Text 2 U: Contacting Wireless Subscribers Using Text Messaging for Mobile Phone Surveys

TEXT 2 U: CONTACTING WIRELESS SUBSCRIBERS USING TEXT MESSAGING

FOR MOBILE PHONE SURVEYS

Trent D. Buskirk, Ph.D. Epidemiology and Statistics Core, Eastern Virginia Medical School

Mario Callegaro, M.S. Graduate program in Survey Research and Methodology,University of Nebraska, Lincoln

Charlotte Steeh, Ph.D., Georgia State University

Keywords: Mobile Phone Surveys, Text Messaging, Short

Text Messaging (SMS), Number Portability, Text-Messaging

Panel, Digital Mobile Phone, Wireless Web

1. Introduction to the U.S. Mobile Phone Context

The increase in the use of mobile phones and the potential rise

in the number of mobile phone only households or primarily

mobile phone households (those for which landlines exist for

the express purpose of internet/fax/data transmission/receipt) is

slowly creating an opportunity and need to supplement

traditional RDD telephone surveys with surveys of mobile

phones. The number of wireless subscribers in the U.S. has

been rising at an average rate of 9.7% a year since December of

1985 and as of July, 2004 the Cellular Telecommunication and

Internet Association estimated a total of 167,236,200 mobile

wireless telephone subscribers in the country. As the number

of subscribers continues to rise and as the number of

subscribers who become “wireless only” or “primarily

wireless” also rises, contact/cooperation rates for traditional

RDD surveys may become adversely affected since many

companies exclude mobile phone numbers from their sampling

frames.

With the continued development of wireless service in

the U.S. comes the improvement in wireless telephone

technology. In fact, 92% of the wireless subscribers in the U.S.

(C.T.I.A., 2004) are now using a digital wireless phone. With

these advanced phones come many new features including text

messaging and wireless web that could be useful in surveying

subscribers, besides making a cold call. While not all of these

subscribers can make use of text messaging capabilities, many

of them can and the number of text messaging users continues

to rise. In fact, the number of text messaging users rose from

18 million in 2002 to 27 million in 2003 (UPOC and Frank N.

Magid & Associates, 2003). A Telephia and Harris Interactive

survey estimated that 20% of all U.S. mobile phone subscribers

either sent or received a text message (TM) during the last

quarter of 2002 (Telephia / Harris Interactive, 2002) compared

to an estimated 12 percent in 2001. A similar estimate for the

number of text message users in the last quarter of 2002 in the

U.S. was found by SIBIS (SIBIS, 2003). Specifically, this

study estimated that 56% of the U.S. population owned a

mobile phone, and of those 24% were TM users. One possible

explanation for the observed increase in the number of users of

text messaging could be the result of the doubling in the

proportion of young adults (18 to 24) who reported using text

messaging from 2001 to 2002 (22% to 45) (SIBIS, 2003).

For users who cannot access text messaging on their

mobile phones, or for those consumers who do not own a

wireless phone, text messages can still be sent to a wireless

subscriber using a Text Messaging Interface (TMI) from

specific wireless provider company websites. These TMI offer

researchers information and may as well develop into a

screening tool that identifies for nonworking/non-assigned

numbers thus improving the efficiency of a survey process in

light of current FCC restrictions prohibiting the use of

autodialers for dialing wireless numbers and a lack of a

wireless telephone directory (Steeh and Buskirk, 2004)..

In this paper we will focus on two methods that can be

implemented for contacting wireless subscribers in the United

States that do not involve calling the subscriber directly,

primarily, text messaging and secondly wireless internet

service. In the next section we will highlight the text

messaging landscape within the U.S. and illustrate some of the

methods that are available from many of the larger domestic

wireless providers. In section three we will describe the use of

wireless internet capabilities for wireless phones and describe

some potential uses for this technology for the survey

researcher in the U.S. In section four we will describe our own

text messaging study of over 800 Nextel wireless phone

numbers. Finally, in the last section we will discuss some

possibilities for the not so distant future for survey research of

wireless subscribers within the U.S.

