on September 20, 2020 suggested that the news media was normalizing death not treating the two milestones (100,000 vs. 200,000) of COVID-19 deaths the same.
closer, we can see that it is a side by side of two front pages of the New York Times. The image on the left shows a full page of names of people who have died due to COVID-19 near the 100,000 milestone. This front page is from May 24, 2020.
The image on the right shows a New York Times front page near the 200,000 milestone. That front page is from September 20, 2020. The milestone story is not featured on this page.
However, the dates of these two newspapers are not the actual milestone dates.
If we evaluate the front page of each web site for other news sources, how did they prioritize these milestones? Are they displayed prominently? How prominently? Are they ranked first? Second? How far into the page does one need to scan before they see a news article mentioning these milestones? Is this different for each source? How does the prominence of the milestone story for each news source compare between the 100,000 deaths milestone and the 200,000 deaths milestone? I explored mementos near these milestone dates to determine the answers to these questions.
When Did We Reach Each of the Two Milestones?
To determine the milestone dates, I used two sources -
Center for Disease Control and Prevention (CDC) and
John Hopkins University (JHU) - Coronavirus Resource Center. Both organizations had been tracking the spread of COVID-19 in the United States.
Center for Disease Control (CDC)
The CDC has a COVID Data Tracker tool which tracks total COVID-19 deaths. A visit to the page at any moment gives you the latest number for total deaths. I needed to find the first date that this tracker listed 100,000 or the closest to that number.
However, I did not find any mementos for the data tracker before August 25, 2020. The Carbon Date tool indicates that the site was born on March 28, 2020.
So I was not going to be able to use the CDC Data Tracker to derive the first milestone. However, the CDC posted a press release on May 28, 2020 indicating that the U.S. had reached 100,000 deaths.
I was able to use the COVID Data Tracker for the second milestone date.
Johns Hopkins University (JHU) - Coronavirus Resource Center
Johns Hopkins University has a COVID-19 Dashboard tool which tracks total COVID-19 deaths. A visit to this page at any moment gives you the latest numbers for total deaths. This site provides global information as well, but you can see total deaths for the U.S. and breakdowns by state.
Dates to Focus Search
Using the CDC press release and the captures from JHU and CDC, I was able to narrow the dates with which to focus my search for mementos.
100,000 deaths - May 28 (JHU & CDC)
200,000 deaths - Sep 22 (JHU), Sep 23 (CDC)
Deciding Upon News Sources to Include in the Study
365 Dots in 2019: Quantifying Attention of News Sources contains a list of news sources and their political leanings. From there, I chose two news sources corresponding to each political leaning.
| Left |
Center |
Right |
| Politics USA |
CNN |
Washington Examiner |
| Vox |
Politico |
Fox News |
| Huffington Post |
ABC News |
Daily Caller |
| MSNBC |
The Hill |
Conservative Tribune |
| New York Times |
Real Clear Politics |
Breitbart |
| Washington Post |
|
The Gateway Pundit |
Considerations for Ranking the News Stories on a Page
I had to create some guidelines for determining the prominence of a story. I also needed to figure out what should be excluded from consideration. I consulted Measuring News Similarity Across Ten U.S. News Sites for guidance on how to determine prominence.
- Things Considered
- Enormous or bigger font
- Inclusion of photo & size of photo
- Position/layout (top of page, centered and large, left to right
- Grouped page content into blocks - "article baskets"
- Things Not Considered
- Live streams
- Video stories/video clips
- Opinion pieces
- Advertisements
In the screenshots if the story discussed reaching the milestone, it is outlined in red.
Determining the Importance of a Story on a Page
I treated each front page like a search engine results page (SERP). And the query was "does the article talk about reaching the COVID-19 milestone?"
With this SERP model, I applied the following metrics to measure a given story's prominence on each news source's front page.
