Researchers in education have found that teachers often differ in their nonverbal behaviors toward children from different social groups and these behaviors correlate with achievement gaps and academic stereotypes about the groups. Early elementary school, when achievement gaps first emerge, is also the time when White, majority children begin to show group-level biases, and when racial minority children are able to detect discrimination and experience anxiety related to their membership in a particular social group. Therefore, if children are sensitive to teacher nonverbal behaviors, these behaviors could contribute to majority children's group biases, and may impact minority children's awareness of being in a negatively stereotyped group. In fact, children are adept at perceiving adult nonverbal behaviors and using these behaviors to guide their own behaviors and to make judgments about others. The primary goal of this research is to examine the effect of biased nonverbal teacher behaviors on group biases for children from positively stereotyped groups, and on affect and anxiety for children from negatively stereotyped groups. The investigators hypothesize that group biases in teacher behaviors will influence children's attitudes about groups, and will result in negative affect and anxiety for students in groups targeted by negative nonverbal teacher behaviors.

In the present study, the investigators test whether children attend to patterns in teachers' nonverbal behaviors toward different groups, and use group-level differences to guide their own evaluations of individuals from different groups. Although existing research has found correlations between teacher behaviors and group-level attitudes (e.g., math/gender stereotypes), it is difficult to identify the causal link between the specific teacher behaviors and children's attitudes. Additionally, with existing groups such as those defined by race or gender, there are a variety of factors that could influence children's stereotype knowledge or bias before participating in the study. Therefore, in the present research participants learned about students who were members of novel social groups (defined by t-shirt color). Additionally, because children's intergroup preferences are influenced by their own group membership, the investigators also assigned children to the novel groups. Some children were assigned to the group that subsequently received positive behaviors from the teacher ('positive group'), some children were assigned to the group that subsequently received negative behaviors from the teacher ('negative group'), and a third group of children participated in a control condition where they were also assigned to a group and received a t-shirt, but did not see any interactions between their ingroup members and the teacher ('no teacher cues').

To test whether teacher nonverbal behaviors play a causal role in children's intergroup attitudes, after assigning participants to one of the three groups, they viewed a series of interactions where students from each of two novel groups interact with a teacher, and the teacher's nonverbal behavior was correlated with group membership: across the interactions, one group always received positive teacher behaviors and the other received negative teacher behaviors. The investigators then assessed children's group-level stereotypes and preferences for new students from the positive and negative groups. Across three sets of test questions, the investigators asked participants who they thought was smarter, who they would prefer as a friend, and who they would select as a partner on an academic task.

Based on the studies of children's stereotyping, it was predicted predicted that children in all three groups would report that new students in the positive group were smarter than students in the negative group. However, when asked whom they would like to befriend, it was hypothesized that children's own group membership would influence their choices: children who were in the group that received positive behaviors or the group that did not interact with the teacher would prefer to befriend students from the positive group, but children in the group that received negative behaviors would show no preference or prefer their own group. When asked who they would prefer as a partner on an academic task, it was once again predicted that students in the positive and no cues groups would select students from the positive groups. However, there were no clear predictions for participants from the negative group; if they were more influenced by stereotypes, they would be more likely to select students from the positive group and if they were more influenced by ingroup preferences, they would be more likely to select students from the negative group.

Participants. Participants were 96 5- to 7-year-old children (32 per condition, Mmean age = 6;6, range = 4;8-8;4, 48 males, 53% Multiracial, 23% White, 21% Asian) living in Hawai'i. The investigators were able to obtain information about current school level for 91 participants; of these participants, 26% were in preschool or pre-kindergarten, 69% were in kindergarten, and 4% were in first grade. One additional child participated but was excluded for not finishing the session. There was a pre-determined a stopping rule of 32 participants per condition based on previous studies examining children's understanding of nonverbal behaviors.

Materials. At the beginning of the familiarization, participants were first presented with a picture of three groups of 12 children gender matched to the participant and representing a variety of races common in Hawaii. Each group wearing a different color shirt (orange, blue, or green). Then, they viewed four 17-second clips featuring an interaction between a female teacher and two students who matched the gender of the participant. In each video, the students wore green and orange t-shirts and the teacher wore black. The teacher and the students were unfamiliar to participants. Throughout each video, the teacher faced the students and the participant; the participant could only see the backs of the students' heads.

In each video, the teacher first faced forward, then turned toward the student on the left while that student read a brief (7 s) passage. Then, she turned and faced the student on the right while that student read the same passage. The audio clips of the students reading were the same with minor adjustments made to tone and pitch so that the voices sounded different. The teacher then faced forward at the end of the clip. Across the videos, the teacher directed positive nonverbal behaviors (i.e., smiles and nods) toward students wearing one color while they read and negative nonverbal behaviors (i.e., frowns and head shakes) toward students wearing the other color while they read.

