Overview of data analysis plan
The data gathered from the teacher’s pet questionnaire were analysed and discussed in four steps analogous to those proposed by Tal and Babad (1989) in the original paper. Therefore, in the first step, open-ended definitions were presented that were proposed by students and teachers. Moreover, respondents’ attitudes towards the teacher’s pet phenomenon and their appraisals concerning its frequency were compared.
In the second step, the subjects’ opinions about three types of students, i.e., pets, leaders, and best students, and their roles in relations with teachers were analysed. In the third step, the evaluations of respondents’ social distance towards pets, leaders, and best students were presented. Subsequently, based on the assumed criteria, two theoretical subtypes of the pet, i.e., pet-leader versus pet-rejected, were distinguished and it was checked whether the proposed classification, as expected, modifies the way students and teachers evaluate the distance from pets. Finally, in the fourth step, the teacher’s pet phenomenon was analysed comprehensively taking into consideration the data gathered by means of 21 questionnaire items focused on pets; behaviour in the classroom and their relations with the teacher and other students.
The open-ended descriptions of pets (B1 item) were analysed by three judges using coding categories proposed by Tal and Babad (1989), i.e. (1) pets’ attributes, (2) pets’ behaviour toward teachers, (3) teachers’ behaviour toward pets, and (4) pets’ relations with classmates (Krippendorff’s alpha for the coding categories, respectively: .888, .934, .932, and .825). The descriptions were divided into positive and negative within each category, placing the 1/-1 value upon them (Krippendorff’s alpha, respectively: 898, .855, .886, and .775).
In total, the respondents presented 2841 open-ended descriptions. Of these descriptions, 60.5% were negative and 39.5% were positive. Most respondents described the teacher’s pet phenomenon in terms of the teacher’s behaviour (51.1%), of which 71.3% were negative and focused on an unfair and preferential teachers’ treatment (e.g., inflating pets’ grades) and the rest (28.7%) concerned normative positive teachers’ behaviour (e.g., providing support in learning).
The second largest category was the pets’ behaviour toward teachers (26.2%), of which 62.0% were negative and concerned ingratiation, manipulating the teacher, taking advantage of one’s special position to promote particular interests, etc., and the rest (38.0%) were positive and focused on supporting teachers in lessons, giving them help in administrative work, etc.
Further categories of descriptions concerned pets’ attributes and relations with classmates (respectively: 19.6 and 3.0%). Within the former category, 72.8% descriptions were positive and focused on cognitive and social characteristics, e.g., interest in studies, responsibility, good manners, etc., and others (27.2%) concerned negative characteristics, e.g., egoism, ingratiation, etc. In the case of the latter, 82.6% descriptions were negative and concerned unfriendly interactions of pets with their classmates, and others (17.4%) were positive and focused on providing help, being polite to classmates, etc.
Of all open-ended descriptions, 2677 (94.2%) responses were generated by students and 164 (5.8%) were proposed by teachers. Descriptions presented by students were mainly negative (62.7%), in contrast to teachers’ responses, most of which were positive (74.3%), χ2(1) = 89.113; p < .001. Defining the phenomenon, students focused on teachers’ behaviours (52.4%), pets’ behaviours (26.7%), their attributes (18.0%), and their relationship with classmates (2.9%). Teachers, on the other hand, associated the phenomenon with pets’ attributes (46.3%), teachers’ behaviour towards pets (30.5%), pets’ behaviour towards teachers (18.9%), and classmates (4.3%).
Two categories of responses were contrasted. When describing the relationships between pets and teachers and classmates, students focused mainly on negative behaviours (respectively: 64.1 and 87.3% of descriptions), and teachers on positive behaviours (respectively: 87.0 and 71.4% of descriptions), χ2(1) = 33.117; p < .001 for pets’ behaviour towards teachers and χ2(1) = 15.424; p < .001 for pets’ relationships with classmates.
