Sorry
That's Incorrect

No. Matching actually helps to decrease error variance.
How? Why? Because error variance is that variation
between individuals in a group (individual differences),
and each variable on which you match the participants
reduces those differences. For example, if we randomly
assign 60 children to two groups of 30, there will be age
differences among the children. Randomization will help
to improve the chances that the age distribution in one
group is much like that of the other group.  However, by
having age-matched pairs of children rather than randomly
assigning them, we can have each child in one group
matched for age by a child in the other group. Thus, age is
reduced as a source of difference between the groups and
a source of variation within each group. Matching helps
to reduce error variance.