Investigating gender variations of using attitude choices in letters of recommendations

Abstract

Purpose

The main objective of the present study is to explore whether there are variations in the employment of evaluative language resources by male and female writers. More specifically, the study focuses on variations, if any, that can be attributed to difference in gender.

Design/methodology/approach

The study compared and contrasted forty recommendation letters written by male academics to the same number of letters written by female recommenders. The study uses both quantitative and qualitative approaches.

Findings

The investigation of three attitudinal resources in letters of recommendations showed that the most employed resource was the judgment sub-system. The appreciation domain was in the second position, and the least frequent was the affect. The results also revealed no statistically significant variations in attitude sub-systems: Affect and appreciation as the writers in both groups (males and females) employed almost the same options in each. In respect with judgment, however, the analysis explored significant differences between the two sets as male academics used more judgment resources than females.

Originality/value

The main contributions of this study may be as follows: first, it is one of very few studies drawing on the attitude-category of appraisal system, as an analytical tool to examine gender differences in recommendation letters very particularly on the ones written by non-native speakers of English. Second, the gender factor is central in the genre of the recommendation letters and hence researchers should be cognizant of its role as certain variations might be impacted by it. Third, the lists of tokens can be offered as heuristics for academics to have most common words or phrases to use in their letters. Finally, the findings can hopefully bear some important pedagogical implications, very specifically for novice and non-native academic writers of recommendations letters.

Citation

Alotaibi, H. (2023), "Investigating gender variations of using attitude choices in letters of recommendations", Saudi Journal of Language Studies, Vol. 3 No. 2, pp. 87-101. https://doi.org/10.1108/SJLS-09-2022-0070

1. Introduction

It has been proved that writing encompasses dialogic language between text participants. In other words, writing and language in general, is not used to transmit facts, but rather users of language perform actions through it. Paltridge (2012, p. 40) pointed out that works in the areas of pragmatics and discourse analysis have changed the traditional view that perceived language as “is always used to describe some fact of state of affairs, and unless a statement can be tested for truth or falsity it is basically meaningless”. Writers of certain text types express explicitly/implicitly their attitudes, emotions and feelings.

The importance of stance-taking as an interpersonal aspect in academic writing has been acknowledged by some scholars and researchers. Hyland (2005), for instance, suggests three factors that enable writers produce a credible work: claiming solidarity with readers, evaluation and accepting others’ viewpoints. In fact, the notion of evaluation and authorial stance as an interpersonal dimension of writing has been the interest of many authors, scholars and researchers (particularly in systemic functional linguistics (SFL) and socio-linguistics). They use different terms to define the nature of inter-personal meanings and language resources pertinent to them. Thompson and Hunston (2000), for example, uses the term evaluation, which they define it as “the expression of the speaker or writer’s attitude or stance towards, view point on, or feelings about the entities or propositions that he or she is talking about” (p. 5). Martin and White (2005) propose Appraisal which is closely related to the concept of evaluation proposed by Thompson and Hunston. Appraisal is developed in SFL and mainly concerns the speaker’s attitudinal positions. For Martin and White, appraisal is one of the semantic resources that interpret interpersonal meanings. More accounts for appraisal system are provided in section 1.3 (The analytical framework).

The main interest shared by all these scholars seems to be the exploration of suitable frameworks and systems via which the language features of inter-personal meanings can be codified and analyzed. In the present study, the appraisal system proposed by Martin and White (2005) is adopted as an analytical framework, since it gives the most comprehensive codifications and categorizations of the linguistic resources involved in evaluation and authorial stance projection.

Due to the growing interest in interpersonal aspects in academic writing, many research studies were conducted to examine the employment of linguistic resources involved in taking stance. Most of the researchers who adopted the appraisal concept as an analytical framework focused on academic research paper (Chang and Schleppegrell, 2011; Cheng and Unsworth, 2016; Hood, 2004; Loi et al., 2016). However, recommendation letter genres received less attention by researcher compared to other academic texts.

Therefore, the main objective of the present study is to fill this gap by examining the general trends of writers’ employment of the linguistics resources pertinent to authorial stance projection in recommendation letters. To this end, this study drew on the appraisal system proposed by Martin and White (2005) to be the analytical framework via which authorial stance features were analyzed.

The genre of recommendation letters involves evaluative and persuasive expressions from academics to recommend their students for joining academic programs and/or applying for job positions. The letters of recommendation generally involve six rhetorical patterns or moves. As pointed out by Maskara et al. (2014), the generic structure of letters of recommendation consists of Move 1: Purpose of writing, Move 2: Context of knowing the applicant, Move 3: Applicant credentials, Move 4: Applicant personal values, Move 5: Applicant social competency and Move 6: Closing remarks.