Positive versus negative pets’ characterisations
The attitude of respondents towards the teacher’s pet phenomenon was estimated by adding up the notes attributed to responses in terms of the four coding categories for open-ended descriptions (B1 item). Teachers created more positive definitions (M = .154; SD = .299) than students (M = −.078; SD = .291), t (2307) = 8.800; p < .001; Cohen’s d = .796. Descriptions were not differentiated by variables: seniority (teachers), class profile (students), and the type of school. More positive phenomena were proposed by boys (M = −.058; SD = .287) than girls (M = −.092; SD = .296), t (2056.470) = 2.709; p < .01; Cohen’s d = .116. No similar differences were observed among teachers.
The attitude towards the phenomenon was also measured directly. Respondents were asked to consider whether it was positive, negative, or both positive and negative (B2 item). Again, in comparison to students (M = −341; SD = .688), teachers’ evaluations (M = .189; SD = .784) were more positive, t (137.695) = 7.453; p < .001; Cohen’s d = .764. Teachers’ estimates were not differentiated by contextual variables. In contrast, the children’s evaluations were higher among junior high school (M = −.298; SD = .714) than high school students (M = −.419; SD = .636), t (1746.472) = 4.048; p < .001; Cohen’s d = .176; and among boys (M = −.237; SD = .726) than girls (M = −.418; SD = .648), t (1875.040) = 5.918; p < .001; Cohen’s d = .265.
Finally, the consistency of respondents’ attitudes towards the phenomenon, both directly and indirectly, was evaluated. The evaluations of open-ended descriptions (B1 item) were positively related to the evaluations of the term (B2 item), r = .420; p < .001; the discovered relationship was nearly twice as strong among teachers (r = .620; p < .001) compared to students (r = .388; p < .001).
Prevalence of the teacher’s pet phenomenon
Respondents’ beliefs about the prevalence of the teacher’s pet phenomenon at school were evaluated by means of two questions. One question was: “How many teachers have pets, in your opinion?” (B4 item), and the second was: “Try to remember whether there are in your class, or were in previous years, students you consider a teacher’s pet” (B8 item). In case of the affirmative response, they were asked to indicate which class it was.
Considering the first question, the evaluations of teachers (M = 3.360; SD = .937) and students (M = 3.220; SD = 1.019) were similar, t (2271) = 1.479; p ns. The responses were not differentiated by contextual variables.
In the case of the second question, the majority of respondents agreed there currently is/was a pet (66.1% for teachers and 85.2% for students) in the classroom. The phenomenon was observed for all grades. Contextual variables did not affect teachers’ responses. Among students, more girls than boys declared experiencing the phenomenon (59.3 vs. 40.7%) and more boys than girls indicated lack of such experience (60.1 vs. 39.9%), χ2(1) = 41.216; p < .001. Furthermore, a greater number of students in scientific profile classes as compared to humanistic profile classes declared the lack of experiences related to the phenomenon (66.8 vs. 33.2%); a similar difference did not occur in a group of students who experienced the phenomenon (49.6 vs. 50.4%); χ2(1) = 18.295; p < .001.
Opinions about pets, leaders, and best students
Traits of pet versus leader and good student
One open-ended question (B3 item) asked subjects to describe the attributes of a student who is the teacher’s pet. In all, 6703 descriptions were coded by the three judges who assigned 1/−1 value for positive/negative characteristics (Krippendorff’s alpha = .921). Of these, 61.6% were positive (e.g., helpful), 37.2% were negative (e.g., conceited), and 1.2% was ambiguous (e.g., sits in the front row).
Respondents differed in how they perceived the features of the teacher’s pet. Again the descriptions of teachers (M = .777; SD = .519) were more positive than of students (M = .106; SD = .792), t (149.828) = 13.289; p < .001; Cohen’s d = .860. Pets were described more positively by teachers with less job seniority (M=.906; SD=.332) than those with more job seniority (M = .733; SD = .552), t (98.992) = 2.055; p = .043; Cohen’s d = .348.
Moreover, subjects were asked to provide any three typical attributes of pets in contrast to the attributes specific for the student who has the role of class leader and best student (B5 item). Responses were coded by three judges who assigned 1/−1 value for positive/negative characteristics (Krippendorff’s alpha = .905). In the case of the teacher’s pet, out of 5995 attributes, 53.3% were positive (e.g., intelligent), 45.1% were negative (e.g., flatterer), and 1.6% was ambiguous (e.g., quiet).
