What I have found from the initial coding process

Image retrieved from Google on January 15, 2011

I just finished discriminating the data from all of the conversations and I found that the results vary depending on each of the students' commitment to their learning process and how much attention they paid to their peers' comments in the forum. For instance, Johana and Fabiola were the students who devoted more time to listen to their partners' conversations and they also had some time to reflect upon their own performances. This caused an impact in terms of progress in the construction of accurate utterances in the past tense especially with Johana who was always demanding with herself throughout the process. On the other hand, Fabiola did great during the fist five conversations, but then in the last two her performance in constructing accurate utterances waned. She told me that she felt tired and she had lots of things to do at her work and that had an influence in her outcomes. Claudia had little change; I can conclude that in spite of improving sentence construction especially at the predicate level as well as the Subject + Verb agreement, she still had problems implementing the verbs in past tense.

Diana and Andres showed no progress at all throughout the process. They did not manage to raise awareness on the importance of constructing accurate full-length utterances. Their comments in the forum were just too shallow (especially Andres's) and they did not consider their partners' advice in order to improve their performances.

End of the Initial Conversation Transcription Analysis

Image retrieved from Google images on December 17, 2010

It took me four weeks to identify and discriminate each one of the five students' performance in terms of grammatical accuracy. The final results show that most of the students are more concerned of conveying meaning at the expense of grammar accuracy. This clearly indicates that there is a high degree of stagnation in four of the five students who were involved in the implementation process. Only one (Johana) showed some improvement in structuring accurate full-length and in using more verbs in past tense. Two of them (Fabiola and Claudia) mostly showed 50% accuracy level during most of the process and the remaining students (Andres and Diana) mostly underperformed throughout the implementation stage.

Types of Coding Procedures

Image retrieved from Googlr images on December 9, 2010



As I continued reading Corbin and Strauss, I wanted to understand the principles that underlie the different types of coding procedures in order to see how I can fit the data I have gathered so far. This is a summary of what I understood.

Open Coding-

It is the process of breaking down, examining, comparing, conceptualizing and categorizing data. This is the first step in the procedure of text interpretation. The goal is to develop concepts based on the data and also on the researcher's contextual knowledge. During the process, comparisons must be made and questions formulated, while open coding starts with the analysis of single passages of text and phrases. The following questions occupy the initial stages of the coding process:

1. What actually happens in the text?
2. What category does the textual passage suggest?

After this, there will be a search for "in vivo" codes (interpretations recognizable in the text itself). Open coding should analyze very exactly and investigate the text with microscopic presicion in order to minimize the risk of overlooking important categories. During this process, theory memos should be written to record the development of concepts, categories and dimensions.

The asking of questions, the precise analysis of words, phases and sentences together with a continuous process of comparison. These will increase the theoretical sensitivity and creativity of the coding process.

Axial coding-

This refers to the procedures which on the basis of the above mentioned coding paradigm reassembles the results of open coding by creating nw relationships between concepts. Threfore, it will assist in the refinement and differentation  of already available concepts whereby these fist acquire the status of categories. After open coding, it is normally not yet clear whether a particular concept has to do with a condition , a strategy or a consequence. Each single concept together with its indicators analysed and assigned according to the points of the coding paradigm.

These are the questions that a researcher should ask during the axial coding process:

1. What are the conditions for the events comprised in the concept?
2. How can the interactions between actors be described?
3. What strategies and tactics can be determined?
4. What are the consequences of the event?

During the coding process. which begins with open coding, axial coding becomes increasingly predominant: either individual text locations are extracted and interpreted axially or several text locations are interpreted in comparison with one another. In such cases, the procedure is similar to open coding this is to formulate a question, dimensionalize. etc.

In axial coding the properties of a category are first elaborated which means that the category is dimensionalized either implicitly or explicitly. Then assumptions about the condition , interactions, strategies and consequences are specified and tested which increases the relationship to other categories. The links between the categories should be recorded in representation of networks and the investigator should capture in the form of coding notes and theory memos as many as possible of the thoughts which have occured during the process.

Selective coding-

This refers to the process of selecting core categories validating this linking process and the filling of other categories which require further requrement and development. A core category is that cental phenomenon around which all other categories are integrated. As the starting point of this final stage, coding lists, memos and network modls should be reviewed and theoretically sorted.

Strauss and Corbin recommend enquiring about the story contained in the data. The essential events  should be brought  together by answering the following questions:

1. What is the most striking feature of the field of investigation?
2. What do I consider to be the main problem?
3. What is the central theme of the story?
4. Which phenomena are represented again and again in the data?

The central story extracted in this way rotates around the core category and shows its relationships to other categories. In the simplest case, the core category may be recognized from the network model and has already been identified, while the most difficult case it must be completely newly identified, refined with regard with its properties and its location established in the network of relationship. The provisional result should be examined continually with reference to the textual data. In this way, step by step, a grounded theory will emerge.

