4.1. El País
For the El País
dataset, a total number of 22,223 tweets are collected from 25 February 2020 to 3 June 2020. After removing retweets, 14,800 original tweets are saved for our in-depth analysis. Eight news frames have been successfully computed, they are “Livelihood” (family life and children), “Public Health Professional” (news about the department of public health), “Pandemic Update” (contagion and death poll), “Madrid” (news about Madrid), “Politics” (general political news), “State of Alarm” (Spanish government and PM’s announcement and policy update), “Economy” (the effect of the pandemic on Spanish economy) and “Covid Information” (general information about the pandemic). Table 1
presents the details of the eight news frames of El País
with their top seven relevant words.
presents the news frame network of the three segmented periods. Each of the nodes represents a news frame and the size of nodes indicates the strength of the node, also known as weighted node degree, it is the sum of the edge weights of the adjacent edges for each node [47
], reflecting the importance of a node in a weighted network. The edges between the nodes represent the connection strength between two frames (normalized by tf-idf), it is the sum of the tf-idf value of the co-words. Table 2
presents the detailed information about the news frames in each of the three periods, with the node strength, number of tweets in each news frame and their proportion of the total number in each segment. Table 3
presents the table of the most weighted edges in the three time segments; it is able to provide us the news frames with the highest similarity ties. Overall, “Livelihood,” “Public Health Professional,” “Pandemic Update” and “Politics” are the most important news frames of El País
. As the crisis is gradually under control, the “Pandemic Update” turned to be less prominent in the recovery period.
“Livelihood” is the most prominent news frame of El País
and it shows a strong connection with “Politics,” “Economy” and “Public Health Professional” in the pre-crisis stage, suggesting a close connection with government policy and economic situation. In the next two periods, it started to have a more significant relation with “Madrid.” This is understandable because the Spanish capital suffered the most during the COVID-19 pandemic. According to the actual policy, the Community of Madrid is one of the last regions that stepped into the recovery plan [48
] and this can also explain why the proportion of “Madrid” increased across the three time segments.
In addition, we indeed observed a news framing change in different stages of the pandemic outbreak. For example, the “Politics” frame is less reported in the second period while the “State of Alarm” and “Covid Information” frames have been paid higher attention during this stage. It is worth noting that although both frames have connections with others, no connections are observed between these two during the three periods, suggesting they are independent from each other. “State of Alarm” is a policy oriented news frame while “Covid Information” focused more on general sanitary information.
As the crisis is gradually controlled, the pandemic related news frames (“Pandemic Update,” “State of Alarm,” “Public Health Professional” and “Covid Information”) are becoming less prominent in the recovery period. The media interests in general political news (“Politics”) decreased during the most difficult time but soon recovered with the crisis situation becoming stable. Regarding the network, the “Politics” frame has the strongest connection with “Livelihood” during all of the three periods. It also has significant relation with “Public Health Professional” (weight: 236.60) and “Economy” (weight: 154.96) during the pre-crisis period but the two connections have been developing in different trends. While “Politics” and “Public Health Professional” remained connected in the other two periods, the connection between “Politics” and “Economy” turned to be less significant. Instead, the “Politics” frame becomes more connected with “State of Alarm” and “Madrid.”
4.2. El Mundo
For the El Mundo
dataset, a total number of 17,577 tweets are collected from 19 February 2020 to 3 June 2020. After removing retweets, 14,290 original tweets are saved for our in-depth analysis. Eight news frames are computed, six of which are considered the same as El País
. They are “Madrid,” “State of Alarm,” “Covid information,” “Economy,” “Pandemic Update,” “Politics.” The two unique El Mundo
frames are “Lockdown” (news about the confinement) and “Hospital” (news related to hospital, doctor and patient). Table 4
presents the news frames with their most relevant keywords.
presents the network of the three segmented periods, Table 5
provides the detailed information of the news frames across time and Table 6
presents the detailed information of the most weighted edges. Generally speaking, “Madrid,” “State of Alarm” and “Lockdown” are the three most prominent news frames during the pre-crisis period, along with the crisis becoming more severe, “Covid Information” is paid more attention by the newspaper. And finally these four frames are the most prominent news frames during the recovery period.
