The United States Congress and the members of it, is determined by the rules of the Constitution, in conjunction with judicial decisions applicable to such, of which the current interpretation basically comes down to the apportionment of the representatives of that Congress, being the census count of all citizens and non-citizens, as well as all armed forces personnel and federal civilian employees that are currently stationed outside the United States, but are able to be allocated to a specific home State. Some people believe that by counting non-citizens into the population of a given State, that those States that have an abundance of non-citizens, which are by definition, denied the enfranchisement of the vote, does not make a whole lot of sense; for after all, why should they be favored with representation? Further to the point and probably of even more relevance, there are those that are citizens of the United States, but have had their voting rights, disenfranchised, for whatever reason, which are also counted for the apportionment of Congressional seats, even though they are unable to vote.
All of the above, basically comes down to a rather big divide, which is that currently, anyone that lives within a given State, is counted for purposes of the apportionment of Congressional seats, citizens and non-citizens alike; with no differentiation between those that are enfranchised to vote or disenfranchised. Not too surprisingly, those States, that have an abundance of non-citizens within their State borders, such as California, in particular, get more seats in the House of Congress, as opposed to a State, such as Pennsylvania, which gets less; because more of the residents of Pennsylvania, actually are citizens, as compared to the great abundance of non-citizen immigrants in California. On the other hand, a strong argument could be made, that the apportionment of Congressional seats, should not be done by the population within a State of citizens and non-citizens, including disenfranchised voters, but rather would more fairly be done by counting only those that are enfranchised to vote, since these are the very people that actually select their Congressional representative at the polling booth.
If, the amount of representatives in each respective State, was actually done by the amount of those registered to vote within those States, as compared to the national population of citizens and non-citizens throughout the United States, and then apportioned, mathematically to each one of those States -- what would happen in all likelihood, is that the States in general, would be much more motivated to see that their citizens within their respective States, were registered to vote, and further that those that have been disenfranchised from voting, would probably find that their State would be far more accommodating towards permitting them to re-qualify to vote.
Of course, some critics of this plan, might rightly complain that such would disenfranchise the value and impact of all those that are under the voting age of eighteen; though this would be true, it could somewhat be mitigated by each State tabulating their enfranchised voters every two years, which matches the Congressional elections which are always biennial, thereupon providing an updated enfranchised voting roll for each one of those States. After all, the apportionment of the Congressional representatives of each State should reflect, the proportional amount per State, of those that are enfranchised to vote, for these are the very people that are the arbiters of who represents the people, and the people should be only those that are enfranchised to vote, no more and no less.