2. Text Messaging as a contact tool for Survey Research in

the United States

2.1: Development of Text Messages in the U.S.:

Text

messages had their origins in the U.S. with the development of

digital technology, either digital cellular, broadband PCS

(Sprint), or digital SMR (radio). In 1999 AT&T wireless,

Sprint PCS, Omnipoint, and Nextel offered one-way text

messaging (Federal Communication Commission, 1999), with

the ability to receive, but not send, text messages. Two-way

messages (i.e. receive and send capability) were not introduced

in the United States until May, 2000 when VoiceStream (now

T-Mobile) offered the service (Federal Communication

Commission, 2001). At the end of 2001, AT&T wireless

became the first carrier to offer cross-carrier network SMS

capabilities (Federal Communication Commission, 2002b). The

process of inter-carrier interoperability was completed by the

first quarter of 2002 by the six nationwide operators and other

mobile data providers (Federal Communication Commission,

2003)

Because of varying technologies among providers, TMs

in the U.S. are limited to a maximum length ranging from 120

to 500 characters (Federal Communication Commission, 2003).

Although most mobile phones in the U.S. were digital at the

end of 2003 (Cellular Telecommunications & Internet

Association, 2004), the ability to receive/send TMs was not

uniformly available to all subscribers- a stark contrast to the

4929

American Association for Public Opinion Research Section on Survey Research Methods

uniform availability for European subscribers. While for

example, every Nextel phone is enabled to receive TMs, it

generally costs users $0.15 to receive and at least $5.00 to

enable the reply capability of their phone plus an additional

$0.10 per sent message. Variation in technologies across

providers creates variation in the types of text messaging

capabilities and options that are available to wireless

subscribers in the U.S. Specifically, depending on the

company, text messaging capabilities have to be activated, and

most of the time there is a fee to receive (and send) TMs.

Table 1 was compiled using information from the operator web

sites regarding text messaging capabilities and fees for the 11

companies with more than one million subscribers at the end of

2002 (the cumulative subscriber base for these 11 companies

accounts for approximately 87% of U.S. mobile phone

subscribers (Federal Communication Commission, 2003). The

table highlights the fact that there are some differences among

companies in both the number of characters per sent/received

message as well as in the cost to send/receive a TM at the time

this paper was written. For example, few operators offer the

ability to receive free messages (AT&T Wireless, Us Cellular,

and partially Cellular One). Nextel, T-Mobile and Verizon let

all customers receive TMs no matter the plan, but for a fee.

Every Nextel phone (new and old) has the built in capability to

receive TMs (and users are charged on a per-message fee

automatically and bundle fees are available as an option).

Other companies require a special add-on to the contract in

order to activate the option to receive TMs for a fee and these

options cost at least $4.99/month. The range of maximum

characters per message for these 11 companies is from 140 (T-

Mobile) to 500 characters (Nextel). Some smaller companies

have a limit of 120 characters per sent/received message.

While there are many differences in the technologies

across providers, continued development to synthesize these

differences in terms of carrier interoperability may help explain

the continued rise in the number of text messaging users

(Federal Communication Commission, 2003).

2.2: Sending text messages to Subscribers without using a

phone:

It is not necessary to have a wireless phone in order to

send a TM. As we mentioned previously, it is fairly

straightforward and generally free to use a company’s website

to send a TM provided you know the number and the provider.

Generally, these websites will limit the number of characters

per message in accordance with the company limit. The degree

of flexibility in other options that are available to a consumer

while using a company’s text messaging interface (TMI) will

often be greater for those who subscribe to a plan from that

particular company. For example, Cellular One customers can

send up to 25 160-character TMs at a time using the online

TMI while non-customers can send 12.

