Data Analysis in research
The Shapiro-Wilk test was used to check if continuous variables were distributed normally. Descriptive data analysis was done, and continuous variables were expressed as either the median, or mean and standard deviation, and inter-quartile range. Also, Categorical variables were presented in the form of proportions, percentages with the 95% confidence intervals. The main outcome calculated was Bivariate logistic regression. It was used to analyze associations between dosing errors and predictor variables identified in the data. Their odds ratio, 95% confidence intervals of the odds ratio, and the related p-values computed.
Multivariable logistic regression was constructed to investigate association while adjusting for possible confounding. The predictor variables associated with medication-related problems were determined using the Forward stepwise model building. STATA version 14.0 software, which was utilized for the purpose of analyzing. P values that were less than 0.05 were deemed to be significant statistically.
Assumptions Based on the Statistical Analysis
With regard to the statistical tests used, assumptions were made to ensure that the data collected were consistent with the nature of the analysis being conducted. These assumptions included normality, independence, linearity, and homosexuality. Failure to comply with these assumptions would have resulted in inaccurate results, which would have been problematic. Testing the data assumptions that violated the statistical test assumptions, depending on the violated assumptions, would have produced both false negatives and false positives.
The normalcy hypothesis meant that the data approximately matched the bell curve before any statistical or regression tests were performed. The hypothesis followed that the study was statistically independent. The hypothesis was that all participants in the sample would only count once. The use of the questionnaire allowed each participant to make independent comments. The independence of observations meant that the variations of each participant affected the overall analysis only once. The assumption of homosexuality ensured that the variance around the regression line was the same for all values of the predictive variables. The study did not break this assumption. The statistical software packages used do not automatically check these hypotheses but assume that they are fulfilled, as they are the conditions underlying the logarithmic functioning of the program.
Determination of Statistical Significance
Fixed standards were used to verify the significance of zero assumptions, and only 0.05 was used as a criterion for assessing or reporting statistical significance.