State and Federal E-Government in the United States, 2003 Click Here to Order E-Govt Data

By Darrell M. West, Brown University (401-863-1163) Email: Darrell_West@brown.edu

September, 2003

Table of Contents

Executive Summary

A Note on Methodology

Readability

Readability by State

Disability Access

Disability Access by State

Online Information

Services Provided

Services by State

Privacy and Security

Security by State

Privacy by State

Foreign Language Access

Ads, User Fees, and Premium Fees

Restricted Areas

Democratic Outreach

Email Responsiveness

Overall State E-Government Ranking

Overall Federal Agency E-Government Ranking

State-Federal Differences

Differences by Branch of Government

Differences by Agency Type

Conclusions

Appendix

Table A-1 Overall State E-Govt Ratings, 2002 and 2003

Table A-2 Overall Federal Agency E-Govt Ratings, 2002 and 2003

Table A-3 Individual State/Fed Profiles for Publications, Databases, and Foreign Language, 2003

Table A-4 Individual State/Fed Profiles for Ads, Premium Fees, Restricted Areas, User Fees, Services, and Readability, 2003

Table A-5 Individual State/Fed Profiles for Disability Access, Privacy, and Security, 2003

Table A-6 Best Practices of Top Federal and State Websites, 2003

Executive Summary

This report presents the fourth annual update on the features that are available online through American state and federal government websites. Using a detailed analysis of 1,663 state and federal government sites, we measure what is online, what variations exist across the country as well as between state and national government sites, and how e-government sites respond to citizen requests for information. We examine the differences that exist across the 50 states and between the state and federal governments as well as compare the Summer, 2003 results to 2000, 2001, and 2002.

We use three new features in our study this year. First, we rely on the Flesch-Kincaid Grade Level Readability test to evaluate the readability of government websites. With half of Americans reading at no higher than the 8th grade level, we wanted to determine how accessible government sites are to the citizenry. Second, we test actual disability accessibility using the "Bobby" evaluation software operated by Watchfire. Relying on both Section 508 and W3C Priority One Level guidelines, we examine every state and federal website for accessibility. Third, we look to see whether websites are accessible through any type of handheld device or personal digital assistant. This included mechanisms such as pagers, mobile phones, or other types of PDAs.

Among the more important findings of the research are the following:

1) 89 percent of government websites are not easily accessible to the citizenry because the sites read at higher than an eighth grade level of literacy. Fully two-thirds of all sites have language consistent with a 12th grade reading level, which is much higher than that of the average American.

2) Only 34 percent of state and federal sites in the United States meet the W3C disability guidelines and only 24 percent satisfy Section 508 of the U.S. Rehabilitation Act of 1973. Federal sites (47 percent) are more likely to meet the W3C standard than states (33 percent).

3) The presence of online services has improved over the last year. This year, 44 percent of state and federal sites have services that are fully executable online, compared to 23 percent last year.

4) One percent of government sites are accessible through personal digital assistants, pagers, or mobile phones.

5) There continue to be high levels of access to publications (98 percent) and data bases (80 percent).

6) A growing number of sites offer privacy and security policy statements. This year, 54 percent have some form of privacy policy on their site, up from 43 percent in 2002. Thirty-seven percent now have a visible security policy, up from 34 percent last year.

7) 13 percent of sites offered any sort of foreign language translation feature, up from the seven percent we found last year.

8) 17 percent of government websites have restricted areas, less than one percent have premium features requiring payment for access, and one percent feature ads.

9) States vary enormously in their overall ranking based on web presence. The highest ranking sites include Massachusetts, Texas, Indiana, Tennessee, California, Michigan, Pennsylvania, New York, Florida, and Kentucky. The lowest ranking states are Alaska, New Mexico, Nebraska, Mississippi, Alabama, Hawaii, and Vermont.

