Product-moment correlation (Pearson). Calculation and interpretation.

Test for independence using tables, assuming

bivariate Normally.

Simple and multiple linear regression. Least

squares estimation. Analysis of variance for
regression models. Examination of residuals.

Knowledge of F tests for regression and partial t

test for regression coefficients.

Linearisation of growth curves and other models.

Problems in which taking logarithms or

reciprocals leads to a straight line relationship.

with
the
bivariate
Normal
Mathematical details are not

14

Paper II: Statistical Methods

The aim of this syllabus is to develop statistical concepts and analysis to a stage where they can be used in
standard problems of estimation and hypothesis testing. Particular emphasis is placed on the need to
summarise and interpret the results of a statistical analysis.
Summarising and interpreting data
Frequency distributions. Numerical and graphical
forms of presentation and statistical interpretation.
Histograms, stem and leaf diagrams, dot-plots,
boxplots (box and whisker plots).
Summary statistics for measures of location,
variability and skewness.

Inference
Sample and population. Concept of a sampling
distribution. Standard error.

Credit will be given for graphs that are well

presented, including neat appearance and
inclusion of appropriate title and labels.
Sample mean, median, mode, quartiles, semiinter-quartile
range,
standard
deviation,
variance, range.

Sampling distribution of the mean.

Point and interval estimates. Construction and

interpretation of confidence limits.

Bias of a point estimator.

Hypothesis tests, test statistic, one- and two-sided

tests. Type I and II errors; power as 1-P(type II
error); power curves, operating characteristic
curves.

Knowledge of p-values and their interpretation.

Use of Normal, t, 2 and F distributions in testing

and interval estimation. Paired and unpaired twosample tests.

Tests and confidence intervals involving means,

variances and proportions.
Use of tables to obtain percentage points.

Statement and use of central limit theorem for

independent, identically distributed random
variables with finite variance.

Proof of CLT not required.

Statistical methods
One- and two-way analysis of variance.

Relationship to completely randomised and

randomised block designs.

Calculation of the standard error of treatment

means and of the difference between two
treatment means.

Including calculation after an analysis of

variance.

Goodness-of-fit tests of standard distributions

listed in Paper I to observed data.

Use of the 2 goodness-of-fit test, including

pooling of classes. The Kolmogorov-Smirnov
procedure is not required.

Analysis of two-way contingency tables; 2 test

for association; Yates' correction in 2 2 table.
McNemars test.
Non-parametric methods
Use of non-parametric and distribution-free
significance tests for paired and unpaired data:
sign test, Wilcoxon rank sum test (Mann-Whitney
U test), Wilcoxon signed-rank test.
Rank correlation: Spearman's coefficient.

Candidates will be expected to be able to use

tables of percentage points but do not need to
know how the tables are obtained.

continued on next page

15

Use of regression and correlation

Standard applications of the regression and
correlation methods covered in Paper I.
Calculation of simple linear regression equations.
Inference for regression coefficients. Residuals
and their use in checking assumptions.

Calculation of prediction intervals will not be

required.

Calculation and interpretation of the multiple

correlation
coefficient
(coefficient
of
determination).

R2 as a measure of the proportion of variation

explained.

16

Paper III: Statistical Applications and Practice

The aim of this syllabus is to develop skills in data analysis, using the theoretical concepts developed in
the syllabuses for the Ordinary Certificate and Papers I and II of the Higher Certificate, to analyse real data
sets and communicate the results comprehensibly.
The questions will require candidates to select and carry out appropriate statistical procedures and to
report the findings and conclusions clearly.
Topics in the syllabuses for papers I and II
The practical application of any of the topics
included in these syllabuses.

Design and analysis of experiments

Principles of randomisation, pairing and blocking.
Factorial treatment structure.

Handling
real
data
and
problems:
summarisation and presentation of data, choice
of model, discussion and checking of
assumptions, appropriate analysis, presentation
of and reporting on results.

Reasons for using these.

Completely randomised and randomised block

designs. Nested designs.

Analysis of variance, hypothesis tests and

confidence intervals for means and differences of
means. Completely randomised designs with
unequal replication. Nested designs with more
than one level of variation.

Advantages
of
factorial
experimentation.
Diagrammatic explanation of interaction.

Two-level factors, simple cases of more than two

levels.
Diagrams of means of treatment
combinations and their use for explaining
interactions of two factors.

Surveys
Censuses, sample surveys and routine collection
of data at intervals of time.
Personal and telephone interviews, postal
enquiries, pilot enquiries.
Sampling frames,
problems arising in the collection of data.
Design of questionnaires and forms for collecting
data.
Problems of non-response and bias among
interviewers, question bias, non-sampling errors.
Systematic sampling. Quota sampling.
Time series
Moving averages for the detection of trend.
Elementary treatment of weighted moving
averages, including exponential smoothing.
Isolation of trend, cycles, seasonals and residual
variation by decomposition. Examination of
residuals. Elementary forecasting.

Principles and practice, with examples from

candidates' knowledge and experience.

Advantages and disadvantages of these methods.

Using simple methods (including graphical) to

forecast a short time ahead on the basis of past
data.

continued on next page

17

Exploratory analysis
Candidates should be prepared to examine a set of
data, to choose and carry out suitable methods of
analysis, to answer questions from non-statistical
users and to present the analysis and conclusions
in the form of a short report. The techniques
required may be of the simplest kind, e.g. plotting,
grouping, transforming or calculating from the
data. Candidates will be expected to use box-plots
and other similar graphical displays.
Report writing
Candidates should be prepared to produce a wellordered, well-reasoned argument in a style
suitable for a designated readership.
(This
readership could be, for example, non-statistical
colleagues, managers, or users of official reports.)
Candidates will be expected to make use of
graphical methods to summarise data and identify
unusual features.
Interpretation of published data

Particular questions may be specified, and also

particular types of non-statistical user (see
below). Otherwise a brief general report on the
main points in the data analysis will be
appropriate.

The use of technical words alone will not be

enough; explanations must be helpful to the nonspecialist.

Graphs and diagrams must be clearly labelled so

that the reader understands them properly.

Candidates should be able to provide critical

comment on tables and graphs.

18

HIGEHER CERTIFICATE IN STATISTICS

MODULAR VERSION
The modular version of the Higher Certificate examination has been introduced by the
Royal Statistical Society and adopted also by the Hong Kong Statistical Society in
response to requests and because it appears to both societies that modularity is now the
correct approach for a professional examination at this level. It gives the opportunity for
candidates to proceed at their own pace. The modular version Higher Certificate as a
whole is intended both as a more advanced qualification in statistics, an end in itself, and
as a basis for further work in statistics. In addition, the individual modules are intended
as valuable free-standing elements in their own right; they may for example be useful
components of a formal or informal continuing professional development programme.
Candidates may enter for any number of modules, either singly or in combination, in any
examination session. Candidates may resit any module any number of times; the highest
mark achieved will stand. The pass mark for each individual module is 50%. Module
marks are "banked" without time limitation.
Each module is examined by a 1-hour written paper containing four questions of which
candidates are asked to answer three. Candidates will be advised of their marks for each
module taken. Candidates will receive a separate certificate for each module in which
they are successful.
Candidates who are successful in six modules including modules 1 to 4 (not necessarily
all in the same session) will be awarded the Higher Certificate in Statistics. Candidates
who wish to proceed to the Graduate Diploma in Statistics should have been successful
in modules 1 to 6 (again, not necessarily all in the same session).
Candidates are advised that entry to any of the modules assumes knowledge of the
material in the Ordinary Certificate in Statistics, or equivalent. Candidates should
therefore ensure that they are familiar with this material before sitting for any modules.
This is particularly important for candidates who wish to attempt to achieve the Higher
Certificate in Statistics by taking six modules (including modules 1 to 4) in the same
session. For candidates, who after applying for academic assessment, have been advised
by the Examination Board to consider sitting for the Ordinary Certificate, are advised to
sit the examination step by step in order to have a solid foundation for better
achievement and examination results.
All modules will be offered in 2008

19

MODULE SYLLABUSES
Syllabuses for the eight modules are presented over the ensuing pages.