Usually academics, through these moves, express their attitudes, reactions and feelings, and these expressions can be influenced by certain factors such as gender, culture, discipline, etc. One of the most useful analytical frameworks to investigate evaluative language is the appraisal system suggested by Martin and White (2005). The model encompasses three sub-systems: attitude, engagement and graduation. However, the present study focuses on the attitude sub-system from the appraisal framework to investigate possible gender variations in the genre of recommendation letters.

1.1 The letters of recommendation

Unlike other academic genres, academics do not receive specific training to write letters of recommendations which consequently leave possible cultural and gender nuances to be investigated (Precht, 1998). The genre of letters of recommendation most noticeably derives its importance from the aspects that academics tend to discuss and explore. As shown earlier in the study by Maskara et al. (2014), there are certain rhetorical structures and patterns that are central in the construction of a letter of recommendation. Recommenders use evaluative resources and express evaluative meanings to comment on their students’ qualifications and to recommend them for and/or graduate studies.

Previous studies on letters of recommendations investigated the aspect of attitude, and specifically they examined the reader engagement (Yang, 2021), the disability disclosure (Vidali, 2009), the disclosure of negative information (Grote et al., 2001), the influence of a relationship between recommenders and applicants (Colarelli et al., 2002). Other studies (Bouton, 1995; Bruland, 2009; Ebadi and Dovaise, 2015; Liu, 2013; Maskara et al., 2014; Qanbari et al., 2014; Precht, 1998) explored the cultural aspects in letters of recommendations. Maskara et al. (2014) in particular used the appraisal framework to contrast the attitudinal resources in British and Indian letters. They found that while appreciation attitudinal resources were significantly less often in the Indian group, affect attitudinal resources were more frequent in the Indian letters compared to the British letters. Concerning judgment attitudinal resources, they were similar in both cultural groups.

Gender variations were also investigated in letters of recommendation. Bell et al. (1992) investigated fifteen pairs of letters written for women and twenty-four pairs written for men. Each pair of letters included one written by a female and one written by a male recommender for the same candidate. The findings showed that there were gender differences in terms of the recommender as well as the applicant. For example, female writers were more likely to discuss administrative skills whereas male recommenders were more likely to discuss intellect, career potential and personality. Similarly, Trix and Psenka (2003) analyzed 300 letters of recommendation for medical faculty at a large American medical school in the mid-1990s. The length of letters was the most salient difference, as the letters written for female applicants were very short compared to those for males. In addition, the authors found that the letters for female applicants included more cases of doubt raisers. The authors defined doubt raisers as those that include negative language, hedges, negative and unexplained comments. Furthermore, they found that most common semantic categories of objects of possessive phrases for female applicants included “her training,” “her teaching,” and “her application,” whereas those for males were “his research,” “his skills and abilities,” and “his career.” (Trix and Psenka, 2003, p. 211). The authors commented on the selection of these references as “the women are portrayed more as students and teachers, whereas the men are portrayed more as researchers and professionals.”

Another study on gender was conducted by Schmader et al. (2007) who investigated recommendation letters for male and female applicants for faculty positions in chemistry and biochemistry departments. Some of gender variations were the tendency to include more phrases pertinent to communication skills in letters for female applicants. In addition, letters for male applicants tended to include more mentions of standout adjectives (e.g. fabulous, magnificent, remarkable, amazing). In contrast, there were no gender differences in the number of grindstone traits (e.g. hard-working, careful), ability traits (e.g. talented, brilliant), research terms (manuscript, study) or teaching terms (advisor, syllabus). Additionally, Madera et al. (2009) examined variations in agentic and communal characteristics in letters of recommendation for male and female applicants and whether such variations influenced selection decisions in academia. They found that female applicants were described as more communal (kind, sympathetic, helpful) and less agentic (independent, assertive, confident) than male applicants. Furthermore, the authors found that communal characteristics had a negative relationship with hiring and selection decisions in academia.

1.2 Previous studies that used the appraisal system

Several studies have used the appraisal system to examine student essays. For example, Hood (2004) compared introductory sections of undergraduate dissertations to published research article introductions. The author found that expert writers used more appreciation resources while student writers used more affect and judgment resources. Therefore, student texts appeared to be more personal and subjective than published articles. Other studies that focused on student writings include Liu and Thompson (2009) who contrasted attitude resources in one English essay and in one Chinese essay written by the same L1 Chinese student. Most of the differences occurred in the affect and judgment domains. Specifically, the Chinese text reflected fewer uses of affect and judgment compared to the English text. According to the author, this result indicates that “there are less disclosure of personal emotions and the avoidance of direct ethical or moral evaluations respectively” in the Chinese culture and rhetoric (Liu and Thompson, 2009, p. 9).

