# Inferential Statistics

INFERENTIAL STATISTICS 5

InferentialStatistics

InferentialStatistics

Abetter comprehension of the major aspects of inferential statisticsis needed to incorporate the results from empirical research into thenursing practice. The focus is to describe statistics terms, describehow it will design studies that full optimize specificity,sensitivity and predictive values and also explain how they will beoptimized to minimize type I and type II error (Vach& Werner, 2013).

*Errors*

TypeI error is the rejecting incorrectly a true null hypothesis, itrepresent a false positive with respect to a non-null hypothesis. Forexample, a test showing a certain disease in a patient is void of thedisease. In the same vein, type II errors are the failure to rejectsa null hypothesis which is null. For example there is blood testingfailure to identify a certain disease that it was presented toidentify (Vach& Werner, 2013).

*Sensitivity*

Sensitivityof a clinical test is the ability of the test to identify correctlythose patients who have a particular ailment or disease. A test witha sensitivity of 100% presents a patient with the ailment. A testwith 80 percent sensitivities detects patients with the disease whilethe 20 percent goes undetected. A high sensitivity is vital intreating a serious ailment like cancer (Chakraborty,Ranajit, Rao, & Sen, 2012).

*Specificity*

Specificityin clinical terms is the ability of the test to identify correctlythose patients without the disease. For instance, a test having 100percent specificities identifies all the patients without theailment. In the same vein, a test with 80 percent of the patientsreports an 80 percent without the disease as a negative test( truenegative) whereas a 20 percent are incorrectly identified as apositive test (false positive) (Chakraborty,Ranajit, Rao, & Sen, 2012).

*Predictivevalues*

Predictivevalue measures the likelihood or the unlikelihood of a given test.Predictive value is categorised as either positive or negative.Positive predictive value answers the question how likely the patientis infected by the disease having the results from the test asnegative. Negative predictive value, on the other hand, answers thequestion how likely the patient does not have the disease having theresults as negative(Broemeling& D, 2007).

*Howinferential statistics will optimize sensitivity, specificity, andpredictive values to minimize type I and type II errors *

Diagnostictest is needed for the assessment of clinical disease and is a goodguide in testing a scientific hypothesis. Specificity and sensitivityare entities of test and also are not predictive of diseases in theindividual patients. Positive and negative predictive values are verynecessary for predicting of the diseases in the patients and itdepends on disease prevalence and diagnostic test. A betterunderstanding of inferential statistics is essential to optimizingspecificity, sensitivity and predictive values as depicted in thedefinitions of the terms(Broemeling,2007).

*Howstatistical power was used to verify the findings*

Poweranalysis isthe ability to detect an effect (caused by the study invention)having that the effect exists (not by chances). It is also thechances that the chance will reject the null hypothesis.Consequentially, it is used to verify the findings by detecting themain difference in the co-existing groups (Burns,Nancy, & Grove, 2009).

Itis of utmost significance used to determine the result of aparticular intervention. The determination in itself cannot beattained when the power is insufficient. Conversely, when the poweris very high, the researcher may be influenced in the to give moreweight to the results than those required (Burns,Nancy, & Grove, 2009).

Conclusion

Itis clear that an intervention can use power analysis. Sample size isimportant in designing statistical results. Sized samples are usefulto give confidence that a particular sample reflects the populationparameters. It can be used to accept or reject a hypothesis of thestudy. In conclusion, a study that is powered efficiently has a scanable chance of answering the questions at hand (Vach& Werner, 2013).

References

Broemeling,D. L. (2007). *.Bayesian Biostatistics and Diagnostic Medicine.*Boca Raton: : Chapman & Hall/CRC,.

Burns,Nancy, & Grove, S. K. (2009). *ThePractice of Nursing Research: Appraisal, Synthesis, and Generation ofEvidence.*St. Louis, MO: : Saunders/Elsevier, .

Chakraborty,Ranajit, Rao, C. R., & Sen, P. K. (2012). *.Bioinformatics in Human Health and Heredity.*Amsterdam: : North Holland, .

Vach,& Werner. ( 2013). *.Regression Models as a Tool in Medical Research.*. Boca Raton,: FL: CRC,.

## Recent Comments