First Amendment Knowledge and Competence in United States First Amendment Knowledge and Competence in United States Residents Residents

Lacking adequate knowledge about one’s rights could inhibit the likelihood of exercising one’s rights or lead one to unwittingly violate laws that place legitimate limits on these rights. Thus, the present research examines First Amendment knowledge as well as competence to apply that knowledge in relevant circumstances. Results revealed that one-quarter of participants failed a test of objective knowledge on First Amendment rights. Furthermore, participants’ belief in their ability varied depending on their level of knowledge, in line with the Dunning-Kruger Effect. Participants also failed to transfer their limited objective knowledge to “real-world” situations, exhibiting impaired First Amendment competence. These findings suggest United States residents’ levels of knowledge and competence related to First Amendment rights and protections could be improved to promote a safe, knowledgeable, and democratic society.


First Amendment Knowledge and Competence in United States Residents
Between January 2017 and June 2020, there were nearly 43,000 collective action events across the United States (Crowd Counting Consortium, n.d.). These events included several different kinds of actions, such as protests and petitions, and concerned a range of issues including women's rights, racial justice, immigration, housing, the COVID-19 pandemic, elections, and more (Chenoweth, Choi-Fitzpatrick, Pressman, et al., 2020). For instance, in 2017, over four million people participated in the first Women's March, which is likely the largest single-day demonstration in recorded United States history (Chenoweth & Pressman, 2017). The next year, more than onethird of Americans reported they felt the urge to protest (Gallup, 2018). More recently, public participation in protests surged in the summer of 2020 after Minneapolis police officer Derek Chauvin's excessive use of force resulted in George Floyd's death, leading hundreds of thousands of people to come together as part of the Black Lives Matter movement in over 2,000 cities and towns in every state across America, despite the ongoing COVID-19 pandemic (Burch et al., 2020).
Petitions to the government calling for justice for George Floyd and two other victims of police violence-Elijah McClain and Breonna Taylor-broke records with a combined 36 million signatures, becoming the petition website Change.org's most-signed petitions of all time (Change.org). In fact, in 2020, Change.org saw a 46% increase in petition creation, a 209% increase in petition signatures, and a 62% increase in declared victories on the platform. Participation in collective action events such as the protests and petitions just described is protected by the speech and assembly clauses of the First Amendment in the United States Constitution (U.S. Const. amend. I.). Although inalienable, these rights are not absolute. At times, they are superseded in favor of others' rights or government interests, allowing jurisdictions to prohibit certain activities deemed harmful to other people or government function. The Supreme Court of the United States has refined a balancing test to determine what is harmful and when this harm surpasses constitutional protections (United States v. Carolene Products Company, 1938). As such, over the years, many cases have been brought before the Court to determine where specific situations lay on the harm-protection spectrum.
Recent events have highlighted the importance of public understanding this spectrum in the context of the First Amendment. Those who are knowledgeable about their rights may be more likely to exercise them, as civic and political knowledge can result in increased public support for these freedoms ("The 2019 State," 2019) and increase the likelihood of participating in public matters via voting (e.g., Galston, 2007). Conversely, lacking adequate knowledge about one's rights could lead to unwittingly violating laws that place legitimate limits on these rights. Both of these ideals-being an informed, active member of society and avoiding entanglement with the justice system-are worthy goals toward which democratic citizens should strive. This raises three research questions: First, what do people think they know about their First Amendment rights? Second, what do people actually know about these rights? And third, can they apply their knowledge and exercise those rights without overstepping the bounds of the law? The present study seeks to answer these questions.

