Would have to make sure that group a roughly matched

would have to make sure that group A roughly matched group B with prior knowledge
before starting the study. Other extraneous variables could include amount of support in
the home, socio-economic income, or temperature of the testing room.
there are four approaches by which the effect of the extraneous variables
can be controlled.
1) Randomization: In this approach, treatments are randomly assigned to
the experimental groups. It is assumed that the extraneous factors are
present equally in all the groups. This technique is only workable when the
sample size is very large.
2) Matching: Another important technique is to match the different groups of
confounding variables. Different confounding variables like gender, age,
income etc. could be distributed equally amongst the group. It sometimes
does become difficult to extend matching within all the groups and another
drawback of the same is that, sometime the matched characteristics may be
irrelevant to the dependant variable.

3) The use experimental designs: In certain studies, the experimental
designs may play a crucial role in reducing or completely removing the role
and impact of the extraneous variables.
4) Statistical Control: There may be situations, when all the above
mentioned methods to control the extraneous variables do not show any
significant outcome. It brings the entire research into question as then causal
inferences are difficult to make. Another method that may work to bring
down the effect of extraneous variables is the method of statistical control.
Among the various statistical tools and techniques, Analysis of Covariance
( ANOVA) helps in reducing the impact of the extraneous factors on the study.
These four methods, in their own way, can be used in the research,
collectively or exclusively to eliminate the relationship impact discussed
above. It is dependent upon the expertise of the researcher to understand
and administer these methods in a way that the best possible results can be
obtained.
For example, imagine that a researcher has devised an experiment to investigate
whether giving study extra study time can help reduce test anxiety. The amount of
study time the students have is the independent variable (since it is what the
experimenter manipulates) and the amount of test anxiety the students experience
is the dependent variable (since it is what the researchers are measuring).
When as experiment is conducted, there are other variables that can affect results if they are not
control them appropriately. The researcher might want to make sure that it is the manipulation of
the independent variable that has made changes in the dependent variable. Therefore, all the
other variables that could affect the dependable variable to change must be controlled. These
other variables are called extraneous or confounding variables.