A New “Quick Look” At CE Reports
Sometimes, what seems like a brilliant idea loses its luster the more people that see it. Our Report Rating - the large number that appears at the top of our new report format - was just such an idea.
We thought it would be great to have a summary number at the top of the report that would tell the Accountability Partner how much attention should be paid to the report. We had “ratings” of zero through five. Basically, a zero meant there was nothing at all of note on the report. A five was not meant to be a “grade” or accusatory, it simply meant that there were sites that scored high, and so the report should be looked at in more detail to discern motivation.
Well, people misunderstood that, and Monday-morning quarterbacking makes it obvious where we went wrong. It is only human nature to look at a numerical system as a “grade.”
So, we implemented a new system that you will find in a box labeled “Quick Look.”
We simply have replaced the numbers with words, descriptive terms that tell the Accountability Partner exactly what the analysis means.
If the Rating used to be zero or one, the Quick Look says “Report Looks Good.” Our Quick Look algorithm seems to indicate that there is no pattern of viewing objectionable material.
If the Rating used to be two or three, the Quick Look says “Review Suggested.” This means just what it says. There are some sites that score high, but not enough to demonstrate a pattern of viewing objectionable sites, so the Accountability Partner should take at least a cursory view of the report, and perhaps ask a simple question, like “I saw some higher scoring sites, how are you doing? Were any of them inappropriate? Etc.”
If the Rating used to be four or five, the Quick Look now says “Close Review Recommended.” Of course, there are legitimate reasons for getting this statement. Anyone who uses the Internet extensively is likely to get such a Quick Look analysis from time to time. It is NOT call for accusations or jumping to conclusions. It may be obvious from the report (and/or the detailed report) what the motivation was. But there may well be need to ask questions regarding the sites, and this is what accountability is all about!
There also is another phrase that might appear in the Quick Look analysis. It might say “Reporting Period Too Short.” The Quick Look analysis is most accurate for a seven-day period. If the period is less than 5 days, this wording appears in the Quick Look box. If you want to see a Quick Look analysis, we suggest generating a report for a period of 7 days.
By the way, the Quick Look algorithm tries to analyze the entire report just as an experienced partner would view it. It asks questions like, how many high scoring, and very high scoring sites are there. Are they different domains? Are they on different days? Is the same high-scoring site visited at different times? Are the high-scoring sites occurring in the middle of the night? Etc.
We hope that members and partners don’t look at this Quick Look as a panacea, an end-all for analyzing the report. If you think about what we’re trying to do, you will realize that we are basically trying summarize all of your Internet use into one out of three concise phrases - you must admit, that’s a pretty daunting task. It really is just a “Quick Look.” It can only be attempted because of the unique Covenant Eyes dynamic scoring procedure, which allows relative comparisons. This simply could not be done if we simply maintained a black list of sites.
Please, feel free to comment - we will refine this tool as necessary. We want it to be as useful as possible. For more information, visit our Help and Support Center and feel free to comment here, or send comments to support@covenanteyes.com.
I noticed that the report contains tons of high-scoring pages that have nothing objectionable on them. The reports I get almost always say “Close Review,” but most of the high scoring hits are nothing.
The main value of the report seems to be that I can glance through a determine whether there seems to be a bad pattern. But the scoring system seems irrelevant.
Is there a way to improve the scoring so that it point out actual problems rather than being 99% false positives?
I would like to echo Frank’s observations. I find many “false positives,” as it were, in the ratings. Sometimes I wonder if many of the ratings are for embedded advertising. While I’ve seen quite a lot of embedded advertising that is less than innocent on otherwise innocent web pages, I’ve very seldom seen any such situations where what I’ve called “less than innocent” really concerns me. I’m FAR more interested in the “main page” being viewed. I may be misinformed about how this works, but I think if the CE system could distinguish more readily between things embedded within a page and the main thrust of the page, that would be helpful.
Hi Frank and Andrew,
Thanks for your questions about the Covenant Eyes report. You are correct that some of the high scores are indeed ads. Let me explain further.
When you visit a website, Covenant Eyes is not giving an overall score to the webpage you view. Actually, we are scoring each part of the page that appears on your screen, including photos, stories, ads and other items. What appears to be one ‘page view’ often consists of several if not dozens of URLs. Each URL has its own web address. As such, each of those URLs shows up on our report because, indeed, the computer did access those URLs.
And yes, we are diligently providing all of this information to show patterns of how the Internet is used. We believe it is necessary to provide all of this information in order to discern those patterns.
Here’s an easy example. If I visit the Sports Illustrated home page, I will see lots of links to great stories, photos and more about football, basketball, baseball and more. But I will also see a link to the Sports Illustrated swimsuit calendar or videos. All of those sports stories are going to receive very low scores, but the ads and links to the swimsuit pages are going to receive scores that suggest mature content.
If I do not click on those links, then my report will show a couple of high scoring hits that show I ran into temptation, just like I do when I watch TV at night or when I walk through the mall.
But if I do click on the swimsuit link, it will take me to a web page that has several, if not dozens of photos, cutlines, stories, and more that will receive scores that show mature content. Each of those URLs will appear on my report, which will provide the web address of the URL, a score, and the time and day I viewed them.
So, in short, Covenant Eyes reports show the temptations we face and the choice we make. Our QuickLook analysis takes into account the number of such sites that were visited, as well as the frequency, times, and other factors as well, in discerning whether or not there is a pattern suggested.
Here are a few additional tips.
