Each line in this dataset represents an annotated learner error. One paragraph line in the essay XML file can contain more than one annotated error.
| Column | Type | Description |
|---|---|---|
| student_id | String | Test taker identification |
| language | String | Learner's L1 |
| overall_score | Float | Combined mark for both tasks |
| exam_score | String | Essay mark |
| raw_sentence | String | Paragraph line extracted from the XML file |
| error_type | String | Tag associated with the error (See Nicholls (2003)) |
| error_length | Integer | How many words are tagged in the error |
| correction_length | Integer | How many words are tagged in the correction |
| correct_sentence | String | Sentence with all the errors replaced by their corrections |
| correct_error_index | Integer | Index of the correction in the sentence |
| incorrect_sentence | String | Sentence with all the errors replaced by their corrections, but the error represented by the row |
| incorrect_error_index | Integer | Index of the error in the sentence |