Analysis of Statistics in Scientific Articles - Article Example

Statistics

Introduction

Research is a fundamental tool for addressing social problems. For them to fulfil this aim, they must provide sufficient information and evidence regarding different aspects of the problems, and this is not possible without the use of statistics (Dowdy, Wearden & Chilko, 2011). Statistics enables the researchers to convey their information and findings in a scientific manner that will require the knowledge in this field to interpret it. Statistics is applied in many fields such as; economics, health, agriculture, and engineering (Adams & Lawrence, 2018). It is, therefore, a fundamental part of any research study, and it is important to know how to interpret various statistical terms to get the deeper meaning of the research. The focal point of this paper, therefore, is to identify and analyze statistics in the given research articles.

Question 1

In the first research article about Alcohol and other drug use at school leavers’ celebrations, Lam, Liang, Chikritzhs & Allsop present two hypotheses;

Hypothesis 1

H0: There is no statistical significance between the levels of AOD used at the end of school celebration and other peer-based social events

H1: There is statistical significance between the levels of AOD used at the end of school celebration and other peer-based social events

Hypothesis 2

H0: There is no statistical significance between harms experienced and the levels of use and engagement in harm-minimization strategy

H1: There is statistical significance between harms experienced and the levels of use and engagement in harm-minimization strategy

The dependent variable in the first hypothesis is AOD used and in the second is harms experienced. Both hypotheses were tested using six independent variables, that is, an average quantity of alcohol used on a single day at the event, use of drugs other than alcohol, use of alcohol-related safety measures as assessed by PBSS, gender, accommodation location and survey administration modality (Liam et al. 2014).

Question 2

The data used was collected using a two-part survey design with self-report methodology. In other words, they used questionnaires to gather the data. Some questionnaires were administered through face to face interviews and others online. The self-report methodology has its advantages and disadvantages. One of its advantages is that it is possible to study a large sample of people easily and cheaply. Besides, the surveys can test a large number of variables. Also, the results can be generalized to the larger population if the sample was selected randomly. In addition, the survey can ask people to reveal behaviour and feelings that have been experienced in real situations (Spector, 1994). The method has its weaknesses, such as; there are high chances that people will lie about certain things either because they do not remember or they want to represent themselves in a certain manner. Also, it may be difficult to get a random sample from the population as a result of selected people refusing to answer questions, or it may be difficult to obtain a full list of the population from which to select a random sample. In addition, cause and effect from the survey data cannot be established since other variables which could have affected may not have been included in the questionnaire or interview (van de Looij‐Jansen, Goedhart, de Wilde & Treffers, 2011).

Question 3

From the article, the respondents were aged between 17 years and 18 years. Despite the legal purchase age of alcohol in Australia being 18 years, majority of the respondents were 17 years. These respondents either intended to attend or attended the 2009 school leavers celebrations on the Rottnest Island in Western Australia. The Island was an ideal location to facilitate the survey administration since most of the visitors entered and exited through a single ferry terminal. In the first survey of the people who intended to attend the celebrations, a sample size of 541 people were collected, 56% being female, 91% 17 years and 9% 18 years of age. 87% of this sample attended an independent school, meaning that most of these respondents came from high-income families. In the second survey of people who attended the celebrations, a sample size of 405 was used. From this sample, 50% were female, 94% 17 years and 6% 18 years and above. About 92% of the people who attended the celebrations enrolled in independent schools, meaning they come from high-income families. Both surveys were available online and face to face. In general, the population surveyed comprised of people between 17 and 18 years (Liam et al. 2014). Majority of these population came from high-income families since they attended an independent school and most were female.

