Please help!! Use "You Can Lead a Horse to Water but You Can't Make Him Edit: Varied Effects of Feedback on Grammar Across Upper-Division Business Students" (Located Below the Page)

Found directly below the assignment instructions under "Attachments," for guidance on completing this.

Develop a 500-750 summary of the article that includes paraphrases of the article, the research problem, questions, method, findings, and implications discussed by the authors. Cite sources.

Reminder: This is an academic exercise that should not be written in first person.

You Can Lead a Horse to Water but You Can't Make Him Edit: Varied Effects of Feedback on Grammar across Upper-Division Business Students Theresa Wilson,* William Provaznik, and Nancy Pigeon Central Washington University Employers have expressed dissatisfaction with business students' basic writ- ing skills, so techniques for improving students' grammar skills should be critically examined. This study investigated the efcacy of using written feedback in a multidraft context as a method of decreasing grammar errors in subsequent submissions. Business students in a principles of marketing class were given the option to receive feedback on drafts. Written feedback on grammar issues was successful in reducing grammar error rates on nal submissions only for highly motivated students with multiple drafts. A dis- cussion of the faculty time commitment necessary to see improvement in students' grammar skills recommends reection on this technique. Keywords: Business Commtmication, Business Writing, Business Education Disciplines of Interest: Business Education, Business Communication, Marketing, Management INTRODUCTION The value of good written communication skills for business students is clear based on feedback from faculty and employers [Enos, 2010, Kellogg and Whiteford, 2009; Parent et al., 2011]. Most college freshmen have condence in their writing skills [Berrett, 2014]. National data, however, suggest that this condence is misplaced. Highschool graduates often lack the necessary basic writing skills to succeed at the college level [Achieve, 2015]. Historically, colleges have required students to gain and demonstrate competency in English grammar through standard freshman composition classes. Even so, employers are concerned about the writing skills of college graduates. Kleckner and Marshall [2014] found that employers rated basic writing mechanics as second in impor tance among communication skills for business college graduates, yet found that the employers' satisfaction level for this skill among business college graduates was the lowest among all communication skills. 180 Journal of the Academy of Business Education In order to correct mistakes in basic writing mechanics, the mistake has to be apparent to the user. Improvements in grammar and spell checking provide visual cues to aid in editing at the sentence level. Still, software programs like Microsoft Word are not proofreaders and fail to catch many grammatical errors. An example of this is in the use of possessives. Such gaps leave users of Microsoft Word, for example, vulnerable to errors such as misplaced apostrophes that change the fundamental meaning of a sentence. College business graduates carmot depend on computerized grammar checking to catch all grammatical errors. They must learn to do this for themselves; thus, editing skills are important. Written feedback is one of the most used methods for improving basic writing skills, yet the time commitment that it takes to give feedback is a problem [Bacon and Anderson, 2004; Bacon et al., 2008, Kellogg and Whiteford, 2009]. More over, universities are being asked to hold down educational costs, which has led to fewer resources for providing the individual feedback that is necessary to improve student grammar skills [Crisp, 2007']. The same pressure on universities to hold down costs also means that simply adding additional communication coursework requirements to the business curriculum may not be an option for xing this problem, and, for students, an increase in required coursework means additional time to graduation. Recent movements toward writing across curricula and within disciplines, however, have shown promise for improving writing skills [Fallahi et al., 2006]. Reframing the view of writing as the responsibility of the whole university rather than the English department pushes students to learn to write across contexts [Bacon et al., 2008]. As most business students are already doing written assignments in their upperdivision coursework, can feedback about writing mechanics in this coursework effectively increase business students' skill levels in grammar? That is the subject of this study. LITERATURE REVIEW Silverman et al. [2005] identied ve individual precursors to accepting feed back: awareness, sense of necessity, confronting change, willingness for feedback, and development orientation. Awareness involves knowing that a problem exists and making the appropriate attributions of its cause. In somewhat of a vicious cycle, lack of awareness leads to a lower level of competence. It also leads to a lower ability to detect problems and subsequently a lower awareness of the value of feedback. Because of this, Silverman et al. [2005] regarded awareness as the most signicant of the ve individual precursors, which is consistent with earlier ndings [Kruger and Dunning, 1999]. The sense of necessity involves the knowl edge that a