4 days agoEnica Saffold Effect Size and Practical and Statistical SignificanceCOLLAPSEEffect Size and Practical and Statistical SignificanceLoss of IGF2 Imprinting: A Potential Marker of Colorectal Cancer RiskMolecular genetics plays a role in cancer risk assessment, for in this instance, colorectal cancer (CRC), but the challenge is developing genetic tests used to identify individuals that might be at an elevated risk of developing cancer. Four hundred twenty-one patients had consented to participation in the study from 1999 to 2001; from that population, 172 patients were evaluated to investigate the utility of loss of imprinting (LOI) as a marker of colorectal cancer patients. Hypothesis testing was performed with multiple logistic regression adjusted for age, sex, race, and family history. Odds ratios, p-value, and confidence intervals were all evaluated in this study as well.For patients where a positive family history, the adjusted odds ratio for LOI in lymphocytes was 5.15 times that of an individual with no family history at a 95% CI; the p-value was measured at 0.002. Odds of LOI in patients with past or future adenomas but no risk for cancer were 3.46 times compared to patients with no past or present colorectal neoplasia (p=0.026). Lastly, patients with a past or present CRC were 21.7 times more likely than patients without (p=0.0005) (Cui et al., 2003). Thus, overall data suggested that LOI is positively correlated with the initiation and progression of colorectal neoplasia. As stated earlier, the sample size was 172 participants. The effect size reflected in the odds ratio was 5.15 with patients with family history, 3.46 with patients with past or future adenomas, and 21.7 with patients with past or present CRC. Overall all the odds ratios were small with even more minor statistical significance. Statistical significance shows an effect exists in a study, whereas practical significance indicates that the effect size is large enough to be meaningful. Statistical significance refers to the unlikelihood that means the difference observed within the sample hasn’t occurred due to sampling error. Practical significance looks at whether the difference is significant enough to value in a realistic setting (Utts, 2005). Based on this article, the effect size was both statistically significant and practically significant. All p values were low (<0.5), which is said to be statistically significant, allowing researchers to reject the null hypothesis of no effect.When you run an experiment, you’re taking a sample of some population (in this instance, individuals at risk for CRC). No matter what you’re studying, the process for evaluating significance is the same. Start by stating a null hypothesis, often something you are trying to disprove, along with an alternative hypothesis. The significant level expresses how to rate the results are, assuming that the null hypothesis is true. It’s usually as p values, and the lower the values, the less likely the results are due purely to chance. Statistical significance is important because it allows researchers to hold a degree of confidence that their findings are accurate and reliable. But, because statistical significance is a way of quantifying how much faith to have in a research finding, researchers are often more interested in a finding’s practical significance than statistical significance (Daniel & Cross, 2019).ReferencesCui, H., Curz-Corea, M., Giardiello, F., Hutcheon, D. F., Kafoneck, D. R., Brandenburg, Wu., Y., He, Z., Powe, N. R., & Feinberg, A. P. (2003). Loss of IGF2 Imprinting: A potential marker of colorectal cancer risk. Science, 299(5613), 1753-1755.Daniel, W. W. & Cross, C. L. (2019). Biostatistics: A foundation for analysis in the health sciences (11th ed.). Wiley. Utts, J. M. (2005). Seeing Through Statistics: Third Edition. Thomson Learning Inc: Toronto, Canada. 2 days agoCarlin Nelson RE: Discussion 1 - Week 2COLLAPSEPost a brief summary of the article you selected.Mendall et al. conducted a case control study of 185 white men aged 45-65 to determine if there is an increased risk of coronary heart disease (CHD) later in life, with those who contracted Helicobacter pylori in childhood (Mendall et al.,1994). The controls were recruited from a health screening clinic, and the cases were recruited from a teaching hospital. The data was collected from diagnostic testing (serum, and high blood pressure reading) and surveys. The data was analyzed by odds ratio through multiple logistic regression (MLR). It was reported that there was a 95% statistically significant confidence 2.28 (p-value:0.007) increase in odds that those who had CHD later in life contacted Helicobacter pylori as a child. Additionally, after controlling for several factors, the odds ratio slightly decreased.Then briefly explain the