Yes. Telephone polls tend to "oversample" older people who tend to stay home, and these tend to be more conservative and pro-republican. Main reasons are that younger people are more mobile, they tend to move more frequently and thus have specific phone numbers registered in their names changed or disconnected, they also more likely not to have land lines altogether and use only cell phones.
Another problem is that telemarketers tend to "spoil" the market and more and more simply refuse to talk to any polling agency. I suspect that this is less of a problem in smaller towns and rural areas - which tend to be more conservative - also among older folks (on a fixed income) who are less frequently targeted by telemarketers and thus less likely to refuse to participate in polls due to the "telemarketer fatigue."
Gallup has been criticized for these biases in connection with their other polls, e.g. their giving and volunteering survey they did for the Independent Sector (http://www.independentsector.org/programs/research/gv01main.html).
According to that survey, about 50% of US adults engaged in volunteer work. However, a recent study by the Bureau of Labor Statistics (_Monthly Labor Review_ 2003 (126(8) http://www.bls.gov/opub/mlr/2003/08/contents.htm that used a very large sample (about 50 thousand) gave the estimates that are only a half (sic!) of the Gallup estimates. In short, Gallup did a rather shitty job. For a critique, see the Symposium on Methodology in Surveying Giving and Volunteering Behavior in _Nonprofit and Voluntary Sector Quarterly_ 2001, 30(3):480-569. Some of that criticism is specific to the field, but much of it applies to Gallup telephone survey methodology in general.
Wojtek