The modules are

Module 1

Data collection and interpretation

Module 2

Probability models

Module 3

Basic statistical methods

Module 4

Linear models

Module 5

Further probability and inference

Module 6

Further applications of statistics

Module 7

Time series and index numbers

Module 8

Survey sampling and estimation

20

MODULE 1: Data collection and interpretation

Summarizing and interpreting data
Frequency distributions. Numerical and graphical
forms of presentation and statistical interpretation.
Histograms, stem and leaf diagrams, dot-plots,
boxplots (box and whisker plots).
Summary statistics for measures of location,
variability and skewness.

Credit will be given for graphs that are well

presented, including neat appearance and
inclusion of appropriate title and labels.
Sample mean, median, mode, quartiles, semiinter-quartile
range,
standard
deviation,
variance, range.

Surveys
Target and study populations. Sampling frames. Principles and practice, with examples from
Problems arising in the collection of data.
candidates' knowledge and experience.
Censuses, sample surveys and routine collection of
data at intervals of time.
Design of questionnaires and forms for collecting
data.
Personal and telephone interviews, postal Advantages and disadvantages of these methods.
enquiries, pilot enquiries.
Problems of non-response, bias among
interviewers, question bias, non-sampling errors.
Simple random sampling. Uses and limitations. Candidates will be expected to be able to use, but
Estimators for means, totals and proportions and not to derive, formulae for estimators and
the variances of these estimators.
variances.
Use of other practical methods of sampling: No formulae are required.
systematic sampling, cluster sampling, quota
sampling, stratified random sampling and multistage sampling.
Exploratory analysis
Candidates should be prepared to examine a set of
data, to choose and carry out suitable methods of
analysis, to answer questions from non-statistical
users and to present the analysis and conclusions
in the form of a short report. The techniques
required may be of the simplest kind, e.g. plotting,
grouping, transforming or calculating from the
data. Candidates will be expected to use boxplots and other similar graphical displays.

Particular questions may be specified, and also

particular types of non-statistical user (see
below). Otherwise a brief general report on the
main points in the data analysis will be
appropriate.

Report writing
Candidates should be prepared to produce a wellordered, well-reasoned argument in a style suitable
for a designated readership. (This readership could
be, for example, non-statistical colleagues,
managers, or users of official reports.) Candidates
will be expected to make use of graphical methods
to summarise data and identify unusual features.

The use of technical words alone will not be

enough; explanations must be helpful to the nonspecialist.

Interpretation of published data

Candidates should be able to provide critical

comment on tables and graphs.

Graphs and diagrams must be clearly labelled so

that the reader understands them properly.

5
21

MODULE 2: Probability models

Probability
Definition of probability. Addition and multi- In this syllabus, a frequentist view of probability is
plication of probabilities, conditional probability, adopted.
statistical independence. Bayes' theorem.
Distributions
Random variables.
Discrete and continuous
probability
distributions.
Probability
mass function, probability density function,
cumulative distribution function. Simple theory
of elementary probability distributions, including
binomial, Poisson, geometric, negative binomial,
hypergeometric, Normal, exponential, gamma and
uniform.

Derivation of cumulative distribution function from

probability density function and vice versa.
Candidates should be able to show that the
probability mass function (probability density
function) of the named distributions sums
(integrates) to unity. Some knowledge of the
derivation and application of these distributions
is expected.
Candidates should know how these distributions
arise in practice and be able to recognise them
from a brief description of a situation.

Properties of distributions
Expectation and variance; their general properties Derivation of the expected value and variance
of random variables with the distributions listed
and values for standard distributions.
above. Questions may be set for other simple
distributions.
Distributions, means and variances of sums of
independent and identically distributed random
variables and simple functions, such as aX + b.
Linear combinations of independent Normally
distributed variables.

Results for distributions of sums of Normal and

exponential random variables should be known.
Distribution of ai Xi when Xi has N (i , i2)
distribution.

Statement and use of central limit theorem for Proof of CLT not required.
independent, identically distributed random
variables with finite variance.
binomial and Poisson distributions.

22

23

MODULE 3: Basic statistical methods

Inference
Sample and population. Concept of a sampling Sampling distribution of the mean.
distribution. Standard error.
Point and interval estimates. Construction and
interpretation of confidence limits.
Hypothesis tests, test statistic, one- and two-sided Knowledge of p-values and their interpretation.
tests.
Type I and II errors. Power as 1P(type II error).
2
Use of Normal, t, and F distributions in testing Tests and confidence intervals involving means,
variances and proportions. Use of tables to
and interval estimation.
Paired and unpaired two-sample tests.
obtain percentage points.

Restricted to cases of testing for the mean of a

Normal distribution with known variance.

Power curves.
goodness-of-fit test
The
distributions to observed data.
2

of

standard

Analysis of two-way contingency tables;

for association; McNemars test.

test

Including pooling of classes. Uniform (discrete

and continuous), binomial, Poisson and Normal
distributions;
distributions
in
specified
proportions.
Use of Yates' correction in 2 2 tables.

Non-parametric methods
significance tests for paired and unpaired data: Candidates will be expected to be able to use
sign test, Wilcoxon rank sum test (Mann-Whitney tables of percentage points but do not need to
know how the tables are obtained.
U test), Wilcoxon signed-rank test.

24

MODULE 4: Linear models

Correlation
Product-moment correlation (Pearson).
Rank Simple tests using tables, with informal
correlation Spearmans coefficient. Calculation understanding of when each measure of
and interpretation.
correlation is appropriate. Association versus
causality.
Design of experiments
Reasons for experimentation, causality.
Principles of replication and
completely randomised design.

randomization,

Regression
Simple linear
estimation.

regression.

Least

squares

Multiple
linear
regression

concepts,
interpretation of computer output, inference for
regression coefficients using estimates and
estimated standard errors from computer output.

Models involving qualitative regressor variables

are not included in this module.
Knowledge of F test for regression and partial t
test
for regression coefficients. Methods for
selecting variables are not included in this
module.

Analysis of variance for regression models.

Calculation and interpretation of the multiple
correlation
coefficient
(coefficient
of
determination).

Simple cases of transforming to linearity.

R as a measure of the proportion of variation

explained.
Calculation of prediction intervals will not be
required.
For example, problems (such as growth curves) in
which taking logarithms or reciprocals leads to
a straight-line relationship.

Analysis of variance
One-way analysis of variance.

Relationship to completely randomised design.

Calculation and use of the standard error of Confidence intervals for differences.
treatment means and of the difference between Multiple comparison procedures will not be
two treatment means.
required.

25

MODULE 5: Further probability and inference

Bivariate distributions
Simple bivariate discrete distributions.
conditional and marginal distributions.

Joint, Distributions with probabilities presented in a

two-way table.

Simple bivariate continuous distributions.

probability density function, conditional
marginal distributions.

Joint Simple cases only rectangular and triangular

and spaces, with one or more sides parallel to a
coordinate axis.

