Examining the Effects of Instructor Advocacy or Neutrality on Students’ Moral Development

Consider the following situation: Dr. Bonitas decides to discuss euthanasia in his college-level ethics course. He is obligated to discuss multiple perspectives on the issue but is not required to remain neutral. Despite his own strong position, he decides to take on a completely neutral stance in the classroom to encourage students’ independent thinking. During class discussion, he refrains from disclosing his own position and attempts to give equal time and treatment to every perspective on the issue. The discussion among students is lively and respectful, and Dr. Bonitas covers all of the relevant arguments and perspectives. However, after class, Dr. Bonitas questions whether he should have placed more emphasis on perspectives and arguments that are more logical and coherent. He also wonders whether his own position on the issue was in any way obvious to the students. Was he completely neutral? And how did this discussion influence students’ moral development?

The next time Dr. Bonitas teaches his ethics course, he decides to not remain neutral in his class discussion of euthanasia so he can model good, logical reasoning for the students. He describes the issue generally and then presents multiple perspectives, emphasizing arguments that are more logical and coherent. He then fully discloses his own perspective before facilitating a class discussion that includes him trying to persuade his students of the strengths of his position. After class is over, he reflects on what happened. Were students too influenced by his own position in that they accepted his position before weighing all other arguments and counterarguments? Did he provide adequate coverage of other logical, coherent perspectives? How did this approach influence students’ moral development?

Dr. Bonitas, in these situations, captures the essence of both neutrality and advocacy in teaching. First, he uses a classic neutrality model whereby he does not disclose his own position or preferences and provides equal treatment to every perspective, encouraging students to consider all options relative to the evidence (Markie, 1994). The intention of such an instructional strategy is to promote the rational autonomy of students (Basinger & Basinger, 1989; Baumgarten, 1980; Bomstad, 1995; Kupperman, 1996) as well as the civil discourse skills of evaluating multiple arguments and evidence, which students will need when entering the larger community (Tomhave, 2015). Many scholars, though, have argued against this strategy. For example, according to Goldman (1981), remaining completely neutral is “an exercise in self-deception” because the instructor’s own position will implicitly emerge in a more manipulative manner (p. 9). O’Neil (1991) argues that instructors having a neutral discussion of a topic (like racism) would merely confirm the students’ preexisting views and biases. In another line of criticism, scholars (Ammon, 1992; Hanson, 1996) have argued that the neutral instructor stimulates moral relativism whereby students walk away with the impression that all values and arguments are of equal worth and hence become moral agnostics. Heybach (2014) claims that teacher neutrality can have an unethical dimension of sacrificing one’s freedom insofar as teachers becoming overly neutral bystanders that fail to question unethical policies and practices (e.g., genocide). Ultimately, according to neutrality critics, this instructional strategy encourages dishonesty, moral relativism, and lack of commitment to values instead of promoting rational autonomy (Bomstad, 1995).

The advocacy model, illustrated by Dr. Bonitas in the latter situation, is at the opposite end of the continuum from neutrality. In the hypothetical scenario, Dr. Bonitas not only fully discloses his own position but also spends more time discussing it in order to actively persuade students of his own view. Those in favor of the advocacy model have argued that students need a model of how to commit to a position that is intellectually and rationally compelling (Ammon, 1992; Goldman, 1981; Hanson, 1996; O’Neil, 1991; Parkyn, 1991; Warnock, 1975a, 1975b). According to Warnock (1975a, 1975b), students need a leader in moral argument, not a fence sitter. Kunkel and Radford-Hill (2011) define advocacy as “the passionate engagement of ideas leading to a principled stance” (p. 97); they argue that advocacy is pedagogically superior because it fosters student engagement and promotes student learning. A common criticism of this model is that the teaching essentially becomes indoctrination and violates student autonomy (Kupperman, 1996). Goldman (1981), however, argues that advocacy is not synonymous with indoctrination —to advocate is to “offer the most intellectually and rationally compelling reasons one can” (p. 8). Bomstad (1995) offers a more nuanced critique of the advocacy model: advocacy that involves a persuasive element does not require a full and fair hearing of all relevant opposing views and thus aims to convince others rather than to discover or to learn.

Other scholars have proposed alternative models that lie along the continuum. One example that Bomstad (1995) and (1986) have advocated is procedural or committed neutrality. Instructors adopting this model disclose their own position while also providing full and fair treatment of all perspectives. Proponents explain that this approach allows the instructor to model a commitment to a position that is well-informed, logical, and coherent. Instructors using procedural neutrality can also show students how to withhold judgment until all the facts and best evidence are in and the need to critically examine all perspectives in a fair-minded manner. O’;Neill (1991) calls this latter modeling a sign of reflective virtue. Moral educational philosopher Nel Noddings (2013) has also supported pedagogical neutrality, defining it as the “willingness to consider all reasonable points of view without endorsing one as the absolute truth” (p. 63). She, in fact, stated that it is an “ethically and strategically effective way” to discuss controversial issues with students. Of course, this model is not without its critics. The arguments against this model are similar to those on the extreme ends. For example, the impartiality aspect may be undermined by the instructor’;s self-disclosure, such that students feel pressured to adopt the instructor’;s position or choose to take the cognitively easier route of focusing only on the instructor’;s reasoning for her position.

The neutrality/advocacy debate pervades all disciplines in education and at every level, from kindergarten through graduate school and within the classroom as well as in broad curriculum decisions. Nonetheless, the issue is particularly relevant in moral education, for values and ethics¹ are the “stuff” of moral education. Not everyone agrees that the teacher should “impose her own values and ethics” upon her students. To see how moral educators have dealt with this issue, we briefly review neutrality/advocacy positions in three well-known moral education approaches.