In contrast to a pet, a student as a class leader was described with 5884 attributes (79.7% were positive, 19.4% were negative, and 0.9% were ambiguous; Krippendorff’s alpha = .909), and the best student with 6321 attributes (95.8% were positive, 2.8% were negative, and 1.4% were ambiguous [B5 item]; Krippendorff’s alpha = .948). The positive attributes of a leader concerned, e.g., intelligence, charisma, etc.; negative, e.g., aggressiveness, bossiness, etc., and ambiguous, e.g., having friends in senior classes, physical build. In contrast, the positive features of the best student concerned, e.g., diligence, responsibility, etc., negative, e.g., being a swot or shy, etc., and ambiguous, e.g., physical build, having few friends, etc.
A two (between-subject factor; respondents: teachers vs. students) ×3 (within-subject factor; type of students: teacher’s pet vs. best student vs. leader) mixed-model ANOVA was used to compare the appraisal of attributes appropriate in the respondent’s opinion for the distinguished types of students (B5 item). The appraisal was influenced by the type of the student, Greenhouse-Geisser corrected F(1.772, 3644.677) = 102.785; p < .001; η2 = .048. The least positive note was given to a pet (M = .059; SD = .807), and more positive one to the best student (M = .930; SD = .272) and the leader (M = .590; SD = .662). Teachers tended to assign more positive attributes to students (M = .852; SD = .365) than students did (M = .511; SD = .577), F(1, 2057) = 83.474; p < .001; η2 = .039.
The analysis also revealed an important effect of interaction, Greenhouse-Geisser corrected F(1.772, 3644.677) = 47.246; p < .001; η2 = .022. The average appraisals of attributes in all comparative groups are shown in Fig. 1.
It can be observed that the teacher’s pet was characterised less positively by students (M = .015; SD = .799) than teachers (M = .784; SD = .556), F(1, 2057) = 104.375; p < .001. The leader and the best student was assigned more complex attributes, although the discovered differences were still significant—students and teachers respectively for the leader M = .590; SD = .662 versus M = .784; SD = .447, F(1, 2057) = 9.675; p < .001 and for the best student M = .927; SD = .272 versus M = .986; SD = .092, F(1, 2057) = 5.318; p < .05.
Teachers perceived the leader and the pet in a very similar way, apparently identifying the indicated types of students (respectively M = .784; SD = .556 vs. M = .784; SD = .447), F(1, 2057) < 1. The leader and the pet were described in a different way by students. Unlike the leader, the pet was assigned radically less favourable attributes (respectively M = .590; SD = .662 vs. M = .015; SD = .799), F(1, 2057) = 743.235; p < .001.
The effect of the student’s type was modified by teachers’ professional experience. More positive attributes were assigned to the pet by respondents with relatively less (M = .939; SD = .256) than greater (M = .731; SD = .615) job seniority, F(1109) = 3.523; p = .063. No similar differences were noticed in the case of the leader and best student, Fs < 1.
Students who change teachers’ decisions
The respondents were asked to indicate a group of students, i.e., pets, leaders, or best students, whose representatives are most often sent to the teacher to influence his/her disadvantageous decisions concerning the aversive consequences of students’ misbehaviour occurring at school (B7 item). Students agreed that the best candidate for a mediator is a pet (61.3%), followed by the leader (25.4%), and the best student (11.5%). Teachers’ responses were distributed evenly; 36.7% indicated the pet, 32.0% best student, and 32.0% leader. In turn, almost 64.0% of students indicated the pet, 25.5% leader, and 10.5% best student.
The discovered difference was significant, χ2(2) = 63.649; p < .001. The choice of pet vs. leader or best student as a mediator was not modified by contextual variables.
Social distance towards the pet, leader, and best student
Respondents were asked to appraise a social distance between the pet, leader, best student, and teachers/peers (C1-C3/D1-D3 items). A 3 (type of students: pets vs. leaders vs. best students) × 2 (rejected by teacher vs. students) ANOVA with repeated measures on both factors was used to evaluate the scope of pets’, leaders’, and best students’ rejection by teachers and classmates. The obtained results (cf. Figure 2) proved to be surprisingly compliant with the results described over 25 years ago by Tal and Babad (1989) in the first empirical article addressing the teacher’s pet phenomenon.
