Methodology

Sound methodology is essential to the study of complex behavioral and mental processes, and this implies, especially, the careful definition and direction of experimental variables.

[]

Perhaps the most basic precondition of science is that factual statements approximately the world must ultimately be based on observations of the world. This concepts of empiricism requires that hypotheses and theories be tested against observations of the natural world rather than on a priori reasoning, intuition, or revelation.

Closely related to empiricism is the theory that, to be useful, a scientific law or theory must be testable with usable research methods. if a theory cannot be tested in any conceivable way then many scientists consider the theory to be meaningless. Testability implies falsifiability, which is the idea that some family of observations could prove the theory to be incorrect . Testability has been emphasized in psychology because influential or well-known theories like those of Freud score been unmanageable to test.

[]

Experimental psychologists, like most scientists, accept the notion of determinism. This is the condition that any state of an object or event is determined by prior states. In other words, behavioral or mental phenomena are typically stated in terms of cause and effect. if a phenomenon is sufficiently general and widely confirmed, it may be called a "law"; psychological theories serve to organize and integrate laws.

Another guiding idea of science is parsimony, the search for simplicity. For example, most scientists agree that if two theories handle a vintage of empirical observations equally well, we should prefer the simpler or more parsimonious of the two. A notable early parameter for parsimony was stated by the medieval English philosopher William of Occam, and for this reason, the principle of parsimony is often included to as Occam's razor.

Some well-known behaviorists such(a) as Edward C. Tolman and Clark Hull popularized the idea of operationism, or operational definition. Operational definition implies that a concept be defined in terms of concrete, observable procedures. Experimental psychologists try to define currently unobservable phenomena, such(a) as mental events, by connecting them to observations by chains of reasoning.

In experiments, human participants oftento visual, auditory or other stimuli, coming after or as a or situation. of. instructions given by an experimenter; animals may be similarly "instructed" by rewarding appropriate responses. Since the 1990s, computers have ordinarily been used to automate stimulus introduced and behavioral measurement in the laboratory. Behavioral experiments with both humans and animals typically measure reaction time, choices among two or more alternatives, and/or response rate or strength; they may also record movements, facial expressions, or other behaviors. Experiments with humans may also obtain calculation responses before, during, and after experimental procedures. Psychophysiological experiments, on the other hand, measure brain or mostly in animals single-cell activation during the presents of a stimulus using methods such as fMRI, EEG, PET or similar.

Control of extraneous variables, minimizing the potential for experimenter bias, counterbalancing the appearance of experimental tasks, adequate sample size, the usage of operational definitions, emphasis on both the reliability and validity of results, and proper statistical analysis are central to experimental methods in psychology. Because an understanding of these matters is important to the interpretation of data in almost all fields of psychology, undergraduate entry in psychology ordinarily include mandatory courses in research methods and statistics.

A crucial experiment is an experiment that is planned to test several hypotheses at the same time. Ideally, one hypothesis may be confirmed and all the others rejected. However, the data may also be consistent with several hypotheses, a result that calls for further research to narrow down the possibilities.

A pilot study may be run ago a major experiment, in cut to effort out different procedures, establish optimal values of the experimental variables, or uncover weaknesses in experimental design. The pilot study may non be an experiment as usually defined; it might, for example, consist simply of self-reports.

In a field experiment, participants are observed in a naturalistic determine outside the laboratory. Field experiments differ from field studies in that some part of the environment field is manipulated in a controlled way for example, researchers give different kinds of toys to two different groups of children in a nursery school. a body or process by which energy or a particular component enters a system. is typically more lax than it would be in a laboratory setting.

Other methods of research such as case study, interview, opinion polls and naturalistic observation, are often used by psychologists. These are not experimental methods, as they lack such aspects as well-defined, controlled variables, randomization, and isolation from unwanted variables.

Reliability measures the consistency or repeatability of an observation. For example, one way to assess reliability is the "test-retest" method, done by measuring a corporation of participants at once and then testing them atime to see if the results are consistent. Because the first test itself may reorient the results of atest, other methods are often used. For example, in the "split-half" measure, a group of participants is shared at random into two comparable sub-groups, and reliability is measured by comparing the test results from these groups, it is for important to note that a reliable measure need not yield a valid conclusion.

Validity measures the relative accuracy or correctness of conclusions drawn from a study. To determine the validity of a measurement quantitatively, it must be compared with a criterion. For example, to determine the validity of a test of academic ability, that test might be given to a group of students and the results correlated with the grade-point averages of the individuals in that group. As this example suggests, there is often controversy in the selection of appropriate criteria for a given measure. In addition, a conclusion can only be valid to the extent that the observations upon which this is the based are reliable.