First Amendment Protections
The First Amendment of the United States Constitution explicitly prohibits Congress from creating and enacting legislation that infringes upon the rights to assembly and petitioning the government, as described previously, as well as speech, religion, and press. These rights are considered so fundamental to the American experience that many Founding Fathers refused to sign the Constitution until these protections were first added to the document via the First Amendment (National Constitution Center Staff, 2019). Since Chief Justice John Marshall declared interpreting and applying the Constitution the "very essence of judicial duty" in Marbury v. Madison (1803, p. 178), the Supreme Court of the United States (SCOTUS) has defined the extent of the protections provided by the Constitution by weighing fundamental individual rights against government interests and evolving standards of decency. That is, although the Constitution protects people from having their fundamental rights infringed upon by the government, there are circumstances in which an individual's expression of his or her rights impedes the rights of other people or the function of the government. It is the role of the judiciary, through case law, to use varying standards of review to debate these circumstances and determine when individual rights versus government interests are paramount (United States v. Carolene Products Company, 1938). Specifically, the Court states that legislation which facially appears to restrict political processes, such as those outlined in the Bill of Rights, are subject to more exacting judicial scrutiny under the Fourteenth Amendment. This strict scrutiny standard requires that the legislation be created in furtherance of a compelling government interest, and be as narrowly tailored as possible to achieve that interest.
As such, most laws that impede on one's First Amendment rights fail to reach this burden and are thus struck down by the courts. For example, when the government attempted to enjoin the New York Times and Washington Post from publishing classified documents concerning the Vietnam War-the Pentagon Papers-for fear of negative foreign policy outcomes, the Court found that these concerns were too speculative to pass muster and overcome the importance of freedom of the press (New York Times Company v. United States, 1971). Importantly, the First Amendment protects against detractors as much as it does supporters; those of the minority opinion have just as much right to dissent as those of the majority do to praise. As found in Wooley v. Maynard (1977), "[t]he First Amendment protects the right of individuals to hold a point of view different from the majority and to refuse to foster…an idea they find morally objectionable" (p. 715).
The debate regarding the extent of First Amendment protections goes beyond tangible speech and assembly and into the broader issue of how speech disseminates thoughts and ideas throughout society. One theoretical basis for free speech is known as the marketplace of ideas, first discussed by Justice Oliver Wendell Holmes, Jr. in his dissent for Abrams v. United States (1919). This ideahailed as "the most eloquent statement of the principle of free speech in American jurisprudence" (Novick, 1992(Novick, , p. 1229 and "one of the central organizing pronouncements for our contemporary vision of free speech" (Bollinger, 1986, p. 18)-centers on the philosophical premise that the best way to achieve truth is to allow for the maximum flow of thoughts and ideas in a free marketplace, and thus from this competition of ideas, truth will out. As stated by Justice Brandeis in Whitney v. California (1927), "If there be time to expose through discussion the falsehood and fallacies, to avert the evil by the processes of education, the remedy to be applied is more speech, not enforced silence" (p. 377) When this open marketplace of ideas is threatened by the government (i.e., through laws limiting freedom of speech), the result is a chilling effect. First mentioned by SCOTUS in Wieman v. Updegraff (1952), a chilling effect occurs when speech or conduct is suppressed due to fear of penalization, to the detriment of individual interests.

Limitations on First Amendment Protections
When examining questions of fundamental rights and government interests, these two standards-keeping an open free marketplace for thoughts and ideas and quelling the possibility of chilling speech-are often the basis for arguments for individuals and their constitutional rights; by allowing individuals their fundamental right to free speech, society succeeds in maintaining an open marketplace. However, in certain instances speech may be regulated for the continuing functioning of the government or for the protection and safety of society, if doing so is the least restrictive option.
Laws generally cannot regulate the political content of speech, but they can regulate nonpolitical speech in certain circumstances. The government has compelling interests that allow it to thwart certain kinds of speech to act successfully in its many prescribed roles, such as educator (Bethel School District. No. 403 v. Fraser, 1986), head of the military (Parker v. Levy, 1974), or head of prisons (Thornburgh v. Abbott, 1989). In other cases, preventing injury or crime takes precedence over promoting the free expression of thoughts and ideas. For instance, fighting words-words that a reasonable person would expect to incite violence-are not protected (Chaplinsky v. New Hampshire, 1942); nor are certain kinds of false speech such as defamation (e.g., New York Times Company v. Sullivan, 1964), speech that places one in imminent danger (Schenck v. United States, 1919), nor speech that is directed at inciting or producing imminent lawless action and is likely to incite or produce such action (Brandenburg v. Ohio, 1969). Although they limit speech, these restrictions are constitutional as they allow other citizens the superior rights of life and liberty.
As long as the prohibition of speech is not determined by the content of the speech-in other words the prohibitions remain neutral-they may be considered constitutional (R.A.V. v. St. Paul, 1992).
The assembly clause of the First Amendment is tied closely to speech in that it often concerns where speech occurs. The essence of freedom of assembly is, in part, that free speech is not limited to the privacy of one's home, but that this open marketplace is encouraged to flourish in public as well.
In Perry Education Association v. Perry Local Educators' Association (1983), the Supreme Court divided forums into three types: traditional public forums (e.g., public parks and sidewalks), designated forums (e.g., meeting rooms or community theaters) that can be designated as public forums for a limited period of time, and nonpublic forums (e.g., polling places and airports).
Like speech, the freedom of assembly can be overridden to achieve goals such as preventing nuisances (Martin v. City of Struthers, 1943), shielding the judicial system from the influence of protesters (Cox v. Louisiana, 1965b), and preventing crime (Schneider v. State of New Jersey, 1939).
The type of speech that falls outside the First Amendment protections can depend on the forum. In public and designated forums, time, place, and manner restrictions may stipulate when, where, and how speech occurs (Cox v. New Hampshire, 1941). For example, laws limiting protests during rush hour are constitutional (Cox v. Louisiana, 1965a), as are laws enforcing fixed buffer zones created around buildings within which protests or solicitations cannot occur (McCullen v. Coakley, 2014), and laws regulating volume levels of public events (Ward v. Rock Against Racism, 1989). In nonpublic forums, the government may restrict the content of speech as long as the restriction is reasonable and does not discriminate based on speakers' viewpoints (Minnesota Voters Alliance v. Mansky, 2018). Regardless of forum, laws limiting the freedom of assembly must be unbiased and not substantially broader than necessary to achieve a government interest (Ward v. Rock Against Racism, 1989).

