r Designing Effective Web Surveys in .NET Include Data Matrix ECC200 in .NET r Designing Effective Web Surveys

How to generate, print barcode using .NET, Java sdk library control with example project source code free download:
328 r Designing Effective Web Surveys generate, create datamatrix none with .net projects What is GS1 DataBar Figure 6.5 Examples of Man visual .net Data Matrix barcode ual and Automatic Authentication.

embedded in the URL, or a manual authentication requiring entry of a four-digit PIN. They obtained response rates (completes+partials) of 52% for the automatic and 54% for the manual group. While not statistically signi cant, the trend is in the opposite direction to ours, suggesting the manual login is more effective.

In addition, they found that those in the manual authentication group were slightly more likely to complete the entire survey (i.e., not drop out), and they found fewer nonsubstantive answers (less missing data) in the manual login condition.

In a second study, again with University of Leuven students and on the same topic, Heerwegh and Loosveldt (2003) tested a semiautomatic login, in addition to the manual and automatic conditions. Those in the manual login condition entered a four-digit PIN and ve-digit user name, while those in the semiautomatic condition entered just the PIN, while the user name was embedded in the URL. In the automatic condition, the user name and PIN were embedded in the URL.

The manual login achieved a response rate of 57.3%, compared to 65.7% for the semiautomatic and 60.

3% for the automatic login. While these differences don t reach statistical signi cance, they do suggest that some combination of automatic (passive) and manual (active) login methods may be most effective. We can summarize the conclusions from the three studies with respect to response rates as follows: 1.

Crawford, Couper, and Lamias (2001): automatic > manual 2. Heerwegh and Loosveldt (2002a): semiautomatic > automatic 3. Heerwegh and Loosveldt (2003): semiautomatic > automatic > manual.

Implementing the Design r 329 In addition, Heerwegh and Loosveldt (2003) found the manual version to produce high ratings of perceived con dentiality among respondents, followed by the semiautomatic version. The semiautomatic version also produced the lowest rates of nonsubstantive answers. While more research is needed to see if these results hold for different populations2 and for different survey topics (varying on sensitivity), I concur with Heerwegh and Loosveldt s (2003) conclusion that the semiautomatic authentication approach is the most effective.

This allows the password to be kept relatively short, making entry a little easier. Given that one might require respondents to enter an ID or password, it is important that this be free of any ambiguities. In one of our early Web surveys (Couper, Traugott, and Lamias, 2001) we included an alphanumeric password.

Several telephone calls and e-mail messages from sample persons complaining about problems accessing the survey led us to investigate further. Those passwords that included ambiguous characters (e.g.

, 1, l, O, 0) produced signi cantly fewer successful logins (43.7%) than those that did not (50.4%).

If respondents are to type in the ID and/or password, make sure it is relatively easy to do. Of course, we re all warned that longer passwords, and those containing a mix of upper and lower case letters and numbers are more secure, but what good does it do if one cannot get into the survey On the other hand, the ID should be suf ciently long to prevent accidental access to someone else s survey. For example, a four-digit ID gives one only 9,999 unique codes.

With a sample size of 1,000, this means that 1 in 5 of the possible numbers are assigned. If the sample size increased to 5,000, one in every two IDs is an active ID, and the likelihood of entering another survey through deliberate or accidental mistyping is increased. Partly for this reason, the U.

S. Decennial Census uses a fourteen-digit ID and the Canadian census a fteen-digit ID. This is another reason why the semiautomatic approach the combination of an embedded ID and a manually entered numeric password may strike the best balance between security and ease of access.

Finally, while none of the studies have examined the use of e-mail addresses as an authentication method, it is an approach used by some vendors of opt-in panels. I m not a big fan of this approach. Respondents may be reluctant to provide an e-mail address for con dentiality and security reasons.

Entering an e-mail address serves to highlight the identi ability of the responses relative to entering a string of numbers provided by the survey organization. In terms of security, entry of. For example, in opt-in pan barcode data matrix for .NET els, trust is built up over several successive surveys, and the convenience of an automatic login may trump other considerations..

330 r Designing Effective Web Surveys information in a recognizable format such as an e-mail address runs the risk of interception by spambots, spiders, or other programs trolling the Web to harvest such information, unless a secure site is used. On the other hand, using an e-mail address as a login is easy for respondents to remember, and is a common practice in e-commerce in combination with a user-generated password. In summary, the login and authentication process should strike a balance between making it easy for respondents to get to the survey, and providing appropriate levels of security to reassure respondents that one takes issues of con dentiality seriously.

The particular approach one uses may then depend on the nature of the relationship with the potential respondent and the sensitivity of the information being collected..
Copyright © . All rights reserved.