p.179 How can organizational experiments be run effectively? ...Experiments should be designed to pursue specific goals.

p.179 Analyses of experimental results should not be restricted to questions which were posed in advance (Dachler and Wilpert, 1978: 32). Experiments always yield unexpected results: the issue is not whether surprises occur but whether they are perceived.
 Indeed, surprises which disconfirm prior expectations and so stimulate new causal theories are the most important results of experiments (Box and Draper, 1969). However, such surprises become less likely when experimenters focus upon the questions which guided experimental design. Effective experimenters exhibit a healthy regard for serendipity, the accidental, and the unplanned.