3.3.4. Demographic Questionnaire

There were 8 questions. These items asked for participants' age, gender, residence, state of residence, race and ethnicity, education, income, and employment status.

#### 3.3.5. Attention Check Questions

Three attention check questions were included in the survey. They were added to help identify inattentive participants and to provide progress status, as they appeared after each quarter (1/4, 1/2, and 3/4) in the survey.

#### *3.4. Analytic Strategy*

A few additional variables were calculated to answer the proposed research questions: (1) Types of ICT devices used on a daily basis: participants could check all that apply from 11 options (e.g., Computer for non-internet use, Computer for internet use, Cable TV, etc.) and write in additional items. These types were then added to reflect the overall total types of devices used. These steps were used for the before and during pandemic periods. (2) Types of ICT as news sources about what is happening on a daily basis: participants could check all that apply from 8 options (e.g., Social media, TV news channels, Radio, etc.) and write in additional items. These types were then added to reflect the overall total types of sources. These steps were used for the before and during pandemic periods. (3) Average hours spent on using applications and systems for the purpose of staying connected with social networks on a daily basis: participants used a sliding bar to indicate the hours for social media, telecommunication, and email and had the options to write in two additional items and then indicate the hours. These hours were then averaged across the items to reflect the average hours spent daily for virtually staying connected with social networks.

To answer our first research question, a logistic regression was used to model the relationship between the perceived importance of social connectedness (low vs. high) and the psychosocial, ICT use, and demographic variables. To answer our second research question, a logistic regression was used to model the relationship between participants' feeling about the future (positive vs. negative) and the psychosocial, ICT use, and demographic variables.

A non-parametric Mann–Whitney U test was used to compare the distributions of responses due to the non-normality of our data distribution. Correlations of the variables were checked and there was no evidence of multicollinearity (all the Spearman correlation coefficients were smaller than 0.6). SPSS version 26 was used for the analyses.

#### **4. Results**

#### *4.1. Sample Characteristics*

The sample consisted of 219 men and 175 women in the U.S., with ages ranging from 20 to 76 and the average being 40.89 (SD = 11.21) years. The participants came from all of the states, except Alaska, Arkansas, North Dakota, South Dakota, and Vermont. In terms of

primary residence, 190 indicated suburban areas, 125 indicated urban areas, 77 indicated rural areas, and 2 chose other. The majority of the participants identified their race and ethnicity as White (*n* = 307) (61 as Asian, 20 as Black, 12 as Hispanic/Latino/Spanish origin, 11 as American Indian/Alaska Native, 1 as Native Hawaiian/Other Pacific Islander, and 2 as Other). As for education level, most of participants reported having a college degree (*n =* 219), followed by having some college (*n =* 70), having a graduate degree (*n =* 56), having a high school diploma (*n =* 47), and having some high school education (*n =* 2). The annual household income item included five options: most of participants selected the 45 K–70 K (*n =* 115) and 25 K–45 K (*n =* 105) options, followed by the 70 K–110 K option (*n =* 69), <25 K option (*n =* 62), and >110K option (*n =* 43). Most of them currently had a full-time job (*n =* 258) (22 worked part-time, 82 were self-employed, 3 were a student, 27 were unemployed). The statistics of the psychosocial scales are listed in Table 1.

**Table 1.** Statistics of the psychosocial scales.


Almost all of the participants (*n =* 391) indicated that they actively took actions about the current pandemic situation (e.g., social distancing, working from home, etc.). The majority of them thought that the news correctly conveyed the current pandemic situation (*n =* 308) and that the current pandemic situation was a threat to their health and safety (*n =* 315). Slightly less than half of the participants rated their feeling about the pandemic upon reading the news as "Positive—it's going to be ok" (*n =* 181), while the rest felt "Negative—it's not going to be ok."

#### *4.2. Importance of Social Connectedness*

In answering our first research question, one survey item asked about the importance of staying connected with friends, family members, and social networks. Participants rated from not at all important to extremely important. Given the uneven distribution of the rated responses (see Figure 1), this variable was dichotomized to reflect two levels of importance: low (combined from "not at all important," "slightly important," and "moderately important") and high (combined from "very important" and "extremely important") importance, having *n =* 192 and 202, respectively, in each level. Using this dichotomized importance as the grouping variable, Mann–Whitney U test suggested that types of ICT devices used and types of ICT as news source were higher in the highimportance group for both before and during pandemic periods. Hours spent on ICTs for obtaining news before the pandemic were not different between the low-and highimportance groups; however, hours were higher in the high-importance group during the pandemic (see Table 2).

**Figure 1.** Response distributions for the importance of the social connectedness variable.

**Table 2.** Comparison of information and communications technology (ICT) use variables between low- and high-importance groups.


Subsequently, a logistic regression was used to model the relationship between the perceived importance of social connectedness (low vs. high) and the psychosocial, ICT use, and demographic variables. These variables were entered in three blocks, with block 1 consisting of psychosocial variables, block 2 consisting of ICT use variables, and block 3 consisting of demographic variables. Insignificant variables were removed with each iteration. The final model had a Nagelkerke R of 0.36 and a Hosmer and Lemeshow Test of *χ*<sup>2</sup> (8, *N =* 394) = 11.55, *p* = 0.17, indicating a good fit to the data. The classification accuracy was 68.80% for predicting low-importance and 76.70% for predicting high-importance, with the overall accuracy being 72.80%. Table 3 shows the logistic regression coefficient, Wald test, and odds ratio for each of the predictors for the importance of the social connectedness variable. Employing a 0.05 criterion of statistical significance, ratings of extraversion, conscientiousness, need to belong, and perceived attachment to phone, types of ICT as news source, and gender had significant partial effects. For each one-point increase on the five-point extraversion and conscientiousness scales there were odds of higher rating of importance of social connectedness by a multiplicative factor of 1.34, and 1.47, respectively. Similarly, for each one-point increase on the five-point need to belong and perceived attachment to phones scales there were odds of higher rating by a multiplicative factor of 1.11 and 1.33, respectively. For each one additional type of ICT that was used for obtaining news there were odds of higher rating by a factor of 1.31. As for the gender variable (women coded as one), women were 2.17 times more likely than men to report higher rating on importance of social connectedness.

**Table 3.** Significant predictors for the perceived importance of social connectedness variable (*n* = 394).