change must occur. This relies on an ability to undergo the unpleasant emotions involved with change. Confronting change similarly involves potentially threatening phenomena that require determination to scrutinize problems in order to resolve them. Willingness for feedback involves an ability to be vuhierable and Spring 2018 I81 the courage to change, even if it involves uncomfortable emotions and cognitions. The nal factor is development orientation, which is a positive outlook on growth. The literature on improving students' writing echoes the importance of using feedback to establish awareness of grammatical errors to stimulate the learning process. Kellogg and Whiteford [2009] argue that feedback is an essential part of the necessary deliberate practice needed to improve student writing. This practice requires a substantial time commitment by the learner. Jorgensen and Marek [2013, p. 174] state that \"students do not become excellent writers overnight.\" Dealing with the potential emotional distress of change and growth takes time. In addition, business college graduates need to be able to perform on the job, so they need to have the rituals of good writing mechanics in their longterm memory. Longterm student retention of correct grammar and sentence structure is much more likely when practice is done repeatedly over time, especially because serious writing puts substantial strain on cognitive abilities [Kellogg and Whiteford, 2009]. Students need to mindilly apply themselves to a practice with feedback from an instructor to improve and excel. While convention holds that writing be taught in an English class, multiple empirical studies have shown that the integration of writing feedback into courses with other content can be effective (e.g., marketing [Bacon et al., 2008] and psychology [Jorgensen and Marek, 2013, Stellmack et al., 2012]). Kellogg and Whiteford [2009] argue that this method is actually preferable, as it distributes the practice of writing and thus encourages longterm retention of writing mechanics. Without this practice, skills begin to deteriorate. By the time college students graduate, they may no longer possess the skill level they did at the end of freshman composition classes [Parent et al., 2011]. Past research on the use of written feedback for improving the basic writing mechanics of college students shows mixed results [Ferris, 1995; Price et al., 2010]. A closer examination of the results suggests that there are some factors that may increase subsequent student writing improvement after receiving written feedback. These include the timing of the feedback [Fallahi et al., 2006; Ferris, 1995; Jorgensen and Marek, 2013], the motivation of students [Bacon and Anderson, 2004, Bacon et al., 2008], the specicity of the feedback [Quible, 2006a; Shintani et al., 2013], and the student's condence in his or her ability to respond to the feedback [Crisp, 2007'; Shintani et al., 2013]. While some have found that feedback on nal submissions alone provides modest gains in writing mechanics when followed systematically over time (e. g., Fallahi et al., 2006, Jorgensen and Marek, 2013), others have seen no signicant improvement, largely because students always have a choice about whether they are going to use the feedback or not [Crisp, 2007'; Price et al., 2010]. Ferris [1995] reported that students spent less time reading and focusing on feedback when it came on nal submissions than they did when a system of writing and revising multiple drafts was in place. Students reported rereading feedback many times in the latter case. Students may be more likely to be motivated in a system of 182 Journal of the Academy of Business Education reviewing, revising, and resubmitting, especially when it relates to writing mis takes that are relatively easy to x, such as errors in writing mechanics [Vardi, 2012]. They are able to see the immediate impact of responding to the feedback [Stelhnack et al., 2012]. Finally, Shintani et al. [2012] note that revision aer feedback may enable students to consolidate their knowledge into action, which reinforces the learning. All of this implies that a review, revise, and resubmit draing system is likely to increase the positive effect of written feedback. The incorporation of draing is not necessarily sufcient by itself to improve basic writing skills in courses not solely focused on writing. Bacon and Anderson [2004] found that business students needed a grade incentive of 5% to show signicant improvement in basic writing skills when mixed with evaluation of other aspects of the assignment. They argue that students need to be incentivized to perform on grammar, or they may not pay sufficient attention to that part of the task. In their study, even with the opportunity to revise and resubmit over multiple assignments, students without a signicant grade incentive to pay attention to writing mechanics showed no signicant improvement. Improvement in basic writing skills in courses not specically designed for that task means that feedback must point out student errors in basic writing skills while still integrating it into the disciplinespecic content [Bacon et al., 2008; Kellogg and Whiteford, 2008]. It is important that students receive specic feedback on the basic writing errors they are making and the grammar issues involved, even if faculty perceive them to be minor in comparison with content issues. Bacon and Anderson [2004, p. 443] state, \"Without feedback on minor errors, students may not feel motivated to improve their writing skills.