- Mean Reciprocal Rank (MRR) to measure how "fast" I got to the expected result (i.e. was the top item the expected result to my query)
- Precision to measure how much coverage my query got (i.e. how many of the results were about my query
- Normalized Discounted Cumulative Gain (nDCG) to measure how far away the results page was from the ideal scenario (the COVID-19 milestones as the top stories
The Process of Finding the First Memento
I started with the Wayback Machine page for a news source. Using the established milestone dates above, I randomly chose a memento. I kept randomly choosing a memento until I discovered the first occurrence of a story about the milestone. So basically, I picked two mementos wildly outside of the range and then moved towards the middle of those.
Milestone #1 - 100,000 Deaths
Breaking the News - Milestone #1
The following table lists the occurrence of the first break of the story on each of the news outlets in order. The outlets released the news within a five hour time period.
| News Outlet |
Political Leaning |
First Matching Memento |
| New York Times |
Left |
May 27, 2020 @ 20:01 UTC |
| ABC News |
Center |
May 27, 2020 @ 22:10 UTC |
| Fox News |
Right |
May 27, 2020 @ 22:13 UTC |
| Washington Post |
Left |
May 27, 2020 @ 23:03 UTC |
| Washington Examiner |
Right |
May 28, 2020 @ 01:14 UTC |
| CNN |
Center |
May 28, 2020 @ 01:21 UTC |
The New York Times - May 27, 2020 @ 20:01 UTC
Box #1 is the one story on the page that is about the milestone. The object in the top center is about COVID-19, but functions as an information dashboard with options for seeing data for world cases, U.S. cases, etc.
| Leaning |
Left |
| MRR |
1 |
| Precision |
1/6 |
| nDCG |
1.00 |
Washington Post - May 27, 2020 @ 23:03 UTC
Boxes #1 and #2 are the two stories on the page that are about the milestone. Box #1 is a headline that is also a story. It is also part of the story image below it.
| Leaning |
Left |
| MRR |
1 |
| Precision |
2/4 |
| nDCG |
1.00 |
CNN – May 28, 2020 @ 01:21 UTC
Boxes #2 and #5 are stories on the page that are about the milestone. Box #5 is an article basket that contains a live update about the milestone.
| Leaning |
Center |
| MRR |
1/2 |
| Precision |
2/6 |
| nDCG |
0.76 |
ABC News – May 27, 2020 @ 22:10 UTC
Boxes #1 and #5 are stories on the page that are about the milestone. Box #5 is an article basket that contains a story about the milestone.
| Leaning |
Center |
| MRR |
1 |
| Precision |
2/5 |
| nDCG |
0.77 |
Washington Examiner – May 28, 2020 @ 01:14 UTC
Box #1 is a story on the page that is about the milestone.
| Leaning |
Right |
| MRR |
1 |
| Precision |
1/2 |
| nDCG |
1.00 |
Fox News – May 27, 2020 @ 22:13 UTC
Box #1 is a story on the page that is about the milestone.
| Leaning |
Right |
| MRR |
1 |
| Precision |
1/7 |
| nDCG |
1.00 |
Milestone #2 - 200,000 Deaths
Breaking the News - Milestone #2
The following table lists the occurrence of the first break of the story on each of the news outlets in order. The outlets released the news within a four hour time period.
| News Outlet |
Political Leaning |
First Matching Memento |
| CNN |
Center |
Sep 22, 2020 @ 15:50 UTC |
| Fox News |
Right |
Sep 22, 2020 @ 16:02 UTC |
| ABC News |
Center |
Sep 22, 2020 @ 16:07 UTC |
| Washington Examiner |
Right |
Sep 22, 2020 @ 16:09 UTC |
| New York Times |
Left |
Sep 22, 2020 @ 16:52 UTC |
| Washington Post |
Left |
Sep 22, 2020 @ 18:55 UTC |
New York Times - Sep 22, 2020 @ 16:52 UTC
Box #1 is a story on the page that is about the milestone.
| Leaning |
Left |
| MRR |
1 |
| Precision |
1/3 |
| nDCG |
1.00 |
Washington Post – Sep 22, 2020 @ 18:55 UTC
This page had a lot of damage. If you wait for the whole thing to finish nearly all the content disappears. I had to take a screenshot of it just before that happened. Unfortunately, this is the first memento that had the milestone story and it was listed under "Live Updates". It was not very prominent on the page. It is box #4.
