Recent research has shown strong support for science among the public in the US (National Science Board 2006). At the same time, this research shows that this same public is generally not well-informed about scientific issues (National Science Board 2006). In fact, the NSB report concludes that “the public’s lack of knowledge about basic scientific facts and the scientific process can have far reaching implications” (National Science Board 2006). This problem is not limited to adults, as tests of scientific literacy rate US students below the level of their counterparts in many other countries (National Science Board 2006). In particular, understanding of evolutionary biology is especially poor among Americans (Miller and others 2006), and it seems to be an issue from grade school (Michigan House Civics Commission 2006) to college (Holden 2006a). While this issue exists in other countries, the United States is the arguably the developed nation where the problem is most severe (Lazcano 2005; Miller and others 2006). Clearly, public perception of evolutionary biology is out of line with the actual state of science, and efforts to correct this should be a high priority.
One potential source of help is the World Wide Web, a venue that allows the dissemination of information to a wide audience quickly and cheaply. By any measure, the growth of the Web has been explosive (Zakon 2005) resulting in the ability of an individual to put up a site dedicated to any topic. Due to this growth, the current generation of students has grown up with the Web as a major part of their lives (Day and others 2005). In fact, with the advent of search engines, the Web has become the place to begin finding information on just about any topic (Barrie and Presti 1996; Underwood 2004). As access to the Internet has increased, particularly in schools, the Web has come to be used more and more as an educational resource, where students will turn to find the answers to questions on exams, term papers, and class assignments (Day and others 2005). Because the Web can be a cheaper way to disseminate and access material compared to traditional forms of publishing (Ciolek 1997), it has become mandatory that any group with information to share should have a presence on line.
As in other areas, the controversy between those who subscribe to various forms of creationism and those who support evolutionary science has moved onto the Internet. While the most visible area of the creationism/evolution (C/E) debate is the ongoing struggle to use political or legal action to disrupt the teaching of evolution in classrooms (Pew Forum nd; Associated Press 2004; Mervis 2005; Annas 2006; Bhattarcharjee 2006; Holden 2006a), a large part of the ideological debate is presented on the Web as well. The purpose of this paper is to develop a basic understanding of the state of the C/E debate on the Internet and make some observations as to how the Web portion of this debate has changed in the last several years. The vast amount of information available on the Internet and its constantly changing nature make a complete review of the state of the Web fruitless. Instead, my purpose is to examine what a naïve individual might find when searching for information on creationism or evolutionary biology. For this reason, there is no content analysis of particular sites to rate their accuracy or objectivity; this would not be something that a naïve individual would be able to ascertain. Given the political nature of this debate, the individuals involved on either side are unlikely to be swayed by opposing arguments, but the information presented on-line could become the basis for an individual’s developing a better (or worse) understanding of the nature of evolutionary science. For that reason, I have focused on the websites that would be found using particular queries that students might use at the beginning of a search for information. These methods were first applied in 1999 and then repeated with minor changes in 2005. This paper will focus on the results from 2005, but I will also discuss comparisons between the results from the two years.
All searches were run using the Metacrawler internet search engine because it engages several different search engines to provide hits from a larger proportion of websites than would be possible with a single search engine (Lawrence and Giles 1998). Today’s more-popular Google.com™ has existed in one form or another since 1998, (Google, Inc 2006), but searches on Google often return million of hits for a search, while Metacrawler returns a much smaller number. For example, a Google search for “Charles Darwin” generated approximately 11 400 000 hits, while Metacrawler listed only 96. Furthermore, the search algorithm used by each search engine and its particular method of ranking and reporting hits to each search introduce bias into the results, making a direct comparison nearly impossible: there is no way to determine exactly what methodology a search engine uses. Therefore, these data should not be taken as a representation of the “true” state of the Internet. Instead, these results should be taken as a sampling of information that could be found when searching the Web — as someone unfamiliar with evolutionary science might experience.
After running each query, the first results page was saved to allow me to browse the sites in it. Only the first 20 sites listed by each query were examined both to decrease the number of sites to examine and to get a list of sites that were the easiest to find (and most relevant to the experience I was trying to simulate). I conducted searches on a number of different search terms that consisted of phrases that relate to the C/E debate as well as the names of prominent individuals on both sides (Table 1). The search terms were chosen arbitrarily, but an attempt was made to include the basic terms that apply (for example, “creationism”and “evolution”) as well as finding sites that were specifically related to teaching these concepts (such as, “teaching creationism”). The search engine was told to report a match only if the exact phrase in the query was found. Many of the results potentially overlapped, as a site found for one query might also be listed in response to several others. Because search engines rate and order the sites that match the query based on a variety of factors, including the number of times the search term is found on the page and the proximity of multiple search terms, there is no guarantee that closely related searches would identify the same sites as the top 20.