Knowledge and Competence Related to First Amendment Rights
An understanding of one's First Amendment rights and limitations on those rights is imperative to allow individuals to effectively and legally exercise them. Unfortunately, research has shown that most Americans do not have a high level of political knowledge, that is, factual knowledge relating to the presidency, political institutions, and political processes (Carpini & Keeter, 1993;Neuman, 1986). In general, research shows that school-age samples are uninformed about political and government-related systems (e.g., the role of different political actors, the legislative process, etc.), with about 75% of high school students testing at or below a "basic" level of political knowledge (Galston, 2001). Civic education programs are meant to improve this low level of understanding, but there is mixed evidence as to their effectiveness (Dudley & Gitelson, 2002;Galston, 2001Galston, , 2007. Moreover, even a program's minimal success may be short-lived as civic education curricula has significant but short-term effects on students' knowledge about civil liberties (Green et al., 2011). Thus, even if students do learn from successful programs, the knowledge gained is unlikely to be retained. This is consistent with research showing that many adults across all ages have significant gaps in knowledge. For example, the Pew Research Center (2015,2017) found that nearly half of survey respondents were unable to correctly identify high-profile Massachusetts Senator Elizabeth Warren or the balance of Democratic versus Republican representatives in the Senate, and fewer than half of respondents knew that Neil Gorsuch is a SCOTUS Justice, despite being nominated and confirmed the same year in which the survey was conducted. A poll by Gallup (2015) showed that a full 50% of adult respondents correctly answered zero or only one of five fact-based questions about Congress, while only 6% answered all five questions correctly. Approximately one-third of American adults failed Bennett's (1988) general political knowledge test-he labeled these people "know-nothings" (see also Hyman & Sheatsley, 1947).
Less research has assessed political knowledge specific to the Constitution, but different polls of Americans suggest similarly low levels of knowledge. For example, although 73% of respondents claimed they knew some or a great deal of the Constitution's contents ("U.S. Constitution," 2003), over 90% did not know that the Bill of Rights protects against abuses by the federal government State," 2019). In addition, 65% agreed that private social media companies violate users' First Amendment rights when they ban users based on objectionable content they post. In light of these findings, the authors of the report concluded that many misconceptions surrounding the First Amendment remain prevalent ("The 2019 State," 2019). The present study builds on this limited body of work by examining First Amendment knowledge on speech and assembly specifically.
It is also important to understand what people subjectively believe they know, as most people overestimate their level of knowledge, with the exception of those who have the highest levels of knowledge about specific topics, who, in contrast, tend to underestimate their knowledge (Kruger & Dunning, 1999). This "Dunning-Kruger Effect" has been found on tests of knowledge generally (e.g., Soll & Klayman, 2004), and American history (The American Revolution Center, 2009) and political behavior specifically (Ortoleva & Snowberg, 2015). Anson (2018) found this effect in a study of knowledge related to basic political institutions, awareness of current political conditions, and ideological differentiation. Our research examines how this effect applies to First Amendment knowledge. This is an important extension of the literature as overestimating one's knowledge could potentially lead to adverse consequences, such as inadvertent violations of laws that limit rights. Yet, underestimating one's knowledge may cause those who are the most knowledgeable to hesitate to exercise their rights.

Distinguishing between Knowledge and Competence
In several of the aforementioned studies, the concepts of political knowledge and competence were used interchangeably or discussed without distinction (e.g., Beaumont, Colby, Ehrlich, & Torney-Purta, 2006;Kenski & Stroud, 2006). However, they are not equivalent (Shaker, 2012).
Knowledge indicates an ability to answer questions about political facts, whereas competence is the ability to apply such knowledge through comprehension and evaluation of public affairs and politics (Bennett, 1997;Lupia, 2006). Bennett (1997) argues that it is competence, not knowledge, which is essential to democracy.
The current study accounts for this important distinction by examining knowledge and competence related to First Amendment rights as two distinct constructs. This allows for the comparison of people's ability to recall facts against their ability to understand their protections in relevant contexts. This distinction aligns with educational psychological theory on knowledge transfer (for review, see Goldstone & Day, 2012). The application of new knowledge and skills to daily life (Simons, 1990 as cited in Simons, 1999) exemplifies the process of taking learned knowledge about the protections provided by the First Amendment and applying it to events such as protests. The current study assesses both First Amendment knowledge and competence to show American residents' ability to effectively and appropriately exercise their rights.