10) In terms of federal agencies, top-rated websites include FirstGov (the U.S. portal), Federal Communications Commission, Social Security Administration, Internal Revenue Service, Library of Congress, Postal Service, Dept. of Treasury, and Securities and Exchange Commission. The lowest-rated sites are the various federal circuit courts of appeals. The new Homeland Security Department scores in the lower third of federal agencies.

11) Government officials are more responsive this year in answering email queries. Whereas 55 percent answered our sample query last year, 68 percent did so this year. This is just below what other studies have found in the private sector. Emails sent to America's 100 largest corporations have a 70 percent response rate, according to a recent study done by the Customer Respect Group.

A Note on Methodology

This project is based on two sources of data. First, we undertake a comprehensive analysis of 1,663 government websites (1,603 state government websites, the federal portal firstgov.gov, 47 federal government legislative and executive sites, and 13 federal court sites). The list of web addresses for the 50 states can be found at www.InsidePolitics.org/states.html, while the federal government sites are located through the national portal, FirstGov.gov. Among the sites analyzed are portal or gateway sites as well as those developed by court offices, legislatures, elected officials, major departments, and state and federal agencies serving crucial functions of government, such as health, human services, taxation, education, corrections, economic development, administration, natural resources, transportation, elections, and agriculture. Websites for obscure state boards and commissions, local government, and municipal offices are excluded from the study. An average of 32 websites is studied for each individual state so we could get a full picture of what is available to the general public, plus all the major federal government sites. Tabulation for this project was completed at Brown University in Providence, Rhode Island by Joanne Chiu and Erica Dreisbach during June and July, 2003.

Websites are evaluated for the presence of a number of different features, such as online publications, online databases, audio clips, video clips, foreign language or language translation, advertisements, premium fees, restricted areas, user payments or fees, two measures of actual disability access (W3C and Section 508 guidelines), several measures of privacy policy, multiple indicators of security policy, presence of online services, the number of online services, digital signatures, credit card payments, email addresses, comment forms, automatic email updates, website personalization, PDA accessibility, and readability level.

In addition, in order to examine responsiveness to citizen requests, we sent an email to the human services department within each state (or a comparable department if there was no human services division). The message was short, asking the question, "I would like to know what hours your agency is open during the week. Thanks for your help." Email responses are recorded based on whether the office responded and how long it took in days for the agency to respond. The remainder of this report outlines the detailed results that came out of this research.

Readability

Literacy is the ability to read and understand written information. According to national statistics, about half of the American population reads at the eighth grade level or lower. A number of writers have evaluated text from health warning labels to government documents to see if they are written at a level that can be understood by citizens. The fear, of course, is that too many government documents and information sources are written at too high of a level for citizens to comprehend.

To see how government websites fare, we use a test of the grade-level readability of the front page of each state and federal government website that we studied. Our procedure is to employ the Flesch-Kincaid standard to judge each site's readability level. The Flesch-Kincaid test is a standard reading tool evaluator and is the one used by the United States Department of Defense. It is computed by dividing the average sentence length (number of words divided by number of sentences) by the average number of syllables per word (number of syllables divided by the number of words).

As shown below, the average grade readability level of American state and federal websites is at the 11th grade, which is well above the comprehension of the typical American. Sixty-seven percent of sites read at the 12th grade level. Only 12 percent fell at the eighth grade level or below, which is the reading level of half the American public.

 

Percentage Falling within Each Grade Level

Third Grade

1%

Fourth Grade

1

Fifth Grade

1

Sixth Grade

2

Seventh Grade

2

Eighth Grade

5

Ninth Grade

5

Tenth Grade

9

Eleventh Grade

7

Twelve Grade

67

 

 

Mean Grade Level

11th grade

There are some differences between state and federal sites. Sixty-eight percent of state sites read at the 12th grade level, while 63 percent of the federal sites do so. It matters a bit what the branch of government is. Sixty-nine percent of executive branch sites are written at the 12th grade level, compared to 65 percent of legislative sites, 60 percent of judicial sites, and 56 percent of portal sites.

Agency type matters much more, although not always in a manner consistent with the particular audience served by the website. One might expect that agencies serving more educated clientele would gear their website to a higher level than those serving more poorly educated people.

