search query: @journal_id 780 / total: 133
reference: 1 / 133
« previous | next »
| Author: | Kane, S. |
| Title: | Scientific methods in finance |
| Journal: | International Review of Financial Analysis
2004 : VOL. 13:1, p. 105-118 |
| Index terms: | Finance Research Methodology |
| Language: | eng |
| Abstract: | Computer software can easily produce many financial models with technical skill, but investigators need to interpret and adapt computer output. Such analysis requires a deep knowledge of scientific fundamentals. This article argues that an evidential perspective is more appropriate than a decision framework for model selection. Four necessary conditions for any scientific investigation are discussed: observables, interpretables, replicables, and robustness in the context of examples familiar to finance professionals. Carefully planning the design of an experiment is the best way to address many econometric maladies. In addition, the need for more sensitivity testing and for independent replication of empirical results within the finance profession is emphasized. |
« previous | next »
SCIMA