Anchor tenant's influence on consumer shopping mall choice

Minimalists	Grazing	Traditionalists	Enthusiasts
Basic minimalistic needs; no impulse buying	Hedonistic and social needs, mostly entertainment and leisure-driven, prone to impulse buying	Pragmatic consumers, need of safety and atmosphere, very rarely prone to impulse buying	Demanding consumers, who pay attention and interest to all spheres of shopping center, however, are rather economical

The first segment consists of consumers who are modest and efficient in their purchases and strive to satisfy their most basic needs; this group is not prone to spontaneous spending. These people view purchases as a functional process. This group is characterized by a lack of desire to have fun in a shopping center. For them, the primary task is to purchase only the goods they need.

The second group considers the shopping center as a place for social communication and leisure; this group was called grazing, these are calm, relaxed consumers, satisfying their hedonistic needs. Such entertainment lovers most often visit the entertainment area of ??the shopping center and the atmosphere prevailing there plays a large role for their shopping center experience (Sit et al., 2003). They are mainly customers of food courts, cinemas and other entertainment venues located in the shopping center. These people often compare the same product in different stores looking for the best value for money. Moreover, among them shopaholics are often found. Not shopping for a long time, they are dissatisfied. However, they often spend more money than they planned.

The third group of consumers is traditionalists, real pragmatists. This group as a whole attaches great importance to the atmosphere, appearance, safety and tranquility of the shopping center. Their main activity is precisely shopping and not visiting a movie theater or food court. Although they experience the pleasure of shopping, but in small quantities. People of this group are very economical and are always looking for the most profitable purchase.

Finally, the fourth type of consumer is a customer, who is also called an enthusiast or demanding customer. This category is interested in all aspects of the shopping center, from shops to all possible services, presenting a number of quality requirements to them.

Research question

In this section the objectives of the paper will be discussed as well as research question and hypotheses. Also, methods that we chose to adhere to will be mentioned as well as the advantages that these methods have.

As seen from the literature review section, the literature on the topic of shopping malls' tenant mix and other characteristics is vast, however, there is no study which would focus on the connection between anchor tenants in the shopping centre and customer's choice if the particular centre.

The general task of the thesis is to find out where there is a connection and if so, what it is like in particular. In order to achieve this goal a survey should be conducted which will help to obtain all the useful data about consumer preferences. Also, all the necessary characteristics like GLA, location of the mall, its anchor tenants number and percentage should be gathered manually from the Internet and be consolidated for the further use. Then analysis is needed to be conducted in order to come up with the results and try to find any correlation between the data obtained.

Considering the methods used in our research we follow the general rules for the quantitative study. Regarding the purpose of the study, qualitative methods are aimed at to find and gain a deeper understanding of the problem being studied, while quantitative methods are aimed at testing hypotheses and coming up with general conclusions based on which certain results can be predicted (Chrysochou, 2017). Quantitative research involves obtaining clearly structured information about a large number of research objects. This type of studies provides accurate quantitative (numerical) values of the studied indicators. As a rule, quantitative research involves the use of the selective method, that is, not all objects included in the target group are examined, but only a part of them (which makes up the research sample).

We chose to adhere to the survey method in order to get more as much structured data as possible on the subject of consumer shopping mall choice. The purpose of the survey study is to outline the characteristics of the target sample, as well as understanding people's attitude, perception, motives, choices and, in general, collecting their opinions on the matter of interest. Survey research integrates sampling, question design and data collection.

According to Chrysochou (2017), a survey study has many pluses. It can gather opinions on a large sample in a short time and at a low price per participant. It is relatively easy to create and analyze, and allows the researcher to be flexible regarding the type and number of questions used. A survey study is also more suitable for generalizing the matter in question. However, it is important to understand that a survey study only provides estimates, not exactly accurate measurements of the population.

Another advantages of the survey method include:

· The influence of the personality of the interviewer is eliminated and it does not affect the data received;

· Frankness of the respondent. Respondents via the Internet give more honest answers, being anonymous;

· Convenience. The respondent decides when to take part in the study, and this improves the quality of answers;

· Speed. It takes less time to fill out and process the questionnaire than with a traditional interview.

Cost reduction in conducting online surveys that give an idea of ??the majority opinion is possible thanks to a selective method that ensures representativeness of the study. The use of the selective method reduces the number of respondents, but not at the expense of the quality of the information received. The sample population (the number of people who will be interviewed) reflects the characteristics of the general population (all people who could potentially be interviewed). A properly compiled sample ensures reliable information while minimizing the cost of the survey.

It is the survey that provides information in the most convenient form for further processing. The main goals of the survey is testing the hypotheses. The first hypothesis is derived directly from the research question and sounds as follows:

H0. People's average spendings per month do not influence their choice of the anchor store.

In this hypothesis we will try to investigate the influence which spendings have or do not have on the customer preferences.