However, as shown below, agencies potentially geared toward the less educated do not have lower grade-level readability levels. For example, corrections departments reports the highest percentage (83 percent) of websites written at the 12th grade level. Other agencies that have a high percentage of sites written at the 12th grade level are budget (81 percent), economic development (79 percent), elementary education (74 percent), housing (69 percent), health (69 percent), human services (67 percent), and taxation (46 percent).

 

Elem Educ

Hum Serv

Health

Hous

Correc

Budge

Tax

Econ Dev

Read at 12th grade level

74%

67%

69%

69%

83%

81%

46%

79%

Readability by State

Readability levels vary significantly across individual states and the federal government. The state whose sites are geared to the highest grade level (meaning the least accessible in terms of average readability) is Utah with a mean grade level of 11.7 across its websites. It is followed by Mississippi (11.5), Texas (11.5), Virginia (11.5), Minnesota (11.4), Arkansas (11.4), and Idaho (11.4). The state whose websites are geared to the lowest grade level (meaning the most accessible) is Rhode Island, which has an average readability level of grade 10.1.

Average Readability of State and Federal Government Websites

UT

11.7

MS

11.5

TX

11.5

VA

11.5

MN

11.4

AR

11.4

ID

11.4

AZ

11.3

CO

11.3

VT

11.3

WA

11.3

OR

11.2

CA

11.2

HI

11.2

CT

11.2

LA

11.2

IN

11.2

WV

11.2

ND

11.2

KS

11.1

SC

11.1

IL

11.1

KY

11.1

DE

11.1

MI

11.1

NE

11.1

FL

11.1

MT

11.1

TN

11.1

GA

11.1

PA

11.0

OH

11.0

NC

11.0

MA

10.9

OK

10.9

WY

10.8

IA

10.8

NV

10.8

MD

10.8

NM

10.7

NY

10.7

NJ

10.6

AL

10.6

US

10.6

AK

10.5

WI

10.5

MO

10.4

NH

10.4

ME

10.4

SD

10.3

RI

10.1

 

 

Disability Access

This year, we altered our test of disability access by examining the actual accessibility of government websites, not just claims of accessibility. In the past, we looked at whether sites displayed TTY (Text Telephone) or TDD (Telephonic Device for the Deaf) phone numbers which allows hearing-impaired individuals to contact the agency by phone, provided text labels for graphics, or claimed that they were disability-accessible. This approach has the obvious disadvantage of not providing an actual test of accessibility so this year we use the online "Bobby" service at http://bobby.watchfire.com to test actual accessibility.

We rely on two different standards of website accessibility: compliance with the Priority Level One standards recommended by the World Wide Web Consortium (W3C) and compliance with the legal requirements of Section 508 of the U.S. Rehabilitation Act of 1973. For each test, we enter the URL of the particular agency being evaluated and use this "Bobby" analysis to determine whether the website complies with either the W3C or the Section 508 guidelines. Sites are judged to be either in compliance or not in compliance based on the results of these two tests.

In this year's study, 33 percent of state and federal sites satisfy the W3C standard of accessibility and 24 percent meet the guidelines for Section 508. Federal sites (47 percent) are more likely than state sites (33 percent) to meet the W3C standard of accessibility. There are few differences between states (24 percent) and federal sites (22 percent) when it comes to meeting Section 508 accessibility standards, as measured by our Bobby analysis.

 

Federal Sites

State Sites

W3C Accessibility

47%

33%

Section 508 Accessibility

22

24

There are some agencies that indicate on their website that they are in compliance with the Bobby standard yet do not pass the test. These agencies include the South Caroline Dept of Education, the North Carolina Division of Aging, the Delaware Dept of Revenue, the Georgia Dept of Education, and the North Dakota's Governor's office. It is possible that these sites were in compliance at one point in time, but that later changes to the site removed them from compliance. As we suggest in our conclusion, it would be useful for agencies to list the date of Bobby compliance so visitors know when the site passed the test.