1. Covenant Eyes’ age-based scoring automatically scores some web domains a little higher, because of their more mature target audience. Web domains such as Myspace, YouTube, Facebook, and social networking sites are scored somewhat higher. Of course, the individual text, ads, and other portions receive individual scores.
2. Image and video searches automatically receive higher scores. Why you ask? If you have ever done an image or video search you likely already know the answer. Often innocent searches bring up mature items that were neither intended nor expected. Videos and images are impossible to score accurately, so we recommend that such sites be visited by children only under the supervision of their parents.
Finally, we are not saying there is anything wrong with watching videos on YouTube, visiting a friend on Facebook, or searching for an image on Google. But we do think Accountability Members and their Accountability Partners should have open and honest discussions about how these resources are used and the temptations they face.
Hi Guys,
I appreciate your answers to the questions given, but they seem to be the same stock, “Well, you see. Our trade secret algorithms mean that some sites will be scored higher than others.” This STILL produces an extremely high volume of false positives.
From my perspective, it ends up being a very precious waste of time. I hope you realize that after taking the words “Close review suggested” seriously over and over again, I am coming to the point where those words don’t mean anything anymore. My Close Review reveals the 99% False Positives others have asked about in this blog.
That of course leads me to object to the time wasted, and little desire to conduct even a cursory review. This if course isn’t what your software is intended to do, so I thought I’d pass this along. I may be alone, but others I’m talking to echo my sentiments.
Please, I hope you can take this as positive criticism, in the spirit it’s given! I believe in what you are trying to do, I just don’t think this particular issue is helping….
Unfortunately, after contacting Rod, he would not give me permission to view his Accountability Log, so I’m not able to address his concerns directly.
Given that limitation, here are some thoughts, however. I certainly do take his comments as positive criticism, and we are working hard at making the scoring system as perfect as possible.
First, it is important to realize that the ‘close review suggested’ is more or less the default position… in fact, without the ‘Quick Look’ what other option is there? Before we had the Quick Look, close review was the only option besides ignoring the report. Close review is also the only option with any other program… take for example keystroke loggers that require extensive research to maintain accountability (akin to trying to go through old VCR tapes to find one particular scene). The Quick Look was designed to shorten the task of accountability when our system can confidently say there is nothing bad about a particular report. When it is absoulutely obvious, even for our servers, that there is nothing of concern on a report, the Quick Look reports that there is generally no need for further review.
Perhaps the wording ‘Close Review Suggested’ sounds too accusatory? If anyone has any wording that would better convey what we’re trying to get across, I would definitely be glad to hear from you. We’re trying to help, not hinder, the process of accountability, and we want the Quick Look to be a shortcut when a report is obviously devoid of anything questionable.
As for the false positives, this is something that we are constantly working on. Our scoring system is unique, and the Covenant Eyes filter is the only filter that allows effective choice of sensitivity settings.
Nevertheless, because our filter is based on these sensitivity settings, we do try to make sure that we err on the side of safety… we’d rather see a site get too high a score than have sites that are bad get too low a score. Of course, we’d like to be perfect…
In our efforts to improve, we have recently more than doubled our scoring staff, and you should start seeing the fruits of that addition in weeks to come. In addition, our programmers have provided some really great tools for database analysis, which should greatly decrease the number of false positives.
The bottom line of all of this is, we’re working on the false positives, and I hope readers will have some suggestions on how to make the Quick Look not appear accusatory.
What is especially malicious about this approach is that at the root of addictive tendency is this struggle between the Id (lust/flesh) and the Superego (conscience). The exhausted ego (will/natural man/soul/wretched man) in the middle ends up giving in to temptation because it is completely wiped out in preserving sanity in the middle of the struggle. If CE wasn’t in the background empowering and prolonging the struggle - the Ego could long ago have resolved the conflict with a better outcome and eventually (not initially) integrate the individual (bring peace between the demands of the conscience and the desires of the flesh) into a coherent whole. Instead - the “damn the consequences and let the devil take the hindmost” strategy that CE uses - simply locks the addicted struggler into a protracted engagement from which true healing seems increasingly elusive. Keep it up long enough and this is exactly how people lose their faith altogether. No matter though - the longer the struggle - the longer CE gets paid.
AJ,
I’m afraid your analysis doesn’t hold up under the evidence. Using your terminology, the software is actually a tool facilitating the ego’s mediation between the id and super-ego. I could see what you mean if it was simply a matter of blocking or filtering software, which simply stops the Internet user from seeing the questionable Web pages (and thus, for some, prolonging the struggle with addiction). But we are talking about “accountability software.”
The value of the software is determined, in large part, by the value of the accountability relationships the user has between him/herself and others. Thus, the “exhausted ego” isn’t trying to preserve sanity amidst the struggle, nor locking a person in a protracted engagement with the temptation; rather, the accountability relationship becomes a resource to help the ego to synthesize reality and the id’s primitive drives.
I’m not sure what you mean about “let the devil take the hindmost.” Our software, if anything, is the exact opposite of this attitude. By encouraging people into accountability relationships we are calling others to assist “the hindmost” in their struggle.
The bottom line is that many people share with us that the software is a great help to them in overcoming addiction. Often people keep the software not as a means of beating addiction, but because they enjoy how the software helps to preserve their integrity on the Internet without blocking where they go or what they see. They enjoy the software as one facet of their now open and honest accountability relationships that have been strengthened amidst the struggle.
Please, let me know if you have any further thoughts or questions.