Current Study
Understanding constitutionally guaranteed freedoms and the exceptions to those freedoms is key to being an active, involved resident of the United States. And understanding the current state of First Amendment knowledge and competence can pave the way for molding a more knowledgeable and competent and, consequently, engaged citizenry. Therefore, this study explores the current level of First Amendment knowledge on speech and assembly in American residents, their subjective estimate of their knowledge on the subject, and their competence. Drawing from arguments by Bennett (1997) and Shaker (2012), as well as educational psychological theory on knowledge transfer, we operationalize First Amendment knowledge as "textbook knowledge," or knowledge of specific facts, whereas we define and examine First Amendment competence as the ability to transfer this knowledge by applying it to four "real-world" situations.
• Hypothesis 1: The proportion of First Amendment "know-nothings," examined as both knowledge and competence, will be low similar to rates of political knowledge found in previous research (i.e., Bennett, 1988).
• Hypothesis 2: Patterns of political knowledge will follow the Dunning-Kruger Effect (Kruger & Dunning, 1999) such that o Hypothesis 2a: those who are less knowledgeable are expected to overestimate their level of knowledge, whereas o Hypothesis 2b: those who are more knowledgeable will underestimate how knowledgeable they are.
• Hypothesis 3: A failure in knowledge transfer will be illustrated by a null association between First Amendment knowledge and competence.

Participants
Participants were 231 adult residents of the United States, including 150 Amazon Mechanical Turk Prime (mTurk) workers and 81 community members. As Table 1 shows, the two samples were similar in age and racial makeup, but the online sample included fewer women and was less educated and less politically active.

Materials and Measures
A survey presented all measures and materials described next, in the order presented here.
Demographics. Participants reported their age, gender, education, and race. Four objective knowledge items-modeled on those used in previous academic and media studies of political and civic knowledge (e.g., Carpini & Keeter, 1996;Galston, 2001; "U.S.
Constitution," 2003)-measured what participants actually know about their First Amendment rights. Two were single-response multiple choice questions (i.e., "Which of the following best describes the Bill of Rights?"; "What is the purpose of the First Amendment?") with four possible response options but only one correct answer. Two items were multiple-response items. The first, "Which of the rights listed are protected by the First Amendment?", had 13 possible choices (e.g., Right to freedom of speech; Right to bear arms, etc.), five of which were correct. The second item, "Which of the following types of speech are NOT protected by the First Amendment?", had 10 possible responses (e.g., Offensive speech, Hate speech, etc.), six of which were correct. An objective knowledge scale was computed as a sum of correctly identified responses to these four items (range = 0 to 13), with higher scores reflecting more objective knowledge ( = .71).

Vignettes. Participants read one of four vignettes describing different types of First
Amendment protections and restrictions based on precedent-setting SCOTUS cases. Two vignettes about Flag Burning (Texas v. Johnson, 1989) and Offensive Speech (Cohen v. California, 1971) describe protected speech, whereas vignettes about Draft Protest by burning a draft card (United States v. O'Brien, 1968) andFighting Words (Chaplinsky v. New Hampshire, 1942) describe nonprotected speech.
Competence Scale. First Amendment competence was assessed using multiple choice and True/False questions. Participants first reported whether the actions described are protected by the Constitution (Yes, No, Unsure). Next, participants responded True or False to four statements relating to whether the actions described in the vignette are protected by the Constitution (e.g., "The behaviors described inhibited a reasonable government interest"). The competence scale was computed as a sum of correct responses to the five items (range = 0-5; s = .64-.78 across vignettes).

Political Activism Measures. Political activism experience was measured with seven Yes or
No items asking about participants' previous involvement with politics (e.g., "Did you vote in the last presidential election?"). A political activism scale was computed as a sum of Yes responses to these six items (range = 0 to 6), with higher scores reflecting more objective political activism experience.
One item measured participants' subjective political activism ("Do you consider yourself to be politically active?") on a scale from 1 (Not at all) to 3 (Very).

Procedure
All participants were required to be at least 18 years old and United States residents to ensure they were living under the country's laws. mTurk participants were recruited with an advertisement placed online, which directed interested individuals to an online survey administered with Qualtrics survey software. After confirming eligibility, participants provided informed consent and completed the survey in the previously described order. Participants were then debriefed, thanked, and compensated with $3.50.
Community participants were recruited from various local places, including on campus, in public parks, and around shopping areas. They were invited to participate and eligible individuals were randomly assigned to receive a paper survey including one of the four vignettes. The paper surveys were identical to the online survey, with all measures and stimuli presented in the order described above. Upon completion, participants were debriefed and compensated with candy. All data were collected in the spring, summer, and fall of 2018.

