Hypothesis How to test a deterministic relation with a case study

This hypothesis can also be formulated as a sufficient condition, in the following form: Hypothesis: For all pairs of cases, if the value of the independent variable in case 1 is higher than the value of the independent variable in case 2 condition, then the value of the dependent variable in case 1 will also be higher than the value of the dependent variable in case 2. The logic of testing is the same as when we test a sufficient condition. If the condition is present, the hypothesis predicts that the effect is present as well, or if the effect is not present, the hypothesis predicts that the condition is not present as well.

6.1.5 Measurement

In a longitudinal case study it must be determined how many meas- urements of the two variables have to be conducted and on which moments within the longitudinal time frame. In a comparative case study there are similar cases with different values on the independent variable. This implies that the value of the independent variable was already measured in the earlier stage of case selection. Therefore the measurement in a comparative case study consists of measuring the value of the dependent variable in each case.

6.1.6 Data presentation

For a predicted continuously increasing or decreasing relation the measurement points in the longitudinal case study or cases in the comparative case study are ranked according to the value of the inde- pendent variable. Independently from this ranking, these points or cases are ranked according to the value of the dependent variable.

6.1.7 Data analysis

The data analysis for a predicted continuously increasing or decreas- ing relation consists of comparing the rank orders of the independent variable with the dependent variable. The hypothesis is confirmed if the two rank orders are exactly the same, assuming that both rank orders have been compared in the direction from low to high, or in the reverse direction, that is predicted by the hypothesis. If the rank orders differ, the hypothesis is rejected. Rank orders might differ considerably or only slightly e.g. when the rank orders differ only for two measurements out of a large number. If the rank order differs only slightly it is tempting to conclude that the hypothesis is almost confirmed. This is only acceptable if, in a large number of instances, only a few exceptions occur, and a pragmatic deterministic view is chosen. Normally the hypothesis is rejected if the predicted pattern does not match with the measured pattern.

6.1.8 Implications for the theory

The issues regarding the implications for theory are the same as for all theory-testing case studies. See 5.1.8 for a discussion of this topic.

6.1.9 Replication strategy

The issues regarding the replication strategy are the same as for all theory-testing case studies. See 5.1.9 for a discussion of this topic.

6.2 Case Study 3: Theory-testing research: testing a deterministic relation

The influences of urban time access windows on retailers’ distribution costs 1 by Hans Quak

6.2.1 Introduction

Urban freight transport is crucial to maintain the current urbanized way of living. It is vital to trade and leisure activities in cities as well as to the liveability in these areas. However, transport also causes noise, emissions, congestion, decreased city accessibility, fossil fuel use, visual intrusion, vibration, consequences of emissions on public health, injuries 1 This chapter is based on: Quak, H.J. and De Koster, M.B.M., Exploring retailers’ sensitivity to local sustain- ability policies, Journal of Operations Management 2007, doi:10.1016j. jom.2007.01.020.