Development of a Coding Scheme
The occupational therapist’s worksheets contained information regarding a) the types of challenges that the women chose to focus on during the problem-solving sessions (step one of problem-solving), b) the types of goals they set regarding those challenges (step two), and c) the number and breadth of solutions they identified during brainstorming (step three). The goals of this analysis were to summarize these data in meaningful categories. This is best performed through content analysis (Berelson, 1952
). The approach uses a systematic and reproducible process to code the manifest content of a piece of text, then count the data attached to each code.
The occupational therapist entered the challenges, goals, and solutions identified by the women into a database and exported it into a spreadsheet for the coding process. The occupational therapist read and coded each entry, recording an operational definition of each code as it was created. A research intern (second author) who was previously uninvolved in the study then reviewed the data and attempted to use the coding system. The occupational therapist, research intern, and principal investigator (third author) met on three occasions to discuss the codes and the challenges of applying them to the data. Based upon these discussions, the occupational therapist finalized the operational definitions and coded all the data. The research intern also used the finalized definitions to independently code the data, in order to assess the reliability of the coding scheme. The research intern was blind to the occupational therapist’s code choices at all times and did not ask questions or attempt further clarification during the reliability coding.
The coding process identified 11 categories of activities that presented challenges: exercise, instrumental activities of daily living (IADL), work, stress management, nutrition, sleep, leisure, childcare, upper extremity exercise, social activities, and miscellaneous (one woman wanted to find a way to incorporate the problem-solving steps into her daily life and habits). There were four categories of goals, reflecting the overall purpose of what the women were trying to accomplish within a session: adapt an activity, find a new activity, plan the steps of an activity, and gather information about making a future decision. The “adapt” category indicated that a woman was trying to figure out how to do a familiar activity in a new way, perhaps to make it easier, less tiring, more enjoyable, or less stressful. The sessions in which the goal was to adapt offered the opportunity to assess the number and breadth of the adaptive strategies generated by the women. It was in these sessions that the occupational therapist prompted the women to identify ways to change something about themselves, the environment, or the way in which they do an activity. Initially, we attempted to code the adaptive strategies in terms of whether they involved an adaptation of the person, the environment, or the occupation. This coding scheme was hard to operationalize. For example, if a woman was adapting her usual activity of walking outside with a friend by walking inside on the treadmill, she has changed both the environment and the occupation. The final coding scheme for the adaptive strategies had five codes: who, what, where, when, and how.
We exported the coded data into SPSS Statistics 19 and used the crosstabs function to calculate the kappa statistic and percent agreement between the occupational therapist’s coding and the intern’s coding for the activity challenges, goals, and the types of adaptive strategies.
For the activity challenges, we calculated the frequencies of the occupational therapist’s categories, first by the proportion of participants whose data included the category and then by the number of sessions. We calculated the average number of activity challenges covered over each participant’s sessions. We repeated these analyses for the goal. We then split the file and explored the frequency of the activity challenges according to the goals.
Next, we calculated the frequencies of the various types of adaptive strategies identified by the occupational therapist in sessions where the goal was to “adapt” an activity. We calculated the average number of adaptive strategies generated by the women per session and the average number of different types of strategies generated per session. For each of the women who had at least one session focused on adapting an activity, we counted how many of the five categories of strategies she used over all of her sessions and identified the mean and standard deviation.