Disability Access by State

When looking at disability access by individual states, there is tremendous variation in the percentage of each state's sites that are accessible. The states doing the best job on disability access are North Dakota (84 percent of its sites are accessible using the W3C standard), Kansas (74 percent), New Hampshire (68 percent), and Texas (67 percent). The poorest states when it came to W3C accessibility are New Jersey (none of its sites met the Bobby test), Mississippi (3 percent were accessible), and Iowa (10 percent compliance).

Percentage of State and Federal Government Websites with W3C Disability Access

ND

84%

KS

74

NH

68

TX

67

ME

59

MO

55

PA

53

MI

52

NY

48

SD

48

US

47

KY

46

HI

44

RI

43

WY

43

MN

41

CT

39

FL

39

CO

38

NC

38

DE

35

TN

35

AK

33

NV

32

SC

32

WI

31

MA

30

IN

28

VA

27

WV

27

IL

26

CA

25

GA

24

NE

24

AZ

22

ID

22

OH

22

WA

22

AR

39

MD

19

MT

19

NM

19

OR

19

VT

19

LA

16

AL

16

OK

16

UT

14

IA

10

MS

3

NJ

0

 

 

Online Information

In looking at the availability of basic information at American government websites, we find that access to publications and databases are excellent. Ninety-eight percent of sites provide access to publications (up from 93 percent last year), while 80 percent have databases (up from 57 percent in 2002).

Similar to the patterns found in previous years, most websites do not incorporate audio clips or video clips into their sites. Only eight percent provide audio clips, up from six percent last year, and 10 percent have video clips (up from eight percent last year).

Percentage of Websites Offering Publications and Databases

 

2000

2001

2002

2003

Phone Contact Info.

91%

94%

96%

--

Address Info

88

93

95

--

Links to Other Sites

80

69

71

--

Publications

74

93

93

98

Databases

42

54

57

80

Audio Clips

5

6

6

8

Video Clips

4

9

8

10

Services Provided

Fully executable, online service delivery benefits both government and its constituents. In the long run, such services offer the potential for lower cost of service delivery and it makes services more widely accessible to the general public, who no longer have to visit, write, or call an agency in order to execute a specific service. As more and more services are put online, e-government will revolutionize the relationship between government and citizens.

Of the web sites examined this year, 44 percent offer services that are fully executable online. This is nearly double the 25 percent that had online services last year. Of the sites this year, 56 percent have no services, 15 percent offer one service, 8 percent have two services, 5 percent have three services, 3 percent have four services, 2 percent have five services, and 11 percent have six or more services. Clearly, both state and federal governments are making significant progress at placing fully executable services online.

Percentage of Government Sites Offering Online Services

 

2000

2001

2002

2003

No Services

78%

75%

77%

56%

One Service

16

15

12

15

Two Services

3

4

4

8

Three or More Services

2

6

7

21

 

Common services that are available online include: ordering a copy of birth or death certificates; filing consumer complaints; filing business and payroll taxes; placing new hire reports; updating professional licenses; filing Uniform Commercial Code reports online; reserving a hotel or campsite; purchasing hunting, fishing, or sporting license; renewing motorboat/snowmobile/all-terrain vehicle registrations; renewing driver's licenses; paying speeding tickets; renewing car registration; ordering duplicate driver's license; ordering special plates; purchasing transportation passes; ordering duplicate registration for motor vehicles; registering to vote, and subscribing to national "Do Not Call" registry.

Some states offer unique and innovative services online. For example, Maine and Virginia have a "live" help feature in which citizens can get instant help from a real person if they encounter a problem at that website. The Washington portal has six foreign language options, including Cambodian, Vietnamese, and Mandarin Chinese. Florida offers documents in English, Spanish, and Creole. The Alaska DMV waiting room has an online webcam so people can see a real-time measure of how crowded lines are and thereby judge when a good time to go would be. Massachusetts has an "online services" banner for each site. Washington offers online services right off its portal in an easy to find place. The Virginia portal has wireless access on its portal site.