H1. The respondents' gender does not have an effect on their perception of the different stores and its importance in the overall center's tenant mix.

The second hypothesis is related to the age group which a particular respondent belongs to. We believe that there will be an influence of gender on certain store choices.

H2. People's age does have an effect on the tenant mix. Young respondents tend to rate the food courts and mass market sores way higher than all the other tenants in the mall.

The last hypothesis is also related to the age of the consumer. We believe that it is the youngest generation that is most prone to the influence of anchor tenants, especially considering mass market clothing store profiles and food tenants.

anchor tenant consumer shopping

Methodology

Although the previous assumptions from the studies performed on the related topic concerning the anchor stores' influence were proven to be positive there are still yet some points to be investigated such as the specific demographic characteristics of a customer and its effect on his or her choice in order for businesses to construct a perfect tenant mix and employ the results for further marketing strategy development.

Highlighting the objectives of the study that were formulated before, the general task of the thesis was to find out where there is a connection and if so, what it is like in particular. In order to achieve this goal a survey was conducted which has helped to obtain all the needed data about consumer preferences. Also, all the necessary data like the types of anchor tenants and main attributes were gathered manually from the official websites of the malls and were consolidated for the further use. Then analysis was be conducted in order to come up with the results and try to find any correlation between these data in terms of age, gender, and average spendings per month.

As it has been stated earlier the paper aims to investigate whether the specific anchor stores can influence the consumers' choice of the shopping center based on their demographic profile and if there are any trends concerning the popularity of some anchor tenants. Further it can be possible to come up with the results that can be useful for many businesses including the shopping centers and anchor stores for marketing use. The consumers' perception of the specific anchor stores in the shopping malls was observed. In this chapter, there will be discussed main steps dedicated to quantitative part of the research including the research strategy, research type, data collection tools and the processes, the research approach, analysis tools, sample selection and the limitations to the analysis. In the previous section, it has been indicated that the filed lacks the literature on the topic of consumers' shopping center experience and the influence of anchor tenants the topic itself lacks the studies on the related topic. Consequently, the proposed study will try to complement the existing literature on this topic.

The proposed research type is set to be both qualitative and quantitative. Both methods were applied towards different aims of the research and required multiple research tools. According to McLeod (2019) quantitative approach is more applicable whenever a social phenomena is to be investigated and measured, it uses the numerical data mostly and the gathered data can be converted into graphs (tables) and be statistically tested. In most cases, quantitative research tries to reject or prove some hypotheses, whereas qualitative approach is more applicable for the studies which focus on a deeper analyzing and conceptualizing of the problem gathering multiple opinions which generally do not include numbers (Shona McCombes 2019). Moreover, the approaches are used with different methods of data collection tools. As for this study, as it has already been mentioned there was a survey conducted, which contained the specifically prescribed questions on demographic characteristics and personal preferences.

In the study there were several methods used which include the search for the consumer behavior patterns in order to construct a questionnaire, survey itself and a statistical analysis of gathered data in SPSS. The hypotheses testing process was performed using the 4 different tools: first, there was a descriptive statistics conducted for all of the demographic characteristics and their answers towards other questions in order to generalize the information and find out the main trends, after that the relationships between the variables were checked with the use of t-test and ANOVA, further for some variables it was needed to carry out a deeper analysis (multiple comparisons). Such research design was previously employed by Muhammad F. Ibrahim and Ng C. Wee in 2002. The study analyzed the importance of mall's attributes among the consumers. The researcher used a mixed method design, where they have gathered the data through the analysis of existing literature and with the help of survey, they further proceeded to summarizing the gathered data and identifying the main patterns and trends between the specific variables. Further they have employed Factor analysis and analysis of variances in order to identify the significance of difference between the identified groups from the descriptive statistics. When analyzing the consumer behavior for the proposed study the pattern remains partly the same.

The paper employed two different data collection methods: primary (survey) and secondary. Survey is used to capture the perceptions and attitudes of a certain population on the specific issues, it helps to easily gather the information on the demographic profile of the polled. In general, such questionnaires are considered to be one of the most used methods for a data collection as it allows to gather the qualitative and quantitative data as groups so it is convenient to use it further for the data analysis.(Fatima Furaiji, 2012). At the rise of tech intervention in companies' operation processes it was possible to get access to a broader volume of various data which allowed to analyze the market better, the consumers' perception which overall provided the companies a higher quality statistically approved information for a business decision making process, mostly for identifying marketing strategies that will fit the company. Overall, the main impact of the implementation of this renovation led to a more reliable and in-depth insights on the consumer preferences and its behavior (Chrysochou, 2017). A structured survey is commonly employed tool for data collection in most studies when investigating the issue of consumer behavior whether it is the process of online shopping or an in-store shopping in the malls. Sujo Thomas (2012) in his study on the customers' approach towards the various malls' attractiveness applied the survey as a main data collection tool designing it based on the semantic differential scale. Such scale is widely used when researches conduct a questionnaires and interviews for marketing researches (considering business oriented spheres) in order to analyze the person's perception towards a certain product or brand in general (Osgood, 1957).

