The Landmark Blog

Practices from the Field: Telephone Surveys and Mobile Phones

 Email Article |  Twitter |  Facebook |  digg it |   delicious |   LinkedIn |   Technorati 

I'm dating myself here, but do any of you remember party lines?  And no, I'm not referring to dialing up your friend and organizing a party by calling everyone that you know...and that your friend knows...and that his friend knows...etc.?  Though social meanings change drastically in short periods of time, I'm talking about something else.  If you do an internet search on "party line," the first response actually generates an adult content web site.  Look past that to the Wikipedia posting of "party line" for the "traditional" meaning of the word. 

Party lines date back to the days when underground lines (Local Loops) weren't available to everyone, and households had to share telephone lines rather than being directly connected to digital lines in a million different possible ways.  A party line was when two or more households shared one telephone line.  In a party line, different houses had unique rings (now that's an old-school ringtone) that identified which household was being dialed.  For instance household #1 would have two short rings, while household #2 would have a long and a short ring, household #3 would have two long rings, etc.  Though almost obsolete now, party lines held out in the "country" much longer than they did in urban areas and may still exist in some very rural areas that aren't easily connected directly to the "local loop."

The real disadvantage to a party line is the ability to eavesdrop on your neighbor (some would call that an advantage).  If you were quiet enough about it, you could listen in on any conversation you wanted (take that all you folks whose entire jobs are aimed at protecting PII).  Remember when your younger brother or sister, or your parents would listen in on your phone conversation (before you had a cell phone), party lines took "listening in" to a whole new level. 

Aside from the eavesdropping aspect there were other disadvantages.  For instance, in the early days of internet...could you imagine what your dial-up would have done to the party line?  For that matter, a long conversation (a simple phone interview) could tie up the other household's line for what would seem like years.  In the heydays of party lines, early market research interviewing was mostly conducted door-to-door (face-to-face) and the market research sampling was address based.  As telephone technology evolved, better digital capabilities emerged, and households were able to connect directly to a line without sharing it with their neighbors, RDD Sampling was a more feasible option and telephone interviewing took off.

As a market research industry, we're back to a tipping point for survey interviewing and collection.  Online interview growth has started to level off (an exciting development for those of us that have huge banks of phones for telephone interviewing).  But, wireless households are growing very quickly.  The last CDC measurement of wireless only households fell out at 25%, while wireless "mostly" households add another 16%, totaling 41% of households who are wireless dominant.  Contrary to the last few years that they tracked this, the Pew Research Center has started to see some coverage response bias related to the data they collect, though still relatively minor depending on the demographic characteristics of those being interviewed.  For phone interviewing to stay relevant to the future of our industry, we must deal with cell phone interviewing and representation NOW.  It's our responsibility as providers of these services to provide solutions and best practices for cell phone interviewing.

Under that assumption, here are a few sampling solutions that are gaining traction in this arena:

• A return to the address-based sampling methods of old.  With the ability to port numbers, eliminate landlines, etc. address based sampling is emerging as a very good option for telephone interviewing and is obtainable through many sample providers.
• Supplementing RDD samples with Cell Phone samples is a good idea to consider.  Although, it's worth mentioning that just because a number is a cell phone, it doesn't necessarily mean that it comes from a cell-only or cell-phone dominant household.
• Client lists that include cell phone sample, seem to be more prevalent than in previous years and a much larger part of the industry.  This solution though used more and more, doesn't address market research projects whose methodologies rely on RDD based sampling frames.

Assuming that we will see an increase in cell phones in our market research samples, here are some suggestions for telephone interviewing cell phones moving forward:

• Ensure TCPA compliance by running samples up against cell-phone lists like Neustar (ported numbers from home to cell) and Telcordia (provides information on cell phones and allows you to identify cell phones against your sample list).
• Hand dial records that are identified as cell phones through these sources.
• Provide multiple methods for completing surveys (telephone, online, IVR, inbound, etc.).  This will help you reduce costs associated with telephone interviewer-only methods.
• Consider incentives for samples that are heavily aimed at cell phones.

And, here are some evolving suggestions for telephone interviewing as a result of this trend toward increasing cell-phone representation:

• Limit the length of the interview.  You've heard it before, you'll hear it again.  The shorter the interview, the less it will cost to complete for you, your supplier, and the respondent.  
• Screen for respondent safety.  Ask a quick question about whether the respondent is in a safe place where they are free to talk.  If they are driving or are operating heavy machinery, terminate the interview.
• Be sensitive to subject matter.  Respondents who are participating in surveys over cell phones are more likely to be in public places than when using landlines.  Be considerate of their location when doing surveys where PII (personal identifiable information) is a concern.
• Screen for age.  Many cell phones are owned or dominated by under-age users.  Consider the ethical issues associated with interviewing respondents who are under 18 years of age.  Ask the respondent before you consider the interview.  
• Consider interviewing during "free-periods"...nights and weekends.
• Provide local numbers.  Many forward thinking telephone long distance providers are now supplying the option to provide a local number on caller id.  It's worth considering a local number as opposed to a toll-free 800 based number.  They're both free to the respondent if called back and the local one is more likely to be answered.

Surprisingly, in the study linked to above, Pew is finding that response rates for cell samples are very similar to the response rates for landline samples.  In addition, they are finding that in some cases, cell samples are more cooperative than are their landline counterparts.  Cell phone interviewing can be done as effectively as "standard" historical land-line interviewing.  It will take some thought, some evolution, and some commitment to doing it right.  What's new?

I'm sure I've missed many suggestions.  Please provide them if you'd like to add to this list.