In addition, Lee (2008) compared two highest-graded essays and two lowest-graded essays. The author found that high-graded essays reflected more inclusions of attitude choices. Also, high-rated essays were full of judgment items. The author argued that such choices reflect the nature of argumentative genre. The narrative genre was also explored by Putriyantina and Said (2018). The authors focused on narrative texts written by male and female students. They found that affect items were the most frequent in both sets of texts. Concerning gender variations, they found that female students used more attitude items.

Ngo and Unsworth (2015) used the attitude system to “investigate Vietnamese students’ repertoires of English and Vietnamese appraisal resources in oral discussions about popular topics that they are usually involved in when living and studying in Australia and their common difficulties in expressions of attitude in English” (Ngo and Unsworth, 2015, p. 6). The authors proposed refinements to the system of attitude. In the first stage, they re-categorized dis/inclination and in/security in the affect system, while in the second stage, they extended judgment and appreciation sub-systems. The judgment in the modified version was extended to include normality and capacity, while the appreciation was extended to include aesthetics, appropriateness, effectiveness and convenience (Ngo and Unsworth, 2015, p. 22).

Other studies have used the appraisal system to investigate research articles by focusing on the introduction (Hood, 2004; Chang and Schleppegrell, 2011) the discussion (Cheng and Unsworth, 2016) and the conclusion (Loi et al., 2016). For example, Loi et al. (2016) examined the conclusion sections of English and Malay research articles written by academic writers. The authors applied the three domains in appraisal system, namely attitude, engagement and graduation. Regarding the results of attitudinal resources, the findings showed that in contrast to Malay conclusions, English papers used more inscribed attitude and graduation evoked attitude resources. It is important to note that inscribe attitude refers to explicit and direct evaluations, while evoked attitude means implicit and indirect evaluations.

Another study focused on doctoral discussions, as the study by Geng and Wharton (2016) who used the engagement sub-system of appraisal theory to compare six of doctoral discussions written by L1 Chinese writers and six discussions written by L1 English writers. In addition, Xie (2016) applied the three domains of the appraisal system to examine the evaluative language in Chinese masters (MA) thesis literature reviews. Concerning the investigation of attitude, the author found that appreciation was the most dominant type while the affect was the least frequent type. This result was consistent with that by Hood (2004) in research article introductions. In addition, Xie found that the inscribed attitudes occurred more than evoked attitudes. The author attributed the preference of using inscribed attitudes, hence explicit evaluation, more than invoked attitudes, hence implicit evaluation, to students’ desire to be shown as confident in the eyes of committee members. This result of using more inscribed attitudes was the opposite in Hood (2004) who found that evoked attitudes were more preferable in research article introductions.

As indicated earlier, Maskara et al. (2014) investigated cultural variations in Indian and British letters by focusing on the use of attitudinal resources. They found that resources of affect occurred more often in the Indian group than in the British letters. In addition, they found appreciation resources occurred more in the British letters compared to the Indian ones. Both groups, however, were similar in the use of judgment resources.

1.3 The analytical framework: the appraisal system

The appraisal framework, introduced by Martin and White (2005), has been used as an analytical framework to analyze linguistic semantic resources. The approach, which was developed from SFL, has been divided into three categories: attitude, engagement and graduation. The attitude domain “is concerned with our feelings, including emotional reactions, judgment of behavior and evaluation of things” (Martin and White, 2005, p. 35). The engagement category “deals with sourcing attitudes and the play of voices around opinions in discourse” (p. 35). Finally, the graduation category “attends to grading phenomena whereby feelings are amplified and categories blurred” (p. 35).

As shown in Figure 1, the attitude domain is divided into three categories: affect, judgment and appreciation. The engagement domain falls into two main categories: monogloss and heterogloss. Finally, the graduation domain consists of two categories: force and focus. The present study uses the attitude system to examine the attitudes and feelings in recommendation letters by male and female academics.