In the last fifty years, Kohlberg’s (1984) stage theory of moral reasoning² has become a foundation for a large contingency of moral development researchers and educators. The most common educational approach directly tied to his theory is the moral dilemma discussion method. Its objectives are to (1) expose students to the next higher stage of Kohlbergian moral reasoning, (2) expose students to contradictions of their current moral stage, which should lead to student dissatisfaction with their current level, and (3) create an atmosphere of dialogue that is open and reflective (Kohlberg, 1976). In achieving these objectives, the teacher should not stress a right or wrong answer; rather, the teacher should pose nonthreatening, open-ended questions and keep the students focused on the moral dilemma at hand (Galbraith & Jones, 1976). Though the goal of the moral education strategy is to have students move to a higher stage of reasoning, the student should do so without any explicit advocacy on the teacher’;s part: the teacher does not expressly advocate for any position or reason but only exposes students to higher stages through the use of open-ended questions to allow them to construct their own new reasons of a higher stage. Thus, teachers using the moral dilemma discussion method would be on the neutral end of the neutrality/advocacy continuum.

Values clarification has been another popular moral education approach in the last 5 decades. The intention of this program is to clarify students’; own values. Importantly, the teacher’;s role is not to advance any particular values (Simon, 1976) but to respond to the student’;s words or actions, resulting in the student’;s thinking about what he has chosen, what he prizes, and what he is doing (Raths et al., 1976). Clearly, then, the teacher is more neutrality than advocacy-oriented. However, the neutrality of the teacher, according to the values clarification approach, “does not mean that

the teacher should not maintain or even express his/her own values; rather, it means that the personal values held by the teacher should not interfere with the more basic function of developing the valuing process” (Chazan, 1985, p. 61). Thus, this approach allows the teacher to use the procedural neutrality model rather than requiring the classic neutrality one.

Traditional character education, following Wynne and Ryan (1993), has explicitly criticized both values clarification and the moral dilemma discussion method for having teachers as neutral “moral bystanders” (p. 121), arguing instead for direct moral instruction. Their approach calls for teachers to use explanation and exhortation, among other strategies, for authentic character education (Ryan & Bohlin, 1999). The teacher’s use of explanation requires describing to students what good habits are and why they are important. Exhortation by teachers is needed to inspire, or persuade, students to be moral. This approach to character education is clearly more oriented toward the advocacy model. The proponents of this approach, though, do qualify the extent to which teachers should use advocacy. According to Ryan and Bohlin (1999), teachers should advocate for their larger community’s agreed-upon core values —not for their own particular view on a highly controversial issue. Thus, they argue for an advocacy model but with some limitations.

The teacher’s stance of neutrality/advocacy within these three popular moral education approaches shows a broad range on the continuum. The moral dilemma discussion and values clarification methods fall on the more neutral end of the continuum, while the traditional education approach is more advocacy-oriented. Regarding more modern moral education approaches, it is quite frankly unclear exactly where they lie on the neutrality/advocacy continuum. Most current approaches (e.g., Child Development Project, Smart & Good Schools, integrative ethical education model, domain theory applications to moral education, the PRIMED model, all described in Nucci et al., 2014) have a much wider scope for moral education that includes aspects like classroom climate, school atmosphere, student-teacher relationships, student discipline, and motivation. With such all-encompassing approaches, moral education scholars of the current programs seldom offer explicit information about the teacher’s role in classroom. Nonetheless, most modern approaches assume or encourage the teacher to have a primarily constructivist classroom whereby the teachers act as guides, facilitators, and models and provide activities and create environments conducive to self-analysis and metacognition. In doing such, constructivist educators require students to become active learners, with students (and the teacher) coconstructing new understandings and ultimately becoming more critically aware and ethically agentic (Gordon, 2009). Constructivist educators, then, would probably take on a more procedurally neutral role in classroom discussions. However, they could also be said to participate in advocacy of a sort, insofar as they promote the goals and values inherent in the program as a whole.

Philosophical education in particular has often emphasized the Socratic method of teaching through questions designed to improve critical thinking and encourage students to make their knowledge and assumptions explicit. While at first glance it may seem that such an approach lends itself primarily to neutral methods such as dilemma discussions and values clarification, it can also —depending upon the nature of the questions —be used to advocate particular virtues or values. In fact, Socrates’ own discussions (e.g., Plato, n.d.-b n.d.-c) were often centered around virtues and values such as justice, wisdom, and love.

The neutrality/advocacy debate has flourished most predominantly in the domain of philosophy, more specifically in the teaching of philosophical ethics. On the one hand, Baumgarten (1980) has argued that philosophy teachers should adopt the classic neutrality model in the classroom. Schaupp (2015) also argues for a classically neutral approach in teaching ethics, especially when disciplinary experts have rational yet conflicting views on an issue. From her perspective, instructors’ commitment to intellectual humility and rational inquiry necessitates presenting conflicting perspectives as fairly as possible —even when it might result in students’ adopting beliefs that the instructor deems incorrect. Goldman (1981) and others (e.g., Brod, 1986), on the other hand, have asserted that philosophy teachers should use a more advocacy-oriented approach, promoting particular values and positions as traditional character education endorses. Others, like Bomstad (1995), have promoted a model that combines elements of the two endpoints: the procedural neutrality approach, which is similar to Besong’s (2016) argument that instructors should engage in ethical debates while teaching them —in contrast to classical neutrality, he argues, this approach “can effectively avoid eliciting skeptical attitudes among students without sacrificing desirable pedagogical outcomes” (p. 401). Cooper (2009) interviewed 40 instructors of ethics at six leading English-speaking universities. He found that instructors were evenly divided on their advocacy/neutrality attitude, with nearly the same percentage of faculty believing that instructors should be neutral (35%) and should take a stand and reveal their personal dispositions (33%). It seems that consensus has not been reached among ethics teachers; thus the debate continues.

Throughout the years, scholars such as Wilson (1975) and Bomstad (1995) have expressed the need for empirical investigations of whether neutrality or advocacy has any effect and the extent to which criticisms against the models are warranted. Unfortunately, little work has been done here. Research most relevant to instructor/advocacy effects comes from three very different projects: one from Britain in the 1960s and 70s, another from the Netherlands within the last 2 decades, and a third more recent British study.

