Customer Satisfaction Levels of Surveyed Customers - Essay Example

Satisfaction Levels of Surveyed s at Four Store Locations Introduction: What was measured This report describes the results of a survey conducted at four store locations conducted between 1 May and 31 May 2011. Interviews were conducted face-to-face with visitors to the store as they exited. As well as asking visitors about their impressions of the stores, the respondents were categorised by gender and age to allow an assessment of whether and how different population groups perceived the stores. The variables are listed in Table 1. The categorical variables are those that divide respondents into distinct classes. Ordinal and interval variables collect measurements. Interval data is actually measured on a continuous scale (actual quantities of some quality like height or age) while ordinal data is numerical form of classification, where whole numbers are used to denote order but the numbers themselves are not measures but a form of classification (GraphPad.com). Table 1: Variables Measured in the Survey Interval variables Ordinal variables Categorical variables Age Age category Gender Distance travelled Distance category Reason 1 Regularity of visits Reason 2 Satisfaction with: price Department Number of items Purchase Service Payment Quality Follow up Overall Store Contact The variables in the top row are italicized to indicate that they are independent variables. In this survey, it was hypothesised that demographic factors such as age and gender (pre-existing qualities or ‘independent variables) might influence opinions and behaviour of respondents (dependent variables). For men and women might differ in the distance they are prepared to travel to a store. Description of the Data Table 2 shows the number of women and men in the sample and various measures of their age profile. Table 2: Demographics of the Sample Gender All Women Men Number of people 582 373 (64%) 209 (26%) Mean age 42.6 42.8 42.3 Minimum age 17 17 17 Median age 42 42 42 Maximum age 75 75 74 The sample comprises 582 shoppers between the ages of 17 and 75, nearly two-thirds of who are women and just over third men. The age profiles of the men and women are very similar. Analysis of the distance travelled by respondents to the store where they were interviewed revealed a wide disparity. The modal distance (the most common length or trip) was less than a mile, but many had travelled much further, up to 53 miles. The median distance travelled was 5 miles and the mean just under 10. This indicates a positively skewed distribution where it is difficult to say what is the ‘typical’ distance travelled to the company’s stores. Inferential Statistics Table 3 shows the results for all shoppers, with men and women grouped separately. Separating women’s and men’s responses in this way allows a preliminary assessment of whether the independent variable (in this case gender) is influencing the dependent variable (distance travelled to the store). Table 3: Distance Travelled to the Store where Interviewed Distance travelled Less than 1 mile 1-5 miles 5-10 miles 10-30 miles Over 30 miles Total Women 49 (13%) 149 (40%) 83(22%) 69 (19%) 23 (6%) 373 Men 23(11%) 74 (35%) 51 (24%) 52 (25%) 9 (4%) 209 Total 72 223 134 121 32 582 The message is mixed: a higher proportion of the women than of the men travelled the shortest distances, but at the other end of the scale women were also more likely than men to have travelled the longest distances. A possible means of determining whether there is a difference between the distances men and women are prepared to travel to the company’s shops is to compare the mean raw distance (using the actual mileages rather than the categories) travelled by respondents of each gender. The mean distance travelled by the female respondents was 9.54 miles compared with 10.26 miles by the men. The standard deviations of the two samples are similar (11.1 and 10.6), so it is appropriate to conduct a ‘type 2’ test, but since the samples are independent and of different sizes we use an independent t-test (Burton). As we are unsure whether to predict that men and women travel significantly different distances to shop, our t-test is 2-tailed. The result of the t-test conducted on the male and female self-reported travelled distances is a p-value of 0.45, indicating that there is no observable statistically significant difference between the mean distances travelled by men and women. It is worth noting with regard to this exercise that our setting of gender as the ‘independent variable’ that might influence the distance travelled to shop might be flawed. Suppose a couple were discussing which retail store to shop at together, and the female partner preferred a more distant shop but was unwilling to drive the distance. In these circumstances the man might offer to drive them both, in which case it would be the distance influencing the gender of the driver rather than gender determining the distance travelled as we have supposed. Our survey asked how far respondents had travelled, but not whether they had actually driven themselves, so we cannot tell. Our second inquiry will concern whether men and women differ in the level of overall satisfaction they reported at the store they visited. Satisfaction was rated on an ordinal scale of 1 (strong disapproval) thru 5 (strong approval) and can be treated as a scale variable since it is a measure of ‘quantity of satisfaction’ and the mean score is meaningful. The results reported by men and women appear in Table 4. Table 4: Overall Satisfaction Levels Reported, Grouped by Gender. Gender Mean satisfaction Median satisfaction Modal satisfaction Standard deviation (sample) Female 3.08 3 3 1.25 Male 3.10 3 4 1.23 Men and women appear similar in their approval ratings and taken as a whole the sample expressed very mild approval of the stores. In the context of this exercise the higher modal score of the men is interesting enough to make it worth investigating whether a t-test will show that men are more approving than women. This time our t-test is a one-tailed test as we have a hypothesis that men in fact approve more than women. In other respects the test will be conducted the same as in the previous exercise, since the variances of the two groups as measured by standard deviation are similar. On this test the p-value is 0.44, allowing the conclusion that men and women are effectively the same in their overall ratings of the stores they were asked about. References Burton, R. Using Excel to do Basic Statistical Tests http://depts.alverno.edu/nsmt/stats.htm Web. Accessed 10.6.11 Carr, R., 2010, XLStatistics 10.05.30, XLent Works, Australia GraphPad.com. Web. Accessed 11.6.11. Web. Accessed 10.6.11 Jupp, V (1989) Methods of Criminological Research Guildford, UK: Unwin Hyman Read More