The bivariate Normal distribution.

Familiarity
with
the
bivariate
Normal
distribution as a model (knowledge of joint pdf is
not required). Proofs of results are not required.

Generating functions
No limiting results.
Probability and moment generating functions.
Use to find means and variances. Use to establish
the distribution of sums of random variables.
Inference
The likelihood function.

Plots of likelihood against the parameter.

Estimation of a single parameter of a distribution Maximization of likelihood using graphical

using the method of moments and the method of methods and calculus. Questions may be set
maximum likelihood.
involving standard distributions listed in Module
2 and other simple cases.
Properties
of
point
estimators.
Range, Introductory treatment of these concepts in simple
unbiasedness,
consistency. Efficiency and
cases.
relative efficiency.
Calculation of approximate variance of a Use in constructing a confidence
maximum likelihood estimator using second Regularity conditions may be assumed.
derivative of log likelihood.

interval.

26

MODULE 6: Further applications of statistics

Design and analysis of experiments
Principles of design including randomization, Reasons for using these.
blinding, pairing and blocking.
Analysis of variance, hypothesis tests for means
and differences of means.

Randomised block designs. Latin squares.

Factorial treatment structure with two factors. Diagrams of means of treatment combinations
Advantages
of
factorial
experimentation. and their use for explaining interactions of two
Diagrammatic explanation of interaction. Two- factors.
way analysis of variance.
Multiple regression
Least squares estimation for multiple regression.

Extension of material in Module 4. Derivation

of normal equations in simple cases. Matrix
notation will not be required. Candidates will
not be expected to solve three or more
simultaneous equations.

Regression through the origin.

Including multiple regression with zero intercept.

Use of backwards
regression.

elimination

in

multiple Use of F tests.

Polynomial regression.

Simple cases only.

Use of indicator variables to model factors or Simple cases only.

qualitative variables.
Residuals and their use in checking assumptions.
Quality control and acceptance sampling
Charts for mean and range for Normal data. Charts Construction and use of Shewhart charts,
including use of warning and action lines.
for proportions.
Cusum charts.
Attribute sampling.
schemes.

Construction and use.

Single and double sampling

27

MODULE 7: Time Series and Index Numbers

Time series
Moving averages.

The use of moving averages for the detection of

trend.

Elementary treatment of weighted moving In-depth coverage of ARIMA modelling is not

averages, including exponential smoothing as an included.
introduction to ARIMA Models.
Isolation of trend, cycles, seasonals and residual
variation by decomposition.
For example, Christmas, Sundays, Easter, car
registrations, etc. Detection and control for these
in seasonal adjustment.

Examination of seasonals.

Examination of residuals.
Elementary forecasting.

Using Holt-Winters.

Interpretation of computer output.

Analysis and commentary on tables and graphs

produced by seasonal adjustment programs.
Confirmation of some results through hand
calculation.

Index numbers
Introduction to index numbers
Index numbers and their uses. Simple price
relatives, Laspeyres and Paasche.
Differences
between Laspeyres and Paasche; circumstances
when each may be used. Further index numbers
Tornqvist, Walsh, Fisher and
Geometric
Laspeyres.

Deflation
Why deflation is used and how it works; what
makes a good deflator; how deflation is carried
out.
Re-basing
Why, when and how re-basing is done.

Consideration of both price and volume indices,

with reference to specific examples within country
of interest.

With reference to specific examples within country

of interest.

With reference to specific examples within country

of interest.

Chain linking
Chain linking of simple price relatives, and
chain linking using Laspeyres.

With reference to specific examples within country

of interest.

Seasonal adjustment of indices

Why seasonal adjustment of indices is necessary;
how it is carried out.

With reference to specific examples within country

of interest.
With reference to specific examples within
country of interest, for example within National
Accounts.

Use of index numbers

28

MODULE 8: Survey Sampling and Estimation

Populations and frames
Target and study populations, uses of frames,
errors involved.
Non-probability sampling methods
Haphazard and quota sampling.
limitations.

Uses and Consideration of different scenarios. For

example: there is no frame and only limited
resources but quick results are needed what
would be done and how, how would bias be
minimised?

Sampling methods
Simple random sampling, systematic random Assessment of practical utility of different
sampling, stratified random sampling (under sampling methods in the context of specific
equal, proportional and optimal allocation), examples.
cluster and multi-stage sampling.
When each
might be used, benefits and limitations of each.
Simple random sampling
Uses, limitations, applications to different data
types, practical examples.
Formulae: estimates for totals and means using
number-raised estimation only, variance,
confidence intervals.

Number-raised (or expansion) estimation is the

application of weights (given by N/n) to individual
survey records to indicate how many units in
the population each unit in the sample represents.

Stratified random sampling

Uses, limitations, applications to different data
types, practical examples.
Discussion
concerning
Neyman/optimal
allocation: the concept, importance of reduced
variance, etc.
Formulae: estimates for totals and means using
number-raised
estimation
only,
variance,
confidence intervals.
Calibration techniques for estimation
Ratio and regression.

Discussion basis, consideration of how or why

one may want to use other survey information
in estimation.
Discussion concerning decisions that, in reality,
need to be made between the sampling methods,
estimators, etc, that are chosen to be used
within
a survey. There is often a trade-off when
considering the bias and variance incurred.
Discussion of sampling problems, seeking
solutions to conducting a survey, linking to
actual surveys carried out. Different types of
study should be discussed panel, longitudinal,
cross- sectional sample survey, experimental
studies, drug trials, etc.

Bias / variance trade off

Application of sampling methods

29

ADVISED ACADEMIC PROGRESSION

THROUGH THE MODULES
The Society does not operate any system of formal prerequisites for the modules.
Any candidate may enter for any module at any time.
The Society nevertheless strongly encourages candidates to ensure that they are
properly prepared for the modules they intend to take. The advice stated above is
repeated here, that entry to any of the modules assumes knowledge of the material
in the Ordinary Certificate in Statistics, or equivalent. Candidates should therefore
ensure that they are familiar with this material before sitting for any modules. This
is particularly important for candidates who wish to attempt to achieve the Higher
Certificate in Statistics by taking six modules (including modules 1 to 4) in the
same session. Such familiarity with the material need not have been acquired by
formally passing the Ordinary Certificate. It could have been achieved in other
ways, including by experiential learning through a candidate's work.
Candidates are very strongly advised to study the published specimen papers and
solutions. These give a good indication of the sorts of questions that will be asked,
their general level, and the breadth and depth of answers that will be expected.
For candidates, who after applying for academic assessment, have been advised by
the Examination Board to consider sitting for the Ordinary Certificate, are advised to
sit the examination step by step in order to have a solid foundation for better
achievement and examination results.
It is an inherent feature of the subject that some modules require knowledge of
some topics in other modules. Candidates may of course have already acquired
such knowledge through other routes, or may be willing to acquire it as part of their
studies for the module(s) at which they are aiming. As guidance, the Society offers
the following advice, but this should not be seen as exhaustive:
Module
1
2
3
4
5
6
7
8

Some dependence on topics in Module or Modules

2
2, 3
2, 3, 4
1, 2, 3, 4
1, 2, 4
1, 2, 3, 4

The form and extent of any dependence varies considerably from case to case. For
some modules, detailed knowledge of some of the topics contained in other module(s)
might be needed. In other cases, candidates will just need to be familiar with or have
some knowledge of the concepts and techniques contained in the other module(s).

