Dissertation Creation - The UK's original provider of custom dissertations and dissertation help...

Free Dissertations - Statistics Dissertations

Statistical Methodology As The Survey Questions Were Reported In A Familiar ...


Statistical Methodology
As the survey questions were reported in a familiar Likert Scale yet the research question is aimed at answering the generalized question, Does Sarbanes-Oxley make a difference?, the data was recoded into a scheme for which the comparison of means is more readily conducted. While there are multiple methods by which to achieve this, the following propositions were utilized:
IF there is no difference, the maximum mean of the respondent's answers would be 3.000 reflecting that each of them was uncertain of the outcome.
IF µ < 3.000, this reflects a tendency to be more certain that it did not affect them as evidenced by a response of a 1 or 2 to the survey item(s).
IF there is a difference, then µ > 3.000 as a result of respondents indicating a 4 or 5 to a survey item.
With this method, the research question is answered statistically by comparing the actual mean of the responses to a hypothetical maximum no affect mean of 3.000.
As there are often multiple ways to review data, one alternate methodology considered would be to recode any response of 3 or less to a No which would then be given a value of 0.000 while a response of 4 or 5 would be equated to a Yes which would be represented numerically by 1.000. The null hypothesis then becomes that the sample mean is not statistically significant from a hypothetical mean of 0.000, or expressed in non- binary language, that statistically the same respondents indicated no perception or behavior different than previously. By utilizing the first methodology, though statistically equivalent, the µ of each item remains similar, scoring-wise, to the data as it was collected, thus aiding in interpretation.
In preparation of hypothesis testing, a group mean of the mean of each survey item built around the testing of Hypothesis1 was computed along with Hypothesis2 and Hypothesis3. This procedure presented no complications of recoding as all items are oriented in the same direction on the same scale. That is, a 4 or 5 response always is positive, or alternatively, there were no trick questions such as, Sarbanes-Oxley has NOT influenced me in which the respondent who perceived a change would choose a 1 or 2.

Each item was also examined against the hypothetical mean of 3.000 in order to provide a balanced insight into the testing of each hypothesis as the group could have been rejected based upon an outlier item mean.
Additionally, it is worth noting that there were no test or other baseline measures to screen out test bias. An example of this would be to include a item such as, I am aware of he passage of the Wuggles Smugley Bill in 2004.. The presence of a positive response to this question would allow the researcher to discard this test based on evidence of either testing bias or simply not being a qualified respondent.

Please note: The above dissertation snippet was written by a student and then submitted to us to display and help others. Thanks to all the students who have submitted their work to us.