Let

1
Subject Matter Issues in Statistics, and Ethics
Subject matter stands for looking into topical
issues, themes and domains
2
This lesson is about subjects

• Now we want to inform about how people influence
  statistics.
• The key words are subjects and subject matters
  and subjective.
• Subjective are views that are strongly influenced
  by the opinion of one person. It is a personal
  opinion.
• Inter-subjective is an opinion that is shared
  with many other persons/professionals.

3
The nature of subject matter issues

• Subject in the philosophical sense refers to a
  (human) being that has subjective experience and
  consciousness. It is a word for humans, and for
  many other issues.
• Subject matter, in general, is anything which can
  be studied , described and analyzed.
• The nature of human(s) is to give meaning to
  their lives. A subject is expressing its own
  views by acting, behaving, or expressing
  opinions. This can also be called the inherent
  subjectivity of humans.

4
More on subject matters

• Subject also refers to an area of study or
  research. In research we can speak about the
  subjects that we study. That can be humans or
  other subjects of interest.
• The subjects and their behavior are normally
  the objects of the statistical research.
• Another meaning for the word subject subjects
  are those who are ruled by rulers.

5
The nature of objects

• Objects are no humans, they are things.
• These are mostly physical things that people
  (humans) move around.
• To complicate
• An object can be a topic of study/research.
• Objects, as a topic for research, can become a
  subject matter for study purposes, when humans
  start to study it.
• And subjects can be also topics that can be
  studied.
• To summarize When humans study a topic it
  becomes their subject matter issue.

6
The human subjects behavior

• When humans travel they make use of the transport
  systems.
• Subjects decide the means of transport and how
  they use it trains, directions, frequency and
  distances.
• When we measure the use of transport we look at
  what people are doing.
• Subject matter refers to a matter presented for
  consideration in discussion, thought or study.
• Other words for subject matters are topics,
  themes and domains.

7
People (subjects) make statistics

• Producing statistics is a work done by humans.
  Therefore humans are always part of the outcomes
  of statistics, because they influence the
  results.
• In other words. The statistics you get are also
  the result of the humans that report about them
  and that work on them.
• That fact is influencing the quality of the
  source of the basic data.

8
We speak about subject matter and
inter-subjectivity.

• Subject matters
• When it is about humans (their characteristics)
• About their behaviors (labor, education, health)
• About the things they influence (production,
  nature)
• About the things they move around (goods, etc.)
• In science we speak about inter-subjectivity when
  a group of professionals come to the some
  conclusions.
• The more experts share conclusions the more we
  may assume that that opinion is correct. But
  that is not always true. Scientists need to stay
  alert.

9
How to define statistical topics

• In the VSS we use a classification for the
  subject matter themes
• Social and demographic.
• Economic statistics.
• Environment and Multi-Domain Statistics.
• Methodology.
• Strategy and Managerial issues of Official
  statistics.

10
How to classify subject matter issues

• When we define differences we do this on the
  basis of criteria.
• Experts can have different opinions about these
  criteria.
• Social and demographic is the behavior based on
  the actions of individuals and households.
• Economic behavior is mostly the behavior of
  enterprise units.
• Environment is the behavior of individuals and
  households and enterprise units and government
  that effect the environment.
• Multi-domain issues are a collection of topics
  that were not considered in one of the three
  groups above.
• Expert opinions can differ. Some see poverty as a
  social issue, other might see it as an economic
  issue. Therefore it is classified as a multi
  domain issue.
• Classifications are extremely important, they
  exist in all themes on all subject matter topics.
  

11
Social statistics

• Social statistics are about individuals or
  households with regard to their actions, and what
  and how they do it.
• Time use is the most personal manner of acting.
• Different categories of work time use can be
  defined paid and unpaid, formally and
  informally, unpaid for the own family, unpaid for
  friends and neighbors, unpaid for organizations.
• Other categories are personal time use, cooking
  and eating, etc.
• For each topic we need classifications in order
  to describe that type of behavior.

12
Classifications are key in statistics

• Before we make produce statistics we use
  classifications to categorize the population we
  want to describe, as well as its behavior.
• For demographics age and sex are key.
• For education types of schooling.
• For labor types of professions.
• In economics the units (firms are classified by
  type of unit like economic activity and size).
• Other classifications are about products and
  trade.
• Each topic has its own set of classifications.

13
Use of classifications

• Classifications are used to analyze the
  populations. Therefore a combination of
  classifications are used.
• Classifications need to be based on some
  principles and criteria.
• Different classifications can be used to serve
  different purposes of research. For certain
  topics a range of classifications is used.
• When classifications need to be related to each
  other we speak about coordination.

14
Types of classifications and VSS.

• Economic activities.
• Products produced and services delivered.
• Products, goods and services traded.
• Types of government spending.
• Types of firms and changes of status of firms.
• Types of peoples and households.
• Types of education.
• Types of jobs.

15
Statistical Coordination

• Statistical coordination takes place within and
  between agencies.
• We have institutional and technical coordination.
• Government work in statistics is enhanced by the
  possibilities of information technology.
• And by structuring and harmonizing the content,
  methodologies, dissemination formats and
  documentation.
• One unit in the system needs to take the
  leadership role in the aspects of nationwide
  statistical coordination.