While proceeding to the first step of the research it is essential to gather the information needed for the development of the survey's questions. As the topic of the concern is not widely studied there are not enough number of survey examples which can be used as a pattern for the further research, consequently, most of the questions are going to be self-developed based on the existing literature on the related topic and on the additional data gathered using the survey with open questions related to common perceptions of customers towards the shopping centers. It is of a high importance to carry out the procedure needed for the development of the questions for the primary questionnaire according to Chrysochou (2017), the pre-test survey was conducted with a smaller number of people in order to identify the main trends or gaps which helped to develop the final survey. It is needed to construct the questions for both surveys in a way that the connection with the research question won't be lost and the focus of the study won't be moved to another subject.

As for the proposed study the main open questions for the pre survey procedure were set to be the following:

1. What is the purpose of visiting a shopping mall for you (can be several)?

2. How can you describe a customer that is most commonly seen in the mall (age)?

3. What factors can potentially not make you visit some of the malls in the city?

4. What kind of shop do you usually visit in the mall (can be one or several types)?

These questions were asked to a small group of 7 people of one age group in the university at one period of time.

After gathering the information needed for the main survey it can be concluded that:

85% of respondents usually visit the mall to go shopping. The rest of the observed have an entertainment (cinema and restaurants) as a main purpose. Considering the age group of visitors that most often attend to be at the shopping centers, most of the respondents stated that the population is not strictly fixed between one age group, yet, they claim that the majority is obviously varies between teenagers and millennials (1981-1996).

Stating all the reasons due to which some shopping centers can be less attractive to some respondents, they can be grouped as:

• Remoteness of the mall from the metro station (or city center);

• Poor variety of anchor tenants;

• The size of the mall (GLA)

• Traffic congestion

Moreover, the data for the questionnaire was gathered with a secondary data collection process. The quality of the survey can be enhanced if the additional data is provided. In order to obtain the reliable data for formulating the precise questions for the survey that would reflect the initial concerns of the study it is important to use several sources. Despite running a survey that would identify the main points of evaluation the consumers' perceptions towards shopping centers and which is used for formulating the main questionnaire, additional research of existing studies that used the same methods and analyzed customer behavior was conducted. Spiggle and Sewall (1987) have analyzed the retail aspect of the shopping centers, for their study they have also carried out the questionnaire which included the questions on the purpose of the visit, most desirable stores and attributes of the mall. Such demographic factors as age, gender, income and education are the key parameters that are usually used in the studies of consumer behavior (Zeithaml, 1988; Arnold and Reynolds, 2003; Sinha and Banerjee, 2004; Fox et al., 2004; Carpenter and Moore, 2006). Most studies that used the same data collection method gather the demographic characteristics of the respondent in order to further identify some additional sufficient correlation and tendencies that would help marketers for developing the strategies. Such parameters as age, gender and average spendings (for certain studies) turned out to be the most employed categories by the researches. Subsequently, the information obtained was then used to elicit the further data.

The survey consisted of questions asking respondents about their personal perception of the shopping centers, demographic profile and the perception of different anchor stores. Those questions were designed using the nominal and interval scales. The nominal scale was applied for the questions that cover the demographic characteristics of the study. The information obtained was further converted into groups and labeled as various variables (Cooper & Schindler, 2011). As for the interval data it was decided that in order to capture the perceptions of the polled towards the importance of the particular centers' attributes, people were asked to rate the category on a scale from one to five (important/not important). It was decided to adapt the Likert Scale for the survey's questions as this scale is widely used for measuring the attitudes in the research studies, especially in those that focus on the consumer behavior(Tom Tulles, 2013). The reason why the respondents are also asked to rate the shopping center's attributes which influence the consumer is to understand the common perception of the anchor stores among the other features of the mall. This provided the information on the overall anchor tenants' role in the mall among the people. After that, the respondents rate the importance of a specific type of store that they prefer whenever they visit the mall. The categories of the stores were self-developed after analyzing the tenant mixes of various shopping centers in Saint-Petersburg. Using the official publicly available information from the websites on the variety of different stores in the most popular shopping centers, the 10 categories of stores were identified. The following acronyms are used for the convenient visualization in the tables and outputs of the analyses.