Preliminary Analyses
Preliminary analyses were conducted to determine if there were any effects of sample or vignette type on the outcome variables. There were no statistically significant differences between

Estimates of Objective Knowledge and Competence
To compare objective knowledge and competence, scores were transformed into letter grades with failing grades assigned to those who scored on approximately the bottom 35% of the respective scale, modeling approaches used by Bennett (1988) and Hyman and Sheatsley (1947). Specifically, the fourteen-point objective knowledge scale was transformed so that scores of 0 to 4 were classified as F, scores of 5 to 6 as D, scores of 7 to 8 as C, scores of 9 to 10 as B, and scores of 11 to 13 as A.
The six-point competency scale was transformed so scores of 0 and 1 were classified as F, a score of 2 as D, a score of 3 as C, a score of 4 as B, and a score of 5 as A. (See Table 2 for distributions.) Using this classification scheme, 26% of participants were "know-nothings"-those who scored Fin regard to objective knowledge, and 24% in regard to competence, revealing that a substantial minority of United States residents perform poorly when tested on their knowledge of and ability to interpret their First Amendment rights.

[INSERT TABLE 2 APPROXIMATELY HERE]
Next, we used a chi-square goodness-of-fit test to examine political knowledge for our sample compared to Bennett's (1988) (Hypothesis 1). Results revealed statistically significant differences in the distribution of objective knowledge between Bennett's participants and ours, χ 2 (4, N = 231) = 23.24, p <.001,  = .32 (see Table 2). Further examination using a chi-square binomial test examining the percentage of participants who scored F as compared to all other grades, using a Bonferroni correction for multiple comparisons (i.e., achieving significance at p = .01), revealed no significant difference in the percentage of our participants who fell into the category of "knownothings" as compared to what Bennett found, χ 2 (1, N = 231) = 3.02, p =.08,  = .11. This indicates that when considering objective knowledge, similar numbers of people are uninformed today about First Amendment rights as they were on political institutions when Bennett tested them 30 years ago.
Our participants also did not differ significantly in regard to the rate of A and B grades compared to Bennett's, all χ 2 s(1, N = 231) ≤ 4.61, ps > .03, s ≤ .14. Where the distributions did differ, our participants were more likely to receive C or D grades relative to Bennett's, all χ 2 s(1, N = 231) ≥ 6.99, ps ≤ .01, s > .17.
In regard to competence, overall, a Chi-square goodness of fit test revealed a statistically significant difference in the distribution of scores in the current research compared to Bennett's (1988), χ 2 (4, N = 229) = 61.59, p < .001,  = .52 (see Table 2). Specifically, a binomial Chi-square test with Bonferroni correction indicated no statistically significant differences between the expected and observed frequencies of participants who received C, D, or F, all χ 2 s(1, N = 229) ≤ 4.44, ps ≥ .04, s ≤ .14. Thus, just as many participants were unable to understand their First Amendment rights in real-world scenarios in our study as were uninformed about political institutions in Bennett's research three decades ago. Even so, significantly more participants in this study received an A than anticipated based on Bennett's findings, all χ 2 s(1, N = 229) ≥ 8.65, ps ≤ .003, s > .19. By operationalizing competence independently from objective knowledge, these findings reveal that more participants are highly competent than would be expected based on Bennett's measure of objective political knowledge.
Because preliminary analyses revealed differences as a function of sample and vignette type, we tested Hypothesis 1 for competence separately for community versus online participants and across each separate vignette. These analyses revealed that the difference in overall grade distributions reported previously can be explained by the higher rate of competence among the online sample. Specifically, whereas the grade distribution of our community sample did not significantly differ from Bennett's (1988), χ 2 (1, N = 79) = 9.70, p = .05,  = .35, it did for the online sample, χ 2 (4, N = 150) = 73.45, p < .001,  = .70. Binomial tests indicated that significantly more of our online participants received an A in competence than expected based on Bennett's findings, χ 2 (1, N = 150) = 71.63, p < .001,  = .69, but their rates of receiving a B, C, D, or F did not differ significantly, χ 2 (1, N = 150) ≤ 4.28, ps ≥ .04, s ≤ .17.
Analyzing competence as a function of vignette type showed that grade distributions for all four vignettes differed significantly from Bennett's (1988) study, all χ 2 s(4, N = 54-60) ≥ 14.60, ps ≤ .01, s ≥ 49. However, results differed based on vignette type. The Flag Burning and Offensive Speech vignettes, on which participants scored higher in competence relative to the other two vignettes, yielded similar results, with participants in the current study significantly less likely to receive an F, χ 2 (1, N = 56-59) ≥ 12.99, ps ≤ .001, s ≥ .48, and significantly more likely to receive an A, χ 2 (1, N = 56-59) ≥ 30.53, ps ≤ .001, s ≥ .72, than expected from Bennett's results.
Conversely, analyzing competence for the Fighting Words vignette revealed that significantly more of our participants scored an F than expected based on Bennett's distribution, χ 2 (1, N = 54) = 16.01, p < .001  = .54, whereas the expected number of participants received an A, B, C, or D, all χ 2 s(1, N = 54) ≤ .001, ps ≥ .02, s ≤ .31. This indicates a high rate of incompetence for this type of First Amendment issue. Finally, with regard to the Draft Protest vignette, the rate of "knownothings" who received an F in our samples relative to Bennett's did not differ significantly, nor did rates of receiving an A, C, or D, all χ 2 s(1, N = 60) ≤ 5.08, ps ≥ .02, s ≤ .29. But, relative to Bennett's findings, fewer participants in this study received a B, χ 2 (1, N = 60) = 8.45, p = .004,  = .38.