\" Empirical evidence suggests that specic written feedback that gives an understanding of the grammatical error made decreases the error rate in subsequent work [Quible, 2006b; Shintani et al., 2012, Vardi, 2012], while holistic types of feedback are ineffective [Kellogg and Whiteford, 2009]. More specic feedback gives quality information that adds clarity to the writing standard. In the case of upperdivision business students, most have been through the freshman composition English classes that require attention to grammar issues. Professors may assume that college students know standard conventions for grammar and sentence structure by the time that they are juniors and seniors. At this point, students should not need the kind of support they might have needed when rst learning how to construct sentences. Instead, what they need is to develop editing skills to nd and correct errors. Feedback about certain errors, while still allowing students to nd and x those errors themselves, gives guidance toward meeting a standard without operating as the students' proofreader. Even if incentivized and given specic feedback, in order for students to improve they need to have knowledge of grammar and condence in their writing abilities. While passing freshman composition may demonstrate that they have sufficient knowledge of the basics of English grammar, student selfcondence (selfefcacy) in writing is not a given. Bandura's [1986] social cognitive theory Spring 2018 I83 suggests that a lack of condence in one's ability to succeed may have an impact on his or her ability to improve. In social cognitive theory, three elements interact to facilitate learning: selfefcacy, feedback, and environmental support. Students who have selfefcacy toward a behavior, get appropriate feedback when dem onstrating behavior, and receive environmental support are more likely to learn. Students may avoid incorporating feedback into a subsequent assignment if they feel that they lack the ability to respond to the feedback or if they do not have sufcient environmental support. Thus, it is important that even students who feel selfefcacy with regard to grammar issues have appropriate support materials to aid them in editing for grammar mistakes if they are to respond to written feedback on subsequent submissions. Price et al. [2011] suggests that access to a writing center with writing tutors is one way to make business students feel more supported as they adjust to the more difcult task of writing in upperdivision business classes. Other ways to provide support are to give students inclass review and access to outofclass materials that explain grammar rules. In sum mary, previous research supports the following hypothesis. In summary, previous research supports the following hypothesis. H]: Where students have condence in their grammar skills, are grade motivated, and have access to additional environmental support in basic writing mechanics, specic written feedback on grammar errors will decrease the error rate from drafts to a nal submission. METHODOLOGY Pretreatment Assessment of Student Condence, Attitudes, Behaviors, and Writing Experience In order to assess the previous writing experience, attitudes, and writing behaviors of the subject population, an anonymous survey was administered at the beginning of three separate principles of marketing courses conducted over two quarters at a rural, western university. A total of 71 usable surveys were returned out of a total population of 96 for a response rate of 74%. In the survey, students were asked to respond to statements on their attitudes about their writing skills on a Likert scale (1 : strongly agree, 2 : agree, 3 : neither agree nor disagree, 4 : disagree, and 5 : strongly disagree) and to provide demographic data, as well as data about the amount of writing that they had done in previous college courses. There were no signicant demographic differences found among the classes. All students were business majors and college juniors or seniors. Eightyve percent were ages 20 to 25, with the rest being older than 25. Fiyeight percent were male and 42 percent were female. A clear majority of students (85 percent) transferred to the university as juniors. All of the results in Table l were consistent across major and demographics. 184 Journal of the Academy of Business Education Table 1. Student Attitudes and Behaviors about Writing 1. I am condent in my ability to write -\" sentences with correct rammar. 3. I feel condent in my ability to defend my 0th of view in writin_. 4. A difcult part of writing sentences is knowin_ the ri-ht words to use. 5.1 feel condent in my ability to use section 6.1 wish T had been forced to do more writing revious ears of education. 7.1 temnd to write m nuaers in one sitting. 8. The hardest part of writing is getting 2.27 1.13 started. 