Q-methodological exploration of environmental characteristics affecting residential security in hybrid blocks: a focus on Beijing

Research design

The Q-methodology is a systematic method that originated in the field of psychology (Zhuang and Ye, 2023) but has been widely used in different fields in recent years. The method is used to systematically investigate interviewees’ perspectives, opinions or preferences towards a given subject, and the result depicts a diversity of perspectives, opinions or preferences held by interviewees (Brown, 1996; Minnery and Lim, 2005). The Q-methodology focuses on identifying diverse views or preferences rather than generalizing findings from sample respondents to a larger population(Mouratidis and Poortinga, 2020; Rost, 2021). Other common quantitative research methods, such as R methodology or spatial modeling, seriously ignore subjects’ subjective awareness, and the sample average cannot represent the situation of most people. This research should pay more attention to “individual subjectivity”(Milcu et al., 2014). In addition, the Q-methodology can distinguish and interpret different opinions on a certain subject in a group of subjects, which is difficult to do with traditional questionnaire methods (Mouratidis and Poortinga, 2020; Milcu et al., 2014; Li et al., 2022). Although group interviews and focus groups can provide information on possible different perspectives, they lack quantitative means. Moreover, some methods only review the items by experts, and often ignore the real ideas and needs of the research objects, thus making the research objects and research content disconnected. Through the Q-methodology, the same subject can be classified in different instructions to examine the differences between the subjects from different angles and starting points (Brown, 1996; Chen, 2016).

For hybrid blocks, the sense of residential security is a rather subjective judgment. There are many factors that affect the inhabitants’ evaluation of the sense of security, resulting in obvious differences in the resident’s individual evaluation standards and evaluation results, which cannot be measured by quantitative means (Casteel and Peek-Asa, 2000). Therefore, both the objective conditions of the built environment and the social environment affect residents’ experience. Q interviews could be conducted with residents through Q-methodology to analyze the motivation of respondents’ subjective judgment on residential safety based on the environmental characteristics of hybrid blocks. Based on the advantages of Q-methodology and the purpose of this study, we adopt Q-methodology as our research method.

In an empirical setting, applying the method involves several key steps (Fig. 2). These involve: (a) the definition of a series of statements as prompts that represent the subject of the study (Q-set); (b) selection of interviewees (P-set); (c) administering the ranking of Q-sets by the different interviewees(Q-sorting); and (d) factor analysis of the Q-sorts(Brown, 1996; Milcu et al., 2014; Chen, 2016).

Flow chart of the research.

The Q sample

Firstly, the definition of the statements set is provided. Conceptually, the term “statements” refers to the collection of all opinions and issues related to our research topic. In practical terms, defining the set of opinions involves gathering relevant opinions and statements on the topic, typically through interviews or literature reviews to compile the opinions and statements forming the discourse. In the context of this study, it specifically denotes the academic discourse regarding factors influencing the sense of residential security. This research compiles and establishes its Q statements from the literature on residential security experience (Table 1, Table 2).

Following the definition of the opinions set, the next step is to select the most appropriate and representative statements to form the Q sample. This study constructs a factor design based on literature regarding the manifestation of residential security (Table 3).

Attempting to build an opinion set on the impact of hybrid blocks on the sense of security in residential areas across two dimensions – social elements and environmental design elements – the study proceeds to select the Q statements required for the Q-methodology research. The first dimension, social elements affecting residential security, includes two influencing factors: “interpersonal relationships” and “surrounding building functions”. The second dimension, environmental design elements as the second factor influencing residential security, comprises three influencing factors: “urban space,” “physical facilities,” and “landscape facilities.” “Urban space” includes two sub-factors: “public interaction space” and “road space.” “Physical facilities” consist of four sub-factors: “building,” “lighting facilities,” “public safety facilities,” and “access control.” Lastly, “landscape facilities” encompass two sub-factors: “surrounding environment” and “green spaces”.