All 20 stored hits for each query were examined to classify them into a number of different categories. The classification system I used reflects my own impressions, but as much as possible, I used the information provided by each site to choose its classification. The primary division was into pages that were either “for” or “against” one side of the debate. Specifically, I defined a site as “pro-creationism” if it either rejects evolution entirely or requires that evolution be guided by an intelligent force — this includes young-earth creationists, “intelligent design” proponents, and some theistic evolutionists (for example, Malina 2006). A “pro-evolution” page is one that accepts the evidence in support of the theory of evolution and supports the scientific method as a mechanism for increasing our understanding about the world without trying to include non-scientific ideas. The important factor in the classification was how authors described how the world works. The question of religion was not intended to be a factor in this study, but due to the fact that religion is the driving force behind creationism, it is not possible to ignore religion completely when discussing the results. This classification system did not require that a pro-evolution site espouse atheism, because a page that only tried to prove what could be supported by scientific evidence was still classified as pro-evolution, regardless of the religious beliefs of the author (for example, Morton 2000).
The next category I used to classify web pages was based on the individual(s) responsible for producing and maintaining the websites. A “professional” site was one that was produced by an organization that (in whole or in part) deals with issues of the C/E debate. This is contrasted with “personal” web pages that were developed by individuals without the site’s being officially associated with an organization. This classification was not based on the credentials of the page author, but on the association between the author and any organization that might be supporting the website. For example, a web page written by a practicing biologist could be classified as a personal page if it were not representing the official view of a particular organization. This distinction was difficult to make in some cases due to the fact that the “stance” of the page and the identity of the author were not always clearly identified on the page retrieved by the search. To be conservative, I took the claims of the page author at face value, because an individual who knew nothing about the C/E debate would have no other way to decide on the stance of particular sites. These two classifications were suitable for the majority of pages that I found, but a few required additional categories. In many cases, a page was developed by a group of authors as a sideline to their regular occupations. In this case, I used a category called “collaborative” to indicate a page that is developed by several different authors to address these issues without being the primary job of any of them. The best example of this is the Talk. Origins archive, which includes writings that have been posted to the newsgroup of the same name by different authors over the course of many years (Talk.Origins 2006). While work has clearly been expended to produce a page that has a consistent interface (including a search engine) the majority of information is based on postings from the newsgroup. A similar sort of site is found at About. com, which is a collection of articles and links moderated by individuals referred to as “guides. ”This site displays many of the characteristics that would be associated with a professional page, but given the wide-ranging attitudes of the different moderators and the mission of About.com (About.com 2006) I felt that the site as a whole is more collaborative. Three additional categories were used as well: 1) “library” sites were sites that allowed users to look up reference information on any topic, 2) “links” described pages that consisted solely of a list of hyperlinks to information on the topic, but having no content of their own, and 3) “encyclopedias” included websites (such as Wikipedia.org) that serve as a collection of information about many topics. My initial inclination was to exclude these types of sites, but in my experience as an educator, these are among the most common reference sites that students use in their online searches.
When classifying sites, I wanted to avoid skewing the results by counting the same site multiple times. There are two ways that this could occur, which I called “duplicated”and “repeated”sites. A site was considered a duplicate if it was found multiple times within a single search. In general, one would not expect the same page to be reported as a hit in the same search, but given the fact that most websites consist of a single home page with multiple subpages, it is easy to see how several pages on a single site might be listed as hits for a single query. In this case, only one of the hits would be counted for that search due to the fact that once a particular page on a site is found, it is generally easy to get to the home page for that site, leading to all the different pages it might contain. A repeated site was one that was found by different searches. For each repeated site, only one hit was counted in the final classifications because the classification would only need to be done once, regardless of how many different searches returned that particular site. Thus the number of sites in the final classification was further decreased to count each site only once, no matter how many queries linked to that site. After removing duplicate and repeated sites, I examined the remaining sites to eliminate those that were not relevant to the C/E debate. Search engines have improved their ability to provide results relevant to a user’s queries, but they often still provide results that are not suited to the user’s needs. The presence of a particular search term on a given page is no guarantee that the page actually contains useful content. For that reason, I further narrowed the list of websites by excluding those that were not appropriate using a variety of criteria.