A company TMI assumes that you wish to send a TM to

a subscriber from that company. If however, the provider is

not known, other text messaging software or third-part websites

(e.g. Smseverywhere.com) can be used to send TMs to wireless

phone subscribers. Some of these interfaces will not break up

messages into smaller SMS/TM that are within the provider

limit, but rather, the subscriber will simply receive a truncated

message. For other software/website interfaces, longer

messages will be split into an appropriate number of smaller

sized SMS/TMs and then delivered accordingly. In this case,

messages should be at most 140 characters (or 120 characters-

if a national probability sample from all providers is required)

in order to reduce respondent burden in terms of expected

expense.

All of the 11 major carriers we described in Table 1

offer the possibility to use their website to send TMs to their

customers for free. Table 2 delineates the principal

characteristics and options for each company, including the

maximum number of characters possible per sent TM. As

shown in the previous Table 1, there are some differences

among the companies. For example few TMIs offer the option

to track the status of the TM delivery and few allow receiving

an email confirmation of such delivery.

Table 1: Text Messages Capabilities and Fees for the Top Ten Mobile Companies in the U.S

Company TM Capability Cost to send/receive per message Cheapest bundle text plan / month Max #

of chars

Verizon Yes- Included in every plan

2 cents to receive

2 cents for receiving/10 cents sending $2.99/100 messages Send/Receive 160

Cingular No, activate text messaging

capability; Plan dependent

10 cents Sending/Receiving $2.99/100 messages Send/Receive 160

AT&T Wireless Yes

Free

Free receiving/10 cents sending $1.99 for sending 25 messages 160

Sprint PCS No, activate text messaging

capability

10 cents Sending/Receiving $5.00/ 100 messages Send/Receive 160

Nextel Yes-Included in every plan 15 cents to receive or (10 cents to

send/receive with two-way messaging)

$5.00/300 messages 140

(500)

T-Mobile Yes-Included in every plan 5 cents Sending/Receiving $2.99 /300 messages Send/Receive 140

Alltel No, activate for $6.95 per month Free receiving $6.95 to send 300 messages 160

Us Cellular Yes, included in every plan 10 cents sending $2.95/50 messages 150

Cellular one 50 incoming messages included

with each plan and 5 cents thereafter

5 cents receiving, 10 cents sending $3.99/200 messages 160

Cricket No, activate for $4.99 Once activated, unlimited

sending/receiving

N. A. 160

Qwest Wireless No, activate for $5.95 Once activated, unlimited

sending/receiving

N. A. 160

4930

American Association for Public Opinion Research Section on Survey Research Methods

The majority of providers do allow the use of email to send

messages in the form of “number@company designated site.”

For some TMIs it is possible to create a login and password

and access a personal area (for customers only) where an

address book can be stored together with a log of the messages

sent. Generally the user can reply to that message if a phone

number is entered by the sender, and in some cases the reply

can be sent to an email address.

2.3: Text Messaging Pathways using TMI:

Once you have accessed the TMI to send a TM, one of several

possible events should occur. Since the possible pathways for

sending a TM from one phone to another would be a function

of their interoperability, we will focus on possible events that

may occur when a TM is sent from the carrier’s TMI to one of

its subscribers. For an example of what happens when one

sends a message using a mobile phone rather than a TMI,

consider the service provided by Cellular One. In this case the

subscriber must be in the “digital footprint” of service in order

to send a message to either a T-Mobile, Verizon, Sprint,

Cingular, Alltel and Dobson or U.S. Cellular subscriber. The

phone unit will continue to attempt to deliver/send the TM for

up to 72 hours compared to 24 hours for messages sent using

the TMI, which does not require a digital footprint, only a

Cellular One number. A study by Keynote System done in

December 2003 in New York, Dallas, and Seattle, showed that

out of 25,000 SMS sent between various carriers such as

AT&T Wireless, Cingular, Nextel, Verizon, and T-Mobile, 5%

were never received and 7.5% were not received within 120

seconds (Sterlin, 2003).

