The test for moderation requires exploration of the covariance matrix,
including a combination of regression analysis and analysis of variance,
to probe for interactions between the slopes of linear regression
lines and moderating variables. A simple graphical approach was used
to explore moderation by flight class (to assess H05) following amethod
described by Bowker and Randerson (2007). Graphs were constructed
consisting of two simple linear regression lines, one for the passengers
who travelled in Business Class and the other for the passengers who
travelled in Economy Class (relatively few passengers travelled in First
Class and Special Economy Class, therefore these were not included in
this analysis.). The two regression lines were used to compare visually
the strengths and directions of the relationships between overall satisfaction,
advocacy, or loyalty and the satisfaction ratings of Business
and Economy Class passengers with respect to specific elements of
their journey. If the two lineswere clearly seen to be not parallel, diverging
or crossed at each other, then itwas inferred that the flight class had
a moderating effect, because the strength and/or direction of the relationship
was different for Business Class passengers compared to Economy Class passengers. If the two linear regression lines were seen
to be either parallel or approximately parallel, with the same or similar
slopes then moderation was not supported.