The analysis revealed the main effect of the rejection factor, F(1, 2241) = 509.688; p < .001; η2 = .185 (M = 3.306; SD = .994 for teachers versus M = 3.884; SD = 1.089 for students) and the types of students, Greenhouse-Geisser corrected F(1.960, 4392.410) = 64.853; p < .001; η2 = .028 (M = 3.712; SD = 1.183 for pet versus M = 3.677; SD = 1.278 for leader versus M = 3.397; SD = 1.042 for best student).
A significant effect of the interaction between the two factors once again confirmed the accuracy of the theoretical model of the teacher’s pet phenomenon (Tal and Babad 1990; Babad 1995; Chiu et al. 2013), Greenhouse-Geisser corrected F(1.724; 3862.931) = 2840.861; p < .001; η2 = .559. Leaders were perceived as closer to peers (M = 2.783; SD = 1.799) than teachers (M = 4.571; SD = 1.705), F(1, 2241) = 1235.266; p < .001. In contrast, pets were evaluated as closer to teachers (M = 1.878; SD = 1.562) than to classmates (M = 5.545; SD = 2.036), F(1, 2241) = 4042.787; p < .001. Finally, the best students were perceived as equally close to teachers (M = 3.470; SD = 1.278) and students (M = 3.325; SD = 1.515), although a discovered difference was still significant, F(1, 2241) = 13.400; p < .001. The described results were not modified by the contextual variables.
Two subgroups of pets: social distance towards pet-leader versus pet-rejected
Based on the analysis of open-ended descriptions of the teacher’s pet (B1 item), attributes assigned to a pet (B3 item) and attributes assigned to a pet in contrast to a leader and best student (B5 item)—two groups of students who had special relationship with teachers and classmates—were distinguished, i.e.: (1) pet-leader and (2) pet-rejected by peers.
The following criteria applied: if the attributes concerned only positive statements, of which at least one concerned the cognitive functioning (smart, creative, etc.), and another one social functioning (nice, charismatic, etc.), then the three judges found these characteristics related to a pet-leader. In other cases, it was assumed that the attributes concerned a pet-rejected (Krippendorff’s alpha = .927, .920, and .959 for B1, B3, and B5 items, respectively).
The reasonableness of the adopted method of distinguishing subgroups was tested in two stages. First, it was checked whether it diversified the view of teacher’s pet term (B2 item). In all cases, more positive evaluations were discovered in respondents assigning pets’ attributes specific for pet-leader than pet-rejected (for distinction derived from definitions proposed in B1 item: M = .053; SD = .733 versus M = −.337; SD = .694, t(2120) = 4775; p < .01; Cohen’s d = .546; for distinction derived from attributes proposed in B3 item: M = .133; SD = .663 versus M = −.397; SD = .680, t(448.366) = 13.179; p < .01; Cohen’s d = .789; and B5 item: M = .108; SD = .697 versus M = −.365; SD = .690, t(288.249) = 9.759; p < .01; Cohen’s d = .682).
Second, the consistency of respondents’ opinions on the characteristic features of the pet-leader and pet-rejected was evaluated. For this purpose, (1) the content of open-ended descriptions of the teacher’s pet phenomenon (B1 item), (2) attributes specific for a teacher’s pet (B3 item), and (3) attributes of teacher’s pet as opposed to features specific for leader and best student (B5 item) were correlated. As expected, significant relationships were discovered between (1) and (2), r = .323, (1) and (3), r = .328, and (2) and (3), r = .670; all p < .001.
Bearing in mind the results of the previous investigations (Babad 1995; Babad and Ezer 1993; Chiu et al. 2013; Lu et al. 2015), it was assumed that a pet should be more rejected by peers characterising pets in terms of attributes specific for leaders as opposed to rejected students. This tendency should be reversed or disappear in the group of teachers. As expected, a series of three, 2 (between-subject factor: attributes specific for pet-leader vs. pet-rejected) × 2 (within-subject factor: being rejected by teacher vs. peers) mixed-model ANOVA revealed a consistent pattern of results. The descriptive statistics calculated for all comparative groups are shown in Table 1.