Once a TM is sent from the TMI, the message may end up

being lost in the network. Although some providers will

attempt to deliver the message for up to seven days, some

messages may be lost in the network due to a working number

that is out of area, a phone that is not turned on during the

delivery period. The message may not be delivered or returned

if the recipient subscriber has a phone without text messaging

capabilities or if the recipient has a phone with text capabilities

but TMs can be outside the contract (i.e. the subscriber has not

elected these options). Lastly, is it possible that a mobile phone

was ported to a landline number due to the recent number

portability option (FCC, 2003) , thus impeding the reception of

the message. It may be possible for the TM to be delivered to

a number ported to another wireless provider, but since the

message has been sent under the assumption of a valid

subscriber for the particular carrier, the message may actually

be lost in the network or become undeliverable. Certainly, if

the number has been ported from a wireless phone to a landline

phone the TM will not be delivered (a special case of the

phone that does not have text capabilities).

Once the message reaches the mobile phone, three

outcomes are possible: the respondent reads it, the respondent

does not read it voluntarily or inadvertently, even though s/he

know how, or the respondent does not read it because s/he

does not know how to read a message on his/her mobile

phone. At the present time we do not have numbers to attach to

each of these possibilities, but we hope to have clarified what

may happen as a message attempts to reach its intended

recipient.

3. Wireless Web as a tool for Survey Research involving

U.S. Wireless Subscribers

In Europe, TMs have been used not only to invite people to

participate in a survey, but also to poll subscribers directly via

embedding quesitons within the body of the TM. In particular,

“SMS Surveys” are precisely those surveys sent as TMs from

research organizations. The body of the message contains a

few short questions and respondents can reply to the message

by including their response options. For a methodological

Table 2: Basic Characteristics of Text Messaging Website Interface for Major Providers

Company Tracking /

Delivery

confirmati

Email to

text

message

Other Max #

of char

Verizon Yes/Yes Yes Receiver can reply to a cell phone number or email address 160

Cingular No/No Yes 160

AT&T wireless No/No Yes User must agree to terms and conditions of the service

Digital PCS phones have lower max. # of char. they can receive without truncation

(i.e. 110)

160

110

Sprint PCS Yes No 160

Nextel Yes/Yes Yes Subscribers with reply capabilities have a larger per message limit (i.e. 500

characters).

140

500

T-Mobile No/No Yes User must agree to terms and conditions of the service No TMs to prepaid plans 140

Alltel Yes/No Yes Website alerts if the receivers cannot receive the TM 160

Us Cellular No/No Yes User must agree to terms and conditions of the service 150

Cellular one No/No Yes Cellular One subscribers can use the web interface to send up to 25 160-char.

messages at a time (non-subscribers are limited to 12 such messages at a time).

160

Cricket No/No No 143

Qwest Wireless Yes/Yes No User must agree to terms and conditions of the service . It is possible to set priority:

high, normal

175

1 Max number of characters in the message, excluding the fields: from and subject.

2 The field from is mandatory, so the resulting number of characters is the maximum minus the size of the from and in some cases the subject field[s].

4931

American Association for Public Opinion Research Section on Survey Research Methods

experiment on the number of questions for SMS surveys see

(Cooke, Nielsen, & Strong, 2003). Using this strategy

(Widman & Vogelius, 2002) measured daily newspaper

readership in Norway and (Down & Duke, 2003) measured

bank customer satisfaction and attitudes for measles mumps

rubella vaccination in UK.

Mobile telephones may also be used to access surveys

on the World Wide Web. In one possible application, a user

could be sent a text message specifying the web address of the

survey. Eventually the website could be browsed on the screen

of the mobile phone using the phone’s internet capabilities.