The Humanities Curriculum Project was developed in Great Britain in the late 1960s with the goal of raising the school leaving age (Stenhouse, 1975) by bringing more relevant and engaging topics into the classroom. More specifically, teachers facilitated student discussions of human issues of universal concern, such as poverty, law and order, war and society. The issues were controversial in that they were laden with divergent sets of values. The teacher, in discussing these issues, acted as a neutral chairman, similar to the classic neutrality stance. Research from the project showed that teachers had a difficult time adopting the neutral chairman role and found it extremely demanding. Despite this, a large number of teachers found learning and using the role to be a rewarding skill. Students’ preferences for teachers’ use of the neutral chairman role were mixed, with some having high regard for it and others being skeptical. In terms of student outcome effects, research showed an increase in students’ verbal achievement scores.

In a more recent project, Veugelers (2000) examined the teaching of values in secondary schools in the Netherlands. The researcher asked both teachers and students about specific instructional strategies that are either used or preferred when teaching value-relevant topics. When teachers were asked about which instructional strategies they used in the classroom, for the most part they initially stressed differences in values without expressing their own preferred values, and then ended by expressing the values they find important. When students were asked which instructional strategies they preferred, they most strongly endorsed teachers’ using or modeling critical thinking by explicitly pointing out differences in values. Of the critical thinking strategies, students were asked to rank two: they showed a slightly strong preference for the teacher expressing the values he/she finds important (similar to advocacy) compared to the teacher refraining from expressing their value preferences (akin to neutrality).

Cotton (2006) investigated the beliefs and practices of three secondary-level geography instructors teaching controversial environmental issues over a 2-year period in England. She used data from interviews as well as classroom transcripts. All three instructors favored taking a neutral or procedurally neutral approach. Cotton’s findings, though, revealed that the instructors encountered significant challenges in maintaining a neutral approach. An analysis of classroom interactions “suggest that the influence of the teachers’ own environmental attitudes was greater than they intended, or in all probability, realized” (p. 237). Thus, the study’s results support Goldman’s (1981) critique that remaining neutral is “an exercise in self-deception.”

Though each of these projects offer interesting empirical results related to neutrality/advocacy, we are still lacking empirical research that directly compares student effects with a teacher who adopts a more neutral approach to student effects with a teacher who uses a more advocacy-oriented model. Our research project examined exactly this question. More specifically, the class setting of our research was a college-level philosophical ethics course, and our outcome was student moral development.³ We focused on two specific variables of student moral development: students’ judgments of ethical theories and their moral self-perceptions. Our choice of these two variables was rooted in the course catalog description of the ethics class, which emphasized learning ethical theories and “application of ethical principles to areas of personal conduct.” We employed a pre-post design, assessing the student outcome variables both before the ethics course began and after it was completed.

The first of our two research questions focused on comparing student outcomes from instructors who used a more neutral approach to instructors who adopted more of an advocacy model. Thus, our first research question was: to what extent does an instructor’s position of neutrality/advocacy influence students’ judgments of ethical theories, moral self-perceptions, and ethical and unethical behaviors? Our other research question focused on the possibility that neutrality/advocacy may not have any effect at all and that maybe students’ moral development occurs for all students taking the ethics course. Our second research question, then, was: to what extent do students’ judgments of ethical theories, moral self-perceptions, and ethical and unethical behaviors change, regardless of their ethics instructor’s position of neutrality/advocacy? In other words, we examined whether students had sizeable increases in the two outcome variables as a result of simply completing the ethics course.

A total of 232 students participated in the study. The participants were enrolled in 11 sections of PHIL 214 Introductory Ethics courses at a midsize Midwestern Catholic university. Students ranged in age from 18 to 37 (mean = 20.5, SD = 1.8). Most participants, 66%, were in their second or third year of college, with 7% in their first and 27% in their last year. The majority of participants, 88%, were Caucasian. Ethnicity of the remaining 12% was African American (4%), Asian (5%), Latino/a (2%), and other (1%). The number of males and females was approximately equal (47% male, 53% female). The students’ field of major study varied, but all of the participants had taken at least one previous philosophy course.

Of all the students (n = 339) in the 11 sections of PHIL 214 that we invited to participate in our study, 68% agreed. The demographic make-up of our sample, in terms of age, gender, ethnicity, and college major, are the same as all of whom registered for the 11 sections of PHIL 214, suggesting no selection effects for those who elected to participate in our study.

Six introductory ethics instructors were also participants in our study. With the exception of one instructor, all were male. Every participating instructor had a PhD in philosophy. Five of the six were full-time tenured or tenure-track faculty; one was a long-time adjunct instructor. All instructors were in good standing with the philosophy department, with no outstanding concerns regarding the quality of their teaching. Additional characteristics of the participating instructors will be described more fully in the Results section.

The introductory ethics course is required for all students at the university. It is the second of two philosophy courses they must complete in order to graduate. Because introductory ethics is a required course, one to two dozen sections of it are offered during each term or semester, with about a dozen different instructors. For our study, we collected data in the introductory ethics course during either an intensive 3-and-a-half week term or a more traditional 14-week term. Student participants were enrolled in one of 11 different sections of introductory ethics, which were taught by one of our six participating instructors. Class size ranged from 20 to 60 students.

Instructors have a fair amount of autonomy in what and how to teach introductory ethics. They have the course description from the university’s course catalog as their most basic guide:

The philosophy department also has a set of criteria (a combination of student learning outcomes and general course content) giving further specification to the requirements of the catalog description and applying to all instructors. For example, the course will emphasize logical analysis of ethical arguments and avoidance of fallacious reasoning; that it “is concerned with such fundamental ethical concepts as happiness, pleasure, virtue, moral development, love and friendship, contemplation, the common good, social justice, human rights and obligations”; and that students will read significant primary texts from different ethical traditions, including substantial readings from the work of Aristotle and Thomas Aquinas as well as from other foundational approaches such as utilitarianism and deontology. To gauge how our participating instructors taught their sections, we asked them, using a short written questionnaire, about their instructional goals, what readings they required, and types of methods they used in class. Information from this questionnaire will be described more fully in the Results section.