30

MATHEMATICS BACKGROUND
The mathematics required to support the modules in the Higher Certificate in
Statistics is naturally at a higher level than that required for the Society's Ordinary
Certificate. A summary of the required mathematics is set out here, but
prospective candidates should study the published specimen papers and solutions to
get a detailed understanding of what is required in each module.
The examination papers will not concentrate on mathematics for its own sake. Rather,
it is the ability to apply mathematics within the statistical contexts defined by the
module syllabuses that will be examined.
Candidates should be aware that the general level of the Higher Certificate is,
broadly, that of the first year of a university undergraduate degree in Statistics,
also including some topics that might typically be found in the second year. The
level of mathematics required is commensurate with this.
The detailed mathematical requirements naturally vary from module to module.
As a general guide, a candidate who has knowledge of the topics contained in Alevel Mathematics in the schools examination system in England, or the equivalent
in other systems, will certainly be able to start work on any of the modules.
This does not imply that A-level Mathematics (or equivalent) must necessarily
have been formally taken and passed (or be studied concurrently). It is intended as
a guide to the breadth and depth of mathematical topics with which candidates
should be familiar. Some of the modules do not in fact require as much
mathematics as this. On the other hand, completion of the work in other modules
will require knowledge of additional mathematical topics beyond A-level
Mathematics (or equivalent).
The table on the next page sets out the various mathematical topics in broad headings
and indicates which are required for which modules.

31

Table showing mathematics requirements for each module

(see notes on previous page)

Topic
1
Algebra
Use of notation
Permutations and combinations
Solutions of linear and quadratic equations
Manipulation and solution of simple inequalities
Arithmetic and geometric series
Summation of series
Limits of sequences and functions
Positive numbers raised to any real power.
Exponential and logarithmic functions, including
their expansions in series
Use of the following results:
n
x
1 n
lim(1+n ) = e and lim(1+(x/n)) = e
Use of binomial theorem with any integer index
Solution of simple sets of linear equations
having unique solutions
Differential calculus
Derivatives of polynomial, logarithmic and
exponential functions, and of sums, products,
quotients or functions of these functions
Maxima and minima. Simple examples of
asymptotes
Graphical representation of functions and
simple examples of curve sketching
Simple examples of partial differentiation
Integral calculus
Indefinite and definite integrals, including those
with infinite limits
Integrals of algebraic, exponential and
logarithmic functions
Simple examples of integration by substitution
and by parts, including reduction formulae
Double integration elementary aspects only

32

Module
4 5

INTERPRETATION OF COMPUTER OUTPUT

In any of the modules, candidates will be expected to be able to interpret
computer output from statistical packages. Detailed
knowledge
of
specific
packages is not required.

CALCULATORS
Attention is drawn to the Society's regulations concerning the use of calculators in
examinations, as contained in the Guide to Examinations and Examination
Associate Membership".

SPECIMEN PAPERS AND SOLUTIONS

Specimen papers and solutions for all the modules are available for free download
from the Society's website (see http://www.hkss.org.hk/Exam/download.htm).

BOOK LISTS
Book lists for the modules are available for free download from the Society's
website (see http://www.hkss.org.hk/Exam/download.htm).

33

GRADUATE DIPLOMA IN STATISTICS

Entry to the Graduate Diploma in Statistics assumes knowledge of the material in the
traditional version of the Higher Certificate in Statistics, or in modules 1 to 6
inclusive of the modular version of the Higher Certificate, or equivalent. Candidates
must have passed or been given exemption from the Higher Certificate as appropriate
before sitting for the Graduate Diploma. However, as an exception to this, candidates
resitting one of the traditional Higher Certificate examinations may enter some of
the examinations for the Graduate Diploma at the same time. The Graduate Diploma
is a qualification in applied statistics at a level equivalent to that of a good University
Honours Degree in Statistics.
The Graduate Diploma is examined by five three-hour examination papers. Four of
these papers cover core syllabuses - two each in Statistical Theory and Methods and
in Applied Statistics. Each candidate has to take all four of these papers. Each paper
has a choice of five questions out of eight. The fifth paper is an Options Paper. The
subjects in this take the form of half-syllabuses in the following six subject areas:
Statistics for Economics, Econometrics, Operational Research, Medical Statistics,
Biometry as well as Statistics for Industry and Quality Improvement. Each candidate
has to answer questions from two of these half-syllabuses. Each half-syllabus has a
section of the examination paper devoted to it, with four questions in each of these
sections. Candidates are required to answer five questions, chosen from any two
sections with not more than three questions taken from any one section.
Candidates may take all five papers at one sitting. Alternatively, they may take the
two Statistical Theory and Methods Papers at one sitting (with or without the Options
Paper) and the two Applied Statistics Papers at a separate sitting (again with or
without the Options Paper), provided that the Options Paper is taken at one of these
two sittings.
Candidates may be allowed to resit the Options paper, or the two Statistical Theory
and Methods papers together, or the two Applied Statistics papers together. No
single-paper resits will be granted in either Statistical Theory and Methods or
Applied Statistics.
In the Graduate Diploma, knowledge of all work covered in the Ordinary Certificate
and Higher Certificate (modules 1 to 6, in the case of the modular version) is assumed,
and questions may be set using it whether or not a topic is specifically mentioned in
the individual paper syllabuses that follow. The level of mathematics required for the
Graduate Diploma is above that assumed for the Higher Certificate, in those areas
which are basic to statistical theory, statistical methods and applied statistics.

34

SYLLABUSES FOR GRADUATE DIPLOMA

The Graduate Diploma examination consists of five papers. Two of these are on Statistical Theory and
Methods, two are on Applied Statistics and the fifth is the Options Paper. The papers on Statistical
Theory and Methods and on Applied Statistics each contain eight questions of which five are to be
answered. The Options Paper contains four questions on each of the six options; candidates answer
five questions chosen from any two sections with not more than three questions taken from any one
section.

Candidates will be expected to be able to interpret computer output from statistical packages. Detailed
knowledge of specific packages is not required.

Knowledge of the mathematical topics listed below is required. These include those topics previously
specified for the Higher Certificate. The ability to apply this mathematics will be examined within the
statistical contexts defined by the syllabuses for the four compulsory papers and the Options paper.
Algebra
Permutations and combinations. Partial fractions, including quadratic factors. Solution of linear and
quadratic equations. Manipulations and solution of simple inequalities. Trigonometric functions and
their inverses. Summation of series with notation. Limits of sequences and functions. Geometric
series. Exponential and logarithmic functions, including their expansions in series and the results
n

n
x
lim (1 + n 1 ) = e and lim 1 + = e x . Taylor series expansion. Double summation (ij).
n
n
n

Positive numbers raised to any real power. Use of the binomial theorem with any real index.
Differential calculus
Differentiation. Derivatives of polynomial, trigonometric, inverse trigonometric, logarithmic and
exponential functions, and of sums, products, quotients or functions of these functions. Maxima and
minima; asymptotes; points of inflexion (defined as those points at which a curve crosses its tangent).
Graphical representation of functions and simple examples of curve sketching. Partial differentiation.
Integral calculus
Integration. Indefinite and definite integrals, including those with infinite limits. Integrals of algebraic,
exponential, logarithmic and trigonometric functions. Simple examples of integration by substitution
and by parts, including reduction formulae.
Double integration over rectangular and triangular regions having one or more sides parallel to a coordinate axis. Functions of a single variable and of two variables. Interchange of order of integration.
Jacobians of transformations.
Matrix algebra
Vectors. Square matrices: symmetry, singularity and non-singularity, determinants, inverses, relation
with sets of linear equations. Solution of simple sets of linear equations.
Rectangular matrices: addition, subtraction and multiplication. Rank of matrices via linear dependence
and independence of vectors. Quadratic forms, including their expression in matrix notation.
Numerical methods
Iterative solution of equations, including Newton-Raphson method.