Within the U.S., such technology is still being optimized for

mass consumption. The types of capabilities along with the

level of services provided do vary by Wireless Provider as

illustrated below in Table 3. We see the mobile phones with

wireless web capabilities as a bridge between Internet surveys

and traditional telephone surveys. More research will be

needed in order to understand how people use and respond to

mobile phone wireless web interfaces in the survey setting.

4. Nextel Text Messaging Study

Nextel subscribers have been identified as the target

population for a study on the use of text messaging as a

primary method of contacting wireless subscribers and inviting

them to participate in a short 20 question survey regarding the

use of cellular phones. A probability sample of 898 Nextel

subscribers was selected using a Telecordia file sorted by

working cell phone area-code, prefix and 1000-block

combinations. The distribution of the sample of numbers by

Census Region was 18.96% Midwestern, 22.10%

Northeastern, 25.24% Western and 33.70% Southern.

This distribution was not identical to the distribution of all Nextel

Numbers but well within the bounds of expectation considering

the sampling variability one will encounter by using selecting a

sample size of 898 from the total number of thousand-blocks

allotted to Nextel [over 20,856].

An attempt was made over a period of four to six weeks to

deliver up to four TMs inviting the selected sampling unit to

participate in a survey. In the body of the message we included a

toll-free (800) number that the subscriber could use to contact the

survey research center to complete the survey or decline

participation. The first such TM was not as detailed as the

remaining three messages in that it did not include a schedule of

the research center fielding returned calls. Attempts to send text

messages were initiated on various times and days of the week,

with the first such message sent out on a Saturday since calling is

often free using weekend minutes. Because our main outcome of

interest was the usefulness of text messaging as a means to make

contact, we focused on the contact rate for the study, rather than

the survey response rate. We expected a contact rate of

approximately 5% using this protocol based on prior experience

with cellular surveys (Steeh, 2004).

The Nextel website Text Messaging Interface (NTMI)

was used to send each TM and this site provided an instant status

of the TM, but variation in network resources made this status

less reliable from one attempt to the next. In addition to the

instant status, an email delivery confirmation was requested for

every attempt and this confirmation (that resulted in the efforts of

Nextel to deliver the message for a period up to 7 days) was

deemed the final outcome of any particular text messaging

attempt. Numbers with two undeliverable email confirmations

Table 3: Mobile Phone Internet Capabilities By Company

Company Type of technology/

Name for customers

Capabilities

Costs

Lowest monthly plan

Verizon WAP /Mobile web by MSN called

VzW

Reading Hotmail, send a instant message with MSN

Messenger and browsing wap enabled websites

4.99 plus airtime (minutes

deducted from the plan)

Cingular WAP GPRS/WAP

Cingular Media Wireless Internet

Reading email, browsing wap enabled websites 3.99 plus airtime or 6.99 for 1

MB (no airtime)

AT&T wireless WAP GPRS/WAP / Mmode Reading AOL, Yahoo!, Earthlink, Mindspring,

NetZero, Prodigy or sprynet email account . Sending

instant message using AOL instant messenger or

Yahoo! Messenger Browsing wap enabled websites

2.99 plus 2¢ per Kb

Sprint PCS WAP ? / PCS Vision Accessing email

Browsing up to 95 predetermined websites

Sprint PCS Vision packages for

15$ including unlimited web

access

Nextel Java 2 Micro Edition Nextel Online

Web

Accessing email,

Sending instant message using AOL instant

messenger. Browse up to 80 predetermined websites

3.50 Unlimited web access

T-Mobile WAP GPRS/WAP

T-zones mobile Web

Accessing email, sending instant message using AOL

instant messenger, browsing wap enabled websites

4.99 unlimited web access

Alltel WAP/Mobile Web Service Axcess Accessing email, browsing wap enabled websites 5.95 plus airtime

Us Cellular WAP / Easyedge Accessing email, browsing wap enabled websites 1¢ per Kb or 5.95 per 2 MB

Cellular-One WAP? / Hello 2 Web Accessing email, browsing websites 7.99 including 250 airtime min

Cricket Not possible - -

Qwest Wireless ? / BrowseNow Accessing email. Only text browsing, no images 9.99 unlimited web access

The microbrowser in the WAP-enabled phones can only access sites that are in WML (Wireless Markup Language) format. If you enter an HTML address not WAP

compatible, you will receive an error message.

