UPDATE: I changed the code below in response to Merlyn's critique.

Perl is built to manipulate lists.
I'd tell you to put all your scalars into a list and use a loop like this to find your min and max. Why can only scalars be used? That seems horribly ineffecient for the task at hand.

for $nums (@numset){
$min = $nums unless ($min);
$max = $nums unless ($max);
$max=$nums if ($nums > $max);
$min=$nums if ($nums < $min);
$total += $nums; ++$count;
}
print"Minimum: $min","\nMaximum: $max\n","Average: ",$total/$count;
[download]

if you insist on scalars or literals put them in the for loop.

for $nums ($first,$second,$third,$forth)
or
for $nums (123,32,7,-12,-23,17)
[download]

also please at least use the <br> and <code> tags, so your program is readable