Scientific transparency

It must be obvious from the text what the author has done, i.e., the methodology must be described sufficiently. The term sufficiently here refers to the ability for potential readers to reproduce the project or experiments, to discuss and analyze the results, and to compare the methodology with other similar works based on the information in the text.

Examples
"To obtain an accuracy measure, we tested our model using [tool X]."
It is not clear from this sentence how the testing was done, what parameters or settings were used. It would be difficult to reproduce the test (or tests, not clear how many were done) and it is difficult to analyze the correctness of any obtained results.

"To obtain an accuracy measure, we tested it repeatedly using [tool X]. The results from our test suite, including the values of the C parameter, are presented in Table 2. As can be seen in the table, ..."
The reader is referred to a table where parameter values and information about the different tests are presented. This makes it acceptable for the text itself to be less specific (testing something 'repeatedly') since the details can be found elsewhere.

For the sake of transparency, it is important to compare only comparable entities. If for some reason you absolutely want to compare values that really cannot be compared, it is essential to mention and discuss this. You must also state your reasons for doing so.

Examples
"We applied our method to test data in the form of a list of phone numbers. The results were compared with another study on lists of zip codes."
Here, only the comparison is mentioned; the reader likely wonders whether the two lists are really comparable, or if the results may depend on dissimilarities in the test data used. (The other study is also not properly referred to.)

"We applied our method to test data (a list of phone numbers). As a comparison, another study [8] used lists of zip codes. We believe the results to be comparable, since in both cases, the lists contain only integers."
The author presents the comparison, explain that the test data is different, and then argues that the difference is unimportant in this case and will not affect the results. This enables the reader to discuss and criticize the results (they might not agree that the lists are comparable!).

Back to Academic writing