CP - cinema park

FC - food court and restaurants

BS - beauty retailers

MM - mass market stores

LB - luxury brands

NMM - non mass market stores

TR - tech retailers

CS - cafes, bakeries and other independent coffee shops

S - service points (pharmacy, banks, mobile retailers)

K - goods and entertainment parks for kids

For the proposed study it was also decided to analyze top visited shopping centers in Saint-Petersburg. The list included: «Galeria», «Citymall», «LETO», «Piterland», «GrandCanyon», «Ohta mall», «Europolis», «London mall», «Raduga center», «Nevsky center». The above mentioned malls is based on the Kudago Spb and TripAdvisor portals ranking (2019) as the most high rated places. However, some of the malls were excluded as their parameters didn't fit to the average shopping centers: such big centers as «Mega Parnas» and «Mega Dybenko» as their gross leasable area and the main anchor stores are different from all of the others in the list (e.g. IKEA, OBI and etc.) were not included, moreover, it was decided not to observe these places for the analysis because both of them are located far from the city and not everybody has an easy access to get to these places. Moreover, for including these centers it will be required to implement more criteria for assessment because of their GLA and the specific anchor stores which are located only in these malls. For the main questionnaire the sample size is expected to be approximately200 respondents. The questionnaire was built using Google Survey tool and was spread through main social media platforms (Instagram, VK and Twitter). For the proposed paper, it was possible to gather the data only once, consequently, the study is cross-sectional. Previous studies on the related topic have also chose to make the research cross-sectional as it is rather rare for such phenomenon as consumers perception of shopping mall and even anchor tenants to be changed in a short period of time. (Micheline Naude, 2018). In some of the previous studies on the topic of the consumer behavior the data was also gathered once, so it was concluded that there is no significant need to consider a longitudinal study. Although, it is known that a bigger time horizon represents more valid results.

Given the data obtained from the preliminary data gathering process described above, the final questionnaire is constructed in the following way:

As for the first part of the survey, it is needed to capture the demographic profile of the respondent.

1. Choose your age?

· Under 20

· 20-30

· 30-45

· Above 45

2. What is your gender?

· Male

· Female

3. What are your average spendings per month (Russian ruble)?

· 0-10

· 10-20

· 20-30

· 30-50

· More than 50

Second stage: perception of different attributes and anchor stores.

On a scale from 1 to 5 where:

1 - is not important at all

2 - not important

3 - moderate

4 - important

5 - highly important

8. Rate the key characteristics of the mall which you consider the most important.

• The popularity and good awareness of the center

• The presence of top recognizable brands in the mall (Zara, Mango, Nike, Vans and etc.)

• The range of different types of shops and services (the presence of tech stores, clothing, restaurants, entertainment activities, sportswear, jewelry boutiques, banks and etc.)

• Convenient parking area and a possibility to easily find a parking lot

• The location of the mall (remoteness from the metro station or the city center)

• The absence of traffic congestion during any time of the week

• Well-developed public transport infrastructure around the area of the mall

• The location of the shopping center is in a good, safe and equipped area

• A good number of positive reviews of the mall on social platforms, from friends and relatives

9. Using the same scale rate importance of different anchor store types.

• Rate the importance of the cinema park in the shopping center

• Rate the importance of the food court and a variety of restaurants in the shopping centers

• Rate the importance of the beauty stores (Rive Gauche, L'etual, Jo Malone, Sephora, MAC, NYX, Bobbi Brown, I'le de Baute, Yves Rauche, Organic Shop and etc) in the shopping center

• Rate the importance of the top mass market stores (Zara, Oysho, SportMaster, Snezhnaya Koroleva, Reserved, Bershka, Mango, Nike, H&M, and etc) in the shopping center

• Rate the importance of the luxury brands (Armani, Furla, Chanel, Tous and etc) in the shopping center

• Rate the importance of the independent non-mass market stores for middle class population (Daniel Wellington, all of the jewelry stores like Sunlight, Bukvoed, Casio and etc.) in the shopping center

• Rate the importance of the large tech retailers (MVideo, DNS, Restore etc) in the shopping center

• Rate the importance of the small cafes which are independent from the food court area (Starbucks, Coffee shop, Makaronika, Bushe) in the shopping center

• Rate the importance of the service departments (Bank department, pharmacy, mobile operators) in the shopping center

• Rate the importance of the sores dedicated to kids (clothing, activity parks and etc) in the shopping center

The analysis has been conducted via the SPSS application (version 25). The data analysis process itself can be divided into four stages. In order to test the outlined hypotheses it was decided to opt several methods for the obtained data: compare means, independent samples t-test, analysis of Variances (ANOVA) and further, the multiple comparisons analysis.

The data on demographic profiles of the respondents which include gender, age and spendings were employed as independent variables in the analysis. The variables must be presented as an interval data, consequently, it was grouped in the given way:

1. Gender: 0 - male, 1 - female;

2. Age: 0 - under 18, 1 - 18 -- 30, 2 - 30 -- 45, 3 - above 45;

3. Spendings: 0 - under 10, 1 - 10 -- 20, 2 - 20 -- 30, 3 - 30 -- 50, 4 - above 50.

The dependent variables were all of the obtained rating scores for the various anchor stores.

