Official statistics are quantitative data gathered by the government or other official bodies.
Examples include statistics on births, deaths, marriages and divorces, exam results, school exclusions, crime, suicide, unemployment and health.
The ten-yearly Census of the whole UK population is a major source of official statistics. Examples of many other official statistics can be found on the website of the Office for National Statistics.
There are two ways of collecting official statistics:
Registration – for example, the law requires parents to register births.
Official surveys, such as the Census or the General Household Survey.
They are a free source of huge amounts of data. Only the state can afford to conduct large-scale surveys costing millions such as the Census. Sociologists can make great use of this data for their research saving both time and money.
Statistics allow comparisons between groups. For example, we can compare statistics on educational achievement, crime rates, divorces etc between different social classes, ethnicities, religions, gender etc.
Because official statistics are collected at regular intervals, they show trends and patterns over time. This means sociologists can use them for ‘before and after’ studies to show cause-and-effect relationships. For example, we can compare divorce statistics before and after a change in policy to measure what effect the new legislation has had.
The government collects statistics for its own purpose and not for the benefit of sociologists, so there may be none available on the topic the sociologist is interested in. For example, Durkheim found that there were no official statistics available on the religion of suicide victims (presumably because the state had no use for such information). However, this was crucial for Durkeheim’s study about suicide, religion, and social integration.
The definitions that the state uses in collecting data may be different from those that sociologists would use. For example, the government might define poverty different to how the sociologist might.
If definitions change over time, it may make comparisons difficult. For example, the official definition of unemployment changed over 30 times during the 1980s and early 1990s – so the unemployment statistics are not comparing like with like.
Because official statistics often cover very large numbers (sometimes even the entire population) and because care is taken with sampling procedures, they often provide a more representative sample than social surveys completed with more limited resources available. They therefore may provide a better basis for making generalisations and testing hypotheses.
Official statistics are generally seen as a reliable source of data. They are compiled in a standardised way by trained staff, following set procedures. For example, government statisticians compile death rates for different social classes following a standard procedure that uses the occupation recorded on each person’s death certificate to identify their class.
Official statistics are therefore reliable because, in principle any person properly trained will allocate a given case to the same category.
However, official statistics are not always wholly reliable. For example, respondents – such as the public – may complete forms incorrectly or the coders of the data may make errors.
A major problem with official statistics is that of validity. Do they actually measure the thing that they claim to measure?
Hard and soft statistics: Some ‘hard’ official statistics do succeed in doing this. For example, statistics on the number of births, deaths, marriages and divorces generally give a very accurate picture.
However, ‘soft’ statistics give a much less valid picture. For example, police statistics do not record all crimes. Similarly, educational statistics do not record all racist incidents occurring in schools. Attempts have been made to compensate for the shortcomings of police statistics by using self-report or victim studies to give a more accurate picture of the amount of crime. However, the validity and reliability of these can also be questioned, what one person determines is a crime might be dismissed by another.
Because of this certain types of official statistics (‘soft’ statistics) should be approached with caution. Patterns can be identified and generalisations can be made but these can be open to criticism due to the validity and accuracy of the statistics.
When sociologists want to study the past to look at changes over time, they need to use historical documents. Many sociologists have used historical documents in their research:
• Weber – used historical speeches and writings in ‘The Protestant Ethic and the Spirit of Capitalism’
• Laslett – used parish records in his investigations into changes in the family structure with industrialisation.
• Anderson – used census statistics for the same purpose as Laslett
• Durkheim – used suicide statistics to look at changes over time
They are the only way to obtain information about the past.
They are important when studying changes over time that cannot be revealed by primary data.
An understanding of past events is important in understanding current society.
Scott gives 4 criteria for assessing secondary sources that apply to historical documents:
1. Authenticity – is the document genuine? Are the pages complete? Is it an original or a copy? Who wrote it? In 1983 the Hitler Diaries managed to fool many historians before they were labelled a forgery. If the document is not genuine then there are problems with validity, i.e the historical document is not a true record therefore it should not be used in research.