We employed a quasi-experimental pretest-posttest design. Student participants completed the same set of measures at two different times: before the course started (pretest) and after the course was completed (posttest).

Our independent variable of interest was the extent to which instructors used neutrality or advocacy in the classroom. To measure this, we provided a short written questionnaire to the instructor toward the end of the semester asking several questions, including their course goals, readings, and assignments; self-described teaching style; number of times having taught the course; number of years teaching; their rating of their class’s academic strength, talkativeness, and interest in subject matter; and number of students in their class. The question that determined the independent variable was, “Where would you place your teaching on the continuum between neutrality and advocacy regarding particular ethical theories and/or positions (mark the appropriate place on the line with an ‘x’)?” Below this question, we had a dotted line across the page with “Neutrality” on the far left-hand side and “Advocacy” on the far right-hand side. After they marked an “x” on the dotted line, we asked three follow-up questions: what theories and/or positions they advocate if not completely neutral, what methods they use to advocate these theories and/or positions (if not entirely neutral), and what methods they use to maintain neutrality if they marked the far left-hand side of the continuum. Our construction of the neutrality and advocacy groups, using this assessment, will be described in detail in the Results section.

Student participants completed a packet that contained several different questionnaires along with a number of demographic questions (e.g., age, year of study, sex, major area of concentration). We deem the measures described below to be more psychological than educational in nature. Thus, these measures should not be considered educational outcomes.

Moral Self-Perception. We had two paper-pencil measures of moral self-perception: one assessing specific virtuous qualities and another measuring moral self-concept. In the virtuous qualities measure, we asked participants to respond to several statements characteristic of one of seven different virtues⁴: courage, kindness, honesty, good temper, patience, perseverance, and lack of envy. Each of these qualities is considered morally virtuous under typical accounts of philosophical virtue ethics, including the Aristotelian-Thomistic tradition, which receives special consideration in the introductory ethics course. Participants rated each statement on a 5-point scale (1 = almost always false of me, 2 = usually false of me, 3 = true of me about half the time, 4 = usually true of me, 5 = almost always true of me). Statements were in random order and were not labeled according to virtue in the questionnaire; sample statements of each virtue included:

• Courage: I stand up for what I believe, even when others are against it.

• Kindness: I am too busy to help others much. (reverse scored)

• Honesty: I admit my mistakes, even when it would be easy to hide them.

• Good temper: I rarely raise my voice, even when I am angry.

• Patience: I become frustrated when I have to wait for other people. (reverse scored)

• Perseverance: If I don’t succeed at something the first time, I keep trying.

• Lack of envy: I envy people who are smarter or more talented than I am. (reverse scored)

For each virtue, we averaged participants’ responses, creating a mean score with a range of 1 to 5. Each virtue scale showed good to very good internal consistency, with coefficient alphas ranging from .67 to .85 on the pretest and posttests.

Our moral self-concept measure was based on Arnold’s (1993) good-self assessment. Participants were presented with three concentric circles (similar to a target board), with each ring corresponding to a different degree of centrality to the self. Above the circles were five columns of personality-related adjectives listed in alphabetical order. Examples included caring, intelligent, joyful, and selfish. The instructions asked participants to first cross off the fifteen qualities that least described them. Next, participants were asked to write the three adjectives most central to who they are as a person in the inner circle, write their three next most central qualities in the middle circle, and write their four next most central qualities in the outer-most circle. To score the measure, we classified each adjective as either: (1) morally good, worth 2 points (e.g., caring, fair, generous), (2) usually morally neutral, worth no points (e.g., health-conscious, popular, romantic), or (3) morally bad, worth–2 points (e.g., impatient, lazy, selfish). The inner circle had a weight of 3, the middle circle a weight of 2, and the outer-most circle a weight of 1. The sum of the three circles was calculated to give an overall score, with possible scores ranging from –32 to 38. (There were not enough morally bad adjectives on our list to result in a score below –32.)

Ethical Theory Judgment. We used eight ethical theories: skepticism/positivism, egoism, emotivism, subjectivism, cultural relativism, utilitarianism, deontology, and virtue ethics. These eight approaches provide broad coverage of the range of possible positions one might take regarding ethics, and instructors of survey courses in philosophical ethics typically give explicit attention to most or all of the last three and at least implicit attention to most or all of the first five, which are often treated as objections to the last three theories. For each, we created two to five statements reflecting a tenet of the theory. Examples include:

• Skepticism/positivism: There are no right and wrong answers to ethical questions.

• Egoism: The “right” thing to do is whatever is in your best interest.

• Emotivism: Moral arguments are just ways people express their feelings about issues.

• Subjectivism: The right thing to do is whatever you think is right.

• Cultural relativism: What is right and wrong varies from culture to culture.

• Utilitarianism: We should always do whatever will bring the best results for all concerned.

• Deontology: There are universal human rights that we should always respect and uphold.

• Virtue ethics: Being ethical means becoming a good person.

Participants rated each statement on a 5-point scale (1 = strongly disagree, 2 = disagree, 3 = undecided, 4 = agree, 5 = strongly agree). The statements for all eight ethical theories were randomly ordered within this section of the questionnaire. We collapsed the eight ethical theories into two groups: the less defensible (skepticism, egoism, emotivism, subjectivism, and cultural relativism) and the more defensible (utilitarianism, deontology, and virtue ethics). The Cronbach alphas for the less defensible group were .85 pretest and .88 posttest; the alphas for the more defensible were .62 pretest and .65 posttest.

Ethical and Unethical Behaviors. In this measure we asked participants to think about the most recent month during which they were at school and answer 39 questions regarding the frequency with which they engaged in common ethical and unethical behaviors during that month. Examples of ethical behaviors included “Take time to comfort someone who was having a hard time” and “Clean up after roommates or family members”; unethical behaviors included “Break a promise to a friend or family member” and “Make a rude or obscene gesture toward someone.” We again asked participants to use a 5-point scale to answer each question (1 = never, 2 = once or twice, 3 = three or four times, 4 = five or six times, 5 = more than six times). We averaged participants’ responses to create a mean ethical behavior score and mean unethical behavior score, each having a range of 1 to 5.

