This field study was made to study the size of social groups among passengers travelling with lifts. The group size was observed in three types of buildings located in four countries, totally in nine multi-storey buildings. The observations were carried out manually for 12 hours on a normal weekday. Analysis results show that the daily mean group size was generally low: 1.2–1.3 persons in the offices, 1.3–2.0 persons in the hotels and 1.1–1.4 persons in the residential buildings. Hourly means differed significantly from the daily means for each building. This suggests that group size should be considered on hourly or shorter basis. In addition, goodness-of-fit tests were conducted to determine a statistical model for the group size. A geometric distribution was found to fit the data the best across all building types which simplifies modeling of passenger group arrivals in buildings.

Practical application: This article contributes to the understanding of lift passenger tendency to travel in socially connected groups in multi-storey buildings. A unique data set of group size observations around the world is reported on an hourly basis for office, hotel and residential buildings. A geometric distribution is proposed to model the group size statistically, which simplifies the use of socially connected groups when generating passengers in lift traffic simulations. The real-world data and the statistical model of the group size enable more realistic lift traffic simulations that are routinely conducted during the design phase of multi-storey buildings.

进行这项现场研究是为了研究乘坐电梯的乘客中的社会群体规模。在四个国家的三类建筑物中观察到了组的大小，总共在九个多层建筑物中。在正常的工作日中，手动进行观察12个小时。分析结果表明，每日平均群体规模普遍较低：办公室为1.2–1.3人，酒店为1.3–2.0人，住宅楼为1.1–1.4人。每个建筑物的小时均值与每日均值有很大差异。这表明应该每小时或更短时间考虑小组人数。另外，进行拟合优度检验以确定组规模的统计模型。

实际应用：本文有助于理解电梯乘客在多层建筑物中以社交联系的群体旅行的趋势。每小时都会报告办公室，酒店和住宅楼的一组独特的团体规模数据集。提出了一种几何分布来统计组的大小，从而简化了在电梯交通模拟中生成乘客时社交联系组的使用。现实世界的数据和组规模的统计模型可实现在多层建筑的设计阶段例行进行的更实际的电梯交通模拟。