Introduction
Most researches tend to select significant tests based on the nature of the data under analysis and the process of sample selection. Conclusions made in such papers normally depend upon the value of the corresponding statistical values. The two papers chosen will further illustrate these points.
Whether the procedures used are parametric or non parametric
The first paper on restenosis was non parametric. This was because the authors do not mention any calculation of standard deviation or mean. The distribution of the variable under scrutiny which was detection of instent restenosis was not analyzed or was not considered as the basis for making conclusions about the outcome of the experiment. This paper utilized distribution free methods to reach its conclusion. Furthermore, the sample size selected in the study was rather small i.e. 57 patients. Parametric methods are often utilized when the value of the sample or n used is greater than 100. This is because such large sample selections often imply that the variable will be normally distributed.
In this case 57 is too small a number to assume normal distribution thus justifying the classification as a non parametric method. Lastly, the procedures used in this experiment further qualifies as a non parametric test because analysis of stent image was ranked through a scale of three points ranging from good to moderate to poor quality images. Whenever precise measurements of the concerned variables are difficult to make i.e. when ranking is involved then one cannot make normality assumptions or one cannot use parametric methods. In this case, the assessment of in stent restenosis was considered positive based on existence of stent – edge – stenosis thus showing that some level of interpretation was used rather than absolute measurements which is an ideal condition for parametric methods. (Dodge, 2006)
In the second research involving oral hygiene, it can be said that the latter arguments also hold. The sample size used in the paper was small i.e. 49, there was no mention of standard deviation or probability distribution of the population so it can be assumed that little reliance was given to the normality of the sample and this qualifies the latter study as a non parametric one.(Dodge, 2006)
Types of statistical test used and what the results tell about the data
In the paper concerning oral hygiene the p value was utilized as a statistical test. In the compliance, it was found that there were significant differences between the parameters under consideration when Pu0.001 was reported. The results reveal that the data itself was not demonstrating random chance but that it was actually quite reliable.
Similarly, in the paper on restenosis, it can be argued that the statistical test used was also the p value. It was assumed that when the data exceeded that value then the relationship was not significant. (Dodge, 2006)
Whether hypothesis testing errors were present and consequences of type 1 or type 2 errors in the projects
A type 1 error in the project on oral hygiene would have occurred if the null hypothesis was invalidated when it was actually true. In other words if there was no relation between oral hygiene promotion and compliance and the authors found the reverse that there is a relationship then a type 1 error would have occurred. This is essentially dangerous. In order to test whether such an error occurred, the authors had to determine the probability value (denoted as p) in their work. For rejection of a null hypothesis, researchers often assume that the probability value should be below five percent. As the authors pointed out, they found that the probability value was much less than that percentage i.e. it was 0.001. This means that there is no danger of the type 1 error and that what the authors had done which was to invalidate the null hypothesis was actually accurate. A type 2 error in the oral hygiene paper would have occurred if the null hypothesis is validated when it has been found to be false. In this case, that did not arise because the null hypothesis was invalidated. In other words, the authors affirmed that hygiene oral promotion led to compliance which is not an endorsement of the null hypothesis.
In the second paper which involved restenosis, a type 1 error would have occurred if it had been found that CTA assessment of restenosis was a feasible method for diagnosis of such patients. In fact, this is what the authors affirmed. However, because there were no reported probability values, it is difficult to assess whether this rejection of the null hypothesis was actually true or was a type one error. A type two error would have occurred if the authors claimed that there was no feasibility of CTA assessment when there actually was. However, this did not take place because the converse was actually true.
Corroboration of the Results with other Studies
In 2004, a similar study was carried out in Wuhan City within the Republic of China. An oral health promotion programe was administered in a school and the results wee published in the International Dental Journal, issue no 54. The authors were Petersen Poul, Peng Bin, Tai Baojun, Bian Zhuan and Fan Mingwen. These authors utilized a sample of 803 children and monitored them for three years. It was found that in schools where the oral health promotions were carried out, compliance levels were higher than in the control schools. Furthermore, it was affirmed that there were fewer cases of gingival bleeding amongst the participants of the program than the control group. Also, hygienic practices significantly improved. This study corroborates with the research under analysis and renders credibility to it especially given the fact that the sample selections were carried out in different parts of the world and that the sample size in this particular study was much higher than it was in the research project under analysis as the latter only entailed 49 students. Additionally, not only was compliance affirmed but it was taken a step further and linked to gingival bleeding.
In the second research project under analysis, the authors claimed that CTA was a suitable non invasive alternative for assessment of instent restenosis. However, these findings were contradicted by a number of researches. In a journal published in the American Neuroradiology journal, Turk et al (2007) affirm that CTA imaging misled the team to think patients suffered from in stent thrombus instead of in stent restenosis which had been earlier confirmed. This implies that the CTA was not that reliable and that the latter research opposes findings by Reinhardt et al (2010).
Another study carried out by Zsolt et al (2010) found that although there were significant improvements in the CTA predictive values, their influence over diagnostic value was minimal due to restrictions brought on by patient and stent factors. Other authors (Peloschek et al, 2006) appeared to support Reinhardt et al (2010) but asserted that this would only work once the method was combined with other diagnostic procedures. For instance, if utilized in combination with angiography and ultrasound then the assessment may be reliable at that point. It is exceedingly important to realize that these authors believe that the role of CTA is for quantification purposes or exclusion purposes of in stent restenosis once some other method has been used. To this end, one can assert that the findings made by Reinhardt actually contravene previous and subsequent works and this implies that perhaps an error could have occurred in their analysis.
References
Peloschek, P., Sailer, J., Loewe, C., Schillinger, M. & Lammer, J. (2006). Role of multisclae CT angiography after endovascular therapy. Cardiovascular and interventional radiology, 29(5), 756-761.
Zsolt, S., Kerecsen, G., Maurovich, H., Lux, A., Marosi, E, Kovacs, A., Kiss, R. & Merkely, B. (2010). Determination of coronary instent restenosis using dual source computed tomography angiography. Interventional medicine and applied science, 2(1), 5-9.
Turk, A., Rowley, H., Fiorella, D., Pulfer, K. & Strother, C. (2007). CT Angiographic appearance of in stent restenosis. American Neuroradiology journal, 28(4), 52-54.
Dodge, Y. (2003). Oxford Dictionary for statistical terms. Oxford; OUP.