College of Nursing - Dissertation Guide

33 population used to obtain these psychometric estimates. If the instrument has not previously been used for application to your target population or if the students are developing their own instrument, it must be tested for validity and reliability. Even if it has been used previously with your population, it is a good idea to recalculate the reliability and use your findings to compare to those reported by others. Include permission to use the instruments in your appendix . Instrument 1 – Name of the instrument Provide background information on the instrument, i.e. who developed it and for what purpose. Include previous applications and findings. Do this for each instrument the student will use to collect data, except for the demographic information . Validity. Address all types of validity and if it has been established for the instrument when applied to your population. Reliability. Reliability can be assessed in two ways: (1) reliability over time typically includes a test-retest procedure to see if participant responses are consistent from one testing to another and (2) reliability as internal consistency. Test-retest is optional but internal consistency must be addressed. Internal consistency is determined by analysis of Cronbach’s alpha ( ). The student must state what an acceptable alpha is and the criteria you will use to achieve this. The minimum number of participants necessary to analyze internal consistency is typically 10 subjects per item on the instrument so if one instrument consists of 10 items, the student will need 100 participants for this portion of the analysis. Scoring. Provide a clear explanation as to how the instrument measures the construct and will be scored. Anyone else should be able to collect data and score the instrument the same as the student does. For example, the instrument consists of 10 items to which the participant may respond on a 7-point Likert scale whereby 1 indicates strongly disagree and 7 indicates strongly agree . The responses are summed for a composite score which may range from 10 to 70 with higher scores indicated a stronger degree of the construct. Identify the level of measurement that it produces, i.e., nominal, ordinal, interval, ratio. General Statistical Strategy For example, responses to the instruments will be entered and all analyses will be carried out by means of Statistical Package for Social Sciences (SPSS) version 19 for Windows (2010). The data will be explored through screening for errors. Parametric data considered as dependent variables will be subjected to analyses to assure they have met the basic assumptions of normal distribution and homogeneity of variance. Additional techniques to be employed include frequency histograms, boxplots, descriptive statistics and others depending on your hypotheses. Data Cleaning Explain how the student will review the responses for completeness and how the student will handle missing data. Describe how the student will detect errors in data entry and outliers. How the student will define an outlier and what you will do with them, i.e., will you include outliers and if so, how will you lessen their impact? Descriptives Descriptive statistics will be used to both describe the sample and report the measures of central tendency of the scores obtained on the instruments. The actual report will be provided in chapter 4. Reliability Testing