Research strategy Case Study 4: Theory-testing research: testing a probabilistic relation

7.2.7 Hypotheses

The propositions state that for a specific change of time access window pressure differences between retailers with different levels on the dimen- sions of distribution strategy will be observable. For the present test, we decided to keep time window length scenario A, B, or C constant and to change only the number of time access windows levels 1–6. The following hypotheses were formulated, which must be tested separately for each scenario A, B, and C. Hypothesis 1: Retailers with a higher number of stops per roundtrip have, on average a steeper rise in total distribution costs with an increase in the number of time windows from level 1 to level 6 than retailers with a lower number of stops per roundtrip. Hypothesis 2: Retailers with a vehicle fleet with higher capacity per vehicle have, on average a steeper rise in total distribution costs with an increase in the number of time windows from level 1 to level 6 than retailers with a fleet of lower capacity per vehicle. Hypothesis 3: Retailers with longer stopping times have, on average a steeper rise in total distribution costs with an increase in the number of time windows from level 1 to level 6 than retailers with shorter stopping times. Hypothesis 4: Retailers with longer distances from the distribution centre to their shops have, on average a steeper rise in total distribution costs with an increase in the number of time windows from level 1 to level 6 than retailers with shorter distances from the distribution centre to their shops. Hypothesis 5: Retailers with less strict self-imposed time windows have, on average a relatively higher increase in total distribution costs that occur with an increase in the number of time windows from level 1 to level 6 than retailers with stricter self-imposed time windows.

7.2.8 Measurement

We could make use of the same data that we used in the study reported in Case Study 3. Time access window pressure was determined by the number of shops that are affected by time access windows and the time-window length. Total distribution costs in euros per week were determined by convert- ing our data on the weekly number of vehicle kilometres, the total time