Replication of Software Engineering Experiments - Empirical Software Engineering and Verification - page 71

Table 7. Replications with variations that adhere to the same experimental protocol

Replication Type Author(s)

Conceptual Replication Hunter [31]; Monroe [74]

Demonstrated Replication Monroe [74]

Differentiated Replication Lindsay and Ehrenberg [68]

Direct Replication Schmidt [35]

Empirical Generalization Tsang and Kwan [66]

Generalization and Extension Tsang and Kwan [66]

Independent Replication La Sorte [33]; Monroe [74]

Instrumental Replication Kelly et al. [72]

Nonexperimental Replication Leone and Schultz [73]

Nonindependent Replication Monroe [74]

Operational Replication Lykken [69]; Kelly et al. [72]

Partial Replication Hendrick [70]; Monroe [74]

Reproducibility of an experiment un-
der a fixed theoretical interpretation

Radder [75]

Retest Replication La Sorte [33]

Scientific Replication Hunter [31]

Sequential Replication Monroe [74]

Systematic Replication Kantowitz et al. [65]; Finifter [71]

Types E..H Bahr et al. [62]

Type II Easley et al. [61]

Varied Replication Van IJzendoorn [63]

Virtual Replication Finifter [71]

Table 8. Some identified elements that can vary in the replication

Variable element Replication Type Author(s)

Measurement
instruments

Differentiated Replication Lindsay and Ehrenberg [68]

Measures Operational Replication Kelly et al. [72]

Method Conceptual Replication Schmidt [35]

Place Types B,F,J,N,D,H,L,P Bahr et al. [62]

Populations Empirical Generalization Tsang and Kwan [66]

Research Design Retest Replication La Sorte [33]

Researcher Independent Replication Monroe [74]

Sample Virtual Replication Finifter [71]

Although the overall objective of a replication is to check an experimental
result, we find that different replication types have specific aims or purposes.
For example, according to Lykken [69], the purpose of operational replication
is to check that the experimental recipe outputs the same results with another
researcher. However, Finifter's systematic replication [71] aims to output new
findings using different methods to the baseline experiment.