The first category of attitude system, i.e. affect, deals with feelings and emotions (e.g. pleased, thrilled, happy, worry). Specifically, it is “concerned with registering positive and negative feelings: do we feel happy or sad, confident or anxious, interested or bored” (p. 42). It can be realized as a quality by describing participants (e.g. a happy boy), as a mental process (e.g. her reaction saddened me) and a behavioral process (e.g. the boy cried) and finally as a comment (e.g. Sadly, the student failed the exam). The judgment category deals with the evaluation of human characters and behaviors (e.g. capable, brave, educated, mean). Martin and White (2005) pointed out a list of questions followed by a set of attributes. Table 1 illustrates these realizations.

The appreciation category concerns with the evaluation of things (e.g. captivating, remarkable, effective, precise). Martin and White (2005) pointed out a list of questions to show the realizations of appreciations, as illustrated in Table 2.

1.4 The present study

The short review above shows that the genre of recommendation letter is central for investigating the evaluative language. The literature shows no study that examined gender in letters of recommendation by using the appraisal system. To fill this research gap, the present study highlights evaluative and attitudinal expressions in male and female letters of recommendations. The study uses the attitude sub-system from the appraisal theory to explore affect, judgment and appreciation choices. The main objective of the present study is to explore whether there are variations in the employment of evaluative language resources by male and female writers. More specifically, the study focuses on variations, if any, that can be attributed to difference in gender.

The significance of this study is twofold: first, the importance of recommendation letter as one academic genre commonly used by academics to show attitudes, feelings, judgments, argument for/against their students, etc. More interestingly, expressing their feelings, attitudes and other evaluation aspects towards recommended students, academics need to follow specific rhetorical move structures as mentioned in previous sections of this study. Second, investigating and analyzing the interplay between evaluative language resources via appraisal analytical framework (Martin and White, 2005) and rhetorical move functions may serve formulating persuasive argument endeavor expected in academic communities. The study is, therefore, an attempt to address the following research question:

RQ1.

What are the gender similarities and differences in recommendation letters written by male and female recommenders with respect to the sub-system of attitude in the appraisal theory?

2. Method

The present study compared and contrasted forty recommendation letters written by male academics to the same number of letters written by female recommenders. The letters were given to Saudi students by academics from different universities in Saudi Arabia. The recommenders were from the following countries: Saudi Arabia, Jordan, Egypt, Sudan, Tunisia, India and Pakistan. Most of letters were collected from students, and a few were collected from the academics. Chiefly, they were maintained through personal communication by using social media platforms and a specific application designed for academic communication called MyU. The letters were written from the period of 2010 to 2019. The length of letters varied. Those written by male academics had the average number of 122.3 words, while those written by female academics had the average number of 155.75 words.

The study used the categories of affect, judgment and appreciation in the attitude sub-system from the appraisal theory. In the following example, I highlight the method of analysis, which shows the presence of each category in the attitude sub-system.

I am writing this letter to give my highest possible recommendation [Affect] for Ms. X for the graduate program at your university.

During her time in the department, Ms. X demonstrated a good [Appreciation] work ethic and interpersonal skills. We outlined a scope of work to be completed, and she successfully [Judgment] completed that work in the time required. She put in extra hours as necessary in order to meet specific deadlines that I set.

Ms. X presented outstanding [Appreciation] commitment to her studies, work, and was a quite remarkable [Judgment] student with a strong [Appreciation] research interest. The quality of her was compelling [Appreciation] and she managed to find beautifully [Appreciation] simple solutions to very challenging [Appreciation] problems. During her work at the department, she always made many valuable [Appreciation] contributions to group discussions. [M 27].

3. Results

As indicated earlier, the letters written by female academics were longer with respect to the total number of words than those by male professors. The analysis investigated attitude markers in the appraisal framework. The attitude domain consists of three sub-systems: affect, judgment and appreciation. The following sub-sections deal first with the general findings, then the findings specific to each sub-system are indicated.

3.1 General findings

As displayed in Table 3, the investigation of three attitudinal resources in letters of recommendations showed that the most employed resource was the judgment sub-system (35.3 per 1,000 words). The appreciation domain was in the second position (20.5 per 1,000 words) and the least frequent was the affect (16.5 per 1,000 words).

Table 4 presents the findings concerning the use of attitudinal resources in both gender groups. The most apparent difference between the two gender groups was in the use of judgment resources, as they appeared 39 times per 1,000 words in the male group compared with 32.4 in the female letters. The gap was narrow for both affect and appreciation sub-systems.