35

Statistical Theory and Methods

The two syllabuses in Statistical Theory and Methods aim to treat these topics at University Honours
Degree level. A satisfactory command of this material will also give candidates a sound basis for study
of the topics in the Options syllabuses.

Statistical Theory and Methods Paper I

This syllabus concentrates on probability and distribution theory, extending the coverage of these topics
in the Statistical Theory paper of the Higher Certificate. Applications to simulation and stochastic
processes are also included.

Probability
Sampling with and without replacement.
Elementary problems involving urn models.
Joint probability, marginal and conditional
probability, independence.
Law of total probability. Bayes' Theorem.
Distribution theory
Random variables.
Discrete and continuous
random variables. The probability function and
probability density function.
Cumulative
distribution function.
Expectation as a linear operator. Expectation of
functions of a random variable. Mean and
variance.
Standard distributions and their use in modelling,
including binomial, Poisson, geometric, negative
binomial, hypergeometric, discrete uniform,
Normal, exponential, gamma, continuous uniform,
beta, Weibull, Cauchy, lognormal.

Ability to recognise the appropriate distribution

from a model description.
Relationships
between
distributions
(e.g.
exponential and Weibull).

continued on next page

36

Joint, marginal and conditional distributions.

Independence. Covariance, correlation and partial
correlation.
The multivariate Normal and
multinomial distributions.
Probability generating function.
Moment
generating function. Applications of generating
functions.
Distribution of sums of random
variables, and of sample mean.
Central limit theorem.

Proof for independent and identically distributed

random variables only.

Distributions of functions of several random

variables.
Transformations, including the
probability integral transform. Joint distribution
of mean and variance from a Normal random
sample. Independence for Normal samples.

Univariate and bivariate transformations.

The t, 2 and F distributions, and their use as

sampling distributions.

Including definitions in terms of N(0,1) random

variables.

Statistical methods
Joint distribution of order-statistics. Distribution
of sample range.
Simulation.
Generation of uniform pseudorandom numbers; testing for uniformity.
Methods of generating random numbers from
common distributions.
Applications of
simulation.

Methods for binomial, Poisson, exponential,

Normal.

Markov chains. Elementary ideas. The one-step

and n-step transition matrices. Long-run and
stationary distributions.

Time-homogeneous finite chains only.

37

Statistical Theory and Methods Paper II

This syllabus extends the coverage of the material on statistical inference in the Higher Certificate. It
also introduces Bayesian techniques and decision theory.
Estimation
Unbiasedness, mean square error, consistency,
relative
efficiency,
sufficiency,
minimum
variance. Fisher information for a function of a
parameter, Cramer-Rao lower bound, efficiency.
Fitting standard distributions to discrete and
continuous data. Method of moments. Maximum
likelihood estimation: finding estimators analytically and numerically, invariance, censored data.
Hypothesis testing
Simple and composite hypotheses, types of error,
power, operating characteristic curves, p-value.
Neyman-Pearson method. Generalised likelihood
ratio test.
Use of asymptotic results to construct tests.
Central limit theorem, asymptotic distributions of
maximum likelihood estimator and generalised
likelihood ratio test statistic.

Proof of the asymptotic distributions of the

maximum likelihood estimator and the
generalised likelihood ratio test statistic are not
required.

Confidence intervals and sets

Random intervals and sets. Use of pivotal
quantities.
Relationship between tests and
confidence intervals. Use of asymptotic results.
Sequential tests
Simple multistage procedures.
Sequential
probability ratio tests, results for approximate
significance level and expected sample size.

Proof for expected sample size is not required.

Bayesian inference
Prior and posterior distributions.
Bayesian
intervals for parameters and predictions. Choice
of prior: bets, conjugate families of distributions,
vague and improper priors.
Decision theory
Loss, risk, admissible and inadmissible decisions,
randomised decisions. Minimax decisions and
Bayes' solutions, including simple results.
Non-parametric inference
Use of ranks and randomisation; robustness.
Sign, Wilcoxcon rank sum (Mann-Whitney U),
Wilcoxon signed-rank, Kolmogorov-Smirnov (one
and two samples), goodness-of-fit and rank
correlation tests.
Non-parametric confidence
intervals;
applications using the sign and
Wilcoxon rank sum (Mann-Whitney U) tests.
Comparative inference
Different criteria for choosing good estimators,
tests and confidence intervals.
Different
approaches to inference, including classical,
Bayesian and non-parametric.
38

Applied Statistics Paper I

This syllabus aims to give an understanding of linear models used in statistical analysis, particularly in
regression, time series and multivariate analysis. Some theory of the standard methods is covered.
Candidates are also expected to have knowledge of how these methods are used and of the
interpretation of results from them (including computer output).
General linear model
Least squares. Properties of LS estimators in the
linear model, Gauss-Markov theorem. Models for
simple and multiple regression and for analysis of
variance and covariance. Estimation of variance.
Interval estimates of parameters. Weighted least
squares. Importance of assumptions. Transformation of variables. Linearising other models, e.g.
multiplicative models and growth curves.
Approximations
Approximate means and variances of functions of
random variables. Variance-stabilising transformations.
Analysis of variance and regression analysis
Analysis of variance for cross-classifications, with
replication. Fixed and random effects; variance
components. Nested forms. Application to data
collected in experiments or by sampling.
Multiple regression:
choice of variables,
regression diagnostics, analysis of residuals,
multicollinearity, serial correlation in time series.
Critical evaluation of computer output. Use of
indicator variables.
Detection of influential
observations.
Time Series Analysis
General principles; correlogram; prediction
methods. Simple autoregressive moving average
models, stationarity.
Multivariate methods
Principal components; theory, correlation and
covariance matrices, interpretation of computer
output and of calculated components, use in
regression. Discriminant analysis, choice between
two populations, calculation of discriminant
function and probability of misclassification, idea
of extension to several populations. Multivariate
analysis of variance. Cluster analysis, similarity
measures, hierarchical methods, single-link and
other methods. Multidimensional scaling.
Generalised Linear Models
Theory and use. Logistic regression for the
analysis of proportions. Log-linear models for the
analysis of structured count data.

Proof of standard results for multiple regression

(matrix notation). Application of these to other
situations as listed.

Derivation of E[f(x)] by Taylor series approximation. Use in finding appropriate transformations.

Estimation of variance components, and use in

planning sampling schemes.

Practical appreciation of use of multiple

regression; application in some special cases.
Use of the Durbin-Watson statistic.
Concept of leverage.

Basic theoretical results. Validity, and conditions

for use, of standard methods.

Reasons for using either method; number of

components retained.

Distance measures used in classification;

differences in results due to choice of measure
and method.

Components of a generalised linear model including the exponential family. Adapting to various
problems by choosing suitable link functions etc.
Candidates will be expected to have some
familiarity with the method of iterative
reweighted least squares used in fitting the
models.

