Talk:Ad Fontes Media

Another attempt to remove the Columbia Journalism Review reference

In 2018, a Columbia Journalism Review article questioned the thoroughness of the Media Bias Chart and similar initiatives, stating that "the five to 20 stories typically judged on these sites represent but a drop of mainstream news outlets' production".

References

The most recent removal seems to be based upon WP:OR and a non-independent, primary source. --Hipal (talk) 17:56, 12 November 2021 (UTC)

@Hipal:I see this as a very, very fundamental example of an underlying systemic problem regarding how Wikipedia operates in its treatment of sources. Anonymous - and presumably, we can't know - unqualified "editors" are placing their own personal judgement above that of the professionals, the experts, in this particular matter (among others). Ad Fontes Media has received praise from across the mainstream media landscape. More importantly, it has be cited as an authority in peer-reviewed academic studies which actually delve into the issues with which we're all supposed to be talking about here.

To give but one example:

We compared the Media Bias/Fact Check list with a list of 115 sources from a recent paper from the Berkman Center at Harvard and found a 0.77 Pearson correlation with p < 2.2∗10−16 when comparing bias ratings for all 87 sources that the lists had in common. We also correlated the list from Media Bias/Fact Check with 105 sources listed in the Media Bias Chart from Ad Fontes Media. We found a 0.92 Pearson correlation with p < 2.2 ∗ 10−16 for the 85 sources these lists shared. Finally, we compared the larger list with a set of 200 matching sources from the AllSides Media Bias Ratings.6ere, we found a Pearson correlation of 0.81 with p < 2.2 ∗ 10−16. These correlations are strong enough for us to trust the larger partisanship classification of 1,434 sources for our analysis. It is important to have such a large set of sources in order to accurately determine the broader media consumption diets of our participants — as described below our participants visited hundreds of these sources..

— Frank Bentley, Katie Quehl, Jordan Wirfs-Brock, and Melissa Bica., "Understanding Online News Behaviors.", In CHI Conference on Human Factors in Computing Systems Proceedings (2019)

EnlightenmentNow1792 (talk) 08:18, 13 November 2021 (UTC)

This is not the venue for questioning how Wikipedia works.

Can you address the policy concerns concerning the content removal? --Hipal (talk) 04:00, 15 November 2021 (UTC)

Have done so, as you well know, in the edit summary, and in the academic, peer-reviewed source I quoted above. EnlightenmentNow1792 (talk) 16:43, 26 November 2021 (UTC)

I guess we're not going to find consensus then if you cannot provide a better explanation. --Hipal (talk) 16:52, 29 November 2021 (UTC)

I've restored it. Sorry I didn't get around to it earlier.

Should the example given above be included as well? --Hipal (talk) 22:24, 7 May 2022 (UTC)

Etymology

I just noticed the etymology section was removed the day I created it a year and a half ago as "SOAP." How is this WP:SOAP??? It's just explaining the Latin name so people don't think Ad means advertisement. I'm restoring it. YoPienso (talk) 04:14, 24 March 2022 (UTC)

The only reference is to their own website. Find an independent source that demonstrates it deserves mention. --Hipal (talk) 22:42, 7 May 2022 (UTC)

Who better than the company itself to explain what their name means?

I don't know what the "NOT problem" may be that you mentioned in your edit summary: Undid revision 1078942859 by Yopienso (talk) find an independent source to avoid NOT problems - some mention may be due, but no a section without an independent sources

Anyway, I'm undoing your revert and adding a ref to the Poynter Institute. YoPienso (talk) 06:37, 8 May 2022 (UTC)

WP:NOT.

Thanks for the additional reference. Seems WP:UNDUE as presented, but I appreciate the progress. --Hipal (talk) 02:47, 9 May 2022 (UTC)

Why do you think WP:UNDUE applies to giving the reason Ad Fontes was chosen for the organization's name? YoPienso (talk) 16:20, 14 May 2022 (UTC)

See WP:NOT. I decreased it's prominence per UNDUE, and I think your rewording is an improvement. --Hipal (talk) 20:38, 14 May 2022 (UTC)