Any site that was not related to the topic of evolution or creationism was removed entirely from the results. There were a number of different reasons why such sites would not be useful for the purpose of this study. The largest single cause for a site’s exclusion was that its main purpose was raising money as opposed to providing information. This includes sites such as Amazon. com and other sites that may sell material related to the C/E debate, but are not involved directly. Because the purpose of this study was to describe the information that would be available to someone who knows little about this debate, it was my judgment that it is unlikely that a commercial site would itself be a primary source of information on the C/E debate.
Sites that primarily provided news reports were also rejected, because the purpose of these sites was not usually to inform readers about the scientific issues within the C/E debate. Due to the timing of my web searches, the majority of the news articles dealt with two issues. First, a report released in April 2005 by the Pew Forum on Religion in Public Life (40% of news pages addressed this report). Because this dealt with public opinion polling on a variety of issues concerning politics and religion, and because it only discussed public opinions on evolution or creationism, I felt that these sites were of limited use to someone seeking to learn about the facts of the C/E debate. The second set of reports dealt with ongoing political issues in school systems around the country including decisions by many school districts to challenge the accuracy of evolution in biology classes (40%). Again, these would be of interest to someone seeking information on public opinion and political issues, but these reports provided little or no factual information on the topics. In addition, many of these sites used reports from wire services, meaning that their text was identical or nearly so. The remaining 20% of the news reports linked to sites that were no longer available or required a subscription to access them. Based on this breakdown of sites, it seemed better to exclude them all to focus on other areas where someone researching the C/E debate could find more substantive information.
Similarly, lecture transcripts, interviews, biographies, book reviews, or discussions of the political or social views of the particular individuals named in the search queries were removed from consideration as well, unless they were part of a larger website about the C/E debate. A number of websites that function as discussion boards on various issues were excluded, because they were sites that listed the opinions of the various authors on many topics, not providing specific information on the C/E debate.
Additional difficulty arose due to the ease of publishing on the Web and the lack of oversight on the quality of the information (Barrie and Presti 1996). Due to this fact, some sites were excluded for poor quality information. Such sites might claim to address the C/E debate, but often seemed to be tracts on metaphysics instead of biology. These sites were usually accompanied by esoteric philosophical discourse without having a realistic understanding of the scientific process (for example, Davis 2005; Mamas 2005). Last, between the time I ran the queries and finished checking the web pages, some of the pages that were listed in the search were no longer accessible, a common problem when dealing with websites (Dellavalle and others 2003), so their content could not be examined and those sites had to be excluded. Once the final set of sites was determined, the remaining pages were examined to classify them.
|Query||1999||2005||Change from 1999 to 2005|
|Young earth creationism||55||74||+34.5%|
|Stephen Jay Gould||54||92||+70.4%|
|Total # of hits||779||1321||+69.6%|
TABLE 1:Summary of search results for each query. Only the first 20 hits for each query were examined. These numbers indicate the total number of hits for each query, without regard to duplicates, repeats, or the appropriateness of a particular web page.
aIn 1999 the top 20 hits for this search all dealt with the design of computer networks, so this search term was discarded.
bBecause Dembski’s first work on the C/E debate was published in 1998, he was not included as a search term in 1999.
cDue to the increase in the importance of ideas about “intelligent design” and a concomitant decrease in the importance of young earth creationism, Gish was not used as a search term in 2005. This is also due to the decrease of his importance as new creationists are taking up the battle.
In 1999, 779 hits were reported for 15 searches, and that number had expanded to 1321 hits for 16 searches in 2005. Only the first 20 hits were examined for each query giving a total of 300 sites in 1999 and 320 sites in 2005.