9. If a student gets a bad grade on a paper in 3.48 1.24 college, most instructors will allow the student to rewrite the paper for a better grade. *(1 : strongly agree, 3 : neither agree nor disagree, 5 : strongly disagree). Generally, students felt neutral to condent in their writing skills. They expressed the highest condence in their grammar skills, as statement 1 in Table 1 had the strongest support and the lowest standard deviation. Seventytwo percent of students at least agreed with statement 1, and only 9 percent disagreed with that statement at all. Students also tended to agree that they wrote their papers in one sitting and they had a hard time getting started writing. Fiythree and sixtyve percent of students, respectively, at least agreed with those state ments. The majority of students were not counting on being able to rewrite papers for a better grade. Fiftyve percent at least disagreed with statement 9. Statements 4 and 8 were weakly positively correlated (r : 0.254, pvalue : 0.035). There are no other signicant correlations between writing attitudes in statements 1 through 6 and statements 7, 8, or 9. The weak, positive correlation between the difculty in starting writing and the tendency to write papers in one sitting may indicate that the former is a possible contributing factor to the latter as students are forced to write papers in one sitting as a deadline approaches. The results also suggest that students' condence in their ability to write with good grammar was high. The students felt more positively about this aspect of their writing than any other. This demonstrates that the population was likely to have high selfefcacy when correcting basic errors in writing mechanics. In addition, students were not counting on getting an opportunity to rewrite a paper Spring 2018 I85 for a better grade. This implies that a nal submission grade for which writing was at least 5 percent of the grade would likely sufce as a strong enough incentive for this population based on Bacon and Anderson's [2004] research. The mean number of individual papers that students had written in previous college courses that were at least ve pages in length was 5.6 (standard devia tion : 4.040). Onethird of the respondents had individually written less than three papers of at least ve pages in length in college. Fieen percent reported that they had written individually no papers of at least ve pages in length. On the other hand, onethird of students reported that they had written at least 10 papers of that length individually in college. The mean number of classes in which students had the option of submitting dras for instructor feedback was 2.46. Fifteen students (21 percent) reported never having had any class in college in which they had the ability to submit drafts for instructor feedback. The mean number of classes in which students were required to turn in paper dras for instructor feedback was 1.25. All students in this sample were required to take and pass at least two courses in freshmen composition. The mean number of classes in which students were required to turn in a dra was less than 2, and the mean number of classes in which they had the option of submitting dras was slightly more than 2, which suggests that a majority of students had little exposure to the use of dras in their academic writing. This lack of experience, combined with the fact that most students tended to write papers in one sitting, suggests that students were not following an informal draing system by themselves. This supports the idea that a formal draing system would provide a structure for revision that they are not providing for themselves. In summary, these results demonstrate that this population of students has selfefcacy in grammar, would be motivated when writing is at least 5 percent of a written assignment's grade, and has had little exposure to the process of draing in their academic experience. An examination of the effects of draing on the rate of student errors in subsequent assignments follows. Drafting and Student Writing Error Rates The 96 students previously surveyed were involved in a quasiexperimental design to examine the effect of drafting on basic student writing errors. Each student was required to complete two applied writing assignments based on major concepts in marketing. Each assignment was limited to two pages in length and was worth 15 percent of their nal course grade. Writing mechanics were evaluated in each assignment and valued at 20 percent of the assignment's grade. There is a long list of grammar mistakes that can be made in English. Connors and Lunsford [1988] found 54 general categories of errors in their examination of 300 student papers. In order to focus on those that would be most benecial to students in their careers, ve general categories of errors that are usually not captured by computerized spelling and grammar checks were selected. All of 186 Journal of the Academy of Business Education these categories were among the top nine listed by Connors and Lunsford [1988]. The categories of errors were as follows: 1. Homophone errorsiErrors of this type happen when students use an incorrect word that sounds the same but is spelled differently. Examples of errors of this type are incorrect usages of \"to,\" \"too,\" and \"two\" or \"schools\" versus \"school's\". 2. Misspelled word errors that correctly spell another wordAn example of this is when a student uses the word \"roll\" for the word \"role\" incorrectly. 3. Sentence fragment errorsErrors of this type occur when sentences do not reect a complete thought. 4. Comma splice errorsiThis happens when two independent clauses are combined together with only a comma. 