16
Institutional coordination

• Institutional coordination is needed to make sure
  that the various statistical units in own country
  are able to work together.
• This is reinforced by creating arrangements like
  stats councils, working groups and committees.
• The range of topics that need to be coordinated
  can even include international affairs. In that
  case a separate unit should be involved to
  coordinate these relations. Also in-country
  training needs coordination.
• Other form of coordination is the regional
  coordination between neighboring countries.
  (South-South)

17
Technical coordination

• Technical coordination is needed to make sure
  that the various products of the statistical
  system can be related to each other when needed.
• Technical coordination can best be addressed by
  subject matter working groups or other technical
  working groups.
• These groups can discuss the national and
  international standards that these units need to
  adhere to.
• They deal with concepts, methods, classification
  and technology.
• Especially around the National Accounts it is
  needed that all data that flows into the system
  is well coordinated and harmonized.
• Harmonization is making data fit to be compared
  and related to with other data.

18
Harmonized statistics

• Harmonization is the adjustment of differences
  and inconsistencies among different measurements,
  methods, procedures, schedules, specifications,
  or systems to make them uniform or mutually
  compatible.
• In statistics we adjust data sets to deal with
  inconsistencies for instance based on the use of
  different definitions or different units.
• In statistics we make data sets comparable so
  that they can relate to each other for accounting
  purposes.

19
System definitions

• (1) A set of detailed methods, procedures, and
  routines established or formulated to carry out a
  specific activity, perform a duty, or solve a
  problem.
• (2) An organized, purposeful structure regarded
  as a whole and consisting of interrelated and
  interdependent elements (components, entities,
  factors, members, parts etc.). These elements
  continually influence one another (directly or
  indirectly) to maintain their activity and the
  existence of the system, in order to achieve the
  goal of the system.

20
Statistical system

• An organization of the relations between a well
  defined set of entities.
• An institutional statistical system is defined
  by the units that belong to the system and the
  relations that they have among each other.
• A conceptual system. All units that are described
  by the statistical and the relations that are
  defined between them.

21
Ethics in statistics

• Ethical behavior is human behavior per
  excellence.
• Statistics are based on trust. Trust is about
  believing that statements that are made are
  correct within the limits that exist.
• In statistics there are two kinds of trust.
  First, the trust with the providers of raw data
  that this information only will be used for
  statistical purposes.
• Second, the trust with the users that they can
  believe the correctness of the results, based on
  methods used and the meta data.
• Statistics is the treatment of data sets that may
  present only a portion of the population that it
  describes. The essence of statistics is making
  statements based on limited observations.
• Only image, skills and transparency can make that
  people trust that type of work.

22
Promoting trust with data providers

• Statistical organizations need to have legal
  frameworks that stipulate how these organizations
  have to behave when asking for data from data
  providers and using that data.
• Data needs to be treated in a very confidential
  way.
• Staff need to be faced with sanctions when there
  is a breach of these confidentiality rules.
• Confidentiality rules also apply to publication
  of the data. The data has to be anonymized.
• Individuals and firms should not be able to be
  identified as such in micro data sets and in
  tables.

23
Promoting trust with the users

• Products of statistical offices need to be
  trusted by the users, in order to be used.
• The trust should first of all be in the staff of
  the organizations.
• Staff should be skilled and trained to apply the
  methods that are expected to be used.
• Staff should demonstrate in their work that they
  master these skills. In articles and
  publications.
• Staff should produce meta data which informs
  users about the key features of the data
  presented and the methods applied.
• Staff needs to be transparent in their work to a
  high degree.
• Lack of transparency, of skill and of image will
  lead to a lack of trust an a lack of funding.
• Lack of trust may harm statistics.

24
The underlying assumption on which trust is based

• Trust is based on the assumption that there is a
  truth that we can know and we should look for.
• This trust in the existence of truth is
  universal.
• When there is no truth possible there is no need
  to make statistics. Or we have biased statistics
  by definition.
• There is a truth in the natural sciences and in
  social sciences.
• We can learn more about this truth by applying
  the right methods. Methods learn us to approach
  the truth.

25
More on truth

• The truth only can be reached if we assume that
  all information collected is the correct
  information.
• Since humans provide information and can make
  errors and misreport for various reasons, we can
  only approach the truth.
• By definition in social statistics we can never
  be completely sure that we reach the absolute
  truth.
• The aim of the methods we apply is to come closer
  to the truth, knowing that we never can be
  certain.
• The methods we use have there own uncertainties.
  Sampling has sampling errors and confidence
  intervals. Sampling also has non sampling errors,
  because people err.

26
Truth and probabilities

• Even if we are not certain to reach the absolute
  truth, we can create evidence that makes it
  probable that we are approaching the truth.
• When we have a concept and we know that a feature
  exist we can look for the appropriate number.
• If no other methods exist people can make
  guesstimates.
• With methods we create evidence based
  information.
• When methods are correctly applied we create
  statistical facts.

27
Statistics and facts

• Facts are statements to be believed unless proven
  that these statements are wrong.
• Statistical statements are more like strong
  evidence than hard facts.
• That is why we speak of evidence based.
• Pure statistical statements can by definition be
  wrong.
• But most statistics should be considered to be
  the best possible approach towards the truth.