Indeed, in each of the three analyses, the appraisals of a social distance were lower for students (vs. higher for teachers) describing pets using attributes specific for leaders rather than for pet-rejected; for the effects of the interaction of the factor of rejection and the opinion concerning attributes distinguished based on B1, B3, and B5 items, respectively: F(1, 2100) = 4.129; p < .05; η2 = .002; F(1, 2122) = 28.946; p < .001; η2 = .013; F(1, 2139) = 11.156; p < .001; η2 = .005.
Teacher’s pet in the classroom
There were 21 questionnaire items (B6a–B6u items; Cronbach’s alpha = .736; in a teachers’ and students’ subgroup alpha = .725 and .732, respectively) concerned: academic achievements and social behaviour of pets at school, pets’ relationships with peers and teachers, and the consequences of the occurrence of the teacher’s pet phenomenon in the classroom. Differences in responses provided by students and teachers were evaluated using the t-test for independent samples. Afterwards, it was checked to see which of them were modified by contextual variables included in the study and respondents’ attitudes towards pet-leader versus pet-rejected.
Significant differences were discovered in the case of 15 statements. Descriptive statistics in the compared groups, the results of t-tests and Cohen’s-d measures are presented in Table 2.
Taking into consideration teachers’ attitudes towards pets, students more frequently agreed that “pets receive preferential treatment” (B6a) than teachers and expressed less approval for the statement: “when a teacher feels special affection from a student, he/she returns affection” (B6m). Assessing the awareness of the occurrence of the phenomenon in the classroom, there was more consensus among the former agreed that “students know if the teacher has a pet” (B6b) and less that “teachers are not aware of their giving preferential treatment to some students” (B6h).
Justifying a special relationship between pets and teachers, students to a lesser degree approved of the statements that “pets deserve special treatment because they contribute to the class more than other students” (B6e) and that “a pet can make a good impression—better than other students” (B6o). Furthermore, they agreed more frequently than teachers that “pets flatter the teachers” (B6f).
Consistent differences were also discovered for items concerning students’ attitudes towards pets and teachers having pets. Students agreed less than teachers that “a pet is liked by his/her classmates” (B6p) and to a greater degree agreed that “a pet is rejected by his/her classmates” (B6i). Analogically, students to a lesser degree than teachers approved of the statement that “students show understanding about a teacher’s special affection for certain individuals” (B6n) and agreed that “students are angry at teachers who show preferential treatment” (B6q).
The last group of items concerned the social behaviours of pets. Students agreed more than teachers that “pets allow themselves to behave differently from other students” (B6r), and that “pets obtain from the teacher ‘goods and benefits’ for themselves” (B6t). In addition, students expressed less approval than teachers for the statements: “pets help weaker students in class” (B6u) and “pets obtain from the teacher ‘goods and benefits’ for the entire class” (B6s).
These differences were modified by opinions about attributes specific for pet-leader versus pet-rejected, distinguished on the basis of open-ended descriptions of the teacher’s pet phenomenon (B1 item; four interaction effects), attributions made to pets (B3 item; two interaction effects), and respondents’ gender (one interaction effect). As expected, it turned out that for the statements: B6o, B6p, and B6u, the discovered differences (i.e., teachers > students) were maintained in the subgroups describing pets using attributes typical of rejected students and were reversed among those describing pets using attributes specific for leaders, respectively: F(1,1882) = 4.396; p < .05; η2 = .002, F(1,1882) = 7.702; p < .01; η2 = .004, and F(1, 1907) = 4.238; p < .05; η2 = .002.
Analogically, for the statements: B6a, B6f, and B6i, the indicated differences (i.e., teachers < students) did not change in subgroups characterising pets using attributes specific for rejected students and was reversed among those characterising pets using attributes specific for leaders, respectively: F(1, 1907) = 9.033; p < .01; η2 = .005, F(1,1882) = 8.897; p < .01; η2 = .005, and F(1,1882) = 3.784; p = .052; η2 = .002.
With reference to the gender variable, a discovered difference (i.e., teachers<students) for the B6a statement was maintained among women and disappeared among men, F(1, 1989) = 7.485; p < .01; η2 = .004.