5. Comma with coordinating conjunction errorErrors of this type occur when long independent clauses are combined with a simple conjunction and without a comma. Students were allowed to submit up to two dras per written assignment. Students were required to submit their dras at least 48 hours before the assign ment was due and received feedback on their dras within 24 hours. Both the dras and the feedback were given through a learning management system. The course instructor gave all the feedback on content, and a professional editor gave all the feedback on writing errors. At the end of each submitted dra, students were told how many errors of each of the above types existed in their dra, but they were not told where in the draft those errors were made. The choice was made to use this type of feedback in order to more closely replicate the editing process that the students would need to undertake on the job. In addition to feedback on these errors, students received written feedback on content. Students were given a 15minute review of the types of grammar errors that they would be required to nd and x for themselves in class a week before the rst writing assignment was due. They also had access to an online grammar resource and a free oncampus writing center with writing tutors for additional help. RESULTS First Writing Assignment For the rst assignment, all students had the option of submitting up to two dras to the instructor for feedback on both writing skills and content. Seventy two (75 percent) of 96 students submitted at least one dra, while 18 (19 percent) submitted two drafts. It took the professional editor an average of 6.25 minutes to give feedback about the basic writing error categories on each draft. The dras averaged 565 words in length. The number of errors in each of the ve categories mentioned above was recorded and totaled for all drafts and the nal submission. Spring 2018 I87 Table 2. Basic Writing Errors in Examined Categories in First Writing Assignment First drafts 0.45 0.53 Second drafts 0.42 Mean Writing Errors Standard N ter 100 Words Deviation IE 0.39 Students who did no drafts 0.39 Students who did onl one draft 0.40 Students who did two drafts 0.33 Table 2 Shows that there was a decrease in the average number of errors per 100 words between the rst and second drafts. The mean number of writing errors per 100 words in the nal submission decreased as the number of drafts increased. A Levene's test of equality of error variance and the sample Size indicated that parametric testing was acceptable. A oneway ANOVA showed that there was no Signicant effect of drafting in the nal submission (F(2, 94) : 1.08, pvalue : 0.344). An examination of the difference between the submitted draS and the nal versions provided irther insight. Of the students who had writing errors noted in their dras that they had to x on their own (N : 58), 31 percent did not x any of these errors before their nal submission, as Shown in Table 3. A higher percentage of students who did two dras xed all their noted errors by the nal submission. A Spearman's rankorder correlation was run to examine the relationship between the different levels of error correction shown in Table 3 and a number of other variables. The results are Shown in Table 4. A strong negative correlation was found between the different levels of error correction and the number of errors per word (rs(56) : *0.646, pvalue : 0.000). No other Signicant correlations were found. Second Writing Assignment Due to the high percentage of students submitting draS in the rst assign ment and the limited resources available, only those students who had not received an \"A\" on their rst assignment were eligible to submit dras for the second assignment. Other than this change, all other procedures remained the same. Of the 65 students eligible to submit dras, 34 (52 percent) chose to submit at least one draft, and 15 (23 percent) chose to submit two drafts. It took the professional editor an average of 9.70 minutes to give feedback on each draft. The dras averaged 573 words in length. Table 5 summarizes the data from the second writing assignment. 188 Journal of the Academy of Business Education \fTable 6. Student Basic Writing Errors Fixed between Drafts and Final Submission in Second Writing Assignment Percenta_e of Students Overall One Draft Two Drafts A11 errors 100% noted in drafts xed 52% 43% 66% Most errors 50-99% noted in drafts xed 19% 21% 17% Some errors 1-49% noted in drafts xed 6% 10% 0% No errors (0%) noted in drafts xed 23% 26% 17% among the groups eligible for drafting (F(2, 63) : 5.976, pvalue : 0.004). A Tukey posthoc test revealed that the mean writing error rate for doing two dras was statistically signicantly lower than doing no drafts (pvalue : 0.005) or doing one draft (pvalue : 0.011). There was no statistically signicant result between the groups doing no drafts and those doing one draft (pvalue : 1.000). This result provides limited support for H1. Compared to the rst writing assignment, a higher percentage of students in all categories xed the basic writing errors noted in their drafts (N : 30), as shown in Table 6. In this case, only 23 percent of students overall did not x any of the errors noted in their drafts. Table 7 shows the Spearman rankorder correlations between the different levels of error correction shown in Table 6 and a number of other variables. As in the rst writing assignment, a strong negative correlation was found between the different levels of error correction and the number of errors per word in the second assignment (rs(28) : 0.712, pvalue : 0.000). A weak positive corre lation occurred between the number of drafts done in the second assignment and the level of error correction in the second