3.2 Affect

As indicated earlier, the affect sub-system was the least frequent in both gender groups. Also, the distribution of affect resources was very similar with about 16.5 tokens per 1,000 words in each group. Unsurprisingly, many recommenders commenced their letters with indicating their pleasure to write a recommendation. This preference was more evident in the female group with 25 letters compared to nine letters in the male group, as in the following examples:

I have had the pleasure to know X for a year in my capacity as her teacher in a couple of courses in linguistics. (F 1).

Ms. X has been my student in the Department of English […]. It is a pleasure to inform that she passed with good grades. She is a hard-working, disciplined, and conscientious student. As good as she is in her academic pursuits; it is also a pleasure to introduce Ms. X as a person. She is efficient, has a good rapport … (F 8)

It is my great pleasure to recommend X for admission in X. (F 20).

It is my pleasure to write this recommendation letter for X. (F 24).

It is my great pleasure to provide this letter of recommendation for X. (F 38).

It is my pleasure to comment on Ms. X for admission to a postgraduate study at your university. (F 11).

I had the pleasure of teaching X in her X English class. (F 19).

As an associate professor of English at X university for ten years, I have instructed many students in the subjects of English literature. However, Ms. X’s willingness to learn and dedication to literature have stood out among her colleagues. I have had the pleasure of Ms. X’s presence in three of my classes. (F 18).

Another major affect resource was offering the recommendation. This occurred more in the female group (in 34 letters) than the male group (in 27 letters). These took several forms such as I recommend him, I strongly recommend, I recommend him strongly, I highly recommend, Highest possible recommendation, I offer high recommendation, I confidently recommend him, I give her my highest recommendation, I would like to take this opportunity to recommend.

Other affect resources included the use of happy, impressed, glad, pleased and delighted. These usages occurred similarly in both gender groups, as exemplified in the following excerpts.

I am happy to recommend X. I have known her since … (M 22).

X’s commitments to attendance and to achieving the best results have impressed me. (M 39).

I was very impressed with her performance and her ability to complete all her course work … (M 39).

I was more than glad to have her as an outstanding student … (M 39).

I feel so delighted to write this recommendation letter for … (M 40).

I will be pleased to provide you with further information about X at your request. (M 11).

I am very impressed by his enthusiastic, keen interest in his work … (M 16).

I am more than pleased to recommend her. (F 1).

I am pleased to offer my highest recommendation. (F 36).

She always impressed me as she was a very distinguished and excellent student who scored an A+ in all the courses (F 38).

3.3 Judgment

The judgment resources were the most frequent in both corpora. Also, they displayed evident gender differences as male recommenders used judgment resources more than female academics.

As shown in Table 5, the most common judgment resources were had-working, good, dedicated, well, intelligent, cooperative, active, motivated, best and excellent. The use of these resources, however, varied across the two groups. Male recommenders favored using good, cooperative and best resources. On the other hand, female writers seemed to prefer using hard-working, dedicated and excellent resources (see the example below). The rest of adjectives were similar in both groups.

I had the chance to recognize many of her personal traits through my observations. Ms. X is ambitious, hardworking, and has a respectful and disciplined personality. She is dedicated and well-motivated in her work, and can perform excellent as an individual and as a team member as well. (F 3).

While the previous judgments attitudinal resources were explicit, recommenders also followed implicit ways by referring to applicants’ excellent grades. This technique (as shown in the following example) occurred 14 times in female letters and 9 times in male letters.

This is to certify that I taught X two courses A and B. X was very good, active, and serious student. In the former course, she got grade B+ as one of the highest 15% in the class, and grade A in the latter as one of the highest 20% in the class. (M 21).

3.4 Appreciation

As indicated earlier, the appreciation sub-system was the second in position in terms of frequency. The distribution of appreciation resources was very similar in both groups, as they were 20.7 words and 19.4 tokens per 1,000 words.

The most common appreciation resources, as displayed in Table 6, were excellent, good, strong, pleasant, great, positive, well and outstanding. Female writers seemed to prefer using good, pleasant, positive and creative resources. On the other hand, male recommenders favored using strong, well and high resources. The remaining of resources were similar in both groups.

The following are two examples of the use of appreciation resources.

She was part of a group who submitted a research paper titled X where they investigated various cloud computing threats and countermeasures. The research paper was well-written, and shows advanced research and English language skills. Moreover, she demonstrated excellent communication and presentation skills. Her participation during class, exam answers, project submitted reflects her deep knowledge, and critical thinking skills. Her excellent mark “A+” and performance is a reflection of her outstanding performance in the computer science program. (M 24).

At a personal level, Mr. X is a well-disciplined, industrious student with a pleasant personality, highly intelligent and has excellent communication skills. Mr. X also demonstrated good team working skills in group assignments. Mr. X’s language competence is excellent … (M 2).

