Methods and systems for textual analysis

1. A computer-implemented method for generating models of writing abilities by applying a linguistic computational tool with a master control system server communicatively coupled to a data store and a graphical user interface, the method comprising:

obtaining, from the data store, multiple examinee-generated responses to corresponding test items

generating, with the linguistic computational tool executed by the master control system server, one or more quality-neutral, data-based profiles of writing characteristics from a set of textual data points extracted from the examinee-generated responses by:

identifying indices of linguistic and discourse representations of the test items, wherein the indices comprise a description index, a scale index, a cohesion index, a language conventions index, or a demographics index;

text mining the multiple examinee-generated responses and applying an analytical rubric to each response to determine a quality score of the response for each index;

normalizing the quality scores based on one or more quality control measures;

aggregating the normalized quality scores into one or more aggregated score profile for a specified examinee proficiency, a specified examinee grade, or a specified set of examinee demographics; and

mapping changes of the aggregated score profiles over specified time periods;

generating fine-grain models for identifying characteristics in at least one of: the identified indices, the text mined multiple examinee-generated responses, the normalized quality scores, the aggregated normalized quality scores, and the mapped changes of the aggregated score profiles over time;

correlating, linguistic computational tool executed by the master control system server, one or more quality-neutral, the one or more quality-neutral data-based profiles with one or more qualitative evaluations;

identifying improvements in an examinee's writing by analyzing relationships between the quality-neutral data-based profiles and the one or more qualitative evaluations based on the identified characteristics;

storing the improvements in the examinee's writing in the data store; and

displaying, on the graphical user interface, an improvement recommendation based on the stored improvements in the examinee's writing.

2. The method of claim 1 further comprising:

correlating the one or more quality-neutral data-based profiles with the one or more qualitative evaluations by differentiating data from the one or more test items for at least one grade by qualitative performance.

3. The method of claim 1 further comprising:

correlating the one or more quality-neutral data-based profiles with the one or more qualitative evaluations by entering scores for one or more qualitative evaluations in a database of the one or more quality-neutral data-based profiles as one or more variables.

4. The method of claim 1 , wherein the description index comprises a paragraph length parameter, a sentence length parameter, a word length parameter, a noun count parameter, a verb count parameter, an adjective count parameter, an adverb count parameter, a pronoun count parameter, or a word frequency parameter.

5. The method of claim 1 , wherein the scale index comprises a simplicity and complexity parameter, a narrativity parameter, a word correctness parameter, a connectivity parameter, a syntactic pattern density parameter, or a readability parameter.

6. The method of claim 1 , wherein the cohesion index comprises a referential cohesion parameter, a latent semantic relationship parameter, a casual cohesion parameter, or a lexical diversity parameter.

7. The method of claim 1 , wherein the language conventions index comprises a grammatical error parameter, a word usage parameter, a mechanics parameter, or a style parameter.

8. The method of claim 1 , wherein the normalizing based on quality control measures comprises analyzing a quality score distribution for each rater, a determined reliability for each rater, a mean score for each rater, or a standard deviation for each rater.