One area where government sites are starting to make better progress is in offering the ability to make credit card purchases online. Of the government websites analyzed, 19 percent accept credit cards, nearly double the 10 percent found last year. With the increase in online services, more and more sites have created a means for credit card payments. However, most sites still do not allow digital signatures for financial transactions. We find less than 1 percent (eight sites in all) are set up for digital signatures.

Services by State

Of the 50 states and the federal government analyzed, there is wide variance in the percentage of states' web sites with online services. The following table shows the average number of online services found in various states and in the federal government. Massachusetts is the leader, with an average of 25.4 online services across its websites. This is followed by Kansas (5.4 services), the United States national government (4.8 services), New York (4.7 services), Maine (4.2 services), and Louisiana (3.8 services). The states having the lowest average number of online services are Alaska with an average of 0.2 services, Wyoming (0.3 services), and New Mexico (0.5 services).

Average Number of Online Services at State and Federal Government Websites

MA

25.4

KS

5.4

US

4.8

NY

4.7

ME

4.2

LA

3.8

MN

3.8

CA

3.4

ID

3.3

KY

3.3

FL

3.3

MD

3.2

MI

3.1

IA

3.0

MO

2.8

TX

2.7

IN

2.7

DE

2.5

PA

2.5

VA

2.5

WA

2.3

AR

2.2

NJ

2.1

MS

2.0

CT

2.0

IL

1.8

TN

1.8

SD

1.7

AZ

1.6

GA

1.6

OH

1.5

NC

1.5

SC

1.4

WV

1.1

CO

1.1

OR

1.1

HI

1.0

NE

1.0

NH

1.0

ND

1.9

OK

.9

NV

.8

MT

.8

VT

.8

WI

.7

RI

.7

UT

.7

AL

.6

NM

.5

WY

.3

AK

.2

Privacy and Security

A growing number of sites offer privacy and security statements. In 2003, 54 percent have some form of privacy policy on their site, up from 43 percent in 2002. Thirty-seven percent now have a visible security policy, up from 34 percent last year.

 

2000

2001

2002

2003

Privacy Policies

7%

28%

43%

54%

Security Policies

5

18

34

37

In order to assess particular aspects of privacy and security, we evaluate the content of these publicly posted statements. For privacy policies, we look at several features: whether the privacy statement prohibits commercial marketing of visitor information; use of cookies or individual profiles of visitors; disclosure of personal information without the prior consent of the visitor, or disclosure of visitor information with law enforcement agents. There has been a decrease in the degree to which consumer interests are protected compared to previous years. For example, whereas 39 percent of government websites in 2002 prohibit the commercial marketing on visitor information, this year that number dropped to 32 percent. The same is true for policies that prohibit the disclosure of personal information. In 2002, 36 percent of sites have this feature, but that figure declined to 31 percent this year.

Assessment of E-government Privacy and Security Statements

 

2001

2002

2003

Prohibit Commercial Marketing

12%

39%

32%

Prohibit Cookies

10

6

10

Prohibit Sharing Personal Information

13

36

31

Share Information with Law Enforcement

--

35

35

Use Computer Software to Monitor Traffic

8

37

24

Security by State

Despite the importance of security in the virtual world, there are wide variations across states in the percentage of websites showing a security policy. New Jersey is doing the best job in showing a security policy as 90 percent of its sites have a visible security policy that visitors can read. This is followed by Indiana (88 percent), Utah (83 percent), Michigan (81 percent), Tennessee (81 percent), and South Dakota (77 percent).

Percentage of State and Federal Government Websites Showing Security Policy

NJ

90%

IN

88

UT

83

MI

81

TN

81

SD

77

MA

73

CT

71

CA

69

NH

68

NY

64

US

62

IL

60

PA

59

WY

55

WA

50

AZ

47

WI

44

KY

43

GA

39

NC

38

OH

38

AR

34

VA

30

RI

30

DE

29

ND

28

KS

26

MO

26

MT

26