Grounded Theory Summary

Image retrieved from Google images on December 5, 2010

I have been reading about the Grounded Theory and I have discovered that this type of analysis requires three main elements:

• Asking questions
• Making comparisons
• Defining concepts

I have also found that the Grounded Theory has three levels of concepts:

• Description: Its purpose is to contextualize, persuade and to concince that phenomenons happen within the different variables.
• Conceptual ordering: It organizes data into categories according to the properties and dimensions, types and stages.
• Theorizing: It compares and contrasts the different categories of data in order to draw a valid conclusion  when explaining the phenomenon. Corbin and Strauss state that there should be only one salient category and taken together with other concepts it explains the what, how, when, where and why of the phenomenon.

The following are other important concepts I consider important in order to understand the Grounded Theory:

• Concepts are designations or labels which are attached to individual events or indicators. The researcher must look for indicators of provisional concepts of data.
• The central ideas or sensitizing concepts are transformed into provisional research questions.
• The brainstorming of categories and literature is made, that is, compared and fit in with the research question(s). These are a summary of of collective and abstract concepts which may become releveant for the most varied fields of investigation and types of problems. Only a few indicators will be appropriate and relevant.
• During the process, different indicators are investigated and compared with one another and similarities and differences are considered. On the basis of this analysis of indicators, concepts are finally  specified. Here the theoretical concepts are dimensionalized or determined.
• Changing indicators that generate new properties of a code will proceed only so far before the analysts discover saturation of ideas through the interchangeability of indicators. This means, the more numerous the indicators that are of equal significance for a concept, the higher the degree of saturation of properties of that concept for the emerging theory.
• Concepts are developed, categorized and dimensionalized . They are enriched with indicators  like textual examples or exerpts.
• By means of permanent comparison of the concepts using the associated text units they are categorized, related to each other, ordered, put into hierarchy and dimensionalized. Through this process, the data is broken down into dimensions, while variables are established at different levels.
• During the coding process, the investigator is permanently switching between inductive and deductive thinking, setting up and testing concepts to create hypotheses.

Coding Procedure and Update

I have been detecting and categorizing  the most salient features of my students' speech in terms of accuracy, peer/self correction strategies and the new findings that may become future further research. So far, I have noticed improvement in most of the learners mainly because they have raised awareness on their most concurrent errors and they are starting to refrain them or they are still in the process of correcting them.

The procedure that I am following is to highlight each one of my students' speaking performances according to the five main categories (hits, misses, self-correction, peer-correction and other types of errors to be analyzed in further research). This is what I have been doing so far.

Approaches chosen for my data analysis stage

Image retrieved fro Google images on October 20, 2010

The approaches I have chosen to conduct the data analysis from all my data analysis intruments are based on Corbin and Strauss' (1999) Grounded Theory and Sagor's (2004) Trend's Analysis. Corbin and Strauss stated that grounded theory is "theory that is developed inductively from a corpus of data". This means I will take the data that I have collected during the implementation stage and then I am going to use known facts (from literature) to produce some general principles. To "ground" my theory I must go through a series of steps in order to filter my data so I can find the core categories and relate these among each other with the aim of creating a single story around my findings.

After creating my categories, I will find the trends or patterns from each category that led my students to improve or stagnate speaking accuracy throughout the entire three-month course. Therefore, I have to employ Sagor's "trend analysis". From this analysis, I must answer three vital questions that will eventually answer my research question. The three questions are:

1. What did I do?: In this stage I have to describe the actions I did during the implementation process so I can detect the patterns and create a timeline.
2. What changes occurred regarding the achievement targets?: In this stage I will identify the trends by using descriptive statistics (mode). The idea is to identify the frequency in improvement or stagnation.
3. What was the relationship between actions taken and any changes in performance on the targets?: In this stage is where I do my interpretations of the events by linking my findings (facts) from all my data collection instruments with my knowledge (intuition).  

Prelude to Analysis

Image retrieved from Google images on October 20, 2010

Now that I have all my data collected (finally!) I am going to start identifying the properties and dimensions from each one of my data collection tools in order to create my categories. According to Corbin ans Strauss (1999) properties are "characteristics or components of an object, event or action The characteristics give specificity to and define an object, event or action". Dimensions are the "variations of a property along a range".

To start finding the categories, I will start by open coding my transcripts and I will use the following conventions in order to identify progress and stagnation trends. In terms of students producing accurate full-length uttereances I will use the green highlighter, if students used incorrect accuracy and they did not do any correction I will use the red highlighter. Students who self- correct will use the yellow highlighter, the ones who peer-correct or assist partners to correct accuracy will use the orange highligter and finally, all those errors that indicate salient features that are not being researched as for this project but are likely to become further future research topics are going to be highlighted in pink.