The first stage of the analysis was to summarize the data on the shopping mall's attributes and anchor tenants types using the descriptive statistics by groups. Average mean scores were calculated both for shopping mall's features and anchor types. As it has been mentioned earlier, from the data collected for designing the main survey anchor tenants were identified as one of the key attributes that affect the customers' perception of the mall. The table of mean scores was made for justifying the importance of anchor tenants presence in the mall among other attributes. Further, the same procedure was carried out for the anchor types to identify the most rated positions in general. In order to justify the hypotheses that claim that the demographic characteristics can influence the perception of the various anchor store types there was needed to compare the means of average scores given between each of the groups of independent variables (e.g. male and female). The descriptive statistics of means allowed to prove the assumption which were set out before, yet the hypotheses were not entirely proven on this stage. In order to identify for which dependent variable the difference that was previously depicted from the descriptive statistics turned out to be more significant the methods of t-tests and ANOVA was used.

Before conducting the tests mentioned earlier it is important for the data to meet the required assumptions, so that the outcome of the analyses can be reliable (Bruin, J. 2006).

1. Dependent variables are continuous (5 values of Likert Scale)

2. Independent variable is categorical (male/female)

3. The observations are independent

4. The data sampling for the survey is random

5. There are no significant outliers that can violate the coherence of the analysis

6. The data is normally distributed (all the variables tested through normal Q-Q Plot, due to the fact that the sample size of the survey is big the Shapiro-Wilk test was neglected for normality test)

7. The homogeneity of variances is approved (the samples are equal within the groups)

Moreover, whenever there are multiple questions that are designed using the Likert Scale, it is important to estimate the reliability of the data. This can be achieved when calculating the Cronbach alpha. (Joseph AG, Rosemary RG, 2003). This measure can be calculated manually using the formula, yet, it is easier to conduct it with the SPSS as all of the analysis for the proposed study were done there.

Cronbach alpha is a coefficient that is used for evaluating the internal consistency of the the given data set of scale, in most cases it is used with the data of multiple-item measures. The interval for the coefficient varies between 0 and 1, the higher Cronbach alpha - the more consistent the measurement is. The advised Cronbach alpha coefficient should be higher than 0,6. There is also a condition that can influence the coefficient that is the number of items used for the measurement: a lower value can be identified due to the low number of questions (Adeniran AO, 2019).

With the obtained data sets the coefficient was calculated (the extended table is paced in Appendix 1). The Cronbach alpha for the observed data set items equals to 0,647, which is higher than 0,06. The number obtained can assure that the there is a consistency of the questions which ask the consumers on the importance of various anchor tenant stores.

In order to analyze the people's attitude towards the specific anchor tenants based on the gender of the respondents, the two-sample t-test was conducted. Independent samples t-test is a technique that is used to analyze the data in order to capture the significant differences between the means of two variables. It is set to investigate the influence of independent variable on a dependent (Cynthia Astono, 2014).

The structure of the analysis includes the following conditions:

1. H0: µ1 = µ2 («the two variables' means are equal»)

H1: µ1 ? µ2 («the two variables' means are not equal»)

2. The confidence level for the test is set to be 95 % (б = 0.05)

3. The null hypothesis is rejected if the p-value is less than 0,05

The null hypothesis is accepted if the p-value is greater than 0,05

4. If the H0 is rejected, the variables means are not equal, which means that the observed independent variable influence the dependent variable.

If the alternative hypothesis is rejected, the variables means are equal, consequently the independent variable doesn't affect the dependent.

Considering the variables age and spendings the One-Way ANOVA test was employed. The reason why the independent samples t-test was not used for these two categories was due to the inability of this test to compare more than two groups. ANOVA test is used when the means of more than two groups are to be compared (Mouhamadou Thile Sow, 2014). Analysis of variances is widely used in management studies and social sciences, specifically for marketing research of consumer behavior. The concept of the One-Way ANOVA is relatively similar with the t-test for independent samples, as for ANOVA, it tests if the independent, which includes the groups (more than two) affects the dependent one. The major difference in tests is the number of groups observed (Green & Salkind, 2012). As well as in the t-test, the required assumptions for the data set are met.

The hypotheses formed in the proposed analysis for both of the independent variables left (age, spendings) is the following:

1. Consumers' age does not influence their perception of the specific anchor tenant.

2. Consumers' average spendings per month do not affect their choice of the particular anchor store.

Technically the null hypothesis can be visualized in the following way.

H0 = µ1 = µ2 = µ3 =... = µk (all the means are equal)

The confidence level for the test is 95% with alpha equals to 0,05.

The determination of the results is performed using the significance level from the table.

If the p-value is less than 0,05 then it is reasonable to reject the null hypothesis.

Yet, with the given results from the One-Way ANOVA if there is a significant difference, the analysis can not illustrate the results on which particular groups had differences. The F-test from ANOVA can only show if there is a difference between the groups or not (Bruin, J. 2006).