Social Desirability Bias. Because many of our questions reflected common, desirable social norms, we were concerned that participants would respond as they thought they should and not as they actually believed or did. We assessed participants’ tendency to conform to these pressures, their social desirability bias, using the Marlowe-Crowne Social Desirability Short Form (Reynolds, 1982). Participants rated thirteen different statements related to social desirability as either “true of me” or “false of me” (e.g., “I sometimes feel resentful when I don’t get my way”, and “I am always courteous, even to people who are disagreeable”). For each item, the socially desirable but almost certainly dishonest answer was worth one point while the socially undesirable answer was worth zero points; so mean scores ranged from 0 (no social desirability bias) to 1 (very strong social desirability bias).

Students enrolled in the participating sections of Introductory Ethics were asked on the first day of the term to participate in the study and read a consent form. Those who decided to participate were given a packet consisting of the questionnaires and asked to complete it at that time. Students who did not participate (approximately three in each section) could complete other tasks silently at their desks. In the last week of the term, students were asked to participate again, and if agreeing to do so, completed the same packet of questionnaires.

In all sections and in both the pretest and posttest packets, the questionnaires were completed in the following order: moral self-concept, self-perceived virtuous qualities, ethical theory judgments, and demographic-related questions.

We first report how we determined the independent variable for our first research question, followed by the results of that first question and lastly the results of our second research question. Our first research question involved the independent variable of the instructors’ self-assessed neutrality/advocacy.

The six participating instructors’ mark on the Neutrality/Advocacy continuum was measured in inches. Being that the dotted line of the continuum was 6 inches long, a “0.0” indicated complete Neutrality, whereas a “6.0” reflected very strong Advocacy. Table 1 shows each instructor’s numerical rating in inches, along with his/her answers to our three follow-up questions.

Instructors’ variability in their self-assessed neutrality/advocacy showed the entire range, from 0 to 6. In deciding which instructors to group together for a neutrality group and advocacy group, we used both their numerical rating and responses to the follow-up questions. Instructor A’s responses describe the classic neutrality model in which the instructor does not disclose his/her own position and provides a balanced and impartial treatment of all positions or theories. Instructor B and D’s responses closely align with the procedural neutrality model whereby instructor states his/her position while providing full and fair critical treatment of all positions. Instructor C’s responses, though brief and somewhat vague, hint at using such a model as well. On the other side of the continuum, Instructors E and F demonstrate strong advocacy, stating that they not only make explicit their specific position or preference but also spend time actively persuading students of the strength of their position. With Instructors A through D indicating classic or procedural neutrality, we aggregated their students into one “neutrality” group. For Instructors E and F, we combined their students into one “advocacy” group.

The quasi-experimental design of our study requires assurance that our two instructor groups were approximately equal. We carefully looked for differences between the neutrality and advocacy instructors along the following dimensions: instructional goals, readings, and assignments; self-described teaching style; number of times having taught the course; number of years teaching; instructor’s rating of students’ academic strength, talkativeness, and interest in subject matter; and number of students in their class. There was strong overlap in instructional goals between the two groups, which is not surprising given that instructors, according to specified course criteria, needed to address particular learning outcomes and course content. In addition to departmental criteria such as “prepare students to reflect critically on ethical questions raised in other courses and in daily life,” at least one instructor in each group mentioned goals of being able to understand key ethical concept and theories, analyze philosophical texts, apply course content to real-life issues, construct ethical arguments that show good reasoning and logic, and grow as a good, happy moral person. After accounting for these goals in each group, the remaining differences between the groups were very closely related to the previously stated goals (e.g., valuing place of reasoning in moral life). For course readings, at least one instructor in the neutrality and advocacy group primarily used one text. At least one instructor in each group, though, also used primary sources as readings, including Aristotle’s Nicomachean Ethics. Each group also had the same variation of teaching methods. The percentage of time spent on lecture ranged from 40 to 90 for the neutrality instructors and from 40 to 75 for the advocacy instructors. Neutrality instructors reported a range of 10 to 40 percent of their time using discussion, and advocacy instructors reported a range of 5 to 35. At least one instructor in each group described their teaching style as “informal;” additionally, at least one instructor in each group used “enthusiastic” or “energetic” to describe their teaching style.

The median number of times that neutrality instructors reported having taught the course was 22 (with a range from 2 to 75); the median for advocacy instructors was 12 (ranging from 5 to 18). For number of years having taught at the college level, the median for the neutrality instructors was 12 (ranging from 6 to 40); the advocacy group’s median was 5 (with a range from 2 to 8). Thus, the neutrality group seemed to be slightly more experienced in teaching in general and the specific course.

Instructors also reported the extent to which the students in their PHIL 214 classes were academically strong, talkative, and interested in the course. The neutrality and advocacy groups did not differ in any of their ratings. Instructors in both groups rated their students on these dimensions as either average or above average, with only one exception: one instructor rated his students as below average for talkativeness in one of his classes.

The last variable to examine between the two groups was class size. The median number of students that neutrality instructors had was 29 (with a range from 20 to 58). Advocacy instructors also had a median number of 29 students in their class (ranging from 28 to 33).The neutrality group included a total of 103 students, and the advocacy group was comprised of 129 students. To make sure that the two student groups were approximately equal in ways other than size, we examined differences in the following: gender, ethnicity, age, year of study, grade point average, politically liberal or conservative, religiosity, and number of ethics-related courses previously or currently taken. Using logistic regression, we found that year of study was the only significant predictor for group membership (see Table 2). More specifically, 76% of the advocacy instructors’ students were juniors or seniors, whereas 48% of the neutrality instructors’ students were in these 2 years of study. Interestingly, age was not a significant predictor considering the year-of-study finding. Nonetheless, we statistically controlled for year of study in all analyses hereafter.