4. Discussion

The study addressed the research question about the role of gender in letters of recommendation. As mentioned above, the main objective of this study is to investigate the evaluative language in recommendation letters written by male and female academics. The two sets of recommendations were, therefore, analyzed through the attitude system derived from the appraisal theory presented by Martin and White (2005). As shown in section 1.3 above, the attitude category consists of affect, judgment and appreciation items.

The overall results showed that the most frequent type was the judgment sub-system, and the least frequent was the affect. The results also revealed no statistically significant variations in attitude sub-systems: affect and appreciation as the writers in both groups (males and females) employed almost the same options in each. In respect with judgment, however, the analysis explored significant differences between the two sets as male academics used more judgment resources than females. The pre-dominance of judgment linguistic resources in the analyzed corpora goes in line with the rhetorical purpose of recommendation letters that is the construing of recommenders’ attitudes to the people they recommend and the way they behave. However, preference of judgment options in recommendation letters over other two types of attitude system could reasonably be attributed to the nature of relationships between evaluative language and rhetorical move functions. Put it another way, each rhetorical move structure involves certain lexico-grammatical resources for realizing its function. It is the nature of rhetorical move function in a particular academic genre that determines the type of evaluative linguistic resource. Since commenting on applicants’ (recommended students, the case in the present study) academic qualifications, capabilities or capacity as technically referred to by Martin and White as part of social esteem; and showing the applicants’ tenacity (how they are dependable) is the normal trends in recommendation letters, the frequency of judgment resources seem to be reasonable and justified.

The predominance of judgment resources can, therefore, be taken as a strategy by the recommenders to share and display positive social esteem towards recommended students as sharing such values are considered critical for the formation of social networks. What remains questionable and needs more investigation is the variation in employment of judgment resources found in this study based on gender. It has been found that male recommenders used the lexical patterns pertinent to capacity (how capable) like: good (together with its superlative form: best), well, cooperative, excellent, etc.; while female writers preferred lexical items associated with tenacity (how dependable) like: hard-working, dedicated, etc. to show their admiration for their applicants. This can be interpreted as, that females most likely tend to evaluate the character or personality of their applicants whereas their male counterparts prefer evaluating applicants in relation to their qualifications, academic achievement and behavior. Further investigations are needed to explore if there are other factors behind the gender variations in the employment judgment strategies based for example, on psychology, sociology or culture.

As for the appreciation category, appreciation resources came second in frequency and were almost similarly used by both groups of writers. The only and slight variation is that, male tend to praise their applicants’ skills, performance, participation, etc. by using the lexical items excellent and strong. Females in contrast employ the expressions good and positive. This may indicate that male recommenders’ appreciation is more forceful than females’.

Concerning the affect category, which was the least in frequency, most recommenders of both groups began their letters with lexical patterns such as I am very pleased, It is my pleasure, I am happy, I am more than glad, etc. This trend is expected to be common among recommendation letter writers as the un/happiness set of meanings is the first to come to mind when we think about emotion (Martin and White, 2005). The preference of such affect resources in female letters in particular may be attributed to the natural behavior of females being more emotional than males.

Although the findings of the present study can partially bear some similarities to other research findings mentioned in the previous sections, differences in objectives, analyzed genres and populations may render the comparability difficult to some extent. Hood (2004), for instance, found that expert writers used more appreciation resources while student writers used more affect and judgment resources. In Hood’s conclusion, the high frequency of affect and judgment resources were associated with student (novice) writers, whereas appreciation with published (experienced) writers. Yet, Hood analyzed Introduction sections by two groups of writers (published and novice). The analyzed genre is different as well population of the study. The high proportions judgment resources in the present study can scarcely be compared with the ones stated by Hood (2004). While preference of affect and judgment in Hood’s study associated with little-experience writers, this may not be true for the present results as the examined recommendation letters were written by published or, at least, experienced academics. It could not be said, therefore, that propositions found in recommendation letters are personal and subjective, as Hood (2004) put it, rather, the high frequency of judgment drawn on by writers of recommendation letters seem to be related to the norm and nature of the recommendation genre as mentioned above. The evaluative language involved in introducing, arguing for a research topic, reporting and commenting on previous researches studies and creating a gap (all of which are features commonly found in research Introductions sections) is different from evaluative language patterns required for recommending students/applicants to find jobs or doing post-graduate studies. The findings of this study also are in consistency with the one reported by Lee (2008) who compared two highest-graded essays and two lowest-graded essays. The author found the high frequency of judgment in the high-graded essays. Yet, the analyzed corpus (student essays) by Lee was different from the one investigated in the current study (recommendation letters written by academics). The two studies can hardly be comparable. However, in both studies the authors attributed the inclusions of judgment resources to the nature of genres: argumentative essays (in Lee’s study) and recommendation letters in the present study.