In order to investigate which particular age group or which interval of average spendings has differences in rating the anchor types, the multiple comparisons analysis was performed. Alongside with the Sheffe's, Dunnett's and Bonferroni multiple comparisons tests the chosen toll Tukey post hoc test (also known as Tukey HSD) provides the detailed results on the differences in the means after the ANOVA test has shown the positive outcomes regarding the significance in the differences. The Tukey method is the most commonly used test for simple pairwise comparisons of equal sample sizes, it observes each of the possible pairs grouped from the independent variables (in this case, the groups of different ages and different amount of spendings) and tests them against the control variables (Altman D.G. 1991).

For the needs of the next steps of analysis it is needed to develop more precise hypotheses.

As for the first variable, which is the age, it was decided to rename the groups into letters for a convenient use: A - under 18; B - 18 -- 30; C - 30 -- 45.

H1: Group A and Group B do not have a difference when rating an anchor tenant

H2: Group A and Group C do not have a difference when rating an anchor tenant

H3: Group B and Group C do not have a difference when rating an anchor tenant

Considering the variables spendings the intervals are also converted into shorter acronym:

1 - under 10; 2 - 10 -- 20; 3 - 20 -- 30; 4 - 30 -- 50; 5 - above 50.

For the Tukey analysis of spendings the alternative assumptions are also developed based on all of the possible pairwise comparisons that can be constructed from the four different groups.

The degree of freedom for the analysis is set to 95%, with alpha = 0,05.

Nevertheless, the results from the study that were obtained can not correspond the full reliability. It is highly necessary to also take into consideration all of the possible factors which prevents the full reliability of the research. There are some major limitations to this study. First of all, one of the most common drawback of each most study is a small sample size. Indeed, a wider diverse population that is to be surveyed can represent more points of view and give a more reliable outcome. It is also important to include more variables for the observation which is more shopping malls and a wider dataset, so that the problem is studied from various points of view. To create a better picture of the observed object it is important to take into account all of the aspects that can influence the customer. Since, the research was based only on the Saint-Petersburg region, the results might not be useful foe other regions. Due to inability to increase the area studied, for the future research it is better to analyze other regions as well (for instance Moscow or other big administrative cities, since they are in a way similar to St. Petersburg market). Moreover, it is known that sometimes due to the character of the survey method, there is always a possibility of biased opinions of the polled which can be influenced by unlimited number of factors. For a more complex study it will be beneficial to include more characteristics to the demographic profile and analyze the issue in terms of these new features. Although it was stated that the research will be cross-sectional and for now it is not needed, moreover, impossible to conduct a longitudinal one, the last can be an advancement to this research in order to generalize the main patterns of consumer behavior towards the shopping centers.

Results and discussion

In the previous section, all of the methodological aspects of the research were identified and discussed which included the details regarding the samples, the data collection tool, the analysis tool with the help of which the empirical research was performed.

In this section all of the results from the performed analyses would be presented and further discussed. As it has been stated before, the analysis of the data obtained is divided into 4 stages: the descriptive statistics, independent samples t-test, One-Way ANOVA and multiple comparisons.

Before testing the hypotheses on the relationship between the consumers' demographic profile and their perception of shopping mall's stores, the descriptive statistics of the data obtained from the survey was performed in order to summarize the data gathered, come up with the preliminary results on the differences between some of the groups observed on the overall perception of some attributes.

As it has been discussed in the analysis of the existing literature on the specific shopping mall attributes, anchor tenants do have a sufficient influence on the shopping mall's overall customers flow and their perception of various malls (Tarun Kushwaha, 2017). The descriptive statistics analysis was performed in order to summarize the data obtained and identify the specific features of the variables that are being observed.

Table 1

T1	Attributes	Mean	Std. dev
1	AW	1,82	0,85
2	TRB	4,41*	0,69
3	DTS	4,29*	1,11
4	PA	4,11	1,08
5	R	3,04	1,23
6	TC	3,91*	1,14
7	PT	3,27	1,13
8	SA	2,97	1,33
9	PR	1,86	0,796

Table 1 (T1) depicts the results of the first descriptive statistics performed on the perception of the various shopping mall's attributes among the respondents. The number marked with the asterisk represent the highest mean values. People were asked to rate each of the mall's attributes on a scale from one to five in terms of its value to each person. The table illustrates the average means of the rating scores of each mall's features across all of the respondents without specifying their demographic profile characteristics. Among all of the presented shopping center's characteristics it can be seen that the categories TRB (top recognizable brands), DTS (non mass market, independent brands) and TC (traffic congestion) were rated as 4,41; 4,49 and 3,91 respectively which makes these attributes more valuable for the respondents rather than the rest features. The presence of top recognizable brands, which in theory stands for the anchor tenants' presence, had the highest score among others - 4,41, whilst the least important attribute for the shooting mall turned out to be the popularity of the shopping mall (reviews on social media and etc.) - 1.82. The decryption of each acronym for the attributes of the mall and anchor tenants acronyms are presented in the methodological chapter. Consequently, it can be concluded that the consumers' behavior towards the mall' attributes is influenced by the retail category mostly (specific stores), categories related to transport, location and customers impacts the choice less.

