Group 2. Employee Response to Customer Needs and Requests
Group 3. Unprompted and Unsolicited Employee Actions
A fourth group titled Problematic Customer Behaviour was created to
accommodate some employee responses. Not surprisingly, during the first study,
customers themselves had not mentioned any critical service encounters which
fitted into this category and no comparisons were therefore possible using Group
4.
The comparative employee/customer data for groups one, two and three
(reported in Bitner et al. 1994) have been entered into contingency tables (Figure
5 - 12 ) and chi square (χ^2 ) tests have been carried out to determine on which (if
any) categories of customer and employee frequency of reports are statistically
significantly different.
Group 1 - Employee Response to Service Delivery and System Failures
Satisfactory Dissatisfactory Row total
Employee data 109 195 304
Customer data 81 151 232
Column total 190 346 536
Group 1: χ^2 = 0.05; d.f. = 1; p = 0.823 (Fisher’s exact p = 0.855)
Group 2 - Employee Response to Customer Needs and Requests
Satisfactory Dissatisfactory Row total
Employee data 196 62 258
Customer data 114 55 169
Column total 310 117 427
Group 2: χ^2 = 3.72; d.f. = 1; p = 0.054 (Fisher’s exact p = 0.060)
Group 3 - Unprompted and Unsolicited Employee Actions
Satisfactory Dissatisfactory Row total
Employee data 89 37 126
Customer data 152 146 288
Column total 241 173 414
Group 3: χ^2 = 11.09; d.f. = 1; p = 0.0007 (Fisher’s exact p = 0.0008)
nb – values are based on Pearson’s χ^2 ; p values are also given for Fisher’s exact test.
Fisher’s method is often reported for 2x2 contingency tables although this method
correctly only applies where expected individual cell frequencies are small (Everitt
and Skrondal 2010: 167). The smallest expected value in these analyses is 46.3.
Source: Bitner et al. (1994: 101)