Two decades ago, the advent of the Internet (specifically, the widespread use of the World Wide Web) transformed the PC from a niche product into something that has become commonplace in nearly every household. Similarly, the telephone has seen a similar transformation. Early on (in the late 1990s), the cell phone became affordable to the vast majority of Americans, then it gained additional functionality (like e-mail/Internet access) that was once only available to PCs.
Given the expanded capabilities of the cellphone, people have begun to abandon their traditional “land line” telephones, and this shift has had political consequences as well. Since there are more restrictive laws regarding the calling and/or surveying of “cell phone only” voters, the cost of reaching these voters has increased. Furthermore, the cell phone population is popularly thought to be a predominately younger and more of a non-white segment of the electorate – a personification, so to speak, of the “Obama coalition” that elected and re-elected President Barack Obama in 2008 and 2012.
This trend towards cell phones has another political consequence: those polling firms not having an appropriate demographic mix in their surveys and/or not having an appropriate representation of “cell phone voters” run the risk of inaccurate surveys, and in fact, this was prevalent in 2012, when multiple Republican pollsters were embarrassed with their predictions of a Romney victory.
Is this popular conception about the makeup of “cell phone voters” true? JMC wanted to test this popular perception by analyzing the demographic composition of likely voters in Louisiana with cell phones and land lines (in Louisiana, a voter’s phone number is part of his/her voter record).
Let’s start by discussing the percentages of “land line” and “cell phone” voters. If we use recent voter data for likely voters, data, we find that 22% of the voters with a phone number listed on their voter record used their cell phone number, while 78% are traditional “land line” voters. Now that we know the percentages of each “type” of voter, let’s look at the demographic characteristic of each:
The likely voter electorate is 66-31% white/black. Those with land lines are 67-30% white/black, while “cell phone” voters are 60-34% white/black. So while the population of cell phone voters is slightly more black and Hispanic, the political impact isn’t as substantial as conventional wisdom would have you believe: the 2012 Presidential electorate was 66-31% white/black, while the 2014 Senate (runoff) electorate was 67-30% white/black. In other words, “land line voters” mirrored the racial composition of the electorate in 2012 and 2014.
The partisan breakdown of the likely voter electorate is 48-31% Democrat/Republican (21% Independent). The “land line voters” are 50-31% Democratic (19% Independent), while the “cell phone voters” are 40% Democratic, 31% Independent, and 29% Republican. From the standpoint of likely voters, there is a noticeable difference in the Democrat and Independent percentages between the “land line” and “cell phone” voters, while the Republican percentages are similar between either population.
Should this be a concern to pollsters? Not necessarily, because the 2012 Presidential electorate in Louisiana was 50-31% Democratic/Republican (20% Independent), while the 2014 Senate (runoff) electorate was 50-34% Democratic/Republican (16% Independent). In other words, the “land line voters” closely tracked the partisan composition of the electorate in 2012 and 2014.
Thus far, we have debunked (from a racial and partisan perspective) the notion that “cell phone voters” are substantially different than “land line” voters. However, one part of popular belief has some validity: that those holding onto their land lines (at least from the perspective of the phone number a voter put on his/her voter registration card) are older than those using their cell phones: the average age of a “land line” voter was 53, while the average age of a “cell phone” voter was 41.
Furthermore, if we look at the land/cell distribution by age bracket for likely voters, we find that 62% of voters under the age of 40 have landlines (38% have cell phones), while the distribution for those 40 years old and older is 86% land line/ 14% cell phone. Since the average age of a likely Louisiana voter is 50 (and 30% of voters are under 40 years old), this age discrepancy is not yet an issue with regards to the “accuracy” of polling only land line voters, but will be in future election cycles.
A glimpse into the future
From analyzing the voter file from a racial, partisan, and age perspective, we noticed that while “cell phone voters” were not substantially different than the overall electorate, we did see that cell phone voters offered a glimpse into Louisiana’s future demographic makeup. A review of voter data confirms this trend: those under 40 years old (regardless of whether they were a “land line” or “cell phone” voter) are 58-38% white/black and 37% Democratic, 33% Independent, and 30% Republican (the statewide likely voter electorate is 66-31% white/black and 48-31% Democratic/Republican).
One facet of this trend has noticeable future political implications: Independent voters (regardless of race) are noticeably less likely to vote. To illustrate, in the 2012 Presidential election, the overall voter turnout was 68%, while among Independents, it was 55% (57% for white Independents, 52% for black Independents, and 46% for Asian/Hispanic Independents). The disparity was greater in the 2014 midterm elections, as overall turnout was 44%, while it was 28% for Independents (30% for white Independents, 24% for black Independents, and 21% for Asian/Hispanic Independents).
While popular belief is that cell phone voters are younger, minority, and more Democratic, voter registration and (more importantly) voter participation data do not show that the demographic composition of “cell phone voters” is substantially different from the overall electorate – for now.
What is noticeable about ‘cell phone voters” (as well as voters under 40) is that both they and younger voters are less politically aligned, and given recent elections, this political “de alignment” will have negative implications on voter participation in future elections.