Relation between Objective and Subjective Knowledge
To analyze the relation between objective and subjective knowledge, the 14-point objective knowledge and 6-point subjective knowledge scales were converted into z-scores. Standardized subjective knowledge scores were then subtracted from standardized objective knowledge scores to create a measure of relative estimation of knowledge, with positive scores indicating overestimation and negative scores indicating an underestimated self-appraisal of knowledge. A one-way ANOVA revealed a significant difference on this measure between individuals who scored different grades, F(4, 223) = 37.90, p < .001, η 2 = .40 (see Figure 1). A planned contrast revealed that those who scored above average (A or B) estimated their knowledge significantly differently than those who scored below average (D or F), t(223) = 11.76, p < .001. Consistent with Hypothesis 2a, those who were less knowledgeable (i.e., received a D or F grade) overestimated their knowledge, with scores significantly higher than 0, all ts(48-51) < 6.93, ps < .001. In contrast, those who were more knowledgeable (i.e., received an A or B grade) underestimated their knowledge, with scores significantly lower than 0, all ts(24-38) > -7.18, ps < .001, consistent with Hypothesis 2b. Only those who received a C grade had an average level of estimation that did not differ significantly from 0, t(62) = -1.30, p = .20. Furthermore, a linear contrast revealed a significant linear trend, F(4, 223) = 37.90, p < .001, indicating that increases in objective knowledge grades were associated with steady decreases in subjective knowledge estimations, in line with the Dunning-Kruger Effect.

Relation between Objective Knowledge and Competence
The final set of analyses tested Hypothesis 3, which posited that participants' First Amendment knowledge would be unrelated to their ability to apply that knowledge in relevant realworld scenarios. A linear regression predicted participants' competence with objective knowledge (both measured on full continuous scales for ease of interpretation) as well as sample type, vignette type, and the interactions between objective knowledge and these variables. Although the omnibus model was significant, F(228) = 4.19, p = .001, R 2 = .09, objective knowledge did not have a significant effect on competence, b = .27, t(228) = 1.78, p = .08. Thus, as predicted, participants did not effectively transfer their objective knowledge to situations in which First Amendment protections were threatened. Neither sample, vignette type, nor the interaction of these variables with objective knowledge significantly affected competence either, all bs ≤ |0.36|, ts(228) ≤ |0.82|, ps ≥ .42.
Adding demographic variables into a second model did not change this pattern of findings.

