The Mediating Role of Non-verbal Behaviors in the Relationship Between Interdependent Self-construal and Transformational Leadership

This study examined the relationship between self-construal (IS) and perceptions of transformational leadership. Structural Equation Modeling (SEM) analyses were utilized to determine whether the relationship between construal style and transformational leadership is mediated by non-verbal behaviors, specifically reflecting intention to build relationship (ITBR). We hypothesized that interdependent self-construal style would lead to ITBR, which in turn, would result in stronger perceptions of transformational leadership. One hundred and forty business managers took part in a multiphase study. In phase 1, they delivered speeches while being recorded. In phase 2, their non-verbal behaviors were then coded by independent coders using Laban Movement Analysis (LMA). Next, in phase 3, additional independent coders coded participants’ speeches to measure the degree of transformational leadership exhibited in the managers’ speeches. The results supported our hypothesis that nonverbal movements were linked to ITBR, mediated the relationship between interdependent self-construal and perception of transformational leadership. The implications of the results are discussed within the context of self-construal and leadership theories.<br><br>One hundred and forty managers (77 women; 63 men) who had been with their respective organizations for a minimum of three years were recruited. To qualify as a manager, participants had to be in management position for at least three years or longer with at least one person reporting direct to him/her. Participants were sampled from variety of different industries in Southern California region including information technology, entertainment, automotive, management consulting, healthcare and education.<br><br><b>Measures</b>Interdependent Self-construal: The participants were asked to complete the interdependent self-construal scale consisting of 14 items developed by Gundykunst et al. (1996). <br>Transformational Leadership: To measure the extent to which the participants exerted transformational leadership during the speech, three independent coders watched the participants’ speeches, and evaluated the degree of transformational leadership using the Multifactor Leadership Questionnaire (MLQ). The coders held a graduate level degree in the field of Organizational Leadership, with seven to ten years of practice in the field. Before the study, the three coders were informed about transformational leadership characteristics in a 60-minute training session. This session provided general background information, an opportunity to ask questions regarding the coding process. All coders were blind to the hypotheses. After their training session, the coders independently viewed the recording of each participant’s speech, and completed the MLQ rating sheet. A reliability testing was conducted after the coders completed coding every ten speeches. If the reliability coefficient was less than .80, then the first author provided further training in the area of inconsistencies, and provided an opportunity for the coders to ask additional questions. This process yielded an interrater reliability coefficient of 0.94 <br>Non-verbal Behavioral Codings: The participants’ speeches were coded for nonverbal behaviors using LMA. Three independent coders (who were different than those who coded transformational leadership) were selected. The coders were trained to code non-verbal behaviors using LMA by the first author. The coders watched each speech while on mute, and were asked to rate the effort and shape behavioral elements using a coding scale ranging from 0 (element is never present) through 5 (element is always present). They coded 4 effort elements (weight, flow, space, time), and 3 shape elements (self-oriented, other-oriented, and process-oriented). After the coders completed the coding, we conducted inter-rater reliability analysis. The results indicated an inter-rater reliability of 0.96 on the effort and shape coding; and for effort and shape movement elements index coding sheet yielded Cronbach’s alpha coefficient of .80, indicating that the coders were consistent and that the measures are reliable.<br>We conducted correlational analyses using the Pearson r statistic to test Hypotheses 1 &amp; 2. To test the mediational model as predicted in Hypothesis 3, Model-Fit/Goodness-of-Fit Indices (GFI), Adjusted Goodness of Fit Indies (AGFI), and Root Mean Square Error of Approximation (RMSEA) of Structural Equation Modeling (SEM)/Path Analysis in AMOS were used because these methods are more suitable for slightly to moderately non-normal distributions, minimize the impact of the skewness and kurtosis, and are less dependent on sample size <br>

本研究考察了自我构念（self-construal，IS）与变革型领导（transformational leadership）感知之间的关系。研究采用结构方程模型（Structural Equation Modeling，SEM）分析，以探究构念风格与变革型领导之间的关系是否通过非语言行为（具体为建立关系意图（Intention to Build Relationship，ITBR））起到中介作用。我们提出假设：相互依赖型自我构念会提升建立关系意图，进而强化个体对变革型领导的感知。本研究为多阶段实验，共招募140名企业管理者参与。第一阶段，参与者进行演讲并被录制；第二阶段，由独立编码员采用拉班动作分析（Laban Movement Analysis，LMA）对参与者的非语言行为进行编码；第三阶段，另有三名独立编码员对参与者的演讲内容进行编码，以衡量管理者演讲中展现的变革型领导程度。研究结果验证了我们的假设：非语言动作与建立关系意图相关，并在相互依赖型自我构念与变革型领导感知之间起到中介作用。最后，本文结合自我构念与领导理论的研究背景，讨论了本研究结果的理论启示。

本次招募的140名管理者中，女性77名，男性63名，且均在当前任职机构供职至少三年。要符合管理者的招募标准，参与者需至少担任管理岗位三年及以上，且至少有一名直接下属。研究样本取自南加州地区多个不同行业，包括信息技术、娱乐、汽车、管理咨询、医疗保健与教育领域。

**测量工具**
相互依赖型自我构念：参与者需完成Gudykunst等人（1996）开发的包含14个题项的相互依赖型自我构念量表。
变革型领导：为衡量参与者在演讲中展现的变革型领导程度，三名独立编码员观看参与者的演讲录像，并采用多因素领导问卷（Multifactor Leadership Questionnaire，MLQ）对变革型领导程度进行评分。所有编码员均拥有组织领导学领域的硕士学位，且具备7至10年的相关从业经验。研究开始前，三名编码员接受了时长60分钟的培训，内容涵盖变革型领导的特征、研究背景以及编码流程答疑。所有编码员均对研究假设不知情。培训结束后，编码员独立观看每位参与者的演讲录像，并填写MLQ评分表。每完成10份演讲的编码后，研究人员会进行信度检验：若信度系数低于0.80，则由第一作者针对编码不一致的问题开展进一步培训，并为编码员提供额外答疑机会。最终，编码员间的评分者信度系数达到0.94。
非语言行为编码：研究采用拉班动作分析（LMA）对参与者演讲中的非语言行为进行编码。本次招募的三名独立编码员与负责变革型领导编码的编码员并非同一批人。第一作者对编码员开展了LMA编码培训。编码员在观看静音状态下的演讲录像时，需采用0（该行为从未出现）至5（该行为始终出现）的评分量表，对动作的力度（effort）与形态（shape）维度进行评分。编码维度共包含4项力度要素（重量、流畅性、空间、时间）与3项形态要素（自我导向、他人导向、过程导向）。编码完成后，研究人员开展了编码员间信度分析：结果显示，力度与形态编码的评分者间信度为0.96；力度与形态动作要素编码表的克朗巴哈α系数（Cronbach’s alpha）为0.80，表明编码员间一致性良好，测量工具具备可靠信度。
本研究采用Pearson相关系数检验假设1与假设2。为验证假设3提出的中介模型，研究采用AMOS软件中的结构方程模型（SEM）/路径分析的模型拟合优度指标，包括拟合优度指数（Goodness-of-Fit Indices，GFI）、调整拟合优度指数（Adjusted Goodness of Fit Indices，AGFI）以及近似均方根误差（Root Mean Square Error of Approximation，RMSEA）。上述方法更适用于轻度至中度非正态分布的数据，可最大限度降低偏态与峰度的影响，且对样本量的依赖性更低。