3.4.2 Coding and Rating

Coding of the transcripts and rating of the nursing diagnosis and nursing care planning outcomes was undertaken by the researcher (myself). Two graduate students helped verify the assessments. Coding of the cognitive and cooperative verbal elements was reviewed and verified by a doctoral student in computer applications in education whose own research deals with the cognitive activity of students using computers. The rating of the nursing outcomes was reviewed and verified by the Nursing Unit Manager of a local psychiatric nursing unit who is also a doctoral candidate in sociology.

Coding and rating by independent assessors would have been problematic for a variety of reasons. Persons with experience in the study of cognition or communication might have had difficulty with some of the nursing content. Nurses, on the other hand, might have been too selectively focussed on the nursing content.

Those unfamiliar with nursing jargon, or the content of the nursing curriculum would run into several difficulties when attempting to code these discussions. For example, there was considerable use of abbreviations such as the ones shown in the list in Table @Ref(tagabbr).

Table 3-5: Sample of Abbreviations Found in the Transcripts

	rtl 	- related to

	y/o 	- year(s) old

	e.p.o. 	- expected patient outcome

	pt 	- patient

	amb 	- as manifested by

	NANDA 	- North American Nursing Diagnosis Association

	dx 	- diagnosis

	rx 	- medication

	adl 	- activities of daily living

	HCT 	- health care team

	AA 	- alcoholics anonymous

	SOAPIE 	- subjective data, objective data, assessment,
		  plan, intervention, and evaluation

Further, some understanding of how nursing process and care planning is taught would be needed to properly code this question, "Are there any more columns left?" It is only clear that this has to do with problem formulation, {FRM} when one recognizes that the components of care planning are often laid out in columns in texts and worksheets.

More generally, one of the keys to applying the criteria and coding scheme is an understanding of the levels of content in the transcripts. As shown in Table @Ref(tabcontran), there are essentially three levels. The "task" level refers to discussion about the communication activity itself and is coded as "management of the task" {MGM}. The "problem" level refers to discussion leading to the formulation of the problem in terms of nursing process and is coded as "problem formulation" {FRM}. The "case" level refers to discussion of the content of the case study and is often coded as "opinion" {OPN}, "arguing" {ARG}, "giving information" {INF}, or "clarifying" {CLR}.

Table 3-6: Levels of Content in the Transcripts

Another important coding consideration is the unit of content being measured in the transcripts. Conventional content analysis might require some a priori specification such as "word", "sentence", or "paragraph". More recently, however, in the qualitative analysis of texts, units of meaning relative to the coding categories are not restricted by arbitrary constraints of grammatical structure or style of expression. For this study, the unit of content was any identifiable segment, of any size, that fit the category criteria. Such a verbal element in the transcripts would be demarcated only by the extent to which it continued, or ceased, to reflect the intended meaning of the category into which it was being sorted. A small sample of coded transcript can be seen in the upper window of Figure @Ref(figcod1) on page @PageRef(figcod1) and some coded excerpts from the transcripts appear in Figures @Ref(figexcerp1) and @Ref(figexcerp2) on page @PageRef(figexcerp1) (also see Appendix @Ref).

To sort and count the coded transcripts, standard computing utilities such as "grep" and "locate" were used to scan the documents. Their output, one line per code found, was redirected to new files which were then processed through a line counting program. Only a few lines were found to contain more than one of the same code, and the counts for these were adjusted accordingly.

Finally, in terms of coding and rating, the post-task comments written on the questionnaires or recorded in the interview were categorized and counted in the following groups: 1) general frustration, 2) typing problems, 3) time too short, 4) slow communication process, 5) synchronous would be better, 6) liked the interaction, 7) generally beneficial.

3.4.3 Statistical Procedures

A codebook was prepared to keep track of all variables (see Appendix @Ref(appcodebk)). Once categorized and counted, or rated and scored, all values, including those from the questionnaires, were entered in a data file for investigation and analysis using SPSS-X.¹⁵ An example of the SPSS-X program is provided in Appendix @Ref(appspss).

Using a conventional approach, the data was first explored using basic FREQUENCIES and DESCRIPTIVES. This was followed by contingency analysis (crosstabulations) and breakdowns on the independent variable (GROUP: asynchronous, synchronous). Subsequently, t-tests were applied to raw scores and combined values. Finally, correlations between cognitive/cooperative activities and nursing outcomes were assessed.

---