Kilobyte estimate when browsing the web: check weather, 8 kb; check top news/read an article, 10 kb; send an email (10 5-character words) 4 Kb.

4932

American Association for Public Opinion Research Section on Survey Research Methods

were deemed not working. In our study, 40% of the numbers were

deemed “nonworking” and the overall “delivery hit rate” was

computed to be 57.9% (at least one email confirming delivery).The

overall contact rate for was 3.823% which was not significantly

different from the hypothesized value of 5% (p-value=.2286).

The

overall final disposition of the Nextel sample is given below in

Figure 1.

Figure 1:

Distribution of the Final Disposition by Reply

Capability Status at the onset of the study period (total sample

size=828).

An interesting characteristic of the sample was that 9.1%

of numbers were reported by the TMI to have reply

capabilities (i.e. two-way text messaging capabilities) (RC) at

the time we initiated the text messaging. This proportion

increased significantly to 15% at the end of the study

(McNemar’s test z=6.096, p-value<.0001) even though 7 of the

original reply capable numbers had lost this status.

Incidentally, the distribution of wireless numbers across

census region was not significantly different for those numbers

that had reply capability at either end of the study compared to

those phones that did not (i.e. region was not associated with

the advanced capability). The RC status of the phone number

was an auxiliary variable that was available form the TMI and

since this status is an elected option from Nextel, it should

indicate some type of working status. The distribution of the

final disposition varies significantly between those numbers

that had reply capability before the study and those that did not

(χ

(4)= 226.258, p-value<.0001); see Figure 2.

The proportion of numbers classified as pending is greater

among those numbers with RC category compared to those

without this capability. However, the distribution of the

number of “Sent” messages given by the TMI at the time the

TM was sent was not as positively skewed for reply capable

numbers compared to those without this capability (not

shown). The differences in these values by type of RC are

displayed in Table 4. We note that the second through fourth

columns form a numeric partition of the first column.

Figure 2. Distribution of the Final Disposition by RC Status at the Onset of the Study Period (Total Sample Size=828).

Table 4. Sample Characteristics for Mobile Phone Numbers that either had no RC, Retained RC or Who Added RC .

Grouping Number of text

messages sent

Number of email

delivery confirms

Number of email

undeliverable

confirms

Number of

pending messages

Number of sent

confirms from

Website

No RC (688)

Mean

St. Dev.

3.367

1.246

2.132

1.951

0.888

0.978

0.350

1.218

0.163

0.521

Retained RC (68)

Mean

St Dev.

5.029

1.795

0.926

1.201

0.412

0.815

3.691

2.627

4.235

2.468

Added RC (55)

Mean

St Dev.

3.927

1.412

2.018

1.748

0.745

0.907

1.164

2.025

1.364

2.094

Total (828)

Mean

St Dev.

3.546

1.391

2.027

1.909

0.835

0.969

0.687

1.719

0.595

1.532

4933

American Association for Public Opinion Research Section on Survey Research Methods

5. For the Future

While one advantage of the text messaging/cell phone survey

approach is that you reach the final sampling unit, sending

only TMs and using the current results from company websites

may result in classifying numbers incorrectly. Later email

confirmation is essential. In addition, if numbers are

reassigned over short time spans, it may be possible that the

number of distinct subscribers who read the message exceeded

the number of distinct wireless numbers we were using in our

sample. We do not have data on how long a particular

subscriber has been using the wireless number used to contact

them. As companies respond to the number pooling

regulations imposed by NANPA we may expect faster

recycling of wireless numbers. Number portability will also

have implications in tracking, composing and delivery of

messages.