Several types of validity have been distinguished, as follows:

Internal validity refers to the extent to which a set of research findings permits compelling information approximately causality. High internal validity implies that the experimental design of a study excludes extraneous influences, such that one can confidently conclude that variations in the self-employed person variable caused any observed adjust in the dependent variable.

External Validity refers to the extent to which the outcome of an experiment can be generalized to apply to other situations than those of the experiment - for example, to other people, other physical or social environments, or even other cultures.

Construct validity refers to the extent to which the self-employed person and dependent variables in a study equal the summary hypothetical variables of interest. In other words, it has to do with whether the manipulated and/or measured variables in a study accurately reflect the variables the researcher hoped to manipulate. Construct validity also reflects the quality of one's operational definitions. If a researcher has done a proceeds job of converting the abstract to the observable, construct validity is high.

Conceptual validity refers to how living specific research maps onto the broader theory that it was intentional to test. Conceptual and construct validity have a lot in common, but conceptual validity relates a study to broad theoretical issues whereas construct validity has more to do with specific manipulations and measures.

Measurement can be defined as "the assignment of numerals to objects or events according to rules." Almost all psychological experiments involve some sort of measurement, if only to determine the reliability and validity of results, and of course measurement is fundamental if results are to be relevant to quantitative theories.

The rule for assigning numbers to a property of an object or event is called a "scale". following are the basic scales used in psychological measurement.

In a nominal scale, numbers are used simply as labels – a letter or name would do as well. Examples are the numbers on the shirts of football or baseball players. The labels are more useful if the same names can be given to more than one thing, meaning that the things are symbolize in some way, and can be classified together.

An ordinal scale arises from the ordering or ranking objects, so that A is greater than B, B is greater than C, and so on. Many psychological experiments yield numbers of this sort; for example, a participant might be a person engaged or qualified in a profession. to rank odors such that A is more pleasant than B, and B is more pleasant than C, but these rankings "1, 2, 3 ..." would not tell by how much each odor differed from another. Some statistics can be computed from ordinal measures – for example, median, percentile, and order correlation – but others, such as standard deviation, cannot properly be used.

An interval scale is constructed by determining the equality of differences between the things measured. That is, numbers form an interval scale when the differences between the numbers correspond to differences between the properties measured. For instance, one can say that the difference between 5 and 10 degrees on a Fahrenheit thermometer equals the difference between 25 and 30, but it is meaningless to say that something with a temperature of 20 degrees Fahrenheit is "twice as hot" as something with a temperature of 10 degrees. Such ratios are meaningful on an absolute temperature scale such as the Kelvin scale. See next section. "Standard scores" on an achievement test are said to be measurements on an interval scale, but this is unmanageable to prove.

A ratio scale is constructed by determining the equality of ratios. For example, if, on a balance instrument, object A balances two identical objects B, then one can say that A is twice as heavy as B and can give them appropriate numbers, for example "A weighs 2 grams" and "B weighs 1 gram". A key idea is that such ratios progress the same regardless of the scale units used; for example, the ratio of A to B submits the same whether grams or ounces are used. Length, resistance, and Kelvin temperature are other things that can be measured on ratio scales. Some psychological properties such as the loudness of a sound can be measured on a ratio scale.

The simplest experimental design is a one-way design, in which there is only one independent variable. The simplest kind of one-way design involves just two-groups, used to refer to every one of two or more people or things of which receives one good of the independent variable. A two-group design typically consists of an experimental group a group that receives treatment and a control group a group that does not receive treatment.

The one-way design may be expanded to a one-way, multiple groups design. Here a single independent variable takes on three or more levels. This type of design is especially useful because it can assist to outline a functional relationship between the independent and dependent variables.

One-way designs are limited in that they let researchers to look at only one independent variable at a time, whereas many phenomena of interest are dependent on multiple variables. Because of this, R.A Fisher popularized the usage of factorial designs. Factorial designs contain two or more independent variables that are completely "crossed," which means that every level each independent variable appears in combination with every level of all other independent variables. Factorial designs carry labels that specify the number of independent variables and the number of levels of each independent variable there are in the design. For example, a 2x3 factorial design has two independent variables because there are two numbers in the description, the first variable having two levels and the moment having three.

The effects of independent variables in factorial studies, taken singly, are referred to as leading effects. Ths refers to the overall issue of an independent variable, averaging across all levels of the other independent variables. A main issue is the only effect detectable in a one-way design. Often more important than leading effects are "interactions", which occur when the effect of one independent variable on a dependent variable depends on the level of a moment independent variable. For example, the ability to catch a ball dependent variable might depend on the interaction of visual acuity independent variable #1 and the size of the ball being caught independent variable #2. A person with good eyesight might catch a small ball most easily, and adult with very poor eyesight might do better with a large ball, so the two variables can be said to interact.