39

Applied Statistics Paper II

This syllabus covers the application of statistical methods to censuses, surveys and designed
experiments, and some topics in demography. Deeper theoretical knowledge of surveys will be
required than in the earlier stages of the examinations. Questions on designed experiments may involve
contrasts (comparisons) among several means.
Survey methods
Census and sample survey design. Population,
sample and unit; sample frame. Pilot surveys.
Questionnaire design; interviewing techniques;
postal surveys; panels. Bias and non-sampling
errors;
non-response.
Examples of social,
agricultural and industrial censuses and surveys.
Sampling from central records. Editing and
presentation of results.
Analysis of sample surveys. Estimation of means,
variances, totals and proportions in samples from
finite populations; theory for simple random
sampling, extensions to stratified sampling.
Optimal sampling when total sample size or total
cost is fixed; models for cost functions. Use of
additional information:
theory of ratio and
regression methods of adjustment. Quota, cluster,
systematic and multi-stage methods of sampling.
Design of experiments
Randomisation, replication, blocking. Factorial
treatment structure.
Confounding, partial
confounding and fractional replication in 2n
experiments.
Balanced incomplete blocks
(omitting mathematical methods of construction).

In-depth
knowledge,
preferably
through
experience, of the practical problems of planning
and carrying out censuses and surveys.

Proofs of standard results.

Derivation of
formulae for expectation, bias (if any) and
variance in the methods listed. Use of Lagrange
multipliers.

Reasons for using these. How to construct a valid

randomised layout for each design. Knowledge
of what balanced incomplete block designs exist
for reasonably small block sizes and numbers of
treatments.

Analysis of completely randomised experiments,

randomised complete blocks, Latin squares,
factorial designs, balanced incomplete blocks.
General linear contrasts among treatments.
Assumptions of linear model analysis. Analysis of
residuals. Transformations. Missing values;
iterative methods of estimation.

Analysis of variance appropriate for each design.

Cochran's theorem (proof not required); partition

of sums of squares. Significance tests and
confidence intervals for treatment means and
contrasts.
Interpretation of results.
Understanding relevance of information given in
computer output of an analysis. Robustness of
analysis of variance to non-Normality, nonconstant variance etc.

Single degree of freedom contrasts.

Basic ideas of response surfaces; steepest ascent,

central composite and rotatable designs. Polynomial response functions. Mixture designs.

First- and second-order designs; search for

maxima. Models of curved and asymptotic
responses. Restrictions on factor levels.

Checks on validity of standard models; graphical

methods. Knowledge of general non-orthogonal
designs is not required.

Effect of outlying observations.

Topic in demography
Population pyramids.
Life tables.

Construction and use; derived quantities,

including calculation of life expectancy.

Standardised rates.

Direct and indirect standardisation.

40

Option in Statistics for Economics

The aim of this syllabus is to develop skills in data analysis with particular reference to economic data.
Topics in the syllabuses of the compulsory papers of the Graduate Diploma may be examined with
special reference to economic applications. Solutions will require a depth of knowledge of the
economic application.
Exploratory data analysis and index numbers
Use of exploratory data analysis (EDA) with
economic data. Use of diagrams, including
histograms with unequal intervals, population
pyramids, time charts, diagrams with one or two
logarithmic axes and box plots.

Index numbers of prices and quantities.

Standardisation, especially in the context of
demographic statistics. Population projections.

Time series
Simple methods of time series analysis. Seasonal
adjustment techniques for economic time series.
Simple autoregressive models; model identification and checking. Forecasting and prediction
intervals.
Surveys
The design and analysis of social surveys. Simple
random, stratified, multistage and quota sampling.
Statistical methods
Statistical inference in economic contexts.
Association in two-way tables and the fit of
theoretical distributions. One-way analysis of
variance. Non-parametric tests.
Regression and correlation
Calculations associated with simple regression
and correlation. Interpretation of simple and
multiple regressions, including standard errors of
estimated coefficients, R 2 and R 2 , standard errors
and confidence/prediction intervals of regression
estimates. Autocorrelation and heteroscedasticity.
The use of dummy (indicator) explanatory
variables. Multicollinearity, specification bias,
log linearity.
Significance tests relating to
regression.

The nature of different economic data sets and the

use of appropriate graphical presentations is
essential. A working knowledge of the key
aggregates, e.g. gross domestic product,
consumer prices index and balance of payments,
will be fundamental.
Different definitions, base changes and
combination of indices. The use of price and
other indices to summarise the performance of an
economy.

The mathematics of smoothing techniques. The

autocorrelation and partial autocorrelation
functions of simple autoregressive models; their
use for identification. The aggregates mentioned
above will be studied in a time series context.

Sample selection; control of variance.

Hypothesis testing in small and large samples for

measured and discrete variables. Use of the 2
distribution for association analysis and
goodness of fit.

Routine calculations for simple linear regression

and correlation will be given but not dwelled
upon. The mathematics of regression will be
subordinate to interpretation and making valid
inferences about regression models.
Autocorrelated errors for time series regression.
The use of indicator (dummy) variables, and
potential pitfalls.

Product moment correlation, rank correlation and

partial correlation.
Interpretation of the output from modern
computer packages reflecting the topics covered
in this syllabus.

The important element here is the output, not the

package used. Ability to make sense of computer
output from a statistical point of view is essential.

41

Option in Econometrics
The aim of this syllabus is to develop an understanding of problems that arise in econometric model
building and how these problems can be resolved. Candidates should be able to undertake econometric
analysis of real data sets. The emphasis is on the interpretation and economic implications, rather than
on deriving the theory.
Topics in the syllabuses of the compulsory papers of the Graduate Diploma may be examined with
special reference to econometrics.
Least squares
Problems arising from the use of ordinary least
squares in econometrics.

When the assumptions underlying the linear model

are relaxed, the consequences are investigated
and dealt with from both a theoretical and
practical point of view.

The use of diagnostic testing procedures in

regression analysis.
Aitken's generalised least squares and correction
for failure of standard assumptions.
Econometric modelling
An introduction to large sample theory of
estimation and testing. Lagrange multiplier tests
and other techniques.
Estimation using the
instrumental variables.
Effect of errors in
variables.

Theory and practice when ordinary least squares

no longer provides optimal estimators. The
instrumental variable principle, including choice
of instruments, should be shown to provide
consistent estimators.

The use of dynamic and distributed lag models in

econometrics. An introduction to the use of
models with limited dependent variables. Logit,
probit, Tobit analysis, and duration models. Tests
of linear restriction.

Estimation and testing in the presence of lagged

endogenous variables, including problems that
arise with serially correlated errors. Limited
dependent variables in the form of binary
response models.

Introduction to ARIMA Models, unit roots,

spurious regressions, cointegration and error
correction models. Forecasting.
Issues in simultaneous equation models, including
identification and the reduced form.
Estimation using two stage least squares.

Two stage least squares estimation for

simultaneous equation models and its relationship
to instrumental variable estimation.

Applications
Applications of econometric techniques to areas
such as demand studies and consumption and
investment functions. Introduction to macro
models of the economy.
Interpretation of the output from modern
computer packages reflecting the topics covered
in this syllabus.

The important element here is the output, not the

package used. Ability to make sense of computer
output from a statistical point of view is essential.

