Observation

A lot of different types and sub-types of observation have developed, and now it has the most complex structure of all data collection procedures.

1. Cross-sectional Studies

A cross-sectional study is a type of research that captures information from a population at a single point in time (Figure 1). The information can be obtained from a group of participants with certain characteristics known as variables (e.g. age, gender, origin, education, religion, geographic location, etc). The variables used in a particular study are determined by the research to be carried out, and by the goals of the study. Researchers can use cross-sectional studies to focus on one independent variable, and see how it affects one or more dependent variables. These studies enable researchers to examine several variables at the same time. Cross-sectional studies do not involve the manipulation of variables. They are descriptive studies.

Some examples of cross-sectional studies are given below:

• Descriptive study: conducting market research by collecting data on consumer behaviour within the given period of time with an aim to predict the future trends of consumer behaviour or introduce new sales strategies or launch new products and services. In such a study, the researcher does not analyse the causes, but focuses on the description of the current situation on the market.
• Analytical study: a study examining the lack of calcium in the population of women above 50 years of age. Whereas the focus of the descriptive study would be to determine the level of calcium in women above 50 years of age, the analytical study would focus on examining the factors such as hormonal status, physical activity, nutrition, socio-economic status, education level, etc., in order to explain why some women belonging to such an age group are more likely to develop calcium deficiency than other women, and therefore have a greater probability of developing other illnesses such as osteoporosis.
• Case-control Studies: Case-control studies are a type of observational research often used in the fields such as medical research or environmental health. A case-control study is an experimental design that compares two or more groups of participants. One of these groups is the ‘case’ group, and other groups are ’control’ groups. It is important to note that the case group is selected because it is already characterised by the attribute of interest for the study, and the purpose of the control group is to find out if the case group systematically exhibits that attribute more than the control group does. Therefore, by analysing the data collected from the control group, the research hypothesis is either confirmed or rejected. Commonly, tables are used to present data (Table 1).

The risk ratio (or rate ratio) is calculated by using the exposure odds ratio (OR):

• The odds of exposure among cases: OR = a/c,
• The odds of exposure among controls: OR = b/d,

e.g. OR = (axd)/(cxb).

If OR = 1,  then the odds of risk is the same for the exposed and non-exposed.

If OR > 1, then the exposure increases the odds of risk.

If  OR < 1, then the exposure reduces the odds of risk.

Example: Сase-control study in medicine

The aim of the research was to examine the connection between the contamination of potable water and potental stomack diseases (e.g. gastritis). The case-study group comprised the people diagnosed with a stomach disease, whereas the control group comprised the people without the disease. During the research, the data on the exposure of the participants to the contaminated potable water were collected, with the focus on the variables such as the source of the water, and duration of the exposure, for each group. The aim was to compare the results in order to determine whether there was a correlation between the water contamination and the risk of developing a stomach disease.

Practice tasks:

Researchers are conducting a case-control study of breast cancer using incident cases. The study includes 100 cases and 100 controls. 70% of the cases report smoking. Among the controls, 50% report smoking.

a) Prepare a 2x2 table with the given data.

b) Calculate the exposure odds ratio.

c) Interpret the exposure odds ratio in a sentence.

Answer:

a)

b) Exposure odds ratio = (a/c)/(b/d) = (a*d)/(c*b) = (70*50)/(50*30)=2.33

c) An odds ratio of 2.33 means that the odds of smokers being a case are 2.33 times the odds of non-smokers being a case.

Cohort studies are a type of analytic studies used to test the hypotheses about cause-and-effect relationships. The term ’cohort’ is defined as a group of people, usually comprising 100 or more people who share a common characteristic or experience within a specified period of time (e.g. age, occupation, exposure to a medicine or vaccine, citizenship, etc.). A cohort is divided into 2 parts – those exposed to the supposed risk factor, and those who are not (e.g. smokers and non-smokers). Then the cohort is observed over a certain period of time in order to find out the incidence of death or disease development (which are supposed to be caused by the exposure to the present risk factor), and at the end of the observation period, the comparison between the incidence of death and illness in the exposed and unexposed participants is performed.

Cohort studies are an important method of medical research, convenient for the identifiction of the causes of a disease, because groups of people are observed before they develop a disease. It means that researchers can examine whether there is a cause-and-effect relationship between the participants’ selected way of life and their health (Figure 2).

The ongoing Millenium cohort study is following the lives of 19,000 babies born in Great Britain between 2000 and 2001 in order to demonstrate how circumstances in the very first stages of life can influence later health and development. In addition to collecting the data about the health of these babies and their parents, the study is examining the behaviour and cognitive development of the children, as well as a number of other social factors.

The design of a cohort study is shown in the following table:

• Incidence rates (IR):

Among the exposed: IR = a/(a+b)

Among the unexposed: IR = c/(c+d)

An IR is the indicator of the occurence of an illness in the exposed and unexposed cohort members over the total period of time.

• Relative risk (RR) – the ratio between the risk of disease in exposed individuals and that in the unexposed:

[a/(a+b)]/[c/(c+d)] or a(c+d)/c(a+b)

The relative risk shows the incidence of an event in the exposed over that in the unexposed individuals.

• Attributable risk (AR)=[a/(a+b)-c/(c+d)]/[a/(a+b)]*100 is the difference in the incidence rates in the exposed and unexposed individuals over the incidence in the exposed.

The attributable risk shows the probability of disease prevention provided there is an effective measure of eliminating the exposure.

Naturalistic observation is a method of observing how people or animals behave in their natural settings. It is a kind of field research, which means that researchers collect data outside the laboratory or clinical setting. Researchers tend to immerse with the environment as much as possible so as not to affect the participants’ natural behaviour in any way.