Research Question 1: To what extent does an ethics instructor’s position of neutrality/advocacy influence students’ (a) judgments of ethical theories, (b) moral self-perceptions, and (c) ethical and unethical behaviors?

We performed repeated measures analyses for all variables, focusing specifically on whether pre-post changes varied by neutrality/advocacy group for this research question.

Responses to the pretest and posttest variables were the dependent variables of the within-subjects effect, and neutrality/advocacy group assignment was the independent variable of the between-subject effect. We also had two covariates: pretest social desirability bias and year of study. Per Weinfurt (2000), we were most interested in the interaction effect when examining group differences, expecting groups to be the same at pretest and possibly differing at posttest.

For students’ judgments of ethical theories, we examined pre-posttest mean differences in two different variables: preference for more defensible ethical theories and for less defensible ethical theories. Students’ judgments of more defensible ethical theories showed no statistically significant pre-post change between groups. However, change in students’ judgments of less defensible ethical theories neared statistical significance for the “group x variable” interaction effect, with students of advocacy instructors showing a larger decreasing preference for less defensible ethical theories (see Table 3).

For students’ moral self-perceptions, we examined group differences in pre-post change for each of the following eight variables: courage, kindness, honesty, good temper, patience, perseverance, lack of envy, and moral self-concept. Of the eight repeated measures analyses, four produced a statistically significant “group x variable” interaction effect, shown in Table 3. As seen in Figure 1, the interaction effect for all four variables show a consistent pattern of the neutrality group having slight pretest-to-posttest decreases, while the advocacy group had small pre-to-post increases. We found no interaction effect for the moral self-concept measure.

The other set of variables was ethical and unethical behaviors. Changes in students’ self-reported ethical behaviors neared statistical significance, with students in the neutrality group reporting a decrease in the frequency of their ethical behaviors while students in advocacy group showed no pre-post change (see Table 3). For unethical behaviors, no difference was shown between the two groups.

Research Question 2: To what extent do students’ (a) judgments of ethical theories, (b) moral self-perceptions, and (c) ethical and unethical behaviors change, regardless of their ethics instructor’s position of neutrality/advocacy?

For our second research question, we examined whether taking an ethics course has general effects on the students regardless of the neutrality/advocacy of their instructor. Our statistical analyses for this research question were the same as our first research question; the only difference being that we were now looking for a within-subjects effect (rather than an interaction effect). Of the ethical judgment variables, both showed statistically significant pre-post changes. Students in both groups had a higher posttest mean for more defensible ethical theories and a lower posttest mean for less defensible ethical theories (see Table 3). We also performed a follow-up analysis, comparing students’ differences in their preference for more defensible and less defensible ethical theories, which created two new scores. We computed the first score by subtracting students’ less-defensible mean score from their more-defensible mean score for the pretest only. The second score was the difference between their less and more defensible score for the posttest only. We then compared these two new scores to determine whether the mean difference was greater for the posttest as compared to the pretest. Results of the statistical analysis were statistically significant: F(1, 205) = 9.32,p < .01 (partial q² = .04), whereby the posttest difference between the two theories score was greater compared to the pretest (1.02 posttest mean difference; .72 pretest mean difference). This finding indicates that students in both groups had a stronger preference for more defensible theories after taking an ethics course.

For the rest of the variables, only one other variable showed a within-subjects effect: unethical behaviors. Students in both groups reported engaging in unethical behaviors less often in the posttest as compared to the pretest. None of the moral self-perception variables showed statistically significant within-subjects effects.

The primary objective of this study was to examine the extent to which an ethics instructor’s position of advocacy or neutrality influenced students’ judgments of ethical theories, their moral self-perceptions, and ethical and unethical behaviors. We found that students of advocacy instructors (a) showed a greater decreasing preference for less defensible ethical theories compared to students of neutrality-oriented instructors, (b) showed increases in four different virtues (courage, kindness, good temper, and patience) compared to the neutrality instructors’ students, who showed decreases for these same virtues, and (c) showed no change in the frequency with which they engaged in ethical behaviors compared to the other group, which showed a decrease. We also investigated the extent to which students’ (a) judgments of ethical theories, (b) moral self-perceptions, and (c) ethical and unethical behaviors change, regardless of their ethics instructor’s position of neutrality/ advocacy. We found that all students showed increasing preference for more defensible ethical theories and decreasing preference for less defensible ethical theories. In addition, students of all instructors reported engaging in unethical behaviors less often at the posttest compared to the pretest. We provide interpretations for each of these findings.

We asked student participants to rate their degree of agreement with tenets of various ethical theories, including skepticism/positivism, egoism, emotivism, subjectivism, cultural relativism, utilitarianism, deontology, and virtue ethics. While both groups showed a decrease in ratings of the first five theories listed, students of advocates showed a significantly greater decrease in preference for less defensible theories. The most likely explanation for this difference is that advocate instructors explicitly discuss the merits of the theories they endorse and the contrasting problems with other theories such as the five showing decreased student preference. While more neutral instructors may also discuss problems with these theories, they tend to present all theories as equal contenders rather than presenting some as more defensible than others. So students of advocates probably received a stronger message than did students of more neutral instructors regarding the problematic aspects of the less defensible ethical theories, as well as the more promising aspects of the more defensible theories; if so, they may have been rationally persuaded (see Besong, 2016; Goldman, 1981; Hanson, 1996; Kunkel & Radford-Hill, 2011; Parkyn, 1991), not merely personally influenced, to adopt positions regarding these theories similar to the positions of their instructors.