42

Option in Operational Research

Operational research is primarily concerned with problem solving and process improvement in
organisations. Although it makes extensive use of mathematical and statistical methods and models, the
design and manipulation of these models and methods is not of primary concern. However, students
qualifying in operational research need to have a good understanding of the common models in use, and
to demonstrate, in particular, the ability to select the appropriate model for a given problem situation, to
formulate the problem in a suitable form for the selected model and to interpret the results of the model
back into the problem situation. Candidates for this option must be prepared to display this ability as
well as having an understanding of the strengths and weaknesses of the models based on a clear
understanding of the structure of the appropriate models. The ability to obtain a numerical answer from
the formulae or the algorithm will not, on its own, lead to good marks.
Topics in the syllabuses of the compulsory papers of the Graduate Diploma may be examined with
special reference to operational research.
Linear programming and optimisation
Recognition of problems suitable for solution by
linear programming.
Problem formulation.
Graphical solution of two-variable problems.
Primal and dual problems. The simplex method.
Sensitivity analysis.
Transportation and
assignment problems. Parametric programming.
Interpretation
problems.

of

computer

output

for

LP

Graphical treatment of non-linear constraints for

two-variable problems.

Interpretation of results from an analysis, with

consideration of validity of results in the light of
assumptions made.

The important element here is the output, not the

package used. Knowledge of any particular LP
package is not required.
Simple problems only.

Multi-objective problems: goal programming.

Unconstrained optimisation problems for single
variables and several variables.

Theory of queues
The M/M/1 queue. Differential-difference equations. Conditions for equilibrium. Equilibrium
distributions of queue size and waiting time for
FCFS queues. Extensions to M/M/k and M/M/
queues.
The M/G/1 queue, imbedded Markov chain
analysis.
The Pollaczec-Khintchine formula.
Mean queue length and waiting time.

The Newton-Raphson method, steepest descent,

conjugate
gradient
and
Fletcher-Powell
(= Variable Metric) method.

Equilibrium behaviour for queues with transition

rates dependent on queue size.

Equilibrium treatment only.

Network analysis
Representation of a system as a network. Arrow
diagrams. Determination of critical path, and of
independent, free and total floats.
Gantt charts. Crashing.

Role of and methods for crashing. Determination

of minimum cost and minimum time solution.

continued on next page

43

Stock control and replacement

Deterministic models: economic order quantity,
simple extensions.

Extensions to cover lead times, quantity

discounts,
storage
limitation,
continuous
replenishment.

Probabilistic models: two-bin policy, (s,S) policy

for single-period models only.
Group and individual replacement policies.
Economic time for machine replacement.
Simulation
Methods of sampling from important distributions. Inversion, rejection and table look-up
techniques.

Distributions to include the exponential, Cauchy,

logistic, Normal, binomial, Poisson.

Monte Carlo methods. Use of variance-reduction

techniques.

Including importance sampling, control variables

and antithetic variables.

44

Option in Medical Statistics

This option covers a range of statistical concepts that are used in medical applications of statistics.
Topics in the syllabuses of the compulsory papers of the Graduate Diploma may be examined with
special reference to medical applications.
Clinical trials
Phase I to phase IV clinical trials.
Design of phase III clinical trials. Use of controls
and blinding. Importance of randomisation.
Methods
of
randomisation:
stratification,
minimisation, permuted block.
Trial protocol and clinical record form (CRF);
items to be included in a protocol.
Design and analysis of cross-over trials; carry
over effects.

A priori plausible treatment effect and trial size

calculation for qualitative and quantitative data.

Description of scope of phase I to phase IV trials.

Two period, two treatment trials. t-tests for

continuous responses. Treatment period interaction and carry over. Advantages and disadvantages of cross-over designs relative to
parallel group studies.
Sample size calculations for comparing two
means or two proportions and for achieving
required precision in estimating a mean or a
proportion.

Multiplicity: multiple treatments, multiple endpoints, interim analyses.

Data Safety & Monitoring Board and early
stopping of clinical trial: surrogate endpoints
Statistical overview (meta-analyses).

General ideas on combining results from different

studies. Methodological details not required.

Epidemiology
Incidence and prevalence.
Design and analysis of cohort (prospective)
studies.
Design and analysis of case-control (retrospective) studies.
Matched case control design and analyses, using
McNemar's test.
Causation.
Relative risk and odds ratio.
Confounding and interaction.
Mantel-Haenszel procedure.

Diagnostic tests
Sensitivity, specificity, ROC curves, positive
predictive value.

Distinction between these concepts.

Use of logistic regression not required.
Estimation of odds ratio from 2 2 tables.
Combining 2 2 tables.
Reasons for matching;
advantages and
disadvantages relative to unmatched studies.
Inferring causality from observational studies.
Estimation of and confidence intervals for odds
ratios.
Use of procedure in adjusting for confounding
variables.

Uses in diagnosis and in screening.

continued on next page

45

Analysis of survival data

Hazard and survivor functions.
Kaplan-Meier estimate of survivor function.

Confidence intervals for survivor function using

Greenwood's formula.

Logrank test.
Parametric survival distributions - exponential,
Weibull.

Use of log cumulative hazard plot to check

Weibull and proportional hazards assumptions.
Use of these distributions in modelling survival
data. Fitting methods not required.

Proportional hazards and Cox regression.

An understanding of the assumptions and

interpretation of the fitted model. Details of
partial likelihood and numerical methods for
fitting the model not required. Calculation and
interpretation of hazard ratios and confidence
intervals.

Checking for non-proportionality of hazards.

Checking for non-proportional hazards using a

log cumulative hazard plot and plots of hazard
functions.

Health information
International Classification of Diseases.
Sources and limitations of data on mortality and
morbidity e.g. cancer registration.
Use of health data in the provision of care.
Standardised rates.

Direct and indirect standardisation of mortality

and morbidity rates. Confidence intervals for
standardised mortality ratio.

46

Option in Biometry
The aim of this option is to cover the topics of experimental design, survey methods, regression
modelling and bioassay that feature prominently in biometric applications of statistics.
Topics in the syllabuses of the compulsory papers of the Graduate Diploma may be examined with
special reference to biological and agricultural applications.
Experimental design
Further design and analysis of experiments:
principles of design, treatment comparisons and
interpretation of results.

A thorough knowledge of the use of completely

randomised designs, randomised block designs,
Latin squares and factorial experiments, and the
corresponding analysis of variance tables.

Methods of dealing with several missing values;

unbalanced designs.

Analysis of non-orthogonal
treatment effects.

Covariance analysis.

Incorporating information on a single covariate.

Relationship of analysis of covariance to the
comparison of regression lines.

Nested designs and components of variance. Split

plot experiments.

Estimation of components of variance

equating observed to expected mean squares.

Survey methods
Censuses and sample surveys: planning and
design. Use of maps and aerial surveys; crop
estimation and forecasting; forestry and land use
surveys. Handling large data sets on a computer.

Knowledge of any particular

management package is not required.

Regression modelling
Multiple regression analysis. Comparison of
regressions. Use of indicator variables. Weighted
regression.

Non-linear modelling. Fitting standard growth

curves (including logistic and Gompertz).
Estimation of parameters in non-linear models.
Bioassay
Biological assay.
Continuous and binomial
responses; probit, logistic and angular transformations; median effective dose (ED50), relative
potency, parallel line and slope ratio assays,
confidence limits for ED50. Analysis of bioassay
data using generalised linear models.

data;

adjusted

by

database

Use of residual sum of squares in comparing

alternative models. Comparison of straight lines.
Variable selection methods. Model checking
using residuals. Multicollinearity. Interpretation
of output from packages.
An appreciation of the Newton-Raphson scheme
for fitting non-linear models.

Greatest emphasis will be placed on the use of

logistic regression for analysing binomial data.
Estimation of ED50 from fitted dose response
curves. Use of Fieller's theorem, including
derivation of the result.
Use of general linear model for continuous data
and logistic regression for binary data.

