The Observational or Experimental Method

Observation is based on the ability to perception and decision of the human being. There is an ordinary observation, which constitutes an important source of information in everyday life, and an observation scientific, which provides objective, valid and reliable knowledge to answer a research question raised.

The reliability of the observational method refers to the quality control of the data that informs us if there is a coincidence in the judgments issued by the same observer in the same situation observation but at two different times or by two different observers who independently face the same situation of observation. observation.

Index

1. Categorization of the observational methodology
2. Research planning
3. Measurement and measurement scales of the observational method
4. Reliability and validity
5. Data analysis and design
6. Observation metric

Scientific observation is a way of capturing reality that can be applied rigorously and systematicity, and that ultimately makes it possible to collect relevant information in a study scientific.

• As a data collection technique: it can be involved in any type of research design. For example, it can be the RV measurement form of any design.
• As an observational method: it is characterized by the non-intervention of the researcher in the phenomenon of study and due to the non-restriction of the responses of the subjects through the tasks or the instruments of evaluation.
• The observational method is defined as a modality of the scientific method that, through the systematic and objective recording of the behavior that is spontaneously generated, allows the testing of hypotheses, the replicability of its results and contributes to the theoretical development by providing valid results in a specific field of knowledge.
• The use of observation in a scientific investigation has to do with the degree of structuring of the observation and the degree of participation of the observer.

Degree of structuring of the observation

Naturalistic observation: when the behavior of the subjects is observed in a natural way, in their usual environment or context, and without any type of modification by the researcher. Semi-structured observation: when the researcher introduces some type of modification in the observation situation, to to be able to guarantee the manifestation of the behaviors of interest or constant elements that allow making comparisons later. Structured observation: when the researcher intervenes in a systematic way or makes substantial modifications in the observation situation. The behavior to be observed is no longer generated spontaneously, but can be an effect of the researcher's intervention.

Observer participation degrees

Non-participant or external observation: it is characterized by the fact that the observer is not integrated into the situation that is being observed. Participant observation: the observer is part of the observation situation itself by participating in the activities that are being observed. Participant as an observer or observation by relatives: a natural member of the group or of the situation under study is the one who performs the functions of observer. Self-observation: it is the subject under study who registers her behavior.

Common aspects with any investigation: Identification of the problem and formation of hypotheses. Design (Procedural decisions). Data collection: registration and coding. Data quality control: validity and reliability. Data analysis and interpretation of results.

Specific procedures of the observational method:

• What to watch: is resolved through the choice, adaptation or creation of the category system (determines, through its internal structure and in the definition of its categories, the behaviors that are relevant to the study problem and that therefore should be registered by the observer).
• Who, when, where and how much to observe: decisions that are resolved by a sampling procedure. In this strategy, it is necessary to guarantee the relevance and representativeness of the sample. To do this, the following are planned: the number of observation sessions, the criteria for starting and ending them, when, where it will be observed (intersessional sample) and who will be the subjects observed in each session (sample intrasessional).
• How to observe: form of registration of the data and properties of the behavior that will be taken into account as a source of information (occurrence and / or duration and / or order of appearance).

The researcher observes and records with a sample of the behavior of the individual or the group under study that must reflect the characteristics and the real dynamics of the same. That is, the sample must be representative. The representativeness of the sample will depend fundamentally on: The sampling rules chosen or decided for the investigation. The registration rules specific. Clarification of two concepts:

• Observation session: continuous period of time during which the observer systematically records the behaviors under study.
• Observation period: total time period in which it would make sense to record the subject's behavior based on the objectives of the study.

Sampling rules

Once the problem has been defined, the study sample is selected, which must be representative of the population. Rules of intrasessional sampling (choose which subjects are to be observed and when within each observation session): Sampling of focal subject: This rule indicates that a single subject (or sample unit) becomes the focus of the subject's sustained attention. observer.

Sweep or multifocal sampling: The observer focuses on each of the individuals in very short periods of time, going from one subject to another, in an established order that is repeated several times in the session. Combined sampling (focal subject and sweep): The observer focuses on a single focal individual, and every certain time period performs a full sweep of all group members then returns to your subject focal. Rules of intersessional sampling (establish criteria for the beginning and end of the observation sessions): Fixed selection: a fixed and limited criterion is applied.

Simple random sampling: the criteria are randomly selected. Stratified random sampling: random choice of criteria from each of the available strata or groups. Systematic random sampling: the moment of initiation of the first session is randomly selected and, from there, the a systematic rule to start the following ones, taking into account the duration of the sessions and / or the distance between them.

Registration rules

The observational record consists of the annotations that are made of the observed behavior. Often these annotations are made through codes that represent each of the categories of the system (coding). Transition Activated Recording (RAT): Recording rule that determines the annotation of all transitions occurrences of relevant behaviors (event record) and with information on their duration (record of state).

It is so named because it reflects that the "activation" of the observer to make a new movement of the register occurs with each change or transition in the behavior of the observed subject.

