Determining Adequacy of Statistical Analysis

NB: We do not resell your papers. Upon ordering, we do an original paper exclusively for you.

NB: All your data is kept safe from the public.

Table of Contents

Introduction
Analysis
Recommendation
Conclusion
References

Introduction

This paper evaluates two cases based on which statistical analysis was carried out, and specific results were obtained. This work aims to determine the adequacy of this analysis to disassemble the variables and values. The aim of the work is also to identify recommendations for the researcher that would help improve the results of such experiments.

Analysis

P-value

Any research is carried out based on only a part of the objects. The study of the effectiveness of a drug is carried out on the basis of not generally all patients on the planet, but only a particular group of patients. The p-values indicate the reliability of a hypothesis, or, in other words, how much it can be trusted. The higher the p-value, the less it can be trusted. The p values in both cases are large enough, but this cannot completely rule out the validity or probability of the null hypothesis (Andrade, 2019). These p-values indicate that either the sample size is too small, the effect size is small, or the issue is highly variable (Krueger & Heck, 2019). Although high p-values are not statistically significant and are pretty typical for point biserial correlation, they should not be underestimated: they protect from essential decisions, the performance of which is based on random error.

P for both cases is calculated based on sample-based test statistics. The boundary values of the cumulative distribution function show the most optimal p values, which are often calculated by software or a standard and commonly used value of 0.05 is taken (Kraemer, 2019). In the first question, anasarca was the most significant symptom of deterioration in a patient with renal failure. However, high p values indicate that this practice cannot be used as the only correct one. In this question, it is needed to either increase the sample or clarify other input data, in which this symptom turned out to be indicative.

In the second question, the best solution was to leave the knife in the chest. However, the p-value is even higher here, which increases the error probability many times over. Therefore, it is impractical to decide based on these data since the probability of a random error is high (Bonett, 2020). As a guide to action in these typical emergencies, this experiment is of no interest.

Difficulty

The researcher’s use of point-biserial correlation is somewhat controversial since this type of correlation requires dichotomous variables, that is, those that take only two values. A person entering with a knife in the chest or symptoms of renal failure are rather complex multicriteria questions. In this case, it is possible to limit ourselves to two possibilities only in the case of a long iterative approach, during which many questions will be asked. In any other case, it is needed to choose a different type of regression analysis to find correlations. The Educator used several dichotomous variables to show the complexity of the question. However, having obtained high p-values, it is not easy to reason about the significance of the results. Perhaps there are other external or internal determinants of these processes, the influence of which cannot be ignored even during the search for new dependencies.

The Points Biserial

Most of the values are above zero in the first question, but only two of them show “good” results. It means that nocturia and fatigue are expected and confirmed symptoms of kidney failure, while anasarca is negative, which signals a disorder. In the second case, the radical decision to pull out the knife and treat it with a particular substance is at a zero level, indicating these solutions’ inadequacy. Leaving the knife in place has an extremely high indicator for this analysis, which generally meets first aid requirements (Franke et al., 2017). High p-values do not give statistical power to the results obtained; however, due to high eigenvalues, they should be considered.

The point biserial provides information on the correspondence of the expected symptoms of the dynamics in the patient’s well-being with renal failure in the first case. This fact means that the dichotomous variables that patients filled in with yes or no answers reflect their state compared to the dynamics of health. At the same time, their indicators of renal failure are recorded, and a correlation is made. In the second case, higher values of the bead point show the correspondence of the action to the best outcome, that is, one in which the patient will ultimately survive and receive the least amount of harm to health.

A distractor is a deliberately incorrect answer introduced into a statistical study to determine the adequacy of the chosen technique. In the first case, the distractor is the answer to fatigue; in the second, the immediate removal of the knife from the chest. If, in the second case, the approach of this correlation turned out to be suitable and this answer was immediately recognized as incorrect, then in the first case, the situation is not so transparent. On the one hand, fatigue can be caused by various reasons. On the other hand, the list of other symptoms cited in the study is more severe and requires treatment. In this regard, relatively high indicators of the bead point in fatigue indirectly show a small degree of inconsistency of the method.

Recommendation

The researcher should make a decision only within the group in which it was carried out. It is due, first of all, to the p values and the high level of the general point at the distractor in the first case. It is these two values that do not provide extrapolation of this study. However, the researcher can increase the complexity of the problem, apply an iterative approach, and change the regression analysis model to obtain more accurate and statistically significant results. The analysis must have a more complex causal relationship in the first case. Between the symptoms of kidney failure and worsening of the condition, there must also be qualifying factors and factors that can significantly influence the accidental error. In the second case, which cannot be reproduced in real life, even for the purpose of the experiment, it is needed to rely on the already used practices of first aid. In general, the values of the points biserial show promising results in the second case, which partly even outweighs the high value of p.

Conclusion

Such studies can assess the dependences of symptoms and dynamics of indicators within a tiny group, as in the first case. These approaches have proven themselves well in assessing situations that cannot be reproduced in real life. Nevertheless, it is not worth using the results of these studies as a roadmap in such situations due to the multicriteria and complexity of the questions presented. It is possible to develop methods using an iterative approach that will consider each criterion at each stage. For a complete analysis, it is necessary to use other mathematical models or regression models.

References

Andrade, C. (2019). The P-value and statistical significance: misunderstandings, explanations, challenges, and alternatives. Indian Journal of Psychological Medicine, 41(3), 210-215.

Bonett, D. G. (2020). Point‐biserial correlation: Interval estimation, hypothesis testing, meta‐analysis, and sample size determination. British Journal of Mathematical and Statistical Psychology, 73, 113-144.

Franke, A., Bieler, D., Friemert, B., Schwab, R., Kollig, E., & Güsgen, C. (2017). The first aid and hospital treatment of gunshot and blast injuries. Deutsches Ärzteblatt International, 114(14), 237.

Kraemer, H. C. (2019). Is it time to ban the P-value?. JAMA Psychiatry, 76(12), 1219-1220.

Krueger, J. I., & Heck, P. R. (2019). Putting the p-value in its place. The American Statistician, 73(sup1), 122-128.

Click Here To Order Now!