@Hipal: This is merely an academic question birthed of my curiosity. Since I was perfectly fine with your edit, I didn't think it was worth my time or yours to pursue the question of WP:NOT, but I happen to have a bit of time right now to follow up. I never understood how you applied that policy. What is Wikipedia not that would exclude an etymology? Was it because you categorized it as "self-published"? If so, how else would anybody know why a business chose its name? Thanks. YoPienso (talk) 20:00, 2 June 2022 (UTC)

If it's important to present how a business chose their name, an independent reference will have covered it. We're not here to provide a soapbox for the subjects of our articles. --Hipal (talk) 02:11, 20 October 2022 (UTC)

Columbia Journalism Review

I added the qualifier to the 2018 Columbia Journalism Review to indicate that the criticism was with respect to The Chart when it was based solely on Otero's personal reviews of "5-20 stories" per news organization. This qualification is important since the current version is now based on the collective opinions of a panel of reviewers and each review is now based on substantially more than "5-20 stories". Nowa (talk) 20:38, 30 November 2023 (UTC)

Thanks for starting this discussion.

The content specifically includes the date of 2018, so I'm not clear how the qualifier helps, and it seems misleading. I'd prefer the quote from the edit summary placed in the reference itself. Wouldn't that be enough? --Hipal (talk) 20:46, 30 November 2023 (UTC)

That would be great. The key thing is for readers to realize that the CJR critique is about an earlier version of the chart when it was self published by Otero. Nowa (talk) 23:15, 30 November 2023 (UTC)

New proposed section: "Use in Academic Research"

A new proposed section, "Use in Academic Research", is presented below:

Ad Fontes Media ratings, and associated individual article ratings, are used in "reputable academic research". An advantage of using Ad Fontes data is that “it publishes ratings of news articles on a common topic from the same news cycle”. This allows researchers to get specific ratings for individual articles without having to impute an article’s rating from an overall rating of its published news source. Examples of how Ad Fontes Medias ratings have been used in academic research include:

• Determining if there is a relationship between exposure to US news media political bias and COVID-19 infection rates.
• Measuring to what extent Twitter’s personalized newsfeed algorithms amplify political bias.
• Determining "the difference between the proclaimed ideological bias of a news outlet and the ideology of annotated articles from the outlet".

An earlier version of this section was reverted as being "examplespam, So" (i.e., WP:Example cruft).

I've modified the earlier version and posted it here for further vetting.

WP:Example_cruft states: "Avoid the temptation to engage in original research by finding every example of a phenomenon...If the number of examples in an article become too many, consider pruning them, or creating a separate list at the bottom of the article."

I believe that the new proposed section is consistent with the recommendations of WP:Example_cruft.

The examples are all from the Haq reference (see footnote 2 on page 5) and were specifically cited by Haq to support his position that Ad Fontes ratings are acceptable for use in academic research. Hence the examples in the proposed new section are not original research.

Furthermore, Haq's five examples have been pruned to three since three of them could be combined to a single example related to COVID.

Finally, the list is presented at the end of the new proposed section.

I would appreciate any additional comments so that we can reach consensus on what material, if any, should be added to the article.

I look forward to your additional input. Nowa (talk) 13:10, 9 December 2023 (UTC)

So is the reference reliable for the information and indicate due weight? I'd say no, even if the topic was more strongly related to the subject matter. --Hipal (talk) 23:43, 9 December 2023 (UTC)

The Haq source is a pre-print. Has it been published yet, and is this information in the published version?

The relevant footnote itself cites CrowdTangle to show that Ad Fontes Media is "popular". CrowdTangle is Facebook, which is infamously unreliable for what is and is not popular. Why would its popularity even matter here? Later, the paper derives its information from Ad Fontes Media's own whitepaper. This is not compelling, to put it mildly.

The specific examples are a mixed bag, but 'Algorithmic Amplification of Politics on Twitter' is in part about Ad Fontes itself. Using this as an example seems to be misrepresenting that source, as well as a missed opportunity. There's also this from that source's supplementary material section: Initially, the team had access to only one media bias rating dataset (AllSides). Multiple teams raised concerns about the reliability of this third party source, and about making our findings dependent on the validity of a single underlying dataset. We have therefore obtained a license for a second media bias dataset (Ad Fontes Media). We present both sets of results in our paper with no normative judgment on the validity of either of these underlying sources of data.[1]

