Random digit dialing (RDD) has grown in sophistication from simply adding digits to listings selected from telephone directories to the more advanced methodologies in use today. Until recently, those wanting to use RDD sampling had few options: either use the services of an outside sampling company or develop their own proprietary in-house system, which required a substantial investment of money and expertise.

But with the arrival of Genesys, a custom RDD sample generation system developed by Marketing Systems Group (MSG), more choices have been added to that list.

Now, a company can order Genesys RDD samples directly from MSG (or its affiliated service bureau, Sophisticated Data Resources, Inc.), or they can license the system on an annual basis and have it installed on their own in-house office computer system, allowing them to access the database whenever necessary to generate their own RDD samples.

Dale Kulp, president of MSG, says that he and his partner Gerry Holzbaur had discussed the concept of such a system for years, but they felt its viability was limited because of the large amount of storage required for an RDD sample generation system. That is, until the personal computer explosion and rapid advances in computer memory technology.

The question we always had was: How do you deliver the system? Going back seven or eight years, not many research firms even had minicomputers, but in the past three or four years, it's gotten to the point where virtually every research company has at least one PC. Once the computer base reached that point, we knew it was time to develop a portable sampling system that virtually any research company could use," Kulp says.

(Genesys is available for Hewlett-Packard 300 systems operating under MPE, DEC/VAX systems operating under VMS, IBM AT compatible or Apple Macintosh PC's.)

Kulp says the system offers users two distinct advantages:

1. Quick turnaround. Most samples can be specified and generated in under 15 minutes, without having to pay a premium. "One of our clients called us up recently, and they were ecstatic. They just got finished generating 150,000 random digit numbers for a large national sample and it took them just 27 minutes."

2. Cost. Up to now, an average RDD sample piece could cost as much as 25-30 cents. "Genesys users can generate them for three or four cents a piece. When you consider an average study might use 5,000 pieces of sample, if they've been paying even 25 cents each, an in-house user has suddenly knocked about $1,000 off the cost of that project."

The samples can be based on ZIP codes, state/county, Metropolitan Statistical Area (MSA), Areas of Dominant Influence (ADI), Dominant Market Area (DMA), and other geographic definitions. In addition, sample designs can incorporate demographic variables such as income, age, race, and home ownership.

Types of RDD

Genesys creates samples using the two dominant types of random digit dialing procedures, which can be termed "classic RDD" and "rdd."

A classic, self-weighting RDD sample is one in which every telephone number has an equal chance of being selected. Genesys uses residential "working bank" information (a working bank is determined by the first two digits in a phone number's four digit suffix, e.g. the numbers   beginning with 8-0 in 555-80XX) to increase the proportion of households in an RDD sample.

"The reason we do that," Kulp says, "is that roughly 50% of the potential working banks don't have any residential numbers assigned, so when you generate an RDD sample, you can eliminate the majority of non-working and non-household numbers before the sample is even printed--before interviewers spend time dialing them."

The other methods, "rdd," are more efficient in terms of the proportion of numbers that will ultimately result in a household contact. They increase efficiency by over-sampling exchanges that have higher proportions of households while under-sampling those with lower proportions--but this creates the problem of potential bias because the resulting sample over-represents households in denser exchanges relative to those in less dense exchanges.

But with Genesys, this bias can be offset by using the appropriate weighting factor, says Gerry Holzbaur, vice president of MSG.

"An in-house licensee or custom client simply needs to specify that the sample file should contain the relative probabilities of selection for each sample number. When the survey is completed, the interviews can be weighted to reflect the same proportions one would expect from a classic RDD sample," he says. This provides the user with enhanced dialing efficiency while minimizing the risks of bias in the survey results.

Updating the database

To make sure that the samples continue to be representative and efficient, MSG updates the entire Genesys database twice a year—which usually means, says Holzbaur, vice president of MSG, "wading through close to a gigabyte (or a thousand megabytes) of data"—including demographic, ZIP code, telephone exchange, and working bank information. Genesys users share the cost and benefits of updating, much like they would with an omnibus study.

Fred Soulas, president of ICR, is very familiar with the chore of updating an in-house RDD system. Prior to becoming a Genesys user, his company used its own in-house system.

"It's not just buying one tape from AT&T," he says of the updating process. "It takes a lot of time and it's very costly. Genesys does it for everybody and it's a shared cost, which helps us all a great deal. It helps with updated samples and it helps from the cost aspect, too, so we can pass those savings back to our clients."

This frequent updating is critical to ensuring consistently representative and productive samples. "It's important to know that every time you get a sample it's fresh and that each telephone number had an equal chance of coming into the sample. With Genesys you're creating your own sample so you know that's the case," he says.

Speed and flexibility

Soulas also appreciates the speed and flexibility that Genesys allows his company.

"We're very highly computerized here. All of the surveys we do, with few exceptions, are done on CRT, and we like the sample to be on-line, so that when the interviewer is finished with an interview, the next randomly generated telephone number comes up on their screen right from the computer, and all they have to do is dial it in."

Patrick Baldasare, president of The Response Center, provides an example of how his company was able to utilize the quick turnaround the system provides.

"One day I got a call from Campbell Soup Co. at 4:45 p.m. and they wanted us to do a project for them. At 4:55 p.m. I received a fax of a rough draft questionnaire. By 7:00 p.m., my computer programmers had programmed the questionnaire, my interviewers had been briefed, a random probability sample had been drawn, and we were on the phones, doing the interviews on the CRTs. The next morning I was able to fax cross-tabs to my client. It's hard to beat that. Genesys was definitely part of it. Because I'm a licensee of that software I have that capability."

Baldasare says the system offers the user "ultimate control" because he or she can customize a sample to target specific groups of people.

"The user can go in and say 'I want to find people who make $50,000 a year or more, and I want to talk to people who are between the ages of 25 and 44, and I want to hit the right kind of person 70% of the time.’ With Genesys you can do that, and at the same time be cognizant of and adjust for the associated sample biases.”

Kulp is confident. “I think we’ve put a very powerful tool in the hands of the research industry, something that was never there before. We’ve given research companies the ability to improve their product in terms of design capabilities and made them more competitive because we’ve been able to cut their costs without sacrificing quality. We’ve moved RDD sampling out of the 70s and into the 90s.”