Given that, the anchor stores' effect on people's choice was determined, the analysis of the perfect tenant mix for the mall can be conducted also using the mean scores. Further, the second descriptive statistics analysis was performed for the anchor types without specifying the demographic characteristics of the polled.

Table 2 outlines the mean scores for each of the anchor types. The number marked with the asterisk represent the highest mean values. It can be revealed that food courts, mass market stores, luxury brands, tech retailers and service points were rated as most valuable anchor stores in the mall. Yet, some were estimated higher than all of the others, such as mass market stores (4,48). Food courts follow immediately after the mass-market stores with an average score of 4,45, having the difference with each other by 0,03, which makes these positions almost equally valued. The kids stores and entertainment parks position was rated as the least valuable for the surveyed. Nevertheless, the fact that the demographic profile of the respondents is not entirely diverse in terms of age, the overwhelming majority of the polled can be referred to the younger population (high-schoolers, university students and post-graduates), which can describe the loss of interest in such stores.

For confirming the assumptions of the study and revealing the differences in the scores between the specific groups observed based on their demographic characteristics, it was decided to perform the same procedure with the descriptive statistics. The mean scores are calculated for each of the anchor type, yet the main concern is on the categories which had a higher mean score on the previous table.

Table 2

T2	Anchor type	Mean
1	CP	3,06
2	FC	4,45*
3	BS	4,19
4	MM	4,48*
5	LB	3,93*
6	NMM	3,12
7	TR	3,76*
8	CS	3,04
9	S	3,91*
10	K	2,15

The table 3 illustrates the main tendencies based on the gender of the respondent.

On average, such categories as luxury brands' stores, tech retailers, beauty stores were apparently rated significantly different by people of different genders. Moreover, beauty retailers and tech retailers turned out to be important categories, yet not for both male and female respondents. Apart from all of the rest anchor store categories, these three types of stores had the most substantial difference in average score of importance. The category Beauty store was rated differently by male and female respondents as 3,19 and 4,46 respectively, given that the margin equals to 1,27. While calculating the mean scores for the Luxury Brands, the average scores for women was 4,04 and for men 3,52, consequently the difference is 0,52. As for the last analyzed type of store, which is Tech Retailers, it can be concluded that the margin is 0,63.

Table 3

Groups statistics (means) T3
Anchor type	Gender	Mean
CP	0	2,95
	1	3,09
FC	0	4,43
	1	4,46
BS	0	3,19
	1	4,46
MM	0	4,57
	1	4,46
NMM	0	3,02
	1	3,14
LB	0	3,52
	1	4,04
TR	0	4,26
	1	3,63
CS	0	3,07
	1	3,03
S	0	3,71
	1	3,96
K	0	2,10
	1	2,16

From the results obtained it can be stated that there is a difference between men and women preferences of anchor stores when rating the Beauty retailers, Tech retailers and Luxury brands. Yet, considering the margins of the means scores between both genders, for this step of the analysis it can be concluded that the Beauty Stores category is valued significantly different by men and women as the difference in mean scores is the highest among the rest two. However, for the rest of the anchor types, the differences in mean score are quite minor, which leads to the fact that respondents choice of these categories is not affected by the gender of the surveyed.

Further, the analysis for the second variable which is the age of the customer was performed (the full table is placed in Appendix 2). For the two age groups the results of the average score turned out to be different. The respondents aged under 18 years old and between 30--45 apparently had different perception of the specific anchor stores in particular the Cinema Parks, Food Courts and Mass Market stores. The difference between these age groups were the following 0,85 for the CP, 1,2 for the FC and 1,06 for the MM. It can be seen that, yet these three anchor stores represented the differences in the means, for the Food courts and for the Mass Market stores this difference is higher. Further these data was be tested on the significance.

The last variable that was analyzed is the average spendings per month, there are five groups identified and the means between these groups were analyzed. It was decided not to attach the table with the mean scores for spendings as the results didn't include anything significant, moreover, the size of the table is quite big. From the results obtained, it was concluded that there were no significant differences between the five groups and their score identified. The maximum margin that was identified was equal to 0.66, yet overall it still can be stated that the differences are not as sufficient compared to other variables. Further, the ANOVA test was performed for this variable.

On the previous stages of the analysis the main differences between the groups were determined, then it is planned to determine the statistical significance of those differences between the groups using the Independent Samples T-test and One-Way ANOVA. The assumptions on the reliability of the data and the settings of the analyses (confidence interval, variables and significance level) are considered in the methodological chapter.

In order to identify the significance of the differences in consumers' choice of the particular store type based on their gender, the two-sample t-test was performed.

The hypothesis that was being tested: «The respondents' gender does not have an effect on their perception of the different stores and its importance in the overall center's tenant mix».

