Sociology and Experimenting

Is Sociology a Science?

This is an important question, because a science is understood to have the ability to hold and isolate variables (usually using control conditions in a laboratory) and to test another variable.
With only one variable tested at one time, in each experiment, and to have a control group where the variable is held stable, the potential is there to obtain the same result every time from the same test - reliability.
Psychology sometimes tests individuals in laboratories and keeps unwanted variables stable. Sociology cannot do this, because of its group and collective perspective in what happens in the real world. Sociologists must study where people are.


Sociologists can still experiment. This is then a field experiment. Sociology cannot put people into the laboratory, because it is interested in what happens in society, either in systems or people interacting. This means any sociological "experimentation" is carried out in the field. The problem is that as one aspect of human society is tested, other variables will vary and cannot be isolated, thus affecting the results as people and institutions do different things at different times.
Nor is it possible for sociologists to create mathematical models and diagrams of perfect behaviour as even economists do. Sociologists never assume behaviour nor suggest one behaviour is normative: sociology is humanistic and people orientated.
People often treat a laboratory as a place in which they should co-operate with the research and give the answers wanted. Sometimes sociologists in the field find that people give the answers wanted. If respondents are in a situation of stress or discipline, they may show positive feedback to the researchers. In fact the results may look as if they respond to variables, but might be the result of researcher impact known as the Hawthorne Effect.

What is the Hawthorne Effect?

So sociology is not a science then?

This is named after the Hawthorne Works of the Western Electric Company in Chicago. There were four studies into productivity and the working environment (pay, light, rest breaks) carried out in 1924 up to about 1933 (into the depression years) with small groups of workers. The main definition most often quoted is that the Hawthorne Effect is a response of the people being researched not to the variables under experiment but to the fact that they are in the presence of researchers. This is the interpretation of Elton Mayo (1933) in The Human Problems of an Industrial Civilization, New York: Macmillan, chapter 3. There are other explanations, including those of more variables at work regarding learning (positive feedback regarding skills) and expectations in the future (that could negate positive effects in the present regarding piece rates). H. M. Parsons realised that a learning circle had been happening so that that outcomes altered the workers' behaviour and the researchers had missed this. See Parsons, H. M. (1974), 'What Happened at Hawthorne?' Science, vol. 183, 922-932. This does not include the experimentation on light levels.
So the Hawthorne Effect itself is something of a myth. Rather than the effect of the researchers and gaining positive results regardless, there were positive outcomes to more and unnoticed variables. There are implications for education theory as much as management study, with longitudinal implications too of feedback loops. Management may also have means of improvement - by groups interacting, with supervision providing a positive feedback in an advantageous environment with workers believing that they have some control over their own outcomes. That research effects were missed suggests inadequate research control. Complexity is important, surprises are likely to happen and must be seen and accounted for. It shows the dangers of doing experimenting, and that experimenting is messy!
The implication is that it is messy. It lacks rigour to be scientific. Many a sociologist goes to great lengths to explain the research method, problems encountered, what could be missing, what new research might follow, and to look for unusual and unexpected outcomes.
There are broader issues that compare sociology with science. Although humanistic sociology can share with science one approach of falsifiability and deduction. These ideas derive from Karl Popper and his book The Logic of Scientific Discovery.
Unlike the inductive method, which is to gather lots of data and see what theories emerge, the deductive method is to have a theory first and test the theory through evidence. This is where sociology can be like science.
A theory is not tested by what may agree with it, but is tested by what evidence may disagree with it. If evidence is found that disagrees with it, then the theory has been falsified. Once it is falsified, the theory is dead. In reality, however, a falsified theory is put to a number of further tests to confirm it is dead.
When a theory is prepared for testing, it is called a hypothesis. A theory is hypothesised so see if it is true (until falsified).
A hypothesis might come about because of some hunch, or a moment of inspiration, or even a suggestion of some evidence beforehand. There might even be a pilot study to develop a hypothesis for bigger testing - the pilot study will lack the numbers researched for statistical reliability. A hypothesis might come about through an inductive method. How the hypothesis comes about is irrelevent. What matters is that it can be tested towards falsification.

Difference with Positivists:
Users of the Deductive Approach

Falsification involves time. If a theory has not been falsified, it stands as true for now. However, any theory may one day be tested and falsification be found. Whereas positivists claim permament laws of human behaviour (starting with Comte), those who follow deduction and falsification realise that any theory that seems true is only true in a transient or temporary sense. Conditions may change.

Softly Softly Sociologists

However, whereas scientists deliberately set out to experiment by changing a variable and seeing what happens, many sociologists have a hands off approach to experimenting. They realise the pitfalls too well and realise it is far better to see what happens in the field with as little effect from the sociologist as possible. They may still use a deductive approach, but they let social actors themselves do the actions and sociologists just find out. These sociologists research passively; a sociologist who by his or her changable presence or suggestiveness alters the field of human activity is arguably weakening the sociology and the research findings. The cost of always so many unknown variables changing, means that a sociologist should be disciplined in a way that disturbs the least.
In general, sociologists are practical people. If they have a deductive theory, but research is suggesting a new theory, they are not above using the inductive method to let the data shift focus and alter the hypothesis. The adjusted or even new hypothesis can be further tested - tested by least disturbance. Social research often leads to new questions.

Users of the Inductive Approach

Indeed many a sociologist interested in the micro-phenomenological appraoch of small groups may rely on quite a bit of the inductive method. Unstructured interviews, that let the interviewee talk on their way, and participant observation, in which the group creates its own agenda, involve inductive approaches. A very valid research result of human behaviour can result from inductive methodology, so long as the sociologist keeps in the relative background of events, but a resulting hypothesis may have to be tested by other more reliable means deductively.
A macro-sociological approach will have to rely on deductive methods like questionnaires and structured interviews and making a quantitative result across large numbers of those researched.

Grand Theories

A theory is not just a hypothesis but it also potentially leads to a perspective, or a grand theory. We live by perspectives and the world views that build on these collections of theories. Should significant theories towards the grand theory be falsified, then the world views they sustain will falter. Sociologists do a lot of theorising and explaining between researching. They also, in their literary endeavours, use words like may, could, perhaps and surely more often than a scientist.
Grand theories or world views include Marxism, functionalism and feminism (and more). Usually these perspectives have a robust life, and are never fully falsified. But a number of detailed hypotheses on which they depend may get falsified. It remains very difficult to falsify grand theories because they protect themselves through smaller theoretical development and making adjustments: most Marxists do not follow the precise views and predictions of Karl Marx any more. Arguably Marxism, with the fall of communism in Europe, seems weakened as a grand theory because it predicted that it would supersede capitalism and it has not (yet) and perhaps has gone ito reverse. In science, evolutionary theory has absorbed genetics and mutations whilst evolution remains a dominant grand theory.

Sociology - a Social Science

So sociology is a social science (so is economics, even though it sometimes models perfect behaviour; and so is Politics, as it relies on human interaction in political action). Sociology can never be as reliable as science, therefore, but it can use the deductive method and draw on research for this purpose.


  1. Where do sociologists experiment?
  2. In what way does sociology differ from being a science?
  3. Which methodologies equate to more inductive research methods?
  4. In what way is sociology said to be like a science?
  5. Define a hypothesis.
  6. Does failing to falsify a theory through experiment or other research prove it is true?
  7. Is the Hawthorne effect any less useful even though it may be based on misinterpretation? Give reasons.

Popper, K. (2002), The Logic of Scientific Discovery, first published 1959, London: Routledge.


Adrian Worsfold


Last updated on April 08, 2006