About those face masks: Trust, care, & rights

It all started when my daughter and I found ourselves exchanging ideas about possible predictors of face-mask-wearing compliance. A few days later, I created a little survey designed to address the question, “What are some of the attitudes and beliefs that underlie face-mask-wearing decisions?”

In this article, I describe the sample, variables, analyses, and outcomes of the face-mask survey. If you aren’t particularly interested in the details you can get a good sense of what we learned by reading the next paragraph and the discussion at the end.

People are more likely to decide to wear face masks if the considerations they find most important are (1) expert knowledge and (2) the need to protect themselves and others. They are less likely to wear face masks if the consideration they find most important is the need to protect their personal rights.

You can learn more about the Face-mask Survey itself by checking out the article, The face-mask question: A survey. The survey questions can be viewed on the survey page.

Sampling

As of October 1, 2020, 499 surveys had been completed. People were invited to the survey via a Medium article, through the personal and professional networks of colleagues, family, and friends, and through several days of paid advertising on Facebook.

Samples collected in this way are called convenience samples. When working with convenience samples, we need to be especially careful about making generalizations from results. Consequently, it’s best to think of the results reported here as representing the attitudes and beliefs of individuals who responded to the survey. It is impossible to say how well the results reported here might represent the population as a whole.

Who completed the survey?

What were the variables?

You can see all of the individual variables by taking a look at the survey itself. You won’t see every single variable represented in this article. Variables that are not mentioned here have been eliminated from the analysis because they did not make it through the data reduction process.

Dependent variable

I used the results from 7 items related to face-mask-wearing practices, to create a variable called vigilance.

These items were responses to the prompt: “I wear a face mask when…” The numbers to the right of the items below represent level of vigilance. A score of 4 means high face mask wearing vigilance, whereas a score of 1 represents an absence of face-mask wearing vigilance. The final score awarded to each participant was based on the level of vigilance represented in the most “vigilant” item they marked with “True.”

I wear a face mask when…

• I leave my house (4)
• I’m around any other people (4)
• I get within 6 feet of any other people (3)
• I get within 6 feet of non-family members (3)
• I’m in a place where face-masks are recommended (3)
• face-masks are required by law (2)
• never (1)

Independent variables

The independent variables in this analysis were based on a factor analysis of the responses to the 18 items that followed the prompt, “When making decisions about wearing face-coverings during a pandemic, how important is it to consider…?” Each of these items was rated on a 10-point scale from unimportant to extremely important.

An exploratory factor analysis (using minimum residual extraction with a varimax rotation) was conducted on these items. The analysis yielded two factors composed of 13 of the original items.

Variable 1 (factor 1): trust & care

The first factor, trust & care, was composed of the following 8 items:

• the effectiveness of face masks in preventing COVID-19
• slowing down the infection rate
• the opinions of public health experts
• protecting others
• protecting high-risk individuals
• protecting essential workers
• the need to protect yourself
• being a good citizen

In the list above, items are ordered thematically, from a group related to trust in expert knowledge to a group related to care for self and others. A single variable (trust & care) was created from these items by averaging their values. The Chronbach’s alpha for the scale created from these items is .936, which is excellent.

Variable 2 (factor 2): self & friends

The second variable is composed of the following 5 items:

• how wearing a mask makes you feel about yourself
• personal preferences
• what your friends are doing
• what most other people are doing
• what others may think of your choice

Chronbach’s alpha for this scale was .743, which is acceptable for the analysis of group-level outcomes. Scores on these items were averaged to create the self & friends variable.

Variable 3: personal rights

One of the items in the factor analysis — your rights—was the only item that loaded negatively on the trust & care factor. I decided to treat this item as a third independent variable:

• your rights

Results

Which of the independent variables predicts vigilance in face-mask wearing?

Variable 1: trust & care

The correlation between vigilance and the trust & care variable is .717 (R-square = .514). In other words, the combination of trusting experts and caring for others predicts more than 50% of the variance in face-mask-wearing vigilance.

Variable 2: self & friends

The correlation between vigilance and the self & friends variable is tiny (-.050). This means that the combination of personal preferences and the perceptions of friends does not appear to relate to face-mask wearing vigilance.

Variable 3: personal rights

The correlation between face-mask wearing vigilance and the personal rights variable is -.405 (R-square = .164). This means that 16 percent of the variance in a lack of face-mask vigilance is predicted by concern about one’s personal rights.

Demographic variables & reported mask-wearing vigilance

None of the demographic variables collected, including age, ethnicity, gender, location, and political leanings, appeared to influence face-mask wearing vigilance. However, because this is a convenience sample and many demographic groups were poorly represented, we can draw no conclusions from this finding.

Being a member of a high-risk for COVID-19 group did correlate with face mask wearing vigilance, although the correlation was small (R=.232, R-square = .054).

Male vs. female participation

Fewer than one-third of the participants in this survey were male. Almost two-thirds were female. This ratio remained stable throughout the data collection period, even when we attempted to target males through Facebook advertising. In general, females are more likely than males to complete both online and written surveys. This is a common problem in survey research, and it raises questions about the generalizability of survey results.

Discussion

I’ve sliced and diced these data in many ways, and all of the analyses lead back to the same conclusions.

1. Trust in science, along with care for others (including society), is strongly related to self-reported mask-wearing vigilance;
2. A desire to defend personal rights is associated with low mask-wearing vigilance; and
3. The preferences of self and friends do not appear to be related to mask-wearing vigilance in a systematic way.

One way to frame these findings: Selfish people who think they know more than the experts are less inclined to protect others by wearing masks.

Another way to frame these findings: Individuals who feel the need to defend their personal rights are less likely to be motivated by trust and care than individuals who do not feel the need to defend their personal rights.

The second framing suggests that people who feel the need to defend their rights do so for a reason. Perhaps if we focused on understanding and addressing those reasons it would be possible to build trust and increase care.