interrelationship between the effect size, power, and sample size in the article. Explain whether the effect size was practically significant, statistically significant, or both, and why.Mendall et al. provided the odds ratio, the confidence interval and p-value. The sample size consisted of 185, with 111 identifying consecutive cases and 74 controls. Additionally, 66 of those 111 cases were categorized as exposed (testing seropositive for Helicobacter pylori) and 29 of the 74 controls are considered exposed. The different combination of the sample size influences the effect size. The MLR reported both the unadjusted odds ratio as 2.28 (95% C. I= 1.25-4.15, X2= 7.35, p-value=0.007). When controlling for age and risk factors, the odds ratio was 2.26 (95% C. I=1.15-4.44, X2=5.71, p-value=0.02). Furthermore, controlling for age, risk factors, and current social class, the reported odds ratio was 2.15 (95% C. I= 1.07-4.29, X2= 4.73, p-value=0.03). Including father’s occupation to the model, the odds ratio decreases slightly to 2.08 (95% C. I= 1.03-4.20, X2= 4.23, p-value=0.04). The last set of variables added to the previous model was housing density, and hot water supply in the childhood home which reported an odds ratio of 1.90 (95% C. I= 0.91-3.97, X2= 2.97, p-value=0.09). Although statistical significance assists in quantifying the results of the data being due to chance often conveys as the p-value, practically significance refers to how meaningful the results are in the real world, typically represented by effect size (Peeters, 2016). While the effect size decreased as more variables were being controlled, all models except the last model were statistically significant. All of the models that were considered statistically significant had p-values that were less than 0.05 and their confidence intervals did not include 1.0. With the “rule of thumbs” the odds ratios reported would be classified as small- medium effects. An odds ratio equaling 1.50 but less than 2.50 would be considered small, 2.50-4.30 would be considered medium and 4.30 and above would be considered large (Chen et al., 2010).Using the article you selected as an example, support the importance of incorporating both practical and statistical significance in study conclusions.Incorporating both practical and statistical significance in study conclusions is pivotal for both internal and external validity purposes. The p-value often interpreted as statistical significance should not solely be reported because it does not allow the reader to fully understand the results. A major limitation of the p-value is that it is heavily influenced by the sample size, when the sample size increases, the random error decreases and the p-value decreases (Thiese et al., 2016). The p-value being confounded does not always mean that an outcome has/hasn’t occurred but that there was a huge sample size. While reporting probability of results being due to chance in those with CHD being exposed to Helicobacter pylori important, quantifying the impact is more meaningful. With the effect of this study being identified as small to medium, it allowed the authors to convey that there is more research needed in this population. Combining practical and statistical significance in research not only assist in making the inferential statistics relatable but providing a clear, holistic understanding to stakeholders. ReferencesHenian Chen, Patricia Cohen & Sophie Chen (2010) How Big is a Big Odds Ratio? Interpreting the Magnitudes of Odds Ratios in Epidemiological Studies, Communications in Statistics - Simulation and Computation, 39:4, 860-864, DOI: 10.1080/03610911003650383Mendell, M.A., Goggin, P.M., Molineaux, N., Levy, J., Toosy, T., Schrachan, D., Camm, A., Northfield, T. (1994). Relation of Helicobacter pylon infection and coronary heart disease. British Heart Journal, 71(5), 437-439Peeters, M. J. (2016). Practical significance: Moving beyond statistical significance. Currents in Pharmacy Teaching and Learning, 8(1), 83–89. https://doi.org/10.1016/j.cptl.2015.09.001Thiese, M. S., Ronna, B., & Ott, U. (2016). P value interpretations and considerations. Journal of thoracic disease, 8(9), E928–E931. https://doi.org/10.21037/jtd.2016.08.162 days agoAnne REIS RE: Discussion 1 - Week 2COLLAPSECui, H. (2003). Loss of IGF2 Imprinting: A Potential Marker of Colorectal Cancer Risk. Science, 299(5613), 1753–1755. https://doi.org/10.1126/science.1080902 Summary of StudyCui et al. (2003) conducted a cross-sectional epigenetic pilot study to examine the effectiveness of Loss of imprinting LOI as colorectal cancer (CRC) risk marker. The investigators estimated the presence of LOI (gene alterations) and the association with increased expression of the insulin-like growth factor II gene (IGF2) for people with colorectal neoplasia (adenomas