Based on the above factor design, the study incorporates two dimensions of residential security and their five primary factors, resulting in a total of six combinations (A × B = 2 × 3 = 6). To balance the breadth of the Q sample and avoid an excessively large number of Q sample statements, the study follows the standard proposed by Van Exel (40–50 statements are advisable) (Dieteren et al., 2023). Considering the coverage of sub-factors, the study selects 4-16 statements from the discourse for each combination instead of distributing them evenly. This ensures that the total number of Q sample statements for this study is 46. In this way, 46 Q statements were selected (Table 7, Statement).

The P set

In Q-methodology, the P set refers to the individuals selected by the researcher for Q sorting, aimed at enhancing our understanding of “subjective opinions on the research topic.” Q-methodology does not employ random sampling to select participants. Instead, it purposefully selects a group of participants who are theoretically or experientially believed to be familiar with the research topic and possess insights or specific perspectives. In this study, participants are strategically selected based on two criteria. Firstly, they must reside in any of the three hybrid blocks under investigation and be familiar with the surrounding context. Secondly, they must be residents who prioritize the sense of residential security. All participants are informed in advance that their involvement in this study is entirely voluntary and anonymous. The data of all participants remain anonymous and are exclusively used for this research.

The Q sorting

In this study, Q-sorting analysis was conducted using the KADE (Ken-Q Analysis Desktop Edition) software. Ken-Q is a platform developed for online Q-sorting analysis. Prior to commencing the Q-sorting procedure, participants were acquainted with the study and its objectives. Subsequently, participants were instructed to perform a pre-sorting task. This pre-sorting involved categorizing the 46 statements into “extremely important,” “fundamentally unimportant,” and “neutral” categories related to the perceived significance of these statements to residential security. Although pre-sorting data was not recorded, this phase aimed to provide participants with an overview of all Q-sort statements before proceeding to the main sorting task. This is a common step in the Q-methodology designed to enhance the quality of Q-sorting outcomes.

Once the Q samples and P set have been confirmed, the subsequent step involves inviting respondents to conduct Q sorting. Firstly, it is imperative to establish the Q distribution. Generally, the “quasi-normal distribution” is commonly employed for this purpose. The determination of the Q distribution is subjective and lacks rigid rules to be adhered to; however, normal distribution and quasi-normal distribution offer statistical advantages.

Given that this study’s Q sample comprises 46 Q statements, the range for deciding the Q distribution is set from +4 to −4 (Fig. 3). The Q distribution is specified as 2-4-6-7-8-7-6-4-2, indicating that under +4, 2 statements are placed, under +3, 4 statements are placed, under +2, 6 statements are placed, and so forth.

Following informed consent, participants were provided with Q-statements printed on uniformly sized cards and a reference sheet depicting the initial sorting grid configuration (Fig. 3, Q-sort). Participants were instructed to arrange the statements spatially according to their subjective attitudes and perceived intensity of agreement/disagreement. To ensure procedural reliability, researchers accompanied participants throughout the task and provided immediate clarification when ambiguities in statement interpretation arose. Each Q-sorting session was strictly controlled within a 20–30 min timeframe. At the same time, in order to avoid mutual interference, the Q-sorting process was completed by all respondents individually.

Background information on respondents and the considerations they make while conducting Q sorting are typically addressed by posing questions on the recording sheet, especially when arranging statements based on the most agreement or disagreement. The questions commonly include: (1) Why did you place the statement that you most agree with under +5? (2) Why did you place the statement that you most disagree with under -5? (3) What was the primary consideration for you during the Q sorting process? The primary purpose of these questions is to elicit information about the considerations made by respondents during Q sorting. This information proves crucial in facilitating a comprehensive understanding of the thoughts underlying different opinion categories during the interpretation phase.

Research process

The choice of spatial scale in this study is a critical consideration in understanding the impact of the built environment on residents’ sense of security in hybrid blocks.

Hybrid blocks

Hybrid blocks refer to urban plots located within the block and not separated by urban roads, in which the activities of different groups of people, buildings of different ages, types and styles are concentrated in an urban plot and collide and influence each other to form the unique nature of the plot (Holl, 2014; Igualada, 2018). It is related to but not limited to “informal space” and “mixed residential area” (Igualada, 2018). This work selects three eligible hybrid blocks in Beijing, China as the research object, named Caoyuan Hutong, Xidan-Horizontal second road community and Temple of Heaven West community (Figs. 4–7) (Xu and Zhao, 2021). The research plots, which were the important development and construction area of Beijing from the late Qing Dynasty to the early period of New China, are located in the Second Ring Road of Beijing(Ruoshi et al., 2024). They have the typical characteristics of hybrid blocks, and respectively represent the three main modes of hybrid blocks in current China (Chen and Wu, 2021): (1) Hybrid blocks in the historical era of architecture; (2) Hybrid blocks with commercial, residential and other functions; (3) The highly complex hybrid layout of formal settlements and informal settlements.