assignment (r5128) : 0.384, pvalue : 0.036). A moderate positive correlation was found between the number of dras done in the rst assignment and the level of error correction from drafts to nal submission in the second assignment (rs(28) : 0.491, pvalue : 0.006). DISCUSSION The quasiexperimental design of this research should have predisposed the results to support the hypothesis that written feedback on specic basic writing errors in drafts would lead to a lower rate of those errors in the nal version. The vast majority of students were condent in their ability to write with good grammar, so they should have felt capable to x the errors. While students were not told exactly where their writing errors were, they were given plenty of free, easily accessible resources to get help to nd and x their errors, and as the assignment was limited to two pages, it was not an overwhelming task. They were motivated to pay attention to these errors by making writing worth 20 percent of 190 Journal of the Academy of Business Education Table 7. Spearman Rank-Order Correlations of Table 6 Levels of Error Correction Correlation with Table 6 Cate_ories Number of words in second assinment 0.218 Writin errors ner word in second assi_nment 70675" Number of drafts done in second assi_nment 0.384M 0096 Final rade in class 0.914 Writin errors er word in rst assi_nment 0.085 Number of drafts done in rst assignment 0491* *Correlation is signicant at the 0.01 level (two-tailed). MCorrelation is signicant at the 0.05 level (two-tailed). the grade in the assignment. This was a much higher percentage than previous research had found to be efcacious. Students were given the option of submitting a dra but were not required to do so. This selection option should have biased the results in favor of supporting H1. Presumably, students who went to the trouble of composing a dra were already motivated to want to get feedback and incorporate it into their nal submission. Indeed, if the results had shown that statistically signicant improvement in the noted writing errors was made be tween dras and the nal version, it would not have been surprising. Instead, the results showed that only students on the second writing assignment who submitted two dras showed signicant improvement in error rates. Consistent with Crisp [2007], H1 was shown to have limited support. In both assignments, a substantial number of students who did only one draft failed to x any of the noted writing errors. Given that the students' condence in their grammar skills was high to begin with, they may have been lulled into the expectation that they had good writing skills and neglected to pay attention to information that contradicted this belief in the rst writing assignment. Other students may not have paid attention to the high value placed on good writing mechanics in the evaluation of the rst assignment. However, this does not completely explain the results of the second writing assignment. A necessary, but not sufcient, condition for getting an \"A\" on the rst assignment was to have good basic writing skills. Students who did not meet this standard were the only ones eligible to submit dras for the second written assignment. Thus, these students should have been even more strongly motivated to improve their writing on the second assignment since they had already received feedback that writing was important in the grading of their nal submission. Substantially more errors were xed in all categories in Table 4. Yet, even in the second writing assignment, 26 percent of students who did one draft did not x any noted errors at all. Spring 2018 I91 A possible explanation for this effect is that even though most students expressed condence in their writing abilities, the difculty in responding to both content and basic writing issues in the drafting process may have caused cognitive overload. Kellogg and Whiteford [2009] explain that \"revision is constrained or even nonexistent in developing writers because of working memory.\" There was a strong negative correlation between the number of errors and the levels of errors xed on both assignments. The more errors that the students had to x, the less likely they were to get all the errors xed. If students were overwhelmed with responding to the content issues in their drafts, then they may not have had the cognitive capacity to address basic writing errors as well. This effect would be even more pronounced for students whose belief in their writing abilities did not match their actual abilities. If students struggle with sentence generation to begin with and then have to apply disciplinerelated content on top of that, they are more likely to be incapable of responding to formative feedback, however specic it may be. Writing, like athletics and music, takes practice, and more complex tasks require a mastery of the basics. There is a danger of overcorrecting when giving feedback in writing [Shintani et al., 2013], but in this case students were given feedback on only ve different types of basic writing mistakes. The students had all been exposed to these grammar issues in previous coursework. It is unlikely that overcorrecting was the problem here. The survey data on previous writing experience seem to suggest a different source of possible overload. As juniors and seniors in college, onethird of these students had written less than three papers of at least ve pages in length. Fieen percent reported writing no papers of