In both the pretest and posttest, we asked student participants to consider statements related to several virtuous qualities: courage, kindness, honesty, good temper, patience, lack of envy, and perseverance. We found that the neutrality group had slight decreases or no change in the degree to which they perceived themselves as having four (courage, kindness, good temper, and patience) of the seven virtues, while the advocacy group had small pre-to-post increases with respect to these same four qualities. There are a few possible explanations for the difference in self-perception between the two groups. First and perhaps most obviously, as a proponent of traditional character education might expect (see Ryan & Bohlin, 1999; Wynne & Ryan, 1993), it may be that students of advocates really do tend to become slightly more virtuous over the course of their studies. While ethicists since the time of Aristotle (n.d./1999) have been careful to acknowledge that study does not by itself make the student morally virtuous, it can be helpful in showing what moral virtue looks like and why it is valuable. In our study, the two instructors who identified themselves as strong advocates both seem to emphasize virtue ethics (see Table 1). Students of advocates may have left the course with a clearer idea of how to develop moral virtue than did students of neutral instructors, and those who were already motivated to become more virtuous may have begun to do so by the end of the term. Because the neutral instructors present multiple, and often conflicting, theories of moral excellence as equal competitors, their students may have been less likely to follow the counsel of any particular theory in their own lives (Ammon, 1992; Besong, 2016; Bomstad, 1995; Hanson, 1996), which could result in a lack of increase in virtuous qualities. Although not very likely, it seems possible that students of neutral instructors actually tend to become less virtuous: perhaps the presentation of several competing theories with no clear “winner” can lead students to become cynical about the role of moral virtue in their own lives, so that they are unmotivated even to maintain the level of virtue they possessed at the beginning of the term.

A second explanation for the increase in self-perceived virtue among students of advocates may be that the emphasis on virtue in class readings and discussions led students to an increased awareness of their virtuous qualities but not (yet) to a real increase in the qualities themselves. For example, a student may have been quite kind and courageous all along but not have been aware of his kindness or courage until he studied these qualities in a reflective and systematic way in the context of his ethics course. Similarly, it could be that students of neutral instructors become increasingly aware of their character flaws over the course of the term, so that by the time of the posttest they perceive themselves to be less virtuous. Learning about the definitions and examples of virtuous qualities may have led students to raise their standards and realize that they failed to meet them. If this explanation is correct, however, it seems that students of strong advocates should show similar or greater decreases in self-perceived virtue compared to students of neutral instructors as a result of their more-focused study of virtuous qualities. But, combining this hypothesis with the previous one, perhaps students of advocates also became aware of their lack of virtue but subsequently strove to become more virtuous by adherence to the ethical theories presented by their advocate instructors.

A third explanation for the perceived increase in virtue among students of advocates and perceived decrease among students of neutral instructors is mere perception in one or both groups: the emphasis on virtue in the advocate instructors’ classes might lead students to identify with virtuous qualities so that they inaccurately perceived themselves as possessing them to a greater degree by the end of the term, and the students of neutral instructors might have underrated their virtues. For example, perhaps the advocate instructors’ more extensive treatment of virtue ethics led their students to a more achievable notion of virtue, while neutral instructors’ more brief treatment of virtue may have given the impression of an unrealistically high standard.

As we noted in the Results section, while both groups of students showed a decrease in self-reported unethical behaviors, only the students of more neutral instructors also showed a decrease in reported ethical behaviors. It is important to realize that participant reports of the frequency of their ethical and unethical behaviors may have been inexact for memory-related reasons. We asked student participants to report the frequency of various behaviors in which they may have engaged during the last month in which they were at school. When they took the pretest, most students had not been in school during the immediately preceding month, which may have meant that they estimated (and perhaps overreported) their behaviors during a typical month at school. In contrast, all of the participants had been in school during the month immediately preceding the posttest, which may have led them to rely on specific memories (and likely underreport) when completing the section regarding the frequency of their behaviors. So we would not find it surprising if both groups showed a slight decrease in both ethical and unethical behaviors. Thus the lack of decrease in reported ethical behaviors among students of strong advocates is at least as interesting to us as is the decrease among the students of more neutral instructors. As with the posttest differences between groups in reported virtuous qualities, the difference in reported behaviors may be due to real change, awareness, or mere perception.

Perhaps the most obvious and straightforward explanation of the difference in results between the two groups is that students of more neutral instructors really did decrease their performance of ethical behaviors. Seeing ethical theories presented in a neutral way may have made students skeptical about ethics (Besong, 2016; O’Neil, 1991) and thus less motivated to act ethically by the end of the term. Or, taking the possibility of reporting errors into account, perhaps the difference was due to students of advocate instructors having increased motivation to behave ethically, as character education proponents would expect (Ryan & Bohlin, 1999; Wynne & Ryan, 1993), and thus actually increasing slightly in their performance of ethical behaviors while students of more neutral instructors did not.

Another possible explanation for the interaction effect is that students of more neutral instructors were simply less aware of their performance of ethical behaviors than were students of strong advocates. It may be that by explicitly advocating for or against particular behaviors, advocate instructors made students more aware of their own behaviors than did neutral instructors —or that focus on aspects of ethics other than specific advocated behaviors made students of more neutral instructors less aware of performing such behaviors than they otherwise would have been.

We were pleased to see that although there was a difference in degree between the two groups, students of both strong advocates and more neutral instructors showed an increase in agreement with tenets of more defensible theories and a decrease in agreement with less defensible ones. We believe the most likely explanation for this result is that both types of instructors presented a variety of ethical theories along with some of their strengths and weaknesses. Even when the theories were presented in a more neutral way, the students’ reasoning abilities, along with the defenses and objections presented, were sufficient to increase preference for more defensible theories while decreasing preference for less defensible ones. This result may provide some evidence that classical neutrality does not necessarily breed skepticism (see Besong 2016; Markie, 1994; O’Neil, 1991; Tomhave, 2015) —and, perhaps, that there is some merit to the claim that even “neutral” instructors subtly influence their students’ positions (Goldman, 1981).

The decrease in reported frequency of unethical behaviors, as of other results, may be divided into three main hypotheses: real change, awareness, and mere perception. First, it may be that after taking an ethics class students of both advocates and more neutral instructors are motivated to avoid unethical behaviors and show a real decrease by the end of the term. Alternatively, perhaps in focusing on positive concepts such as virtue and maximizing happiness, students become less aware of their negative behavior (or are even motivated to suppress or deny it). Finally, students may merely perceive themselves to be engaging in negative behavior less frequently, perhaps (as discussed earlier) due to flaws in their memory leading to overreporting in the pretest and/or underreporting in the posttest.