Map of the three experimental blocks selected for this experiment.

a Caoyuan Hutong. b Xidan-Horizontal second road community. c Temple of Heaven West community.

a Caoyuan Hutong. b Xidan-Horizontal second road community. c Temple of Heaven West community.

a Caoyuan Hutong. b Xidan-Horizontal second road community. c Temple of Heaven West community.

Caoyuan Hutong

The Caoyuan Hutong exemplifies a hybrid block typical of the built environment era, featuring an architectural ensemble from various historical periods. Additionally, buildings of differing scales create a disjointed urban fabric, limiting efficient space use and failing to meet residents’ demand for quality public living environments. Historically constrained, the transportation network in Caoyuan Hutong mainly consists of narrow one-way alleys. While this limits road connectivity, it also hampers daily commuting efficiency. The chaotic functional layouts further complicate everyday life, reducing both neighborhood vitality and livability.

Xidan-Horizontal second road community

The Xidan-Horizontal second road community is a hybrid block that combines commercial and residential functions. It features vibrant commerce and high foot traffic, attracting diverse individuals and creating a complex social ecosystem. However, this diversity can lead to conflicts between transient visitors and local residents. The dense pedestrian activity complicates area management and requires advanced residential security strategies in this hybrid block. Additionally, the urban interface shows some discontinuity, detracting from the neighborhood’s esthetic coherence. Public spaces face issues related to limited variety and functionality, hindering their ability to meet residents’ diverse activity needs. This limitation further restricts development potential.

Temple of Heaven West community

The hybrid layout of formal residential areas and informal settlements within the Temple of Heaven West community creates a highly complex regional environment, significantly impacting the residential security. The transportation network in this block suffers from limited accessibility and a dense arrangement of dead-end roads, which severely restricts internal traffic circulation efficiency and hinders seamless connections with the surrounding environment. Moreover, the lack of comprehensive public space system and low community vitality fail to adequately meet the daily activity needs of its residents.

Since actual crime data in neighborhoods are not permitted to be disclosed in the study area, the researchers visited the management personnel of each community to identify the approximate locations prone to crime in recent years and created a security risk zone map (Fig. 7) of high, medium, and low crime probability based on descriptions from professionals. In this way, the basic information of the three hybrid blocks was obtained in detail.

Definition of respondent scope

Individuals residing in various age groups within the study area. People in general show the characteristics of diversity, their occupation, age, social status and so on have a large degree of difference (Marzbali et al., 2012a), as a research scope can better collect and analyze the impact of different groups on the security of urban built environment caused by various factors and judge (Brown, 1996; Rost, 2021; Marzbali et al., 2012a; Marzbali et al., 2012b).

Research ethics

There are human participants in this study, and the overall distribution characteristics of participants are consistent with the distribution characteristics of China’s Hybrid blocks(Zhuang and Ye, 2023; Zhang et al., 2020). The duration of this work spanned from September to November 2021, lasting for 3 months. Before participating in this research, each participant was informed of the purpose, basic research content, process, method and other basic information of this research, and also knew which of their information would be used in this paper. Meanwhile, we promise not to make specific correspondence between participants’ personal information and the research content. All subsequent data analyzes have been anonymized. The study does not involve any technological experiments with human or animal subjects, and there are no associated risks. It is classified as a project with a risk probability and level not exceeding the risk standards of ethical considerations in science and technology. Therefore, the ethics approval was granted by the relevant ethics committee (see “Ethics” statement part) and verbal consent was obtained from participants. At the same time, in order to ensure the accuracy and non-misunderstanding of the data needed for the research, all the questionnaires were written in Chinese (the first language of the place where the research was conducted), which was translated into English for easy understanding in the subsequent expressions of this research.

