Introduction
This report is based on a statistical analysis to identify the key characteristics that define a shopping area as attractive based on data collected through a telephone survey of consumers (n = 150). Among the variables of interest were ease of returns, high-quality merchandise, low prices, product variety, friendliness of staff, the convenience of organized shopping hours, cleanliness of retail spaces, and availability of multiple bargains. Each of the variables was measured on a discrete seven-point Likert scale, where “7” corresponded to the maximum level of importance of the attribute. Respondents have priorities and preferences for what they believe forms an attractive bargaining area. Collecting and descriptively analyzing such data for a large sample size allows us to assess the average trends prevalent in the group.
Descriptive Analysis
Using an Excel spreadsheet, a descriptive analysis was performed for each of the eight variables used, which included determining measures of central tendency (mean, median, mode) as well as measures of variation (sample standard deviation, range, minimum and maximum value, first and third quartiles). Table 1 below shows the results of all calculations, and the Excel file demonstrates the formulas used for all calculations.
Table 1. Results of descriptive analysis for the variables
A cursory descriptive analysis suggests that high product quality (M = 5.7, SD = 1.76) and low prices (M = 5.7, SD = 1.87) were the leaders in importance based on median values, while cleanliness of retail space (M = 4.8, SD = 1.90) was the least important. The leadership of these two categories is also determined by their median values, in both cases equal to 7. We can also look at the results from another perspective: the convenience of shopping hours was the only one most often assigned a value of “6,” meaning that this variable was essential but not as important as the others. The IQR method was used to determine the outliers in the eight distributions. Table 2 shows the calculation of the IQR and the upper (Q3 + 1.5IQR) and lower (Q1 – 1.5IQR) limits for the distributions. From the results, it can be seen that for the variables IMPEXCH, IMPVARIE, IMPHELP, IMPHOURS, IMPCLEAN, and IMPBARGN, there were no outliers, as all values (1-7) fell within the range limits. However, there were outliers for IMPQUALI and IMPPRICE, judged leaders by cursory descriptive analysis, because the values of “1” did not fall within the lower bounds of the range for both variables.
Definition of Outliers
Table 2. Determination of emissions using the IQR
The z-method can also be used to determine outliers. In this method, for each value in the distribution (1-7), the z-score must be calculated using the sample mean and standard deviation. Table 3 shows the results of calculating the z-score for each value in the eight distributions. Each of the values in this table shows how much, on average, the item deviated from the mean for that distribution and in which of the two directions. The critical point of this method is the choice of some threshold that determines whether a value can be classified as an outlier (Mahmood, 2022). The standard limits are -3 and +3, meaning any values smaller or larger than these thresholds, respectively, are outliers. A careful analysis of the table shows that there are no such values, which means that according to the z-score method, there are no outliers in all distributions of the eight variables.
Table 3. Determination of emissions using z-score
Correlation Analysis
It is also possible to perform a pairwise correlation analysis for the data, which allows for determining the strength and direction of the relationship between the variables. However, the standard Pearson correlation test works with continuous data, while in this case, discrete ordinal scales are used (Akoglu, 2018). Instead, one can use Spearman’s rank correlation, which works well with ordinal data. Since Excel already has automatic choices for the best correlation method, use =CORREL() for ordinal data to determine the appropriate correlation coefficient value. In this analysis, the nature of the relationship between IMPQUALI, on the one hand, and IMPVARIE, IMPHELP, IMPHOURS, IMPCLEAN, and IMPBARGN, on the other hand, were identified in pairs. Table 4 shows the calculations of all the coefficients.
Table 4. Results of pairwise correlation analysis for five pairs of variables
All of the coefficients were positive, which means that as perceptions of the importance of high-quality products increased, so did perceptions of the importance of the other variables used. The maximum strength of the relationship was found for the relationship between perceptions of the importance of quality and shopping hours (rS = 0.31). This means that the more critical the high quality of the products was to the respondent, the more critical the convenience of the shopping hours was to them as well. Overall, all of the coefficients are moderate in strength.
Conclusion
The statistical analysis results revealed that high-quality products combined with low prices were the key benefits of an attractive shopping area. In contrast, convenience shopping hours were identified as the least important (but still necessary) attribute. The importance of high-quality merchandise was shown to be the strongest positive correlation with convenience shopping hours. Emissions were not detected using the z-score method but were detected using IQR. These data can be used to fine-tune shopping areas to attract more shoppers.
References
Akoglu, H. (2018). User’s guide to correlation coefficients. Turkish Journal of Emergency Medicine, 18(3), 91-93. Web.
Mahmood, M. S. (2022). Outlier detection (part 1). TDS. Web.