“Madrid” is the most prominent news frame of El Mundo of all the time. The proportion of this topic is greatly changed from the second period to the third. As we have explained in the previous section, Madrid is the last region that stepped into recovery plan, so this change is understandable. The “Madrid” frame has the strongest connection strength with “Lockdown” and “State of Alarm” during the first two periods and the association between “Madrid” and “Covid Information” becomes more and more eye-catching during the last two periods. The second most important news frame is “State of Alarm,” it has been paid less attention during the lockdown period but still, shared a significant proportion of the total news coverage. The “State of Alarm” frame has the highest connection strength with “Madrid” and “Lockdown,” similar to “Madrid,” the relation between “State of Alarm” and “Covid Information” is becoming stronger during the second and third time segments (weight in the 2nd period: 259.72, in the 3rd period: 139.42).
As Spain started to get recovered from the strict national lockdown, the proportion of the relevant news frames “Lockdown” and “Hospital” decrease during the recovery period but the connection between these two topics have been strengthened in this stage. As the “Lockdown” frame is highly associated with “Madrid” and “State of Alarm,” we assume this frame is strongly policy orientated. On the other hand, the “Hospital” frame includes both health and social news, so it is naturally associated the most with “Madrid” and “Lockdown.” Regarding the “Economy” frame, the proportion of this topic arrived its peak at the second period. It is significantly different from the frame “Politics,” which has been less adopted during the same period. Both of them have strong ties with “Madrid” and “State of Alarm” but no significant connections have been exposed between these two frames.
Given that the frames “Covid Information” and “Pandemic Update” have almost no proportion changes during the three time periods, these two news frames are considered as stable news frames, tweets about “Pandemic Update” is slightly fewer than “Covid Information.” Regarding the network, like many other El Mundo news frames, both of the two have the strongest connection with “Madrid” and the tie between these two frames is getting more and more meaningful over time.
4.3. Comparative Discussion
Significant differences are observed between El País (EP) and El Mundo (EM) in the frames used in their Twitter news posts. First, the most prominent news frame of the two Spanish newspapers are different. While EP focused on “Livelihood,” EM tended to adopt the “Madrid” frame most frequently. Despite the fact that “Madrid” is also a frame in the EP dataset, it is considered as a peripherical news frame. Both of the two frames have the strongest connections with other topics in the networks, so these two frames can be seen as the motor themes of their newspapers on Twitter.
Second, both of the newspapers have two unique news frames. While the EP news coverage on Twitter focuses on “Livelihood” and “Public Health Professional,” we observed the “Lockdown” and “Hospital” frames in the EM Twitter posts. The “Livelihood” frame is somewhat similar to “Hospital,” because both of the two news frames contain social and living attributes. Nevertheless, their connection strength with the other common frames are different. While “Livelihood” associates the most with “Politics” and “Public Health Professional” in the EP networks, “Hospital” associates the most with “Madrid” and “Lockdown” in the EM networks. A possible interpretation of this difference is “Livelihood” is linked to government (including relevant government departments) policy but “Hospital” is more linked to the news about specific regions. Also, EP shows higher attention to the Ministry of Health and professional perspective by adopting the “Public Health Professional” frame while EM focuses more on the effect of confinement from social perspectives with the “Lockdown” frame.
Third, although there are six common news frames identified in the Twitter posts of both newspapers but the longitudinal changes in their proportion over time are different. For example, the “Economy” related news tweets are increasingly scarce over time in the EP dataset but for EM, such information is more posted during the second time period (the lockdown period). Another significant example can be seen from the “Politics” frame. The EP Twitter account posted more politics-related news during the recovery period than during the lockdown period. But for EM, the increasing trend during the same periods is not so salient as EP.
“State of Alarm” is the second most important news frame for EM on Twitter but this frame is not so prominent in EP Twitter posts. Although the most relevant keywords of this frame in the two datasets are almost the same but the connections are different in the networks. During the first two periods, “State of Alarm” is considered most associated with “Lockdown” and “Madrid” in the EM network, while it is mostly linked to “Livelihood” and “Public Health Professional” in the EP network. During the recovery period, the link between “State of Alarm” and “Politics” is strengthened in EP network, while the connection between “State of Alarm” and “Covid Information” is more eye-catching in the EM network. This finding implies that, with the pandemic crisis getting under control, Twitter posts about “State of Alarm” is more related to political news on EP but connected to health news more closely in the EM Twitter coverage.