As pointed out earlier, the only study that used the attitude sub-system from the appraisal theory to investigate recommendation letters was Maskara et al. (2014). Their results were similar to the findings of the present study in terms of the high frequency of judgment resources as well as the low frequency of affect items.

To summarize the results indicated above, it could be said that the judgment resources are the most vital component in the attitude sub-system. This importance is not restricted to the genre of recommendation but also includes other genres such as the argumentative type. It is important to note, however, that previous studies found culture/language to be an important factor. This role of culture was cogently noted by Bruland (2009, p. 407) who showed that “[p]art of what makes occluded genres such as the letter of recommendation so important to study are the clues they reveal about a professional culture.” Likewise, Qanbari et al. (2014, p. 311) addressed the notion of culture and context as they showed that the letter of recommendation “is considered as a befitting representation of the rhetorical structures, patterns, and formulas for a specific context.” The factor of culture, therefore, makes it essential to restrict the selection process of letters to one cultural group in order to avoid cultural variations.

For the findings of the current study, however, the preference of judgment options by male over female group, is most likely due to some cultural variations, which, in my own viewpoint, needs more thorough investigation later in future research studies. Using less judgment resources by females may indicate women’s general tendency of being tentative in disclosing personal emotions or avoiding direct moral and ethical evaluations, particularly in Arabic culture.

The results showed that most gender variations occurred in the use of judgment choices as they were more evident in the male group. There were also some minor differences in affect and appreciation choices. In the following lines, there will be a short summary of the main findings of each attitudinal category.

First, the analysis of affect resources showed that female recommenders were more willing to open their letters with indicating their pleasure to give recommendation letters. In addition, female writers showed more tendency to use lexicons related to the word recommendation (e.g. I recommend, I strongly recommend, I give her my highest recommendation). Both gender groups were equal in using other affect resources that reflect their happiness (e.g. I am pleased, I am happy, I am delighted, I am impressed).

Secondly, the analysis of judgment reflected that it was the most frequent sub-system in both corpora, and were more used by the male group. The analysis showed certain resources to be more common, but male recommenders seemed to prefer among them the adjectives: good, cooperative and best, while female writers seemed to favor hard-working, dedicated and excellent. This result is similar to that by Trix and Psenka (2003) who found grindstone adjectives which include hardworking and dedicated to be more common in letters written for female candidates. They attributed this finding to the tendency to associate effort with female individuals while men were associated with ability terms. These results also correlate with those by Bell et al. (1992) who found that male recommenders were more likely to discuss intellect, career potential and personality. Furthermore, in Bell et al.’s (1992) study, men were more likely to describe candidates’ personality than women. In the present study, judgment resources, which include comments on personality, were more common in the male group. The analysis also yielded a common strategy, particularly in the female group, which was the use of implicit judgment resources by referring to applicant’s grades.

Thirdly, the analysis of appreciation resources showed that their presence in the letters of recommendation was similar in both gender groups. This result was not similar to that by Trix and Psenka (2003) who found that men were more willing to use terms of praise about entities (e.g. research) compared to women. Also, the present study showed that while male academics preferred to use strong, well and high resources, female writers favored to use good, pleasant, positive and creative resources. According to the theory of attitude, appreciation items include the reactions and evaluations we make towards things. While the results in this study showed that both gender groups were almost equal in terms of using a total number of appreciations tokens, they had different preferences of selecting the tokens. This finding suggests that male and female recommenders equally see the importance to refer to entities, such as research papers, academic programs and universities. Yet, each gender group had different ways of showing their evaluations.

5. Conclusion

The study drew on the attitude in Martin and White’s appraisal theory to investigate gender variations in letters of recommendations. The analysis showed important gender variations in judgment category. In the other two categories: affect and appreciation, the study revealed no significant statistic variations due to gender. The high proportions of judgment resources found in this study might be attributed to the nature of recommendation genre where the main trends are to show applicants/students’ capacity (how they are qualified and capable) and tenacity (how they are reliable and dependable) in the eyes of the institutions’ officials they are due join. The preference of judgment options of the present study may hardly be attributed to the lack of awareness or experience in manifesting evaluative language resources as shown in some previous studies (Hood, 2004, for example) where the high frequency of judgment associated with novice writers.