| Leaning |
Left |
| MRR |
1/4 |
| Precision |
1/10 |
| nDCG |
0.39 |
CNN – Sep 22, 2020 @ 15:50 UTC
The milestone story was the top story and there were a total of three stories about reaching the milestone.
| Leaning |
Center |
| MRR |
1 |
| Precision |
3/6 |
| nDCG |
0.85 |
ABC News – Sep 22, 2020 @ 16:07 UTC
Box #1, Box #2, and Box #4 are about reaching the milestone.
| Leaning |
Center |
| MRR |
1 |
| Precision |
3/5 |
| nDCG |
0.86 |
Washington Examiner – Sep 22, 2020 @ 16:09 UTC
Box #3 is the milestone story. It may look like Box #6 should be labeled as Box #1, but it seems more like a top ticker. Box #1 has the bigger font.
| Leaning |
Right |
| MRR |
1/3 |
| Precision |
1/6 |
| nDCG |
0.46 |
Fox News – Sep 22, 2020 @ 16:02 UTC
Box #6 is the first and only mention of reaching the milestone.
| Leaning |
Right |
| MRR |
1/6 |
| Precision |
1/7 |
| nDCG |
0.44 |
Comparison of Metrics Across Milestones
Mean Reciprocal Rank (MRR)
| News Outlet |
Leaning |
Milestone #1 |
Milestone #2 |
Judgment |
| New York Times |
Left |
1 |
1 |
Same |
| Washington Post |
Left |
1 |
1/4 |
Worse |
| CNN |
Center |
1/2 |
1 |
Better |
| ABC News |
Center |
1 |
1 |
Same |
| Washington Examiner |
Right |
1 |
1/3 |
Worse |
| Fox News |
Right |
1 |
1/6 |
Worse |
Precision
| News Outlet |
Leaning |
Milestone #1 |
Milestone #2 |
Judgment |
| New York Times |
Left |
1/6 |
1/3 |
Better |
| Washington Post |
Left |
2/4 |
1/10 |
Worse |
| CNN |
Center |
2/6 |
3/6 |
Better |
| ABC News |
Center |
2/5 |
3/5 |
Better |
| Washington Examiner |
Right |
1/2 |
1/6 |
Worse |
| Fox News |
Right |
1/7 |
1/7 |
Same |
Normalized Discounted Cumulative Gain (nDCG)
| News Outlet |
Leaning |
Milestone #1 |
Milestone #2 |
Judgment |
| New York Times |
Left |
1.00 |
1.00 |
Same |
| Washington Post |
Left |
1.00 |
0.39 |
Worse |
| CNN |
Center |
0.76 |
0.85 |
Better |
| ABC News |
Center |
0.77 |
0.86 |
Better |
| Washington Examiner |
Right |
1.00 |
0.46 |
Worse |
| Fox News |
Right |
1.00 |
0.44 |
Worse |
Conclusions
None of the news outlets were excessively later than others with breaking the news. For milestone #1, they were within a five hour period. For milestone #2, they were within a four hour period.
As a top story, five of the six outlets had the story as their top story for milestone #1. Three of the six outlets had the story as their top story for Milestone #2. So the significance of the story had decreased between the two milestones.
As far as coverage (i.e. how many main page stories were about reaching the milestone), the results were mixed.
For milestone #1, three of the six had one story about it; the other three had more than one story about it. For milestone #2, four of the six had one story about it; the other two had more than one story about it. Per the point of the tweets at the beginning of the post, in an ideal world scenario, all of the stories would be about reaching the milestone. Four of the six got a perfect score for milestone #1. Only one of the six got a perfect score for milestone #2. This shows a decrease in prominence.
- Valentina Neblitt-Jones
The source of this post is from my web archiving forensics course presentation (Fall 2020). Materials for the project are located here.
Comments
Post a Comment