13.07.2015 Views

Developing a Self-Assessment Toolfor Culturally - Office of Minority ...

Developing a Self-Assessment Toolfor Culturally - Office of Minority ...

Developing a Self-Assessment Toolfor Culturally - Office of Minority ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

complex and uncertain at the time to employ a stratified sampling strategy <strong>of</strong> any sort. Becausethe health care system was in a fluid and evolutionary state, many MCOs were merging withother organizations, and many were going out <strong>of</strong> business. As such, characteristics or conditionsthat would normally be used to stratify a sample <strong>of</strong> MCOs were transient during the time <strong>of</strong> thisstudy. Therefore, the group concluded that the most appropriate strategy for selecting the studysample was the most parsimonious one, i.e., a national random selection <strong>of</strong> MCOs.To ensure that the sample drawn would be representative <strong>of</strong> all MCOs, calculations wereperformed to determine the degree <strong>of</strong> precision with which generalizations could be made to thetarget national population. The most typical measure <strong>of</strong> survey precision is the widths <strong>of</strong>confidence intervals for simple univariate estimates and for estimates within analyticalsubgroups. An analysis <strong>of</strong> confidence intervals for this study revealed that a sample <strong>of</strong> 240MCOs (from the universe <strong>of</strong> approximately 1,100 at the time) would allow 95 percentconfidence intervals for categorical variables that have widths <strong>of</strong> approximately 4-7 percent. Fortypical subgroup analyses utilizing six subgroups, a sample <strong>of</strong> 240 MCOs would yieldconfidence intervals that range in width from 6-15 percent, depending on the size <strong>of</strong> thesubgroup sample and the particular variable.2. DEVELOPING THE DATA COLLECTION PLANOnce the three respondent types (per MCO) were identified, the next step was to develop adata collection plan which would most likely render a high response rate. Here too, however, therapidly-changing structure <strong>of</strong> MCOs at the time—i.e., the evolution <strong>of</strong> the organizationsthemselves, as well as changes in the staff within the organizations—influenced the decisionsmade during the data collection planning phase.The first point <strong>of</strong> contact with each MCO would be an introductory letter mailed to theorganization’s senior executive, as listed in the AAHP directory, with a follow-up telephone callconducted to confirm receipt <strong>of</strong> the letter and determine the senior executive’s availability forstudy participation. The contact information in the directory provided a starting point forpenetrating the MCOs and for determining the appropriate senior executive to complete thetelephone interview.An interview is the most appropriate method for gathering information on an organization’spolicies, histories, and future plans; because the types <strong>of</strong> information to be collected typicallyrequire the participation <strong>of</strong> senior-level executives who have busy schedules, these respondentsare more likely to agree to a telephone interview than to a face-to-face interview (Marshall andRossman 2000). The telephone interviews for this study would be conducted using ComputerAssisted Telephone Interviewing (CATI) by experienced interviewers who would receivetraining specifically for this data collection effort. The use <strong>of</strong> CATI provides a number <strong>of</strong> timesavingadvantages in the data collection process including: electronic recording <strong>of</strong> aparticipant’s responses (eliminating the manual data entry step); electronic guidance through theinterview (e.g., making appropriate skips automatically based on the respondent’s answers); andlogic checking capability to ensure consistent responses.COSMOS Corporation, December 2003 G-4

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!