Wednesday, November 10, 2010

Experimental Research


Experimental Research is a common method used to test theories in order to prove or disprove them. Experiments carry out certain steps defined in what is called the “scientific method” to further investigate a problem. In an experiment, there are two groups being tested, an experimental group and a control group. The experimental group is the group that is being manipulated in the experiment, whereas the control group is the group that is not being manipulated. This group is the foundational point for which to compare the experimental group against. These experiments can be either design blind experiments meaning the participants in the study are unaware of what group they are in (experimental or control). Or they can be double design blind where the person conducting the experiment is also unaware of what group he or her participants belong (experimental or control). In a double design blind study there is no bias.

In every experiment there is an independent variable and a dependent variable. The independent variable is the variable that is manipulated in the experiment in order to show the effect on the dependent variable. The dependent variable is the variable being observed in the experiment; in order words it is what is affected by the independent variable.

In any given experiment the person conducting it makes either a correlation between variables or finds there to be causation between variables. Correlation is the strength and degree to which two things are related. Causation is when there are two things happening as the same time, which makes one think they are happening because of each other but in actuality there could be another, outside reason, cause or factor that’s determining certain things to happen. In other words, one thing doesn’t necessarily mean another thing.

In experiments its essential to isolate the independent variable in order to have an initial measurement and to provide a baseline measure to compare the control group and experimental group to one’s results. It is also important to isolate the independent variable in order to ensure that the two groups are “equal”. Also, both the experimental and control groups must be chosen at random in order to prevent change in the results. Each experiment should be either a double design blind or just a regular design blind.

When conducting an experiment there are several sources of invalidity you should be aware of. These sources consist of selection, history, maturation, testing effects, instrumentation, statistical regression, and mortality. For selection you must choose your participants at random to prevent selection bias. For history, a lot can change over the course of an experiment therefore; experiments should be completed in a timely fashion. For maturation, over a long period of time the perceptions and attitudes of participants and the participants themselves can change between the pre test and the post test which in turn causes a change in the results. 

As far as testing effects goes, it’s the level of sensitivity a participant has to a topic because of being asked the same questions multiple times. The participants will then start to answer the questions differently because they are paying closer attention to them after seeing them previously. Instrumentation is the opposite of testing effects because if the participants suddenly start changing their answers to the questions it becomes difficult to compare them to the initial set of questions. In other words if the questions in the pre test and the post test are not exactly the same it is considered to be invalid. For statistical regression, if participants’ answer is extreme (on a scale to 1-10, they choose either 1 or 10) the answer can only go up or stay the same if they chose 1, or only go down or stay the same if they chose 10.  In order words you cannot see the affects of the independent variable in the experiment. For mortality, participants may drop out of a given study if it goes for a long span of time. For example, in the pre test you start our with 10 participants but by the time the post test comes around only 7 people show up.

No comments:

Post a Comment