Discussion
The present research examined how much the nation's residents know about their rights, the limitations of these rights, and whether they can correctly identify what these limitations look like in real-world contexts. About one-quarter of our study participants failed tests of relevant objective knowledge and competence as assessed via fact-based questions and hypothetical situations in which First Amendment rights and protections are challenged, respectively. Our results indicate that a substantial minority of U.S. residents are unknowledgeable and incompetent with regard to these rights.
Prior empirical studies of political knowledge, on which we modeled our methods, found similar levels of "know-nothings" (Bennett, 1988;Hyman & Sheatsley, 1947). In fact, statistically similar rates of participants scored an A or F on our measure of objective First Amendment knowledge as did on Bennett's measure of political knowledge, indicating that the same small percentage of people knew a significant amount and the same large percentage were "knownothings." This is not surprising, given both the little change in knowledge found by Bennett (1988) relative to Hyman and Sheatsley (1947), as well as research showing relatively stable rates of such knowledge over time (Galston, 2001;Somin, 1998). Our findings also mirror those of polls and surveys conducted showing low levels of objective knowledge about one's constitutional rights (e.g., "Poll Finds only 33%," 1991; "The 2018 State," 2018; "The 2019 State," 2019). Even so, our findings are disheartening.
Conversely, significantly more of our participants scored an A on competence than for political knowledge in Bennett's (1988) study, highlighting the importance of examining this construct separately from knowledge. This difference was driven by the higher rates of competence in our online sample relative to the community member sample. Although our two samples differed in gender, education, political activism experience, and subjective political activism, a regression including these and other demographics (age and race), and sample type as predictors of competence revealed a significant effect of only sample type. Thus, unlike in previous research on political knowledge (e.g., Shaker, 2012), none of the demographic variables measured accounted for the differences in competence scores, all bs ≤ .26, ts(219) ≤ |1.77|, ps ≥ .08. The sample difference in competence, then, must be driven by some other unmeasured variable(s).
One possibility is that online participants, who had to have internet access to take the survey, used that access to search for correct answers. This seems unlikely, however, as there were no sample differences on objective knowledge scores, and it would arguably be easier to find answers for that clear-cut measure than the competence items, which required participants to interpret and apply their knowledge to a hypothetical situation. Further, neither objective knowledge nor competence scores were significantly correlated with time to completion of the online survey, all rs < .04, ps > .66, indicating that participants who answered correctly did not take more time than others because they were searching for the answers. Further research is needed to determine whether and why online samples might exhibit greater First Amendment competence than community samples, as well as the practical implications of this difference.
In addition to the sample differences in competence, there also were significant differences across the four vignettes. Of interest, participants performed best on the Flag Burning and Offensive Speech vignettes, which described protected speech. Conversely, the two vignettes in response to which participants performed the most poorly (Draft Protest and Fighting Words) described scenarios in which one's rights are limited. This suggests people may be familiar with the protections afforded by the First Amendment even while they are unaware of limitations on those rights, or at least how these limitations may manifest in the real world. Moreover, people may have a default assumption that speech is generally protected by the First Amendment. Future studies should explore these possibilities and whether there is a clear point at which people's understanding of their rights begins to falter by examining additional constitutional protections and limitations.
Just as concerning as the low rates of First Amendment knowledge and competence is the inverse relation we found between objective and subjective knowledge. Overestimating one's knowledge could lead to entanglement with the justice system, as believing one knows more than he or she does could potentially lead one to violate laws that place legitimate limits on those rights.
Conversely, underestimating one's own knowledge could potentially inhibit the most knowledgeable people from exercising their First Amendment rights. Democracy works only when residents make their wishes known to their representatives. If those who are the most knowledgeable underestimate their knowledge regarding their rights, they may not signal their wishes. Future research should seek to examine the consequences of both over-and underestimations of First Amendment knowledge.
Finally, participants did not transfer objective knowledge to application in hypothetical realworld scenarios, as there was no relation between objective knowledge and competence. This lack of effective knowledge transfer poses another obstacle to having a politically active citizenry, which leads to the complex question of how we might increase both First Amendment knowledge and competence. The present research takes a first step in resolving this problem by highlighting the gaps in knowledge and failure in application that currently exist, providing direction regarding where improvements are needed. Potential solutions, such as offering continuous schooling over several years, increasing the availability of easily comprehended materials, and using the best method of education, need to be created and empirically evaluated to determine the best path forward.

Limitations and Future Directions
This study is not without its limitations. First, although our objective knowledge scale (converting raw scores to an A to F scale) aligns with previously used methods (Bennett, 1988), operationalizing competence in a similar way may not be ideal. Competence requires a level of comprehension and understanding that should be measured in ways that capture participants' logic and thought processes. Thus, qualitative research asking participants to explain their reasoning may be more informative, as it would allow researchers to code for details and arguments that reflect an accurate understanding of the relevant issues and knowledge application. Such responses also would be less vulnerable to inflation due to participant guessing on multiple-choice questions.
Another potential limitation of this research is that comparisons were made between First Amendment knowledge and competence scores in the current study's participants to broader political knowledge among Bennett's (1988) participants. Given the lack of similar research with which to compare our findings, Bennett's work provided a starting point to estimate how competence might be distributed. Now, our findings on First Amendment competence provide a measure against which future research can compare, be that for testing educational interventions or examining competence in other areas. This study adds to the literature a baseline and precedent for testing constitutional and political competence, which, despite having been discussed in this area of research as a concept theoretically different from knowledge (e.g., Beaumont, Colby, Ehrlich, & Torney-Purta, 2006;Kenski & Stroud, 2006), has not yet been empirically studied.
The subjective knowledge measure also may have limited validity for understanding the Dunning-Kruger Effect. In Kruger and Dunning's (1999) original work, subjective knowledge was operationalized as either participants' estimation of their score or their achievement compared to their peers, which allowed for a direct comparison between participants' estimated and actual scores.
In the current study, instead of asking participants to estimate their objective knowledge score, which would have revealed how many of the multiple-response answers were correct, we asked participants to report a global estimation of their knowledge. Even so, we standardized subjective knowledge appraisals and objective knowledge scores and analyzed how these estimations differed across levels of objective knowledge. This allowed us to test for statistically meaningful differences between groups, instead of simply looking at patterns of raw scores and describing, nonstatistically, what the relations between estimated and actual knowledge looked like, as previous researchers had done (e.g., Anson, 2018).
Another limitation of this research involves the hypothesis of a null relation between objective knowledge and competence, as it is statistically impossible to prove a null hypothesis.
However, an a priori power analysis indicated that the present sample size would yield sufficient power of .80 to detect a small effect size (partial R 2 > .04). As is, this exploratory research indicates support for the null hypothesis, but it is possible that a very small effect was undetected, or that the measures were not sensitive enough to assess the variables under study. Future research is needed to replicate the null relation between objective knowledge and competence.
Finally, our samples were not random nor nationally representative. Participants were recruited in two different manners to combat some of the limitations that come from convenience and online marketplace sampling, but neither approach is ideal. The community sample was recruited from many different locations to increase diversity and representativeness, but it was still majority White and women and sampled from one small city in a northeastern state. The online sample was more geographically diverse, yet participants were not in our presence as they completed the survey so we cannot be certain that they did not cheat or that they paid close attention throughout the entire survey. Thus, a fully random, national sample is needed to replicate our findings and determine their generalizability.
Such work also should account for participants' political orientation, which was not measured in the present study. Given the hyperpartisan sociopolitical context in which this survey was administered (Iyengar & Westwood, 2015), we did not want to prime political beliefs or suggest a relation between political orientation and First Amendment knowledge or competence. Although participants may have been primed regardless due to the survey examining issues surrounding the Constitution and government, we did not want to highlight political orientation specifically or cause participants to view the survey through a partisan lens. However, this variable may be important for future research to include.