Analysis and interpretation

Through the above steps, a total of 22 valid data pieces were collected. Among them, there are 9 in Caoyuan Hutong, 7 in Xidan-Horizontal second road community and 6 in Temple of Heaven West community. Considering the common view of the Q method(Bertaux, 1981; Creswell and Poth, 2016; Watts and Stenner, 2012), the sample size of 22 participants in this study is in line with the standard and can reflect the population characteristics in this study area (Table 6).

After the necessary organization of the Q sorting data, it can be entered into KADE for factor analysis. Ken-Q offers two methods for factor analysis: Centroid factor analysis and Principal Components Analysis (PCA) (Li, 2015). This study has opted for the Principal Components Analysis method recommended, which is more effective for dimensionality reduction, as it focuses on maximizing variance explanation to efficiently extract a small number of principal components that capture multivariate variation patterns in participant-sorted data, while factor rotation enhances interpretability(Wold et al., 1987). In contrast, centroid factor analysis relies on factor model assumptions and exhibits lower efficiency in dimensionality reduction for complex subjective datasets(Wold et al., 1987; Lawley and Maxwell, 1962). This study employed factor analysis to cluster participants into distinct viewpoint groups (factors) based on similarities in their subjective rankings, followed by an analysis of intra-group differences in statement rankings to explain diversified subjective stances.

Initially, the software is employed to calculate the correlation matrix and eigenvalues of the matrix for the 22 samples. According to the Kaiser criterion, the decision on the number of factors to retain is based on eigenvalues greater than 1. The cumulative percentage of variance for the first six eigenvalues reached 60% (Fig. 8 and Table 4). It meets the basic requirements of Q analysis, indicating that these 6 factors have a certain representativeness of samples. Therefore, the first six factors should be retained and subjected to Varimax rotation to maximize variance.

Weight of each P-set member for corresponding Factor.

The Critical Value formula: N is the number of statements in the Q sample, and 1.96 corresponds to the Z value at the 0.95 significance level. Given that this study utilizes 46 Q statements, the critical value is calculated as (1/√46) * 1.96 = 0.289. When the factor loading for a participant’s Q sorting under a particular factor exceeds or equals 0.289, that participant is classified under that specific factor category.

After the factor sequence behind the rotation axis is obtained, it is necessary to test the factor load of each interviewee, that is, whether the factor load of a certain interviewee ranked Q is significant, in order to classify them. A respondent is classified under a category factor when the number of factor loads in the Q ranking of a category factor is greater than or equal to the critical value (the critical value in this study is 0.289).

Through the above Q-methodology data analysis, this study extracts 6 factors affecting the security of residential security. Table 6 shows the background information of 22 respondents who participated in this study. Notably, residency registration in Beijing is important information that has not been paid attention to in previous related studies. Its significance lies in its rigid linkage to social welfare entitlements: non-local residents face systemic barriers to resource access due to policy restrictions, exacerbating material insecurity. Meanwhile, residency registration can lead to “institutional exclusion” at the psychological level, weakening the sense of belonging of migrant groups to urban security.

If an interviewee is classified into a certain factor type, “X” is added after the factor load. Twenty-two respondents fell into six categories. Among them, factor 1, factor 2 and factor 4 have 4 digits, factor 3 has 3 digits, factor 5 has 5 digits and factor 6 has 2 digits (Table 5).

We use different methods to explain and characterize these factors. We mainly consider the following points to explain, characterize and compare these factors. (1) We characterize each factor according to the nature of the factor load, i.e., the type of respondent with a significant load on a similar factor. Here, respondents loaded onto similar factors have similar preferences. (2) By looking at preference patterns involving residential security in the built environment as well as prioritized social factors and environmental design elements, we characterize each factor in terms of the overall configuration stated under each factor. (3) We take account of statement scores in characterizing each factor. Here, we use the weighted average of each statement to explain the highest and lowest priority of each factor. Thus, the statement with the highest ranking (i.e., +4, +3, +2) is considered the highest priority of the factors, while the statement with the lowest ranking (i.e., −2, −3, −4) is interpreted as the statement with the lowest priority of each factor.