Test trials featured photographs of children who matched the gender of the participant and varied by race to match the racial diversity of classrooms in Hawaii. On each trial, participants viewed a pair of children who were matched according to race, age, and attractiveness (based on adult ratings). In each pair of children, one wore an orange shirt and the other wore a green shirt.

Procedure. Participants were tested individually in the lab. The study was presented on a monitor and an experimenter sat facing the child during the familiarization so that she was unaware of which t-shirt group (orange or green) received positive or negative nonverbal behaviors from the teacher.

Familiarization Phase. At the start of the session, participants were told that they would be learning about a school with three groups of kids: oranges, greens and blues. They were going to pretend like they went to the school and would be in one of the groups. They then received either an orange, green, or blue t-shirt to wear during the study.

After putting on the t-shirt, they were told that they were going to watch some of the kids in the orange and green groups do some reading for their teacher. Participants then watched four familiarization videos in which the teacher directed positive behaviors toward students from one group, and directed negative behaviors toward students from the other group.

Test Phase. After viewing the familiarization videos, participants were told that the experimenter was going to ask them some questions about new kids in the orange and green groups. There were three blocks of four test trials. On smart test trials, participants were asked to select the student they thought was smarter and the investigators defined smart as "someone who is really good at learning stuff." On friend test trials, participants were asked who they would want to be friends with. On partner test trials, participants were asked who would pick as a partner to help them read a hard book. Participants indicated their answers by pointing and never received feedback. In between each block of test trials participants viewed one of the familiarization videos for a second time. After the three blocks of test trials, participants completed some additional measures that are not reported here.

Design. Participants were randomly assigned to the positive, negative, or no cues condition. In the positive condition, they were assigned to wear a t-shirt that was the same color as the student who received the positive nonverbal behaviors from the teacher in the familiarization videos. In the negative condition, participants were assigned to wear a t-shirt that was the same color as the student who received the negative nonverbal behaviors from the teacher in the familiarization videos. In the no cues condition, participants were assigned to wear a t-shirt that was not shown in either the familiarization videos or test trials. The t-shirt color of the positive and negative groups, the order of familiarization videos, the order of test trial blocks, the students in each test trial block, and the students in the positive and negative groups varied across participants.

Scoring. On each test trial, selecting the student who wore the same t-shirt color as the students receiving positive nonverbal behaviors from the teacher in the familiarization was scored as 1 and selecting the student who wore the same t-shirt color as the students receiving negative behaviors from the teacher in the familiarization was scored as 0. For each participant, the investigators created an average positive group score for each trial type by summing across each block and dividing by the total number of completed trials (4 in most cases, except for one participant who did not give a response on two trials, and two participants who did not give a response on one trial, one participant in the no cues condition failed to give any responses on friend trials so they were excluded from analyses of that variable).

Nonverbal Coding. To assess whether being a member of a group that received positive, negative, or no teacher behaviors influenced participants' affect and anxiety, the investigators analyzed footage of the participants during the task. The investigators first created video clips of the participants when they were viewing the teacher directing nonverbal behaviors to the students in the familiarization. For each participant, there was a total of 8 clips: four clips of the participants viewing positive interactions and four clips of participants viewing negative interactions. For participants in the positive condition, the positive interactions were ingroup interactions and the negative interactions were outgroup interactions. For participants in the negative condition, the negative interactions were ingroup interactions and the positive interactions were outgroup interactions. For participants in the no cues condition, both positive and negative interactions of interactions were outgroup interactions. The videos were then coded for the affect and anxiety of the participant on a scale of -3 to 3. For affect, negative codes corresponded to behaviors like furrowing the brows, mouth turned down, or head shaking; for anxiety, negative codes corresponded to behaviors like tense muscles, withdrawn or fidgety, or displaying self-comfort behaviors. For affect, positive codes corresponded to behaviors like relaxed or raised brows, mouth turned up, or head nodding; for anxiety, negative codes corresponded to behaviors like relaxed muscles, open expansive posture, or fluid movement. Each coder rated both kinds of codes for each clip.

Inclusion Criteria:

• children older than 54 months
• children younger than 102 months
• children who are part of an existing database managed by the investigators
• children whose parents agreed to be contacted during community events

Exclusion Criteria:

• consent of the parent
• assent of participant
• parent and participant both fluent in English

The investigator will receive a randomized list of sequence orders for each age and gender group. Within the sequence orders, one third of them are for participants in the positive condition, one third are for participants in the negative condition, and one third are for participants in the no cues condition. The experimenter will not know which sequences relate to each condition.

The nonverbal behavior coders will also be blind to the participants' condition. They will view videos where only the participant is visible but the teacher is not.