Based on the position of the observer, observation can be participant and non-participant. The participant observation is the observation when the researchers immerse themselves in the observed phenomenon they are collecting data about. The observer has to integrate with a social group, and take part in their actions. This type of observation is used for two reasons. The first reason is the fact that a lot of groups, such as religious sects, elites, secret associations, are closed and inaccessible to observation. So, in order to obtain the necessary data, the observer has to become part of the group, and ’lose’ their identity, i.e. hide their role as the observer. This leads to a lot of ethical and technical problems, which almost raise doubts about the possibility of conducting covert observation.

The second reason why participant observation is used is that it helps to overcome the difference between the observed object and dominant subject, thus introducing the dialectic action into the research. Participant observation revolutionises the relationship between the observer and the observed so that all the subjects become one phenomenon. Therefore, it is not surprising that the action research has accepted this type of observation as important for achieving not only a specific scientific, but also a wider social objective – the change of tradional research. The abovementioned ethical problems are eliminated in this case because the observer’s performance is public. However, the methodical and technical difficulties, such as objectivity and immersion with the group persist, but if the traditional research method is neglected, then these problems disappear too. There is also non-participant observation. The non-participant observation implies observing the participants without actively participating. So, the observer enters a community or social system involved, but stay separate from the activities being observed.

According to the number of the observed cases, the distinction between single-case and multiple-case observation can be made. Single-case observation is the observation of a single case or a single type of a phenomenon. The systematic collection of data on a single case has been termed ’Case Study’. The aim of this procedure is to explain an individual reaction within a collective situation and therefore it is also called the development of personal documentation. It has developed into a separate research method. Closely related to single-case observation is clinical observation, which has developed in medicine and psychiatry. It is aimed at establishing a direct contact between the observer and the patient on the one hand, and obtaining thorough information about what affects an illness, on the other hand, i.e. which complex individual and social factors cause a certain mental condition of the patient.  On the other hand, in the multiple-case observation, several cases or groups of the same or different phenomena are the object of observation. It implies registering facts using numerous examples owing to which the structure of phenomena, and even their development are explained. This data collection method is similar to creating social registers on the one hand, and to the mass registration of population – a census, on the other hand

Based on its duration, observation can be snapshot and longitudinal. Snapshot observation is used at a specific moment, which is usually crucial for the development of a process and recorded in a similar manner as a photo is taken. The aim of the snapshot is to spot the main factor that determines further development of a series of actions caused by it. All dynamic activities are observed in this way, because it allows for discerning the important, and eliminating the unimportant characteristics of a phenomenon. Longitudinal observation is used in different continuous sequences of a process with an aim to spot the changes over a longer period of time (Figure 3). Social processes last longer, and take place over a longer period of time and therefore such an observation procedure is important for discovering the dynamics of all events, and the axis they evolve about.

Structured or methodical observation is the observation performed according to a pre-defined plan, and the plan is important because it ensures the insight into those facts that are within the observer’s scope of interest. The observation remains ‘elastic’, but deeper insight into the key data or knowledge thoroughly describing a phenomenon is obtained.

All the above-mentioned observation types and subtypes have different applications depending on the nature of the specific research. It is this nature that almost entirely directs the observer as to which procedure to select. This, of course, depends on the researcher’s intuition and imagination, as well. Therefore, it is very important to be familiar with the specific types of observation, as well as with its scopes and limitations.

Due to the objectivity of the five senses of the researcher, observation is an extremely objective research method. It seems to require little or no interaction between the researcher and those being studied (Angrosino, 2007, p. 37). This procedure does not disturb the natural, continuous flow of events at all or it does so to the smallest extent. The natural dynamics of the procedural unfolding of an event, and all the important elements of the event, are successfully identified by means of observation. Observation can also be directed to some crucial moments, which can be isolated, and thoroughly studied. But more than anything else, observation, and especially its participant forms, directly establishes communication between people, and opens up the closed groups of people, bringing some changes to them that did not exist before (Pečujlić, 1982, p. 105).

In spite of being a very fruitful data collection procedure, observation has some significant limitations, which relate both to the research method, and the observer as a central figure in the procedure. Observation is a long-lasting and complex process. Even when thoroughly planned, it depends on the so-called dispersion of senses. Namely, due to the long-lasting process of observation and its complex organization within a place, senses cannot detect all the facts. So, the unconscious process of suppressing some facts with others, or suppressing old facts with new ones, inevitably takes place.

The success of keeping structured and organised field notes depends on a thoroughly prepared plan for note taking, but even then, if there is not enough time, it is difficult to record all the data. So, this is usually done after the observation, but then it loses the elements of real observation and becomes recollection, which implies that some data processing has been done in the mind that interprets the facts, thus revealing the deceptiveness of human senses. A note taking plan can help to solve the problem, but cannot completely eliminate it.

The observer as the central figure in the observation process is limited by the deceptiveness of human senses, but also by his or her own skills, because skills differ with different people. When it comes to observation, they differ so much that the same two people will not perceive the same fact in the same way. Most people are in fact persuaded to see what they want to see. A series of the observer’s mental characteristics such as speed, precision, mental ability, recollection, mood, together with a series of social characteristics such as sociability and talkativeness, form the so-called ‘personal equation’ of the observer, which affects the observation, and can distort the originality of facts. Bias is even more dangerous than this latent danger accompanying observation. Very often the observer distorts the accuracy of facts not only unconsciously, but quite consciously as well. The traditional methodology considers this not only a serious limitation, but also the breach of the scientific rigour of the procedure. In the dialectic action research model, this is not a problem, and the bias itself, which breaks through the established ideological veil, is the precondition for reaching objectivity and truthfulness. Bias serves as a social engagement that provides a path to good observation (Pečujlić & Milić, 1995, p. 107).