Teenage Health Vulnerabilities with Substance Abuse United States

Discuss how you would use regression to learn more about the nature of the relationship between the variables
Regression is a significantly common technique used for analyzing data. Regression analysis is employed to delineate the relationships between a group of independent variables and the dependent variable. In this case, the dependent variable is substance abuse. On the other hand, the set of independent variables comprise of exposure to substance abuse, ethnicity, age, and gender, poverty, and loneliness. Regression would play a significant role in helping with the learning process regarding the nature of the relationship between the variables. The analysis will particularly entail descriptive statistics as well as regression, correlation, and ANOVA analysis. Descriptive statistics will be used to present demographic data and provide measures of central tendency such as averages and median, while the statistical analyses will be important for testing the relationship between the dependent and independent variables (Draper and Smith, 2014).
One-way ANOVA is delineated as an approach that can be utilized for the comparison of means of two or more samples. There are different necessitated conditions for a one-way ANOVA. First of all, the variances of populations are the same. Secondly, the response variable residuals follow a normal distribution and lastly, the responses for a particular group are independent and they are also random variables that are normally distributed and are identical (Weiers, 2010). The analysis will seek to analyze the F-Statistic. Basically, this is the figure that is attained subsequent to conducting a regression analysis or subsequent to running an ANOVA test. This is purposed to ascertain whether the means of two populations are substantially different (Weiers, 2010).
Specifically, this research study will use multiple regression analysis. This is because there are different sets of variables. This analysis is meant to determine the correlation and trends in these sets of data. Therefore, in this case, the different independent variables will be set in regard to the different predictors including: (x1)1, (x2)1, (x3)1, Y1).
The equation used in the research study will be:
Multiple regression: Y = b0 + b1 x1 + b0 + b1 x2…b0…b1 xn
In this case Y = substance abuse
(x1) will be level of exposure
(x2) will be ethnicity
(x3) will be age
(x4) will be gender
(x5) will be poverty
(x5) will be loneliness
Discuss any issues associated with measurement error that could be plaguing your study
Measurement error takes into account the dissimilarity between a measured quantity and its true value. There are two types of measurement error including random error, which is an error that naturally takes place and is anticipated to be present with any sort of experiment, and also systematic error, which is instigated by misconstrued research instrument that impacts all measurements. There are a number of issues linked with measurement error that could be adversely impacting my research study. Notably, this is an epidemiological research study, which measures characteristics of a population. Bearing in mind that the parameter of interest in this particular case is the prevalence of an exposure, and that the research study is carried out on individuals, it is subject to bias (Trochim, 2006). There are different types of measurement error. One of them is systematic error, which is a kind of error that brings about measurements that incessantly shift away from the true value in the similar direction. The second type of error is the random error, which takes into account the variance between a person’s reported understanding of the research instrument’s reading compared to the actual reading. Third, there is gross errors, which take into account the physical errors in investigation tool or computing and recording measurement results. On the whole, these gross errors will come about all the way through the experiments, wherever the researcher may study or record a worth dissimilar from the actual one, perhaps as a result of a reduced view (Trochim, 2006).
How would you know/identify it, and what can be done about it?
One of the key ways in which I would identify measurement error is by assessing the quality of measurements in the research study. An ideal survey technique is valid. From time to time a reliable standard is obtainable against which the validity of a survey technique can be evaluated. In this case, measurement of the exposure to substance abuse and also measurement of illness is incapable of complete validation. Therefore, from a practical point of view, it would be possible to assess validity indirectly. Two approaches are prevalently employed. A technique that has been streamlined and standardized to make it appropriate for use in surveys might be matched with the best conventional clinical evaluation. A self-administered substance abuse questionnaire, for example, might be contrasted against the popular view of a substance abuse panel. On the other hand, a measurement may be validated by its capability to predict forthcoming instances of substance abuse. Validation by prognostic ability might, nonetheless, necessitate the study of numerous subjects (The BMJ, 2019).