The main contributions of this study may be as follows: first, it is one of very few studies drawing on the attitude-category of appraisal system, as an analytical tool to examine gender differences in recommendation letters very particularly on the ones written by non-native speakers of English. Second, the gender factor is central in the genre of the recommendation letters and hence researchers should be cognizant of its role as certain variations might be impacted by it. Third, the lists of tokens can be offered as heuristics for academics to have most common words or phrases to use in their letters. Finally, the findings can hopefully bear some important pedagogical implications, very specifically for novice and non-native academic writers of recommendations letters. Designers of technical writing materials can make use of these findings for improving novice writers’ abilities of employing linguistic resources bear and formulate evaluation values and meanings. Writing teachers and practitioners can focus on factors affecting recommendation letters’ writing such as variations in gender, culture and language. More importantly, in academic wiring class activities, students’ attention can be turned to the interplay between evaluative language and rhetorical move functions in different academic genres including recommendation letters for the sake of persuasive argument commonly expected in academic domain.

Yet, due to the small corpus, the findings might not be generalized. Future studies are recommended to increase the corpus, and more importantly to use the other two sub-systems of the appraisal theory, namely engagement and graduation. Furthermore, future studies are recommended to investigate gender on the part of both the recommender and the applicant, as this study restricted the analysis to the gender of the writer. They should broaden the investigation to include factors that may cause variations rather than the ones due to gender differences. Finally, the multiple nationalities of writers might have an influence on the way people write, and thus can be a variable, so it is worth investigating the role of the nationality in future studies.

Figures

Figure 1

The appraisal system

Table 1

Judgment realizations

	Positive	Negative
How special?	Lucky, fortunate, normal, familiar, cool, stable, fashionable	Unlucky, odd, peculiar, eccentric, dated, obscure
How capable?	Powerful, mature, experienced, insightful, clever, educated, competent, successful, productive	Sick, immature, childish, helpless, naïve, foolish, uneducated, ignorant
How dependable?	Brave, heroic, cautious, patient, careful, thorough, reliable, dependable, faithful, flexible	Impatient, reckless, weak, distracted, unreliable, disloyal, stubborn
How honest?	Truthful, honest, credible, frank, candid, direct, discrete	Dishonest, deceitful, deceptive, manipulative
How far beyond reproach?	Good, moral, ethical, fair, just, sensitive, kind, caring, modest, polite, generous	Bad, immoral, corrupt, mean, cruel, rude, selfish

Source(s): Adapted from Martin and White (2005, p. 53)

Table 2

Realizations of appreciation

	Positive	Negative
Did it grab me?	Captivating, engaging, fascinating, exciting, moving, remarkable, notable	Dull, boring, predictable, unremarkable
Did I like it?	Okay, fine, good, lovely, beautiful, welcome	Bad, nasty, ugly, repulsive
Did it hang together?	Balanced, symmetrical, consistent, considered, logical	Unbalanced, irregular, flawed, contradictory, distorted
Was it hard to follow?	Simple, pure, elegant, clear, precise, rich, detailed, precise	Extravagant, unclear, plain, monolithic, simplistic
Was it worthwhile?	Profound, deep, innovative, original, creative, timely, exceptional, unique, authentic, real, genuine, worthwhile, appropriate, helpful, effective	Shallow, insignificant, conventional, common, fake, worthless, ineffective, useless

Source(s): Adapted from Martin and White (2005, p. 56)

Table 3

The three attitudinal resources in both groups

	Affect	Judgment	Appreciation
No. of tokens	183	392	224
Per 1,000 words	16.5	35.3	20.5

Table 4

The three attitudinal resources in each gender group

Written by	Affect	Judgment	Appreciation
M	80	190	95
Per 1,000 words	16.4	39	19.4
F	103	202	129
Per 1,000 words	16.5	32.4	20.7

Table 5

The judgment resources in each gender group

	M	F	Total
Hard-working	14	22	36
Good	15	4	19
Dedicated	3	12	15
Well	6	6	12
Intelligent	7	4	11
Cooperative	8	3	11
Active	4	7	11
Motivated	4	7	11
Best	10	1	11
Excellent	3	6	9

Table 6

The appreciation resources in each gender group

	M	F	Total
Excellent	17	15	32
Good	9	19	28
Strong	6	2	8
Pleasant	2	5	7
Great	4	3	7
Positive	1	6	7
Well	4	2	6
outstanding	3	3	6
High	3	1	4
Creative	0	4	4