Method Overview
Observational research, also known as naturalistic observation, is a research technique that involves direct observation of phenomenon in their natural setting. Note that observation is not merely looking, but looking with a purpose. Quantitative forms of observation involve adhering to an objective set of criteria, while qualitative forms examine and analyze behaviours.
Types of observational research include:
1) Covert observational research (Naturalistic)
The researchers do not identify themselves. Either they mix in with the subjects undetected, or they observe from a distance. The advantages of this approach are:
- It is not necessary to get the subjects’ cooperation
- The subjects’ behaviour will not be contaminated by the presence of the researcher.
2) Overt observational research
The researchers identify themselves as researchers and explain the purpose of their observations. The problem with this approach is subjects may modify their behaviour when they know they are being watched. They portray their “ideal self” rather than their true self.
3) Researcher Participation
The researcher participates in what they are observing so as to get a finer appreciation of the phenomena. Researchers that participate tend to lose their objectivity.
In order to deal with reactivity (participants reacting to observers), researchers often either go covert, or conduct overt research using deception. This is controversial in the social sciences, as deception borderlines its ethical boundaries.
Procedure
- Determine the purpose of the research
- Develop standardized way of recording behaviour (see below)
- Observe subjects that are representative of the population
- Have the experimenter be someone unrelated to the experiment
- See link for guidelines
For conducting a quantitative experiment, the researcher is to begin with a clear sense of purpose and a well-developed coding scheme that allows for the reliable and valid accumulation of data. Coding systems can either be sign or category.
- Sign System – researcher waits, noting each time predetermined criterion behaviour occurs.
- Category System – creates a set of mutually exclusive categories (an exhaustive list) that each behaviour could possibly fall under. (i.e. Observing a child you might have categories called pro-social, affectionate, communicative, aggressive, etc.)
In general, the sign system tells how much of an action happened, whereas the category system can create correlations (i.e. Are pro-social greater during mornings? Are aggressive actions greater after watching Power Rangers?).
The main concern with determining coding schemes is defining them in such a way that different individuals can watch the same event and come to same conclusions. For example, if the experiment is recording the number of violent incidents in a TV show, we’re in big trouble if you count 3 and I count 27. Thus, the reliability has to be established before doing the actual study. Some simple ways of improving reliability include: having a coding system that’s clearly defined and non-overlapping, reducing the time between an event and recording its action, and reducing the number of categories.
Data Analysis
Guide your analysis of observational data by the study’s hypotheses. For example, if the goal is to assess improvement in instructors’ oral presentation skills, you might use classroom observations to evaluate these skills before and after oral presentation training.
Quantitative observation gives the very straightforward approach that allows you to quantify your conclusion (remember to factor in confounding variables).
A useful heuristic for quantifying reliability is:
| # of ratings that agree total # of ratings |
x 100 = % agreement |
Advantages
- Observation allows for a great deal of flexibility
Instead of designing a restrictive experiment or survey, you can record relevant data as it naturally occurs
- Observational research is not subject to the Hathorne Effect (subjects modify an aspect of their behaviour that is being experimentally tested)
The makes observational research significantly more valid than other forms.
Disadvantages
- Observational research runs the risk of being unreliable. For example, when John records his perceived action of Cindy, he may record Cindy’s emotions/feelings differently than Stephanie does.
This can be minimized if researchers are unbiased and objective in their observations, strictly adhering to the coding system.
Recommended Usage
Observational research is an unlikely choice for SERA’s mainstream projects, but could be used in the future. Some instances when it could be useful are when sensitive issues are involved, studying student or teacher interactions, and when the Halthorne Effect is present. Also, it can be used for improving the building project. For example, observational research can be used to examine the use of a printer in a bank: the observer notices how, due to the printer’s sloped surface, workers are using make-shift stands or making extra trips. Observational research can be used for problem solving.