this length individually. Quible and Griffen [2007] also note that, in the past 25 years, English teachers have increasingly stopped providing sentencelevel correction and grammar instruction. This implies that a large percentage of these students have had little practice in transforming knowledge into applied compositions to begin with and have not been as exposed to sentencelevel error correction as in generations past. Asking students to use feedback to improve grammar in addition to responding to content issues may have been more than some students were capable of handling given their past educational experience. Another source for a possible explanation of the variance of our results comes from studies on millennials and management. Our study concentrates on largely millennialaged cohorts (i.e., students born between 1980 and 2000). Millennials are described as having a strong need for feedback on their performance [Meister and Willyerd, 2010], but Alexander and Sysko [2012] hold that an environment of abundance and shis in parenting styles have led this generation to have expectations of a future with more abundance, even when missing performance expectations. Compared to other generations, they tend to be optimistic and condent in themselves and their abilities [Blaine, 2008]. This generational tendency is a reection of the positive psychology movement that emerged in the 1990s. In this movement, psychologists advocated an attributional style that explains positive events as a reection of personal and permanent causes, while 192 Journal of the Academy of Business Education negative events should be seen as external and temporary [Seligman, 1991]. Thus, a recurring theme in the literature of the millennial generation is the use of fundamental attributional error (FAB). FAE is the tendency for individuals to attribute their own success to their own intemal characteristics and their failures to external phenomena, while conversely seeing others' successes as conse quences of external phenomena and their failures due to characteristics of the actor [Ross, 1977]. This bias makes it more difcult for some millennial students to respond to negative feedback. Like educators, managers have long observed variations in an individual's willingness and ability to accept feedback on his or her performance. Silverman et al. [2005]'s research suggested that awareness was the most important precur sor to utilizing feedback. Awareness means both knowing the existence of a problem and being able to accurately attribute the cause of the problem. An inability to attribute the cause of the problem correctly will lower a person's ability to use feedback. Millennial students who are given a higher level of criticism than they were expecting will be faced with feedback that is inconsistent to their sense of self. Not only are they less able to appraise the magnitude of their deviation from the expected standard, the inconsistency is also contrary to their overestimated sense of efficacy. Unwilling to endure the potential cognitive difculty and preferring to avoid the emotional labor of investigating the depth of their problems, they will resort to externalizing the errors and undervalue the use of feedback to avoid iture mistakes. Millennial students who have incorporated the optimism of fundamental attribution error into their psychological makeup will thus demonstrate a stronger tendency to ignore the negative feedback that they are given on a draft, as they see it as an external, shortterm problem that they do not have to address. On the other hand, students who receive feedback consistent with their sense of self will incorporate this information into better performance. They will have the insight to respond to the feedback in their iture work and will be able to absorb the emotional and cognitive distress involved in making changes, given that the changes needed are relatively modest. The strong negative correlation between the error rate per word and the level of error correction in both assignments provides support for this explanation. A nal explanation for the variation in response to written feedback on grammar is that repeated exposure over time to a drafting system was necessary to increase the level of error correction. Although the number of drafts in the rst writing assignment was not correlated with the levels of error correction in the rst assignment, by the second writing assignment there was a signicant, albeit weak, positive correlation between the number of drafts and the level of error correction. There was a stronger positive association between the number of drafts done in the rst writing assignment and the level of error correction in the second writing assignment. These results tend to support the notion that students were learning to respond to the feedback and correct their grammar mistakes as they had more exposure to a drafting system. The study suggests that more than one Spring 2018 I93 exposure to draing is necessary to see signicant effects of written feedback on grammar. If millennial generational tendencies are pervasive in a student body, then it may be that repeated exposure to negative feedback on grammar helps to break through FAE. Crisp [2007, p. 572] notes that \". . . there is an implicit assumption that the provision of feedback will necessarily lead to improvements in subsequent pieces of submitted work.