and carcinomas). The investigators performed a multiple logistic regression analysis to assess the association between demographic (age, sex) and clinical factors (familiar and personal history of colorectal neoplasia (predictor variables), and DNA and RNA mechanisms of peripheral blood lymphocytes (PBLs) (response variable). The study population comprised of patients who were undergoing a colonoscopy for any medical indication. Data was collected from 421 participants from medical history data and from proximal and distal colonic mucosal specimens.The Interrelationship Between the Effect Size, Power, and Sample Size. Cui et al. (2003) recruited a convenience sample of 421 patients to participate in the study. This sampling methodology was appropriate because it represents a specific population exposed to specific stimuli (colorectal neoplasm) and but is not generalizable to the general population. While there is no support that the sample size of 421 has sufficient statistical power to represent the over two million reported living with colorectal cancer, the study found a strong relationship, odds ratio (OR)between LOI (exposure) and PBL (outcome). The statistical power of a significance test is contingent on three factors, sample size (n), the significance level (α), and the effect size (Wasserstein, 2016). An increase in sample size corresponds with an increase of statistical power and effect size, while a rise in the effect size and significance level (α) corresponds with an increase in statistical power (Gelman, 2019; Daniel, 2019). The effect size shows how relevant the relationship between two variables (Gelman, 2019). If a study lacks power, then the degree to which the differences and relationships reported are neither meaningful nor applicable to social change (Gelman, 2019).Effect Size: Practically Significant, Statistically Significant, or Both. Cui et al. (2003) found a statistically significant relationship between the familiar history of CRC, personal experience with CRC, and PBL. The multi regression analysis were adjusted for age and sex. While the sample size is an essential determinant of statistical power for any empirical research study in which the objective is to make inferences about the hypothesis for a population from a sample. Small sample sizes lead to a risk of obtaining wider confidence intervals and errors in statistical hypothesis testing(Gelman, 2019; Roth, 2009). Cui et al. (2003) recruited a retrospective study and kept the sample size constant. The investigators found odds ratios greater than 2.5 in several relationships. Technically, when the effect size is 2.5, low samples are sufficient to obtain power greater than 80%. When the effect size is 1, increasing the sample size significantly increases the power of the study (Gelman, 2019; Daniel, 2019). Since the power of a study decreases as the effect size decreases, Cui et al. (2003) study was both practically (meaningful) significant and statistically significant. The importance of Incorporating Both Practical and Statistical Significance in Study Conclusions.Cui et al. (2003) provide the importance and its impact of the study on public health and the benefit of intervention to CRC screening and treatment. Research is about solving complex problems. Reported findings rely on the appropriate interpretation and use of statistical tests. The statistical tests provide confidence (power and sensitivity) that allows general utility and application of the findings. Meaningfulness in research reflects the value derived from research (Frankfort-Nachmias, 2020). The utility of hypothesis testing and Type I errors and Type II errors is widely used in medicine, law, biometrics, and genetics. ReferenceCui, H. (2003). Loss of IGF2 Imprinting: A Potential Marker of Colorectal Cancer Risk. Science, 299(5613), 1753–1755. https://doi.org/10.1126/science.1080902Daniel, W. W. & Cross, C. L. (2019). Biostatistics: A foundation for analysis in the health sciences (11th ed.). Wiley.Gelman, A., & Greenland, S. (2019). Are confidence intervals better termed "uncertainty intervals"? BMJ : British Medical Journal (Online), 366 doi:http://dx.doi.org.ezp.waldenulibrary.org/10.1136/b... Ranstam, J. (2012). Why the p-value culture is bad and confidence intervals a better alternative. Osteoarthritis and Cartilage, 20(8), 805–808. http://doi.org/10.1016/j.joca.2012.04.001Roth, J. V., M.D. (2009). Prediction interval analysis is underutilized and can be more helpful than just confidence interval analysis. Journal of Clinical Monitoring and Computing, 23(3), 181-3. doi:http://dx.doi.org/10.1007/s10877-Wasserstein, R. L., & Lazar, N. A. (2016). The ASA Statement on p-Values: Context, Process, and Purpose. The American Statistician, 70(2), 129–133. https://doi.org/10.1080/00031305.2016.1154108 Requirements: 4