According to Trochim (2006), one of the ways of dealing with or reducing measurement error is conducting a pilot test for research instruments, obtaining feedback from the research participants concerning how simple or difficult the measure was and also information regarding how the environment in which the testing was undertaken impacted their performance. In addition, reducing measurement error can be facilitated by double checking the data collected in an extensive manner. That is, data entry can be done twice to ensure that the data collected is in fact the one entered. Furthermore, statistical processes can be employed for adjusting measurement error. These vary from rather basic formulas that are directly applicable to the data to significantly intricate modeling processes for modeling the measurement error and its impacts (Trochim, 2006).
Discuss the concept of factor analysis and how you can apply it to the development of a questionnaire designed to study your identified public health issue
Factor analysis is an approach that is employed to facilitate the reduction of a large number of variables into smaller numbers of factors. Specifically, this approach excerpts the greatest common variables from all of the selected variables and places them into a common score. Subsequently, this score is utilized as a form of index for all of the selected variables and employed for further analysis. Factor analysis can be applied to the development of a questionnaire that is designed to study the public health issue at hand. Specifically, factor analysis will be beneficial in condensing variables as well as uncovering clusters of responses (Brown, 2014).
For instance, say the questions included in the questionnaire are all driving at dissimilar, but closely associated aspects of substance abuse:
1. Do you partake in substance abuse?
2. Would you recommend it to a fellow teen?
3. How likely are you to purchase such substances in the future?
However, the main objective is to select one variable to represent substance abuse score. In this regard, one option in regard to factor analysis would be to average the responses obtained from the three questions. There is also the option of generating a factor dependent variable and this could be undertaken by carrying out Principal Component Analysis. The strong suit of the latter over obtaining an average of the solutions is that it automatically takes into consideration the weight of each of the variables (Brown, 2014).
Discuss the concept of statistical power as it relates to your study
Power is delineated as the probability that a research study will end up rejecting the null hypothesis when in actual fact it is false. More often than not, research studies that have high statistical power have a greater likelihood of detecting the effects of the interventions or treatments put in place whereas the research studies that have low statistical power can direct the researchers in dismissing prospectively significant effects as sampling error (Murphy, Myors, and Wolach, 2014). Imperatively, the statistical power of a test can be defined as a function of the size of the treatment effect within the target population, the sample size and also the specific criterion that is employed to delineate statistical significance (Murphy et al., 2014).
The concept of statistical power will largely relate to this research study in the sense that it will be a metric of the probability that the researcher will attain statistical significance in the research sample if the impact is existent in the population as a whole. That is, whether the factors pointed out earlier will have an impact on the study sample of the teenagers selected and whether this impacts teenagers as a whole. Imperatively, statistical power is a function of both primary and secondary factors including the sample size, the effect size, the level of significance and also the power of the statistic that is employed. Bearing this in mind, there will be a need to conduct a statistical power analysis in order to ascertain the sample size necessitated for the study and also conduct the analysis subsequent to the completion of the study to ascertain if the reason an effect failed to significant was because there was insufficient power (McHugh, 2008).
Discuss which non- parametric statistical techniques you could use to study your problem. Be certain to explain why the selected technique would be most appropriate
Nonparametric statistics allude to a statistical technique in which the data is not necessitated to fit a normal distribution. According to Grant and Kenton (2019), nonparametric statistics employs data that is more often than not ordinal, which implies that it is not dependent on numbers, but is instead reliant on a ranking or order of sorts. A statistical technique is termed as being non-parametric if it fails to make any sort of supposition on the population distribution or sample size. This is different from majority of the parametric methods in elementary statistics that make the supposition that the data set used is quantitative, the population has a normal distribution and the sample size is satisfactorily large. On the whole, inferences drawn from non-parametric methods fail to be as influential in comparison to the parametric ones. Nevertheless, as non-parametric techniques make lesser suppositions, they have the major advantage of being more flexible, more forceful, and pertinent to non-quantitative data (Buskrik and Willoughby, 2013).