2. Credibility – is the author sincere in what they write, or has it been distorted to deliberately mislead people? Many politicians, for example, may distort accounts of their actions in their memoirs to give a better impression. Many historical documents are not objective, are one-sided, biased and concerned to put across one particular viewpoint.
3. Representativeness – How representative is the document of other documents of the time? This is especially important for historical documents as many have been lost or destroyed. For example, Laslett only used parish records relating to certain villages that had complete records, it is therefore difficult to accept any generalisations made from his findings. Some official documents are kept for 30 years before they can be seen by the public, whilst others are never released. If the document is not representative this will cause problems as the findings from any research cannot be generalised.
4. Meaning – this refers to both the literal meaning of a document – i.e. it may be in a foreign language, or old handwriting etc, and the understanding and interpreting of the underlying meaning of the document. It is difficult to interpret historical documents as they are from a different era, possibly from a different culture, and the person who wrote them may be dead. How can a sociologist know that their interpretation of a document is the same as that intended by the writer? Added to this problem is the fact that different researchers will interpret the same document in different ways depending upon their background and views, e.g. Berger argued that 17th and 18th century paintings show that people were concerned with material possessions. He saw this as linked to capitalism. However art critics disagree with his interpretation.
These documents record details of a person’s life experiences and are mainly qualitative and favoured by interpretivists, as they allow the uncovering of meanings. They can be historical or modern. Examples include diaries, letters, photographs, memoirs.
Examples of the use of personal documents in research
Young and Wilmott’s study of family life in London asked subjects to keep diaries.
Oscar Lewis’ research into poverty in Mexico also made use of diaries.
advantages and disadvantages
For interpretivists, qualitative documents are valuable as they uncover the meanings of those involved.
Again, from an interpretivist point of view this makes them high in validity.
They allow sociologists to get close to the real people rather than just dealing in abstract theories.
For interpretivists, personal documents are the best alternative when participant observation is not possible.
Positivists, who favour quantitative research, tend to see personal documents as unscientific.
They are difficult to obtain.
There is the problem of representativeness, are the ones that exist representative of the population?
Interpreting meanings can also be a problem – how do we know the sociologist and writers meanings are the same?
Credibility – some are written with an audience in mind, e.g. autobiographies and letters – this questions their validity.
Authenticity – who wrote the document? Letters that appear to have been written by one person may in fact have been written by someone else.
Content analysis is used to study media texts.
This method is seen as objective and reliable, and involves counting to produce statistical data. The researcher has to work out a coding scheme – the criteria whereby subjects are coded in terms of belonging to a particular category, e.g. the number of males and females shown as housewives, males and females shown in paid work Once the categories are decided, the researcher then studies their chosen media text and counts the frequency with which their chosen subjects appear.
This type of method would appeal to positivists due to its objective nature. The researcher is not involved in analysing meanings, and so stays objective. Also if a representativesample of media texts is used, then generalisations can be made from the findings.
Examples include Lobban’s study of the portrayal of gender roles in textbooks. She listed and counted certain things, such as: activities children were engaged in, skills they had, roles of adults etc.
2. It can be checked by others, so it is reliable.
3. It claims to be scientific and objective.
4. If a representative sample of text is analysed then generalisations can be made.
5. It gives some indication of the importance of certain themes by counting the number of times they recur.
1. It says little about the meanings of a media text – either for the producer or the audience. They only tell us how frequently a particular image occurs. Therefore it would be criticised by imterpretivists for lacking depth.
2. Content analysis is not as objective as it claims to be. The researcher decides on the categories which they choose to count, and the researcher may decide to be selective in their counting, unconsciously ignoring material that does not fit in with their hypothesis. Therefore the research will be invalid.
3. It can be very complicated in terms of choosing categories and time consuming to study media texts in this way.