First, we chose 20 towns as sites for our data collection.
With the help of a commercial directory, we randomly identified
15 companies per town whose employees personally
and regularly interacted with customers. In unannounced
visits to their workplaces, we then contacted 300 employees
and asked them to join the study, without providing any
incentives. To prevent any performance bias, such that better
performing employees might be more willing to answer our
questions, we explicitly stated that the results would not be
shared with the employing firms. Rather, we assured all participants
that their responses would be used exclusively for
research proposes. In addition, we told them only that we
sought to survey their customers after they had completed
the questionnaire. The frontline employees evaluated customers
with whom they came in contact, not interactions of
customers with their colleagues. We gathered 165 questionnaires
(response rate=55.0%), 150 of which were complete.
Second, outside each location, we solicited responses from
three consecutive customers who interacted with these employees.
We asked them to evaluate the frontline employees’
behaviors in their previous interaction. Of the 495 identified
customers, 428 returned questionnaires, for a response rate of
86.5%, and 388 customer responses referred to one of the 150
employees who had completed questionnaires.
Third, we matched the data at the frontline employee level
by computing the mean of all customers per employee. We
eliminated nine employee (and 13 related customer) questionnaires
because some customer responses were not complete.
We also calculated the index of within-group interrelated
reliability (rwg) to assess agreement among the customers’
judgments (James et al. 1984). The measures of both frontline
employees’ customer-oriented behavior and customer satisfaction
with the frontline employee provided median rwg
values greater than .70 (.96 for customer-oriented behavior,
.95 for customer satisfaction with the frontline employee;
Burke et al. 1999).
Thus, our data collection procedure generated 141 dyads of
frontline employees and average information provided by
their customers. In our final sample, 50.4% of the employee
respondents and 56.3% of the customer respondents were
women. The sample represented the following industries:
retailing (32.1%); gastronomy, hotel, and tourism (30.0%);
crafts and coiffeur (21.4%); automobile (9.3%); and financial
services (7.2%). Detailed information about the dataset appears
in Table 1.