\" Faculty should give written feedback on grammar with the knowledge that it will not necessarily result in lower error rates in subsequent assignments unless they are prepared for a substantial time commitment. In this study, providing feedback on just ve basic writing errors over two dras took a professional editor approximately two minutes per 100 written words. This is roughly in line with the estimation of 10 minutes per 100 words that Connors and Lunsford [1988] gave when they had a group of college English teachers count 20 different errors over 3,000 papers. Thus, to give even a limited amount of writing feedback for a onepage, doublespaced written assignment of approximately 250 words would require at least ve minutes. For a class size of 25 students, that is more than two hours of work for each round of drafts. This research suggests that even when students are highly motivated, it takes two rounds of dras over two assignments before the writing error rate drops signicantly on the nal submis sion. That is eight hours of grading for a onepage assignment before faculty even get to grade a nal submission. The eighthour estimate is a conservative one because most upperdivision faculty members are not professional editors. A faculty member without an editing background would be expected to spend even more time on the task. It may be more reasonable to spread out these efforts over a number of courses and instructors to distribute this burden more evenly. RECOMMENDATIONS Recommendations for Future Research This research was conducted in one upperdivision principles of marketing class. It would be benecial to see if these same results were found over multiple functional areas of business. Also, only ve different types of basic writing errors were examined in this study. Connors and Lunsford [1988] identify 54 types of basic writing errors in their analysis. It may be that the results in this study are a function of the errors studied. Future research that includes a broader range of basic writing errors may yield different results. Another limitation of this study was that it did not test the students' prociency in correcting basic writing errors prior to the treatment. Linking prociency to nal results would determine if the limited success of written feedback in draing is because students lack the ability to x the writing errors noted in a draft. While this study found limited support for H1, it is only a pilot study. The number of students in the study restricted the analysis that could be done. It may be that a bigger sample with a coordinated plan of drafting and written feedback over a series of upperdivision classes would 194 Journal of the Academy of Business Education yield a better result. Related research into the use of peer review [Stelhnack et al., 2012] and automated editing practices [Bacon et al., 2008] in conjunction with draing may yield useil ways to cut down on the number of hours of faculty time required to give written feedback. Additional research in these areas would be helpful. Recommendations for Practice In the right circumstances, a multidraft system was found to be effective at reducing grammar error rates in the nal submission. It is apparent that some students respond well to written feedback when it comes to grammar issues. This study seems to suggest that this group becomes a larger percentage of students when they are motivated and submit an increasing number of drafts. There were always some students who completely avoided responding to the written feedback on grammar. Despite grade motivation and resource availability, these students did not appear to act on this feedback at all. The research suggests that increasing students' response to written feedback on grammar requires the commitment of faculty to give specic, regular, and formative feedback over time. This is a daunting task for many college faculty. Virtually all of the research on written feedback cites the time required to provide it and the resource constraints faced by college faculty. Kellogg and Whiteford [2009, p. 260] state, \"Although there are probably many reasons why more writing is not routinely assigned, the time and effort required by instructors to provide useful feedback surely ranks high on the list.\" In a survey of business professors, the amount of time spent on grading was ranked rst among draw backs to having increased written assignments [Parent et al., 2011]. In addition, \"Improving undergraduate writing skills receives relatively meager rewards com pared with faculty publication, mentoring of graduate students, and sponsored research\" [Kellogg and Whiteford, p. 261]. Finally, there is always the consider ation that faculty members are subject to the students' evaluation of their instruc tion. Students are not likely to complain about an instructor who is not picky enough about grammar. Given the low level of rewards and the enormous investment of time, it is not surprising that individual faculty members reduce the length and number of written assignments, much less go through the increased hassle of reviewing dras. Administrators in higher education should consider changing the reward system if it wants its faculty members to commit to improving student writing and give this level of written feedback. REFERENCES Achieve. (2015). \"2014 Annual Report on the Alignment of State K12 Policies and Practice with the Demands of College and Careers,\" Retrieved from http:f/www.achieve.orgfles/AchieveClosingEXpectGap2014%20Feb5. pdf. Spring 2018 I95