There are different tests that I would utilize in the study. One of them is the Kruskal-Wallis test. Basically, this is a rank-based nonparametric test that can be employed in ascertaining if there are statistically significant dissimilarities or variances between two or more groups of an independent variable. This method would be chosen as the conditions are also ideal. It is imperative to note that in the earlier section, it was pointed out that ANOVA would be used in order to test whether the means, which are more than three are equal. The nonparametric alternative to an ANOVA test for means is the Kruskal-Wallis test (Mckight and Najab, 2010). The conditions for the Kruskal-Wallis test include the following:
1. The assumption made is that the samples drawn from the population are done in a random manner
2. There is also the assumption that the observations made from the sample are independent of one another
3. The measurement scale for the dependent variable ought to be at least ordinal (McKight and Najab, 2010).
Discuss your general analysis plan
An analysis plan ought to be formed before partaking in the data analyses. The analysis plan in this case will contain a comprehensive research question that will be initially formulated. Basically, this is the question that is purposed to be answered or resolved by the different data analyses. In this case, the research question is: What are the different factors that impact teenage substance abuse in America. Secondly, the analysis plan also ought to delineate the different statistical techniques that will be utilized in analyzing the data. In this case, the population included in the study is teenage substance abusers in the United States. Stratified sample is selected to facilitate obtaining the ideal research sample. The dependent variable is substance abuse whereas the independent variables include exposure to substance abuse, ethnicity, age, and gender, poverty, and loneliness. Regression analysis will be employed for analyzing the data collected.
Discuss how you would address the presentation of your study’s results
A key constituent of the scientific and experiential research is the presentation of the research findings to the target audience. First of all, the written report would be presented using different topics. The first section will be the introduction, which will provide succinct background and overview information regarding the problem or phenomenon to be investigated. The introduction will encompass the contemporary state of insight and understanding regarding the phenomenon. The section will close with a properly outlined hypothesis of the study and experiential expectations. Secondly, there will be the methods section, which will delineate the different procedures and processes employed in testing the hypotheses aforementioned. These processes will be delineated in extensive detail that they can be reiterated by others. This section will also comprehensively outline the materials, analytical as well as statistical techniques and procedures.
The following section of the written presentation will be results section. This is where a summary of the experimental results of the research study will be delineated. It will include a proper explanation of the results in addition to well-placed tables and figures, demonstrating statistical outcomes and the experimental error. Specifically, these figures and tables will comprise of the different legends that shed light on what they are outlining. The fourth section of this report will include the discussion of the study. Basically, in this section, the outcomes obtained in the preceding chapter will be interpreted and thereafter inferences will be drawn. In addition, this section will encompass a succinct comparison of the results obtained to the outcomes that were attained from other research studies in the existent literature and thereafter provide the overall significance of the research findings. Lastly, any kind of limitations faced in the research study will be pointed out, in addition to reasons for the measurement errors and also prospects of future research studies.
Presentation of the study will also include an oral and PowerPoint presentation. It is imperative to note that these presentations will be fundamentally similar to that of the written report and will include the aforementioned section comprising: introduction, methods, results, and discussion. Finally, a concise period for any questions will be permitted.




References
Brown, T. A. (2014). Confirmatory factor analysis for applied research. Guilford Publications.
Buskirk, T. D., Willoughby, L. M., & Tomazic, T. J. (2013). Nonparametric statistical techniques. The Oxford handbook of quantitative methods. Statistical analysis, 2, 106-141.
Draper, N. R., & Smith, H. (2014). Applied regression analysis (Vol. 326). Hoboken: John Wiley & Sons.
Grant, M., Kenton, W. (2019). Nonparametric Statistics. Investopedia.
McHugh, M. L. (2008). Power analysis in research. Biochemia medica: Biochemia medica, 18(3), 263-274.
McKight, P. E., & Najab, J. (2010). Kruskal?wallis test. The corsini encyclopedia of psychology, 1-1.
Murphy, K. R., Myors, B., & Wolach, A. (2014). Statistical power analysis: A simple and general model for traditional and modern hypothesis tests. Routledge.
The BMJ. (2019). Measurement error and bias. Retrieved from: https://www.bmj.com/about-bmj/resources-readers/publications/epidemiology-uninitiated/4-measurement-error-and-bias
Trochim, W. M. (2006). Measurement error. Social Research Methods. Retrieved from: https://socialresearchmethods.net/kb/measerr.php
Weiers, R. M. (2010). Introduction to business statistics. New York: Cengage Learning.