Compare the statistical material in the paper to the STROBE checklist items 10, 12-17.
You will use the data you are given to address the following research question.
What does the logarithm predict for GPA?
Answer to Question: 401077 Introduction To Biostatistics
Give your assessment of the strengths/weaknesses of the presentation in Fox et.al. 2010 against STROBE items 10, 12-17
The researchers used a sample consisting of 4746 high school students and middle school students.
However, it is not known how the sampling was done.
The methods of sampling used by the researchers, be they simple random, cluster-, systematic, stratified, have not been mentioned.
To allow respondents to understand the significance of the sampling technique used, it is important to clearly explain what they did.
I think the researchers did a great job in providing demographic information about the respondents.
They also gave information on the proportion of males and females in each age group, as well as the ethnicity, socio-economic status, and gender of respondents.
Researchers did not discuss control for confounding that could have occurred in the course the study.
It is assumed that the confounders used to bias the results were not controlled.
The missing data was not mentioned by the researchers, rather they highlighted it.
Missing data, also known by missing values, occurs when there is no data value for one or multiple variables within an observation.
Research can suffer from missing data. This is a common problem that can affect the conclusions and results.
Parametric tests were used for the inferential analysis.
The test of their assumptions about parametric tests is what is missing.
Parametric tests can’t be distributed freely. They must follow certain assumptions. In order to avoid biased results, it is possible for them to violate any of these assumptions.
Parametric tests require you to assume the following assumptions: linearity assumption; normality assumption; homogeneity assumed (also known under equal variances assumption); independence assumption.
Present the results of your descriptive analysis
This question was intended to answer the research question whether the logarithms for MVPA could predict GPA, after correcting the overweight population of Australian university students.
We began by looking at summary statistics.
LogMVPA had a mean value of 0.44 while the median was 0.52 (a bit higher than its average).
The minimum value was 0.52, while the maximum value was 1.22. A range of 1.75 was possible.
Both the values for skewness, and kurtosis were negative (-0.49 & -0.46), respectively. This could suggest that the data has been negatively skewed.
See Figure 1.
Figure 1: Histogram logMVPA
Students averaged a GPA score 4.76, while the median score was 4.7.
This group of students had the highest GPA score at 6.9 and the lowest at 2.4.
The summary statistics also provided the values for the skewness as well as kurtosis, which were -0.08 & -0.44, respectively.
This indicates that data appears to have come from a normally distributed dataset, since the skewness value is very close to 0.
The figure 2 illustrates this concept.
Figure 2 shows the histogram showing the GPA.
The figure shows clearly that the GPA data are normally distributed (bell-shaped) as can be seen in the figure.
Figure 2: GPA-History Histogram
Expose the relevant inferential and regression analyses (150-200 words, 10 Marks).
To determine whether logMVPA predicts GPA, we used a linear regression model along with a Pearson correlation to test.
A Pearson correlation test was conducted and the coefficient of GPA and logMVPA found to be 0.6648. This suggests that there is an averagely strong positive relationship between these two variables.
You will also find the scatter plot showing GPA versus logMVPA accounting for the overweight
Figure 3: The scatter plot of GPA/logMVPA
The figure clearly shows that GPA and logMVPA have a linear positive relationship.
Regression analysis was used to develop a model capable of predicting the GPA score using logMVPA.
The fit of the model was determined by two important aspects: the coefficient of determination and the significance of the model.
First, we examined the goodness and fit of the model. We found that it is indeed capable of predicting GPA based logMVPA with a 5% level significance (p0.05).
The coefficient for determination (R-squared), which is 44.2% of variation in the dependent variable, GPA (based on logMVPA at 5% level of significance) in this model, is 0.442.
The model also identified logMVPA as significant (p 0.050), with its coefficient being 1.66256. This means a logMVPA change would result in a decrease in GPA by 1.6256.
We expect the GPA increase by 1.66256 if logMVPA is increased by one unit.
In the same way, GPA would drop by 1.6256 if logMVPA is decreased by one unit.
The constant intercept was at 4.03387
Please answer the research question.
This study explored whether the logarithms of MVPA (logMVPA), could predict GPA, after correcting the overweight population of Australian university students.
To answer this question, regression models have been constructed.
Results revealed that the logMVPA accurately predicts the GPA if overweight is controlled.