SPSS Statistic / Data Analysis

Introduction

In this report statistical analysis has been conducted via SPSS to determine the significance of training on the accuracy of malaria detection using virtual microscopy. The data comprises 10 participants who were divided equally into two groups: a control group and the trained group. The initial score and final score for both the groups have been assessed to find out if there is any statistically significant difference before and after training. To test the hypothesis of the undertaken model, the researcher has applied an independent sample T-test. Moreover, other supplementary tests have also been applied such as the normality test, descriptive statistics, and box plots have been obtained to understand the data more comprehensively. 

Data Analysis

Overview of Data

In this research, the data has been obtained from 10 participants who were bifurcated into two groups, namely: a control group and the trained group. The Control group included the individuals who were not provided training regarding virtual microscopy and the trained group included participants who were given proper training after their initial score. The following table gives a summarized view of

The above table states the value of mean and standard deviation for the initial score and final score for the participants of the study. The mean value in the initial score is recorded to be 13.51 which is deviated by 1.64 units. On the other hand, in the case of the final score, the mean score is found to be 56.5 which is much higher than the former. Superficially, it can be stated that overall for both the groups, the final results were significantly better as compared to the initial score. The average final score deviated by 4.26 points. 

Hypothesis

Following is the hypothesis that is being tested by this report: 

H0 = The mean values of the experimental group and control group after undergoing training regarding virtual microscopy will be equal

H1 = The mean values of the experimental group and control group after undergoing training regarding virtual microscopy are not equal

Normality Test

Normality tests were conducted on the data to assess whether or not the data is normally distributed (Park, 2015). Moreover, the normality was to be determined to assess which type of test was to be conducted. The following table shows the results of the normality test: 

The null hypothesis for this test is that the data is normally distributed (Norusis, 2011). For both the Kolmogorov-Smirnova and Shapiro-Wilk tests, the null hypothesis has not been rejected. This indicates that the data is normally distributed. Following is the Q-Q plot for the initial score: 

The above graph shows the observed values for the initial score plotted against the expected values. The graph shows an upward trend where most of the values are plotted within the line except one value which is an outlier. This further validates that the data for initial scores are normally distributed. The following graph shows the Q-Q plot for final scores: 

Figure 2: Q-Q Plot for Final Scores

The above graph shows the observed values for the final score plotted against the expected values. The graph shows an upward trend where most of the values are plotted within the line with no outliers. This further corroborates that the data for final scores are normally distributed.

Moreover, from the results of the normality test, it can also be said that a parametric test can be applied to the model. Hence, to test the main hypothesis of this research, an independent sample T-test will be used.

Box Plots

For the data under consideration, box plots have been used to depict the groups of numerical data based on the quartiles. The box plots also determine the variability in the data along with the outliers in the data (Kerr, Hall, and Kozub, 2002). The following image shows the box plot for the initial scores of the participants: 

The box plot shows the minimum value i.e. 8.33 and the maximum value i.e. 15.83 however there is a presence of an outlier in the data which is denoted by a small dot superscripted with a 3. This indicates that the score at the third number is the outlier in the data i.e. 25.83. It is considered to be an outlier because it is at an abnormal distance from the average values in the initial score. Moreover, the box plot also indicates that the

majority of the scores are more than the median score. The following image shows the box plot for the final scores of the participants:

In the case of the final scores, the minimum value is 35 and the maximum value is 77.5. It is also apparent from this box plot that in the case of final scores, there are no outliers. The median value for the final scores appears to be 60 in the above box plot. Whereas, the majority of the participants scored less than the median value which is 60. As compared to the box plot for initial scores, it can be stated that their extent of variability is significantly less in the case of the final score. 

Independent Sample T-test

When the population of two independent groups is compared to see the existence of difference or similarity, an independent sample T-test is applied (Allen, Bennett, and Heritage, 2018). In the case of the data set that has been considered, the two independent groups are the control group and the trained group. The following table shows the group statistics of the model: 

the data were undertaken: 

Table 3: Group Statistics

The mean values of the initial score indicate that the control group had a slightly higher score as compared to the trained group, however, the deviation from the average value was significantly higher in the control group for the initial score. The mean values of the final score depict that the scores were improved majorly for both the groups, however, the mean score of the trained group was higher i.e. 66.5 as compared to the mean value of the control group i.e. 46.5. The deviation in the average value was again higher for the control group. Overall, from the group statistic, it can be evaluated that the final score has improved significantly after the intervention applied (training). However, at this stage, the significance of the difference between the mean of the two groups can be determined with the help of the following table:

Firstly, in the above table, the sig value for Levene’s test is given which hypothesize that the population of variances are equal (Marshall and Boggis, 2016). In the case of initial scores of the participants, the sig value for this test is 0.273 which is higher than the alpha value at the 95% of significance level hence the null hypothesis is accepted stating that the population of variances is homogenous or equal. This indicates that to test the equality of means, the sig value for ‘equal variances’ will be undertaken. As per this assumption, the sig value appears to be 0.542 which means that the null hypothesis of equality of means between the control group and the trained group cannot be rejected. 

On the other hand, in the case of the final scores of the participants, the sig value for Levene’s test is 0.154 which is higher than the alpha value at a 95% of significance level hence the null hypothesis is accepted stating that the population of variances are homogenous or equal. This indicates that to test the equality of means, the sig value for ‘equal variances’ will be undertaken. As per this assumption, the sig value appears to be 0.008 which means that the null hypothesis of equality of means between the control group and the trained group is rejected. Henceforth, the results have indicated that final scores for the trained group and control group differ significantly.