H0: there is no difference in the means of the scores given to the specific anchor type between the male and female respondents.

The results from the analysis are represented in the table 4. The number marked with the asterisk represent the p-values which represent the significance. As the p-values of the Levene's test on all of the observed categories depicted values that were greater than 0,05 then for the t-statistics equal variances are assumed. It can be seen that for three out of ten positions (anchor types) the difference between the groups are significant. There is a difference in scores given by men and women for Beauty stores, Luxury brands stores and Tech Retailers, the p-values are 0.000, 0.022, 0.001 respectively. The identified values are less than the significance level that was set earlier as 0.05, which means that for these positions the null hypotheses can be rejected. There is a difference in the means of the scores given to the Beauty retailers, Tech retailers and Luxury brands anchor types between the male and female respondents, yet in the previous steps it was determined that for Beauty stores the margin in means was 1,27 which was the highest among the others. Thus, it can be stated that for the category Beauty store the difference is more significant than for the Tech retailers or Luxury Brands.

Table 4 (the output from SPSS is in Appendix 3)

T4	Sig.	t	Sig. (2-tailed)
CP	0,761	-0,671	0,503
FC	0,988	-0,206	0,837
BS	0,025	-6,103	0,000*
MM	0,132	1,039	0,300
NMM	0,705	-0,510	0,610
LB	0,025	-2,359	0,022*
TR	0,092	3,415	0,001*
CS	0,503	0,202	0,840
S	0,078	-1,140	0,256
K	0,800	-0,404	0,687

p-value > 0,05 - no statistically significant difference

*p-value < 0,05 - there is statistically significant difference

Considering other demographic characteristics that were gathered, it was decided to check the statistical differences using the One-Way ANOVA, due to the fact that such variables as age and people's spendings per month are divided into three or more groups.

The assumptions for this method are the same as for the independent samples t-tests. For the given variables such as age and spendings, the following hypotheses were tested:

1. People's average spendings per month do not influence their choice of the anchor store

2. People's age does not influence their choice of the anchor store and doesn't influence the tenant mix

H0 = there is no significant differences between the age groups of the surveys and their score given to the specific anchor store.

Table 5 illustrates the results of the ANOVA for the independent variable `age'.

Table 5 (The output from the SPSS is in Appendix 4)

ANOVA (age)
T5	F	P-value
CP	3,115	0,047*
FC	2,632	0,034*
BS	0,148	0,862
MM	3,038	0,050*
NMM	0,301	0,740
LB	1,899	0,152
TR	0,088	0,916
CS	1,336	0,265
S	1,310	0,272
K	0,028	0,972

*p-value < 0,05 - there is a significant differences between the means of one of the observed groups

p-value > 0,05 - there is no differences between the means

The number marked with the asterisk represent the p-values which represent the significance. Initially, the null hypothesis stated that there are no differences between the means of three age groups. From the results obtained the null hypotheses are accepted for the BS, NMM, LB, TR, CS, S, K categories, while for the CP, FC and MM they are rejected. There was a statistically significant difference identified between the age groups when rating an anchor type as given by ANOVA. The p-values for the categories represented the differences: for the cinema parks - 0,047, for the food courts - 0,034 and 0,050 for the mass market stores. The numbers presented are all less than the set significance level, which is equal to 0,05. Given that, for the two categories mentioned, there is a differences in the means between the two groups of age. Thus, it can't be entirely stated that any type of anchor store has an equal value for person of any age.

Considering the results obtained it is known that between some groups there are some statistical differences. Yet, the analysis of variances does not identify the exact details needed for the identification of the groups which represented the differences.

However, it is possible to identify the exact groups which have the differences in the means using the Tukey post hoc test (Olleveant N.A. 1999). Like in the previous analyses, it was assumed that the data contained the observations which were normally distributed. Table 6 demonstrates the matrix of all pairwise comparisons between the age groups.

Table 6

Multiple Comparisons (T6)
Dependent variable	Age	Age	P-value
FC	0	1 2	0,373 0,027*
	1	0 2	0,373 0,064
	2	0 1	0,027* 0,064
MM	0	1 2	0,079 0,043*
	1	0 2	0,079 0,325
	2	0 1	0,043* 0,325

*P-value < 0,05 for a significant difference between the groups

The number marked with the asterisk represent the p-values which represent the significance. As a result, taking a look at the age groups 0 (under 18) and 2 (30 --45) it can be seen that for both dependent variables, there is a significant differences in the means between the age group 0 and age group 2. The p-value is significant when it is less than 0,05, in this case the p-value is 0,027 (<0,05) for food courts and 0,043 (<0,05) for mass market stores. The above results illustrate that the young respondents have given a higher score for the food courts and mass market stores than the respondents from 30-45 age groups. Considering the descriptive statistics performed before ANOVA and Tukey test the results are more significant to the Food courts than for Mass market as the mean value is higher for the first one.

...