Question 5

According to the article, a series of Wilcoxon signed-rank tests for paired data were performed to compare the AOD use between the contexts of the last event and an average day at leavers celebrations. The reason for the use of this non-parametric method in the first hypothesis is because we do not know the distribution and parameters of the population (Adams & Lawrence, 2018). The Wilcoxon tests revealed that there is no statistical significance between AOD use in the contexts of last peer social event and an average day at leavers celebrations, hence, accepted the null hypothesis. The second hypothesis used a multivariate regression analysis since it wanted to assess the impact of the six independent variables on negative consequences experienced in peer social events. The p-values and confidence intervals were also used since they reveal if there are any significant differences in the data to enable the researchers to accept or reject the null hypothesis (Dowdy, Wearden & Chilko, 2011).

Question 5 & 6

The relationship between association and outcome determines an Odds Ratio. It establishes the possibility that an outcome will occur under a certain exposure, relative to the possibility of the outcome’s occurrence without exposure (Bland & Altman, 2000). According to the article, the OR for engaging in unprotected sex after AOD use compared with those who engaged in safety strategies with the greatest frequency was 10.92. This ratio is much greater than 1 meaning that peers under AOD influence that do not engage in safety strategies are 10.92 more likely to engage in unprotected sex compared to those who engage in safety strategies with the greatest frequency. The sample of the teens used is highly representative in this study since the sample was chosen at random, and more than a third of the population that was highly responsive was surveyed.

In the second article about the rising tide of diabetes mellitus in the Chinese population, Wong, Leung, Tsang, Lo & Griffiths have to aim for their research;

• To investigate the prevalence of self-reported diabetes by territory-wide household surveys representative of the whole Hong Kong population.
• To examine the factors associated with diabetes.

Question two

From the demographic characteristics of the article, the proportion of the respondents surveyed having been diagnosed with diabetes in these years:2001, 2002, 2005 and 2008, and make comparisons on the trends according to the age of respondent, sex of the respondent and the income of the respondent being used in measuring the socioeconomic status of the population under study (Porter & Whitcomb, (2005). The study used the age of the respondent and respondent's sex rates of prevalence by considering the dynamic changes in the demography of the population in different years. A total of 33,609 interviews were conducted in the year 2001 with a standard deviation of 16.35 recording the lowest deviation from its mean, 29,561 were conducted in the year 2002 having a standard deviation of 16.69, 29,802 in the year 2005 with a standard deviation of 17.90 and 28,923 interviews were conducted in the year 2008 having the highest deviation from its mean(SD=21.08). The adults with the age of 15 years and above recorded a total of 103,367. The population displayed some disparities in the demographic features for the respondents, thus implying there was a change in the structure of the population. This is because of the sampling technique used in sampling, which displayed similarity across all the three years. The percentage of the age-adjusted prevalence of diabetes among male adults were 2.80%(2001), 2.87%(2002), 3.32%(2005) and 4.66 %(2008). The female adults recorded the highest prevalence in the year 2008(4.31%) while the lowest level of prevalence was recorded in the year 2001(3.25%). At a glance, it can be seen that there was an increasing upward trend of self-reported diabetes across all the years in the age groups greater or equal to 75 years.

Question three

The inferential statistic used in this study is the Binary Logistic regression. This is because of the binary related variables under study (sex which can either be male or female). Here we have a binary outcome, for instance, male and female, and test the effect of predictors on this outcome. We use logistic regression to fit a straight line of the form Y=A + BX (Ostle, 1963).

Question four

From the article, Table 5 depicts the interviews conducted in Hong Kong in the year 2008 using stratified sampling in the whole territory. The total number of males interviewed were 14,189 males recording a total of 603 diabetes reported while the females interviewed were 14,734, which was higher than the males interviewed. Diabetes recorded by females is 5980 lower cases. Thus, bringing a total of 28,923 interviews. The findings thus show an increasing trend in self-reported diabetes across the years, with approximately 50 % from 2001 to 2008. This upsurge trend in diabetes was among the female residents of Hong Kong about ( 69.3 %)of the total population as compared with male respondents who are about 47.9 % of the total population. Those who are over 65 years of age recorded the highest diabetes cases(636). From the multiple regression analysis using the year 2001 as a reference for age, the relative increase in the rates of diabetes prevalence among males who are 15 years and above are (3.32% with standard deviation(17.50) in 2005 and 4.66 % with standard deviation(21.08) in 2008 were lower than the female adults who are 3.77% with standard deviation(19.04) in the year 2005 and 4.31%, having a standard deviation of 20.32 which is significant with a p-value less than 0.001. The prevalence of diabetes with age- and sex-adjusted shows an upward trend from 2001 to 2008, ranging from the least income earning groups among the males and females interviewed. The main cause of the increasing upward trend in the prevalence of diabetes may be due to increment of incidences over a long period and also the better healthcare system in Hong Kong (Gaskin et al. 2014).

