Extraneous variance

Question – Discuss the various ways to control the extraneous variable.

Answer –

There are four ways to control the extraneous variance –

1) Randomization: An important method of controlling external variables is randomization. This is considered to be the most effective way to control variability due to all possible external sources. If achieved through randomization, treatment groups in the experiment can be considered statistically similar in all possible ways. Randomization is a powerful way to control variables. In other words, it is a process of equalizing groups in relation to secondary variables. Randomization means the random selection of experimental units with a large population. Random assignment means that there is a possibility of any treatment status or in the group in each experimental unit. There may be some problems in using the randomization method. It is possible to select random sampling from the population, but then the assignment of experimental units to the groups can be biased. Random assignment of topics is important for internal validity. If the subjects are not randomly assigned, then Confusion can occur. Randomized group design and randomized block design are examples of research design, in which randomization is used to control extrusion variables.
2) Elimination: This process is the easiest way to control the variable through the abolition of unwanted elimination variables. Suppose, the gender of the subject is in the form of an unwanted secondary variable, it is found to affect the variable in an experiment. Therefore, the variables of sex must be controlled. The researcher can decide to take all the men and all women in one experiment and thus, can eliminate the variability due to the sex variable. Using the abolition to control the extinct variable, the researcher loses the power of normalization. If the researcher selects the subject from the restricted range, then the researcher can normalize the results within the restricted range and not outside it. The elimination process is used in non-experimental design.
3) Matching: There is also a non-experimental design process, which is used to control the external source of variance. In order to control this process, organisms and background variables are used, the relevant variables are kept equal or constant in all situations of experiments. Suppose if the researcher finds that the variable of intelligence is highly correlated with the dependent variable, then it is better to control the variance through matching on the variation of intelligence. Although limiting the availability of matching topics as a method of control. If the researcher decides to match the subjects on two or three variables, then they can not get enough subjects for the experiment. Apart from this Milan’s methods fulfilled the principles of randomization.
4) Statistical control: In this approach, no effort is made to prevent the effect of secondary variables. In this technique, one or more concomitant secondary variables (covalent) are measured and the dependent variables are statistically adjusted to remove the effects of uncontrolled sources of variation. Analysis of covalent coagulators is such a technique. It is used to remove the possible amount of variation in the converged secondary variable.

Other articles –