After duplicates were eliminated, so that each site was included for a particular search only once, there were 249 sites in 1999 (17. 0% duplicates) and 207 in 2005 (35. 3% duplicates). Further excluding repeated sites so that each site was only counted once resulted in 212 sites in 1999 (14. 9% of non-duplicate sites were repeated) and 140 in 2005 (32. 4% were repeated sites). When applying the criteria that were used to exclude sites that were not appropriate for the purposes of this study, a total of 138 sites (65. 1% of the remaining sites) were excluded in 1999 while 39 were excluded in 2005 (27. 9%). Of particular interest in 2005 was the large number of commercial sites (including on-line retailers and auction sites) because most of these sites did not seem to have any relationship to evolution at all. It is not clear why these sites ended up in the top 20 results for some of the queries. The end result of this process was to give 74 “acceptable” sites in 1999 and 101 in 2005. Because only the first 20 hits were examined, the analysis is properly restricted to examining trends between the two samples. For example, there is a general trend for an increase in the number of responses to the queries. As Tables 1 and 2 show, the number of hits for individual queries, as well as the total set for all queries, significantly increased between 1999 and 2005. There was an increase of 69. 6% for the total number of hits and an increase of 36. 5% when only examining the acceptable hits for each search. This increase in the number of sites is not particularly surprising, given the growth of the Web in the same time (Zakon 2005). In addition, for both years the total number of pro-creationist sites was higher, due to the fact that the number of professional pro-creationist sites is significantly higher than professional pro-evolution sites.
|Type of Web Page||1999||2005||Change From 1999 to 2005|
|Personal Web page |
|Personal Web page |
|Collaborative Web page |
|Collaborative Web page |
|Professional Web page |
|Professional Web page |
TABLE 2:Classification of Web pages after removing duplicated, repeated, and inappropriate sites.
The general increase in the number of hits to the various queries is probably affected by a number of factors, including the general growth of the Web, the expansion of C/E sites onto the Web, and changes in search engines. There were both an increase in the number of acceptable sites and also those that were excluded as unacceptable. The fluid nature of the Web makes any analysis on particular searches at particular times inexact, but the differences between the two samples make some qualitative trends discernible.
The first is the greater number of duplicate and repeated sites reported in the top 20 hits in 2005. This could have been due to consolidation among these websites so that there are fewer sites available to find. Another possibility is that there has been no change in the sites, but that there has been a change in the search engines, so that the sites they report are giving a different representation of the Internet. In fact, these options are not mutually exclusive, and it may well be that both the websites and the search engines are changing to produce this trend. When looking at sites that were not duplicated or repeated, there were more acceptable sites in 2005 (101, which is 31. 6% of the 320 sites that I originally recorded) than 1999 (74, which is 24. 7% of the 300 recorded), which may indicate that the sites that were being reported were in fact more useful than those that had been reported in 1999.
When examining the acceptable sites, there was a greater number of creationist sites in both years, but in 1999 there were roughly twice as many pro-creationist sites as pro-evolution sites. By 2005, there were four pro-evolution sites for every five pro-creationism sites. This is an encouraging trend because it suggests that there has been a general increase in the number of pro-evolution websites or at least an increase in the likelihood that these sites will be found by the search engines. This may mean that people searching the web will find more evolution sites than they would have in the past. Of particular interest is the fact that the total number of pro-creationist websites that were found did not change between the two years while the number of pro-evolution sites increased.
While more sites were reported to the queries in 2005, the actual usefulness of the queries is affected by the presence of repeated sites. In 1999, 57 sites were only found by one of the search queries, while the remaining sites were reported by as many as six different searches. Of the repeated sites, the vast majority were found by two or three queries. In 2005, the majority of the sites were also only found once, but one site (Wikipedia) was repeated for every search while another site (Talk.Origins) was repeated 10 times. These results are probably due to the fact that these particular sites both consist of large collections of pages that cover many of the topics that were used as search queries. The remaining sites were repeated no more than six times. This difference between the two years might be a result of changes in the makeup of the websites, or it could be due to the fact that the search engines classified the pages differently in the two years. It is also affected by the fact that Wikipedia wasn’t online until 2001, so that the results for that site cannot be compared between the two years.
