For some reason when it comes to computers and digitized records, genealogists forget that digitization does nothing to make records better than they were before. It is easy to forget about existing errors or where the records came from when working with digitized records, especially when they are easier to read than the original record.
Know What You Are Looking At
When I give a lecture on online sources, I always point out that it is up to the researcher to know the limitations of the source that has been digitized. This means understanding the pros and cons of the resource before it was digitized.
Be sure to know the record.
For instance, census records that have been digitized are no more accurate or better than their microfilmed equivalent. For example, you still have to deal with the fact that you can't be sure of who answered the enumerator's questions. In the case of a published family history or a compiled genealogy, it is important to know who the author was and if there are any major issues in accuracy.
It is up to me, as the researcher, to know what the potential errors are if I am using Pope's Pioneers of Massachusetts online. It is up to me to recognize that the original book has some mistakes and it is also up to me to understand how it was digitized and whether that digitization could introduce any errors to the text. In this example, the pages of Pope's book were digitized as images. The other type of digitization involves running the pages of a book through an Optical Character Recognition (OCR) program that "reads" the text and translates it into searchable text.
Understanding What Records Can and Can't Do
Some of the records that genealogists use online are simply misunderstood in general. By far the most misunderstood are the compiled family trees, such as are found World Family Tree. Some researchers expect that the information made available through the World Family Tree be 100% accurate. They expect the publisher to verify all the facts shared in these trees.
Really, you should work with the World Family Tree, for example, the same way you would a published family history. When working with a published family history, we hold the compiler of the information, rather than the publisher, responsible for any errors in the data. The same should be true of compiled trees made available on CD-ROM and online.
One source that has always been computerized, but is misunderstood is the Social Security Death Index. So often I hear from people who are concerned about how their ancestors aren't in it. It is important to remember that the SSDI is not an all encompassing index to deaths in the United States. In fact, while Social Security began in the late 1930s, the index that we know as the SSDI was not begun until 1962. Those who died before 1962, even if they were receiving social security, are not likely to be in there. This does not mean that the organization offering the SSDI should be questioned for their accuracy. Instead we should be aware of this limitation and accept it when we use this database.
Digitization has made genealogical research faster and more convenient. Of course, with those blessings have come a few problems. Understanding that digitization a record doesn't magically make the resource more accurate is the first, and in my opinion, most important step in working with information online.