I take this to mean that these researches had exactly the same problem with AllSides and Ad Fontes as Wikipedia does. They are both arbitrary and subjective. From this it seems like the best they could do was use both of them averaged together and then hope the biases at least partially canceled-out. This is a reasonable approach to a recurring problem for this specific research paper, but it's absolutely not a ringing endorsement of Ad Fontes Media! It would be a disservice to this paper to misrepresent it as such. Grayfell (talk) 22:58, 9 December 2023 (UTC)

Thanks for looking closer at the ref. It looks poor. I wouldn't give it the weight being proposed, nor provide the examples. --Hipal (talk) 23:46, 9 December 2023 (UTC)

I'm not wedded to the examples. If you can propose alternative language that gives proper weight to the reference, that would be great. Nowa (talk) 19:18, 11 December 2023 (UTC)

These are great comments! Thanks for doing the additional follow up. I don't have any fundamental disagreements with what you are saying. If you can propose more appropriate language for the section, that would be great.

Regarding your specific comments:

• is this information in the published version? It is. I've updated the link to the accepted version. The material related to Ad Fontes use in academic research is the same.
• Why would its (Ad Fontes) popularity even matter here? It doesn't as far as whether or not Ad Fontes ratings are used for academic research. That's why its "popularity" is not mentioned in the proposed section.
• Later, the paper derives its information from Ad Fontes Media's own whitepaper. True, but this is in Appendix A where it is giving an overview of AFM and summarizing its methodology. Citing AFM's white paper is appropriate in that context. In any event, the paper is not citing AFM's white paper to support its assertion that AFM ratings are used in academic research.
• "We present both sets of results in our paper with no normative judgment on the validity of either of these underlying sources of data." That's an excellent point! I think it would be an important caveat to include in an updated version of the proposed section on Academic Usage.

Nowa (talk) 19:17, 11 December 2023 (UTC)

I'm still not really convinced this belongs at all, tbh. I mention the popularity thing because it's in the same footnote you cite. The paper is obliquely indicating that despite being 'newer' than other platforms, Ad Fontes is usable because it is popular. I don't accept that. The use by other academic works is almost in passing within the footnote. That footnote also specifically directs readers to that appendix entry which in turn directs readers to the Ad Fontes website. It is what it is, but as I said, it's not really persuasive. I also note that the previous footnote on the page (citing the previous sentence in the paragraph) is to Wikipedia. Specifically: "We select Gun Control as our political topic of interest, as it is a commonly studied topic for political bias and receives wide media coverage in the U.S. spreading across the entire political scale (left to right)."(doi:10.1145/3555636) cites Gun politics in the United States. That Wikipedia article barely mentions the left-right spectrum, and the only direct mention I see there is a paragraph on pro-gun leftists. Since the paper is using gun control as a proxy for left-wing bias, this seems like a very poor cite for this specific point, and in my opinion, it weakens the reliability of the source. Grayfell (talk) 20:45, 11 December 2023 (UTC)

"I'm still not really convinced this belongs at all, tbh" I guess I am a little confused. When you say "this", do you mean the article should not say, in effect, "AFMs ratings are used in academic research"? Nowa (talk) 13:16, 12 December 2023 (UTC)

More or less that's what I'm saying. Being used in academic research isn't, by itself, inherently noteworthy. That particular study uses an article in The Daily Caller, also, but it wouldn't make sense to say "the The Daily Caller is used in academic research". Even though this is technically true, it would need a lot more context to be informative to readers, otherwise it's just too vague. For this article, having looked at some of these sources, I don't see an obvious way to summarize this informatively. That doesn't mean there isn't a way, of course, but sources which are substantially about Ad Fontes would be a much better starting point, IMO. Grayfell (talk) 19:55, 12 December 2023 (UTC)

sources which are substantially about Ad Fontes would be a much better starting Excellent point. I think for now, we can let this one lie until if/when there are more sources about the use of AFM ratings in academic research. Nowa (talk) 22:22, 12 December 2023 (UTC)

References

Update to article History section using Wall Street Journal article

The History section of the article can be updated using the following article from the Wall Street Journal

I would normally just do that, but my first attempt was reverted, citing "rv - grossly UNDUE - SOAP, NOTNEWS".

I would like to request that a less involved editor take a shot at it and I will support their efforts. Nowa (talk) 17:25, 16 December 2023 (UTC)

References