Record Activated by Units of Time (RAUT): Registration rule that imposes on the observer a registration pattern determined by periods of time. Either the specific moments in time in which what is happening is recorded through the categories (punctual or instantaneous sampling) are identified, or divides the session into short consecutive periods and at the end of each one the categories of behavior that have occurred during the session are recorded (sampling of intervals).

Reliability by consensus is the progressive adjustment of the records of two or more observers (often one of them is the researcher) that is carried out by joint and negotiated work in the application of the category system in some sessions of observation. It is often part of the observer training process and also serves to test and debug the category system.

Sources of error and forms of control in observation

The subject of study as a source of error: reactivity Reactivity is the name given to the changes that occur, voluntarily or involuntarily, in subjects simply by feeling observed. Factors that influence reactivity are: the visibility of the observer and certain qualities of the observer.

The observer as a source of error The observer can be a source of error in the study data due to the application of the system of categories, due to their inadequate interpretation or their errors in the registry due to lack of attention to certain elements. Let's look at some of these errors:

• Observer drift: his own experience can lead him to develop personal interpretations and adaptations of the original definitions of the categories, systematically deviating from them in the registry of data.
• Observer Expectations: or expectations of what should happen or appear in the situation studied, altering the behaviors to be recorded. A control strategy is the blind procedure (the observer does not know the objective and the hypothesis of the study).

The category and code system

Problems of definition of the categories, a breadth or complexity are also a source of errors. excessive system or the application of arbitrary codes, too far removed from the meaning of the categories.

Agreement indices

Percentage of agreement: expresses, in terms of percentage, the comparison of the records of the two observers and the count of the occurrences of behavior recorded in which they coincide (Number of Agreements), to divide it by the total number of registered events (Number of Agreements + Number of Disagreements).

(P = (No. Agreements) / (No. Agreements + No. Disagreements) x 100)

Kappa index (Cohen, 1960): expresses the ratio between actual agreements and possible agreements corrected by subtracting agreements due to chance. (K = Po - Pe) / (1 - Pe) x 100)

The statistical techniques applicable for the analysis of the data will vary depending on the type of observational study proposed and are, in definitively, the specific objectives of the research which determine the structure of the study and the procedural decisions consistent with the same. Let's see the different possibilities of data analysis:

• According to the objectives of the analysis: Exploratory analysis: analysis of specific relationships to find significant presences or relationships.
• Confirmatory analysis: checking and contrasting some hypotheses or predictions.
• According to the inclusion of the time variable: Synchronous analysis: when measures that have been taken simultaneously are described and related.
• Synchronous analysis: when measures that have been taken simultaneously are described and related.
• Synchronous analysis: when measures that have been taken simultaneously are described and related.
• Macro-analysis: when the description and study of relationships is carried out through global measures.
• Design or cross-sectional analysis: if they are carried out at the same time (synchronous analysis).
• Longitudinal design or analysis: if they are carried out at different times (diachronic analysis).
• Micro-analysis: when studying the relationships between the units of conduct that, as links in a chain, reflect the behavior of the individual or group in an orderly manner over time.
• Sequential analysis: allows detecting the existence of temporal contingency relationships in the occurrences of the categories of behavior, and discover the existence of systematic patterns in the dynamics of behaviour.
• Synchrony or co-occurrence analysis: It is interesting to know the probability that certain behaviors occur at the same time as others.

Frequency: it will be obtained by counting the number of times a certain category occurs in an observation session. It is a discrete quantitative variable (it does not admit intermediate values) that is measured on a ratio scale (it has an absolute zero at its origin).

It has as secondary measures:

• Category rate: it is obtained by dividing its frequency by the total observation time (session or sum of sessions) and can be considered as a measure of the temporal density of the category of the conduct.
• Relative frequency or relative proportion: it is the result of dividing the frequency of the category by the total of events recorded in that observation period (the sum of the frequencies of all the categories of the system), and it is a way of knowing the greater or lesser predominance of the categories of behavior under certain conditions (those of the observation sessions analyzed).

Duration:

• indicates the total number of time units that all occurrences of the category occupy during the observation period. It is a continuous quantitative variable (admits intermediate values) that is measured on a ratio scale (it has absolute zero at its origin).

It has as secondary measures:

• Average duration: calculated by dividing its duration by its frequency.
• Relative duration or prevalence: calculated by dividing the duration of a category by the total observation time.
• Transition frequency: it is the number of times that a certain category of behavior is followed by another during the same observation session.
• Its secondary measure is: Relative frequency of transition: estimation of the probability that a certain category occurs and another has occurred.

It is obtained by dividing the transition frequency of the pair of categories under study by the frequency of the antecedent behavior category. Intensity: reports the degree to which a certain behavior is present in a subject. Different degrees must be established and it may have a subjective assessment component that increases the risks of bias.

This article is merely informative, in Psychology-Online we do not have the power to make a diagnosis or recommend a treatment. We invite you to go to a psychologist to treat your particular case.

If you want to read more articles similar to The Observational or Experimental Method, we recommend that you enter our category of Experimental psychology.