Nextel was selected as the primary source for wireless

numbers primarily because text messaging is available to all

their subscribers and because the advanced Nextel TMI

provided email confirmations. Using other providers may

make difficult the interpretation of non-deliverability since

some companies (including Verizon and Cingular) provide

ambiguous undeliverable messages for some subscribers who

simply do not have the ability to receive TMs. Many wireless

providers are currently in the process of developing and

improving text messaging options for their subscribers. For

example, in a recent phone conversation (January, 2004) with a

representative of Verizon Wireless, it was revealed that

Verizon just completed a round of upgrades to their processing

algorithms for text messaging interfaces and delivery systems.

We believe that as the capabilities of text messaging and

TMI’s advance, the generalizability of the usefulness of the

results of this study may extend to a text messaging experience

from not only Nextel but other companies as well. The future

use of text messaging in the survey sampling context is sure to

be influenced by legislation/regulations on SPAM,

technological advancements and our understanding of the full

spectrum of communication using mobile devices.

References

Cellular Telecommunications & Internet Association.

(2004). CTIA Semi-annual wireless industry

survey. Retrieved 06/01/2004, from

http://files.ctia.org/pdf/CTIA_Semiannual_Surv

ey_YE2003.pdf

Cooke, M., Nielsen, A., & Strong, C. (2003). The use

of SMS as a research tool. In R. Banks, J.

Currall, J. Francis, L. Gerrard, R. Khan, T.

Macer, M. Rigg, E. Ross, S. Taylor & A.

Westlake (Eds.), ASC 2003. The impact of

technology on the survey process (pp. 267-276).

Chesham, UK: Association for Survey

Computing.

Down, J., & Duke, S. (2003). SMS polling. A

methodological review. In R. Banks, J. Currall,

J. Francis, L. Gerrard, R. Khan, T. Macer, M.

Rigg, E. Ross, S. Taylor & A. Westlake (Eds.),

ASC 2003. The impact of technology on the

survey process (pp. 277-286). Chesham, UK:

Association for Survey Computing.

Federal Communication Commission. (2002,

07/27/02). FCC Consumers Facts: Keeping

Your Phone Number When Your Change Your

Service Provider. Retrieved 04/30/2003, from

http://www.fcc.gov/cgb/consumerfacts/numbpor

t.html

Federal Communication Commission. (2003). Eight

annual CMRS competition report (No. FCC 03-

150). Washington DC.

SIBIS. (2003). Benchmarking telecommunication and

access in the information society. Retrieved

04/01/2004, from

http://www.empirica.biz/sibis/files/WP5_No1_T

_and_A_2.pdf

Steeh, Charlotte (2003). Surveys Using Cellular

Phones: A Feasibility Study. Paper presented at

the Fifty-eighth Annual Conference of the

American Association of Public Opinion

Research.

Steeh, Charlote and Buskirk, Trent D. (2004) R U

There: Using Text Messaging in U.S. Wireless

Phone Surveys. Presentation at the Sixth

International Conference of German Online

Research, Duisburg-Essen, Germany, March

2004.

Telephia / Harris Interactive. (2002). Young adults set

to upgrade phones, drive usage of mobile

messaging applications in New Year. Retrieved

04/01/2004, from

http://www.harrisinteractive.com/news/newscats

.asp?NewsID=562

UPOC and Frank N. Magid & Associates. (2003).

More mobile owners turning to text messaging.

Retrieved 04/01/2004, from

http://www.upoc.com/corp/news/news-

emarketer.html

Widman, L., & Vogelius, L. (2002, 13 June). Daily

reach using SMS. Measuring and reporting

reach in real time. Paper presented at the WAM

- Week of Audience Measurement 2002,

Cannes, France.

4934

American Association for Public Opinion Research Section on Survey Research Methods