Students of both more neutral and more advocacy-oriented instructors did show an increase in moral perception scores (see Table 3) approximately equivalent to replacing a neutral quality in the outer ring of the “target board” measure with a moral virtue or moving a virtue one ring closer to the center (representing the most central qualities of one’s identity). This change, however, was not found to be statistically significant. A possible explanation for the lack of significant change in self-concept, despite significant changes in perceived virtuous qualities, is that the self-concept measure asked participants to choose only the ten qualities most central to their identities. So changes in less-central virtues would not be detected by this measure. Another explanation may be that identity tends to be fairly stable over time and thus cannot be expected to change in easily measurable ways over the course of one academic term. In a K-12 setting, in which students typically remain with the same instructor for the entire academic year —and, in the case of the elementary grades, for the better part of the day —teachers’ impact on students’ identities may be more dramatic.

The implications of our study, if any, for university-level ethics instructors’ pedagogical methods will depend upon the instructors’ goals for their courses. As we noted earlier, we did not study educational outcomes of advocacy versus neutrality; we measured only moral self-concept, endorsement of ethical theories, and self-reported behaviors.⁵ The main differences between students of strong advocates and students of more neutral instructors in our results seemed to favor advocacy: as defenders of advocacy in ethics education would likely expect, students of strong advocates (a) perceived themselves to have grown in virtue while other students evaluated themselves as having decreased in virtue and (b) showed a stronger tendency to favor more defensible ethical theories over less defensible ones, and (unlike students of more neutral instructors) (c) reported no decrease in ethical behaviors. If, as (b) may seem to imply, critics of neutrality in ethics education are correct in claiming that it promotes ethical relativism, agnosticism, and/or a bystander approach to civic engagement (Ammon, 1992; Besong, 2016; Bomstad, 1995; Hanson, 1996; Hey back, 2014), then educators aiming to avoid such outcomes may do well to embrace advocacy.⁶ As already discussed, however, these results are subject to interpretation: perhaps students of more neutral instructors have become more honest or realistic, which would provide an alternative explanation for results (a) and (c). We were of course unable to measure whether the students of advocates really became more virtuous or better-behaved than students of more neutral instructors; we only know that they perceived themselves to have increased in several moral virtues and not to have decreased in ethical behaviors. We turn now to this and other limitations of our results.

The content of our study is obviously limited in that we didn’t measure every variable that may be considered relevant to the question of student effects of advocacy versus neutrality in ethics instruction. As noted just above, we measured only self-perceptions rather than other sources of information such as observed behaviors, the perceptions of peers, or educational outcomes.⁷ We also did not measure perceptions of every possible moral virtue, and we did not directly address moral vice. These content-related limitations open up many interesting areas for further investigation, which are discussed in more detail in the next subsection of this article.

Some of the self-perceived virtues we attempted to measure yielded low consistency in participants’ responses (with individual participants rating themselves very differently with respect to various questions that were supposed to address the same virtue). This problem points to a second main limitation in our study: our measures are, for the most part, newly developed and their reliability hasn’t been independently verified. Our results often pointed to ways we might fine-tune our measures for future use, and we look forward to doing so.

A third main limitation of our study relates to sample size. While we had hundreds of student participants, we had only six instructor participants: two strong advocates and four who identified themselves as more neutral in pedagogical style. Given the small size of our instructor sample, we cannot completely rule out the possibility that the differences in student groups’ results stems from instructor differences unrelated to neutrality versus advocacy. We think we’ve done all we reasonably could do at this point to minimize it, however: as discussed in the Results section, we made sure the two instructor groups were similar in terms of course goals, readings, and assignments; self-described teaching style; number of times having taught the course; number of years teaching; instructor’s rating of students’ academic strength, talkativeness, and interest in subject matter; and number of students in their classes. Further steps one might take to isolate the effects of advocacy versus neutrality might include obtaining a larger sample of instructors or looking at data other than self-reports (e.g., student evaluations or researchers’ observations). We hope to do at least the first of those in future studies.

As mentioned earlier in this section, there are a few ways in which our methods could be adjusted to further address our present research questions. A larger sample size would increase our confidence that we have isolated our independent variable of advocacy versus neutrality. Further, it would be beneficial to replicate our study in universities of various sizes, regions, and affiliations⁸ and in secondary schools. It might also be helpful to ask instructor participants to list (or select from our list) the philosophical theories that they explicitly address in their courses.

There are many directions in which our study might be built upon. It would be interesting to see the effects of advocacy versus neutrality in instruction on other aspects of moral development such as moral reasoning (on which the more neutral Kohlbergian approach focuses [see Kohlberg 1976, 1984]), sensitivity to morally relevant features of situations, or empathy. Educational researchers might also want to investigate educational outcomes of advocacy versus neutrality; that sort of research might be more easily carried out among K-12 instructors, who tend to have more similarities among themselves in instructional materials and goals.

Another possible direction in which to take our research would be longitudinal: one could study the longer term effects of advocacy versus neutrality, perhaps even in multiple ethics-related courses, on student moral development. For example, participants might take the pretest as incoming first-year students and the posttest as graduating seniors.

Further, one might study the effects of neutrality versus advocacy in comparison (and possible combination) with other variations in pedagogical style or course design. For example, do the perceived increases in virtue also result from pedagogies such as service-learning? If so, does combining advocacy with these pedagogies compound the increase? Does combining neutrality with the pedagogies reduce or eliminate the increase?

As we noted at the beginning of this article, despite longstanding debate regarding the relative merits of neutrality and advocacy in ethics instruction, there has not been any empirical research on the effects of these pedagogical methods on students’ moral development. While our research project represents a significant step toward closing this large gap in the literature, much work remains to be done. It is our hope that psychological and educational researchers, together with instructors of philosophical ethics, will continue to investigate systematically the effects of these and other pedagogies on students’ learning and their moral lives.