Conclusion
This study examined the current state of First Amendment knowledge and competence in adults residing in the United States. Unfortunately, the affirmation of the hypotheses paints a disappointing picture of an unknowledgeable, incompetent American people, most of whom cannot accurately estimate their level of First Amendment knowledge nor effectively transfer that knowledge to relevant situations. These results fall in line with previous literature examining general political knowledge (e.g., Bennett, 1988), the Dunning-Kruger Effect (Kruger & Dunning, 1999), and knowledge transfer (Simons, 1990as cited in Simons, 1999.
Beyond academia, our findings echo real concerns and experiences from American residents.
In 2018, fourteen high school students filed a federal class-action lawsuit against the state of Rhode Island for failing to provide adequate civic education and prepare students to be "capable citizens" (Marco, 2018). Calls for increased civic education were made after the insurrection at the U.S.
Capitol Building on January 6, 2021, too (Heim, 2021). Despite beliefs that they were acting within the bounds of their First Amendment rights as "patriots" (Barry et al., 2021, par. 22), individuals involved in the insurrection have been arrested and charged with a number of crimes, including disorderly conduct in a Capitol building; parading, demonstrating, or picketing in a Capitol building; civil disorder; obstruction of law enforcement during civil disorder; obstruction of justice/Congress; obstruction of an official proceeding; and many others (NPR, 2021). The lines between protected and unprotected speech blur even further for on-and off-duty police and peace officers, including some who were part of the 2021 insurrection but who have narrower free speech rights than the general public in certain circumstances (Cohen, 2021). The actions taken by these individuals further point to a need for more research and education on this topic. Now that we have an understanding of the current level of U.S. residents' First Amendment knowledge and competence, as prior researchers have called for (Bennett, 1997;Lupia, 2006), we can start the work of increasing knowledge, tackling the problems related to overestimating knowledge in those who are unknowledgeable, and improve individuals' ability to successfully apply First Amendment knowledge. This is particularly important because knowledge transfer failures of this type might lead to unwanted contact with the justice system and an inability to effectively identify and fight plausibly unconstitutional legislation. Calls have already been made to begin the important work of increasing public understanding of how the freedoms of the First Amendment are applied in daily life ("The 2019 State," 2019). Future research should aim to develop evidence-based interventions that remedy the issues revealed in this research and build a more knowledgeable, competent, and better America. 1 One community participant identified as "Other" gender. 2 Education was coded as 1 = Some high school; 2 = High school graduate; 3 = Some college; 4 = College graduate; and 5 = Postgraduate. 3 Objective Political Activism was scored as the number of experiences participants reported (range = 0 to 6). 4 Subjective Political Activism was coded as 1 = Not at All; 2 = Somewhat; and 3 = Very. Note. Asterisks indicate significant divergences in the percentage of people who scored in this letter category from the percentages expected based on Bennett's (1988) findings. Significance is achieved at the p = .05 level for objective knowledge, overall competence, and competence for the online and community samples, and at p = .01 for the vignettes based on Bonferroni correction for four comparisons.