I had particular interest in hits that resulted to queries that included “teaching” as part of the search term because they would seem to address the idea of providing instruction as opposed to simply refuting the opposing side of the debate. A number of sites included teaching materials that could be used to teach in schools or as part of a home schooling curriculum. In both the case of pro-evolution (National Academy of Sciences 1998; WGBH Educational Foundation 2001) and pro-creationist sites (Answers in Genesis 2006a; Let Us Teach Kids 2002), the teaching material available on the web often included general curricula and study guides as well as online videos and/or DVDs that can be used in the classroom. Overall, there was a larger total number of creationist sites, due to the large number of pro-creationism websites that were classified as professional. As might be expected, most of these sites are associated with organizations that have an explicit religious agenda, such as Answers in Genesis (Answers in Genesis 2006b) and the Institute for Creation Research (Institute for Creation Research 2006). Since 1999, however, there has been an increase in sites that attack evolution but claim to do so without reference to a particular religious belief (for example, Access Research Network nd; Discovery Institute nd). These organizations are most likely to be attacking evolution using the ideas of “intelligent design”. As this is the form of creationism that is popular at this time (Mervis 2005, 2006; Bhattarcharjee 2006) it comes as no surprise that there are many sites devoted to this topic. Given recent events favoring evolution over “intelligent design” (Bhattarcharjee 2006; Mervis 2006), it would not be surprising if the anti-evolution sites were espousing a new idea in a few years. Another interesting observation was the number of personal websites dealing with this issue. There was a large number of personal pro-evolution sites in 2005 that helped balance the greater number of professional pro-creationist web pages. While these “personal” sites are maintained by individuals without any ties to an organization, these sites often provided content that matches or exceeds what is available on some of the professional pages (for example, Babinski 2005). Unfortunately, given the financial resources available to many creationist organizations, it is unlikely that personal pages will be able to match professional creationist pages, but such personal pages still provide a useful way to cover the C/E debate. Another important development was the introduction of Wikipedia in 2001 (Wikipedia 2006). This is a website that serves as an encyclopedia that can be edited by anyone with Internet access. Because the content of Wikipedia is determined by consensus among many individuals, that material can change without warning (Fisher 2005). At the current time, Wikipedia contains over 1 million articles in English and thousands more in other languages. This site is easy to search and contains information on the ideas of creationism and evolution. Unfortunately, the ability of anyone to edit these pages also means that they are of varying quality. While browsing the hits to my queries that came from Wikipedia, I found them to be fairly accurate, which matches results of a study published by the journal Nature (Giles 2005), but there is no guarantee that this will be maintained in the future. Given the opinion among a majority of Americans that creationism is equal or superior to evolution (Associated Press 2005), the modification of Wikipedia in consensus with majority opinion could easily lead to it containing incorrect information (Fisher 2005).
This study has examined only the hits produced by the search engine and not the appeal or utilization of the sites themselves. A further analysis of this aspect of the search results could serve to improve the presentation of evolutionary biology on the Web so that we can be more effective at reaching those who are seeking information. While it is unlikely that any change in presentation will convince someone who has already determined which “side” he or she supports, it might still serve to convince those who have not made such a determination.
Due to the visual nature of the Web, sites that present the material in a way that is visually appealing may be more likely to attract the attention of someone looking for information (Zhang 2000; Becker and Mottay 2001; Lindgaard and others 2006). Obviously, it would be best if all sites present information accurately, but the methods used to present that information may be as important as factual accuracy. Good web design is becoming more important because Web users are coming to expect certain characteristics if a website is going to keep their attention (Skaalid 1999; Nielsen 2006). If some sites are more pleasing to view, then they may get more attention from users, leading to the impression that they have more validity. For this reason, future research should be directed at analyzing the sites to determine which designs are more effective.
Because this issue is a debate between two polarized camps, it should not come as a surprise that some sites specifically aim themselves at attacking the opposing viewpoint (for example, Discovery Institute nd, attacking evolution, or New Mexicans for Science and Reason nd, attacking creationism). A recent study showed that attacking false claims may actually increase how strongly people believe them (Schwartz and other 2007). Refuting creationism is a natural outgrowth of explaining how evolution works, but if too much time is spent attacking anti-evolutionary ideas, it can give the impression of being defensive, suggesting that evolution is a weaker idea. Given the generally low level of scientific literacy of the American public, there should be more online material that makes learning evolution easier (for example Brain nd), as it is imperative that people be educated about scientific methodology as a necessary step towards becoming informed citizens (Nowotny 2005).
Fortunately, it is possible to present information in an interesting and appealing way that still preserves its scientific integrity, otherwise, the books of Carl Sagan, Richard Dawkins, Jared Diamond, and Stephen Jay Gould would not be as popular as they are. Scientists need to be sure that we are working to make science more accessible while also defeating creationist ideas. If we spend excessive time refuting creationism, we may find that the time has been wasted, because resisting one form of creationism is a short-term benefit. Like Hercules facing the hydra, for every brand of creationism that is defeated, a new one develops. The recent successes in Kansas (Bhattarcharjee 2006) and Pennsylvania (Mervis 2005, 2006) have dealt a setback to the proponents of anti-evolutionary ideas, but it would be foolish to believe that the fight is over.
I would like to thank two anonymous reviewers for helpful comments on an earlier version of this manuscript. I would also like to thank Andrea Wolfe and the students from EEOB710 for helpful discussions during the initial portion of this project.