47

Option in Statistics for Industry and Quality Improvement

The aim of this option is to cover statistical methods needed by engineers and industrialists. The
approach should be such that statistical thinking is developed in the context of industrial/engineering
problems. Topics such as failure mode, effects and criticality analysis (FMECA), quality function
development (QFD) and fault tree analysis (FTA) have been included in the syllabus in order to show
how the techniques of experimental design, reliability engineering and process control can be employed
effectively.
Topics in the syllabuses of the compulsory papers of the Graduate Diploma may be examined with
special reference to industrial applications.
Process control and quality improvement
There are many methods appropriate for statistical
process control and the following provide
guidelines:
Ishakawa's tools. Shewhart charts. Special and
common cause variability. Statistical process
control and control charts: charts for x , R; x , s;
non-conforming items and proportions (np, p);
non-conformities (events) and rates of recurrence
(c, u). 3-sigma and probability based limits.
Moving average, range and s.d. Zone charts,
exponentially weighted moving averages and
adaptive control. Cusums - monitoring and data
exploration.
Inherent (common cause) and
assignable (special cause) sources of variation.
Design of control schemes and sampling methods.
Process capability indices.
Experimentation
Directed experiments, informed observation and
testing. The plan-do-study-act (PDSA) cycle for
designing, executing, analysing and predicting.

A paper based approach will be used for control

charts before using computers.
Careful
consideration will be given to the interpretation
of charts. An awareness of the importance of
measuring and monitoring for continuous quality
improvement will be expected.

Use of Deming PDSA cycle in designing any

experiment, with clear objectives, purpose,
sample size, power, and statistical and
engineering conclusions.

Experimental design.
Two and three level
experiments, the use of orthogonal arrays,
contrasts, probability plots and confirmatory trials.
Confirmatory trials and comparative testing.
Response surface methods.

Knowledge
of
the
differences
enumerative and analytical studies.

between

Parameter design (Taguchi-type methods) and

measuring quality characteristics in robust
engineering.
Orthogonal arrays. Designing for robustness control and noise factors. Signal to noise ratios
including advantages and limitations.

The distinction between control and noise factors

will be emphasised. Larger-the-better, smallerthe-better and nominal-is-best characteristics will
be examined as potential for S/N analyses.
Candidates will be expected to have knowledge of
real applications of these methods.

continued on next page

48

Acceptance sampling. Sample size determination

and the use of producers' and consumers' risks.
The collection of data according to design.

The traditional approach of acceptance sampling

with producer's and consumer's risk will be
placed in the context of quality by inspection.
Advanced quality planning differs by promoting
designmanufacture with experiments which have
engineering improvements, measures of current
process variation and error rates (producer and
consumer risks) contained.

Reliability engineering, time to failure, time

between failures. Reliability and life testing.
Probability plotting including Weibull and hazard
plots.
Reliability prediction and modelling.
Failure mode, effects and criticality analysis
(FMECA). Fault tree analysis (FTA) and quality
function deployment (QFD).
The use of
regression models to improve reliability;
residuals analysis.

Reliability processes and methods should be

studied with component and system reliability and
with the Weibull and lognormal distributions.
Relationship with FMECA and FTA should be
made. Knowledge of methods of dealing with
censored data is expected.

49

Option in Social, Economic and Financial Statistics

The aim of this syllabus is to develop knowledge and skills in utilizing published data that are generally
available to government statistical offices to study practical problems in social, economic and financial
areas. Candidates should be able to understand the purposes for which the data are collected and to
identify the limitations, if any, in using the data in an analysis. It is also expected that candidates should
be competent to recommend the type of statistical data that should be considered in studying a given
specific problem that is likely to have a social, economic or financial impact on society.
Statistics included in this syllabus are those that are most widely used to formulate strategies by
governmental and other public organizations, and in business and finance. These statistics are
categorised into three groups, namely, social statistics, economic statistics and financial statistics. Their
categorization will be, as far as possible, consistent with those being adopted by international
organizations, in particular the Statistics Division of the United Nations and the International Monetary
Fund. Students are required to have good knowledge of the appropriate methods of data collection and
applications of statistical methods, and the limitations of both.
This subject is an extra Option in addition to the existing six Options for the fifth paper (Options Paper)
of the Graduate Diploma Examination. The Options paper contains four questions on each of the
Options. Candidates are required to answer FIVE questions altogether; these questions must be taken
from TWO options, with not more than THREE questions taken from any one Option.
Four questions will therefore be set on the syllabus of this Option. The syllabus consists of topics in
social statistics, economic statistics and financial statistics.
SOCIAL STATISTICS
Population Statistics
Population Census, population pyramids,
population estimates and projections.

Working knowledge of methods of compiling

population estimates. The effect of changes in the
population composition on social planning and
business forecasting.

Vital Statistics
Birth rate, mortality rate, fertility rate,
standardization, life tables and
life expectancy.

Methods of conducting longitudinal and intercountry comparisons. The effect of changes in

birth rate/mortality rate/fertility rate on
population composition, social planning and
business forecasting.

Manpower Statistics
Estimation of labour force, labour force
participation rates, unemployment,
underemployment, nominal and
real wage indices.

Working knowledge of methods of estimation of

the number of employed, unemployed and
underemployed persons. Methods of construction
of wage indices. Applications in social planning
and economic forecasting.

ECONOMIC STATISTICS
Statistics on National Accounts
National income statistics with focus on
gross domestic product, gross national
product and real income measures,
e.g. real gross domestic income (RGDI).

Knowledge
of
relevant
concepts
and
methodologies of compilation and data
collection. Analysis and interpretation of
published official statistics and applications in
economic research and business planning.

continued on next page

50

Balance of Payments and International

Investment Position Statistics
Current Account (i.e. goods, services,
income and current transfers) and Captial
and Financial Account (i.e. capital transfers,
direct investment, portfolio investment,
financial derivatives, other investment and
reserve assets).

Knowledge
of
relevant
concepts
and
methodologies of compilation and data
collection. Analysis and interpretation of
published official statistics and applications in
economic research and business planning.

Merchandise Trade
Exports, re-exports, imports, trade indices
including terms of trade.

Working knowledge of methods of data

collection. Methods of construction of trade
indices. Applications in economic research and
business forecasting.

Production Statistics
Producer price indices for manufacturing
industries, index of industrial production.

Knowledge of concepts and methodologies of

compilation and data collection. Problems in
data collection. Applications in economic
research and business forecasting.

Price Indices
Consumer price index, wholesale price index.

Methods of construction of these indices.

Knowledge of concepts and methodologies of
compilation and data collection. Applications in
economic research and business forecasting.

FINANCIAL STATISTICS
Analytical Accounts of the Banking Sector
Money supply, linked exchange rate system,
domestic credit by public and private sectors,
reserve assets, external position.

Knowledge of concepts and methodologies of

compilation and data collection. Knowledge of
the nature of the statistics that are accessible to
the public.

Interest Rates
Short-term rates, long-term rates,
prime rate, inter-bank rate.

Knowledge of the nature of the statistics that are

accessible to the public.

Stock and Derivatives

Stock price indices.

Knowledge of the nature of the statistics that are

accessible to the public. Working knowledge of
methods of compilation. Business applications.

Index Futures and Index Options.

Knowledge of the nature of statistics that are

51

accessible to the public.

Fixed Income and Debt
Commercial and government bonds,
mortgage loans and backed securities.

Knowledge of the nature of statistics that are

accessible to the public.

Exchange Rates
Exchange rates of foreign currencies,
nominal and real effective exchange rate
indices.

Knowledge of the methods of estimation. Business

applications.

52