为了检测有意义的基本尺寸，可以解释所观察到的相似性七社会网络之间，我们使用QAP相关聚合为多维尺度（MDS）程序输入。一个三维的MDS表示最适合的数据，90.5%的变异和应力核算（适合指示）在接受上的规范，即15（博尔加蒂，1997）。我们可视化的二维MDS表示在2。这种二维表示占缩放数据的方差的75.7%，但应力超过上应力极限（应力= 26）。为了便于讨论我们的研究结果，我们选择了描绘的二维表示。同时应谨慎解释MDS表示，超过极限应力时，更长的距离往往更准确地显示比短距离（博尔加蒂，1997），因此总体格局依然可见即使应力高于合理范围内保证良好的配合。为了确保可信性的调查结果，结果进行了检查和确认的三维解决方案。从MDS分析的两个研究结果脱颖而出：首先，MDS的结果的二维表示的工作相关和个人之间的关系显示出明显的分裂。在右上象限，我们发现三个网络，是指与工作有关的问题，即讨论工作，协作，并征求意见。在左下象限外，我们发现三个网络代表更多的个人关系，即接触工作外，友谊和个人指导。因此，横向尺寸可以反映仪器的表现区别了伊瓦拉（1993，1995）。“花钱休息”的网络可以在右下角找到，与工作相关的个人网络不同。因此，这个网络可以被认为是一个明确的表达，也不是一个明确的工具性社会网络。更具洞察力的是，围绕工作和友谊外的社交网络在MDS维空间中处于接近的位置。这个发现让这两个表达网络之间的最高的QAP相关。
新西兰动画学论文代写:观察到的相似性
In order to detect meaningful underlying dimensions that may explain the observed similarities between the seven social networks, we used the aggregated QAP correlations as input for a Multidimensional Scaling (MDS) procedure. A three-dimensional MDS representation fit the data best, accounting for 90.5 % of the variance and the stress (an indication of fit) at the upper norm of acceptability, namely .15 (Borgatti, 1997). We visualize the two-dimensional MDS representation in 2. This two-dimensional representation accounted for 75.7 % of the variance of the scaled data, but the stress exceeds the upper stress limit (stress = .26). To facilitate the discussion of our findings, we chose to depict the two-dimensional representation. While caution should be exercised in interpreting a MDS representation that exceeds stress limits, longer distances tend to be more accurately displayed than shorter distances (Borgatti, 1997), and so the overall pattern is still visible even when stress is above reasonable limits to guarantee good fit. To ensure the trustworthiness of the findings, the results were checked and confirmed with the three-dimensional solution. Two findings stand out from the MDS analysis.First, the two-dimensional representation of MDS results shows a noticeable split between work related and personal relationships. In the upper right quadrant, we find three networks that refer to work related issues, namely discussing work, collaboration, and asking for advice. In and just outside the lower left quadrant, we find three networks that represent more personal relationships, namely contact outside work, friendship, and personal guidance. As such, the horizontal dimension may reflect the instrumental-expressive distinction as proposed by Ibarra (1993, 1995). The ‘spending breaks' network can be found in the lower right corner and as such differs from the work related and personal networks. This network can thus be considered nor an explicit expressive nor a clearly instrumental social network. What is furthermore insightful is that the social networks around contact outside work and friendship are located in close proximity in the MDS dimensional space. This finding mirrors the highest QAP correlation between these two expressive networks.