Question five

After running the multiple regression analysis on the effect of respondents demographic features to the rates of diabetes, the respondents with 40-65 years of age had adjusted odds ratios of 32.2 with 95% confidence interval lying between 20.59 and 50.37 (Wong et al. 2013). It is significant in explaining the diabetes levels having a p-value which is less than 0.001 while those aged 65 years and above had an odd ration of 120.08 using a confidence interval of 95% lying between 76.6 to 188.26 which is significant since the p-value is less than 0.001. All the age groups were significantly associated with diabetes (Wang et al. 2017). The respondents' sex has no association with the prevalence of diabetes. By using a sample of 14,189 males, those with 65 years and above recorded the highest diabetes (299) with an adjusted odd ratio of 141.08 with 95% confidence interval lying between 69/47 and 286.48 which is significant(p-value<0.001).

Question six

The impact of the limitations described by the researchers on the self-reported information form the basis of the discussion and conclusion because the information gotten from the respondents are the one to give results on the prevalence of diabetes. It is from this information we are able to draw a valid conclusion; otherwise, it will yield spurious results.

The limitations are also very important by placing the research findings in context, interpreting the validity of the scientific work, and ascribing a credibility level to the conclusion of published research (Lowry, 2014). This goes beyond listing the magnitude and direction of random and systematic errors and validity problems.

Question 5

According to the article, a series of Wilcoxon signed-rank tests for paired data were performed to compare the AOD use between the contexts of the last event and an average day at leavers celebrations. The reason for the use of this non-parametric method in the first hypothesis is because we do not know the distribution and parameters of the population (Adams &amp; Lawrence, 2018). The Wilcoxon tests revealed that there is no statistical significance between AOD use in the contexts of last peer social event and an average day at leavers celebrations, hence, accepted the null hypothesis. The second hypothesis used a multivariate regression analysis since it wanted to assess the impact of the six independent variables on negative consequences experienced in peer social events. The p-values and confidence intervals were also used since they reveal if there are any significant differences in the data to enable the researchers to accept or reject the null hypothesis (Dowdy, Wearden &amp; Chilko, 2011).

Question 5 &amp; 6

The relationship between association and outcome determines an Odds Ratio. It establishes the possibility that an outcome will occur under a certain exposure, relative to the possibility of the outcome’s occurrence without exposure (Bland &amp; Altman, 2000). According to the article, the OR for engaging in unprotected sex after AOD use compared with those who engaged in safety strategies with the greatest frequency was 10.92. This ratio is much greater than 1 meaning that peers under AOD influence that do not engage in safety strategies are 10.92 more likely to engage in unprotected sex compared to those who engage in safety strategies with the greatest frequency. The sample of the teens used is highly representative in this study since the sample was chosen at random, and more than a third of the population that was highly responsive was surveyed.

In the second article about the rising tide of diabetes mellitus in the Chinese population, Wong, Leung, Tsang, Lo &amp; Griffiths have to aim for their research;

• To investigate the prevalence of self-reported diabetes by territory-wide household surveys representative of the whole Hong Kong population.
• To examine the factors associated with diabetes.

Question two

From the demographic characteristics of the article, the proportion of the respondents surveyed having been diagnosed with diabetes in these years:2001, 2002, 2005 and 2008, and make comparisons on the trends according to the age of respondent, sex of the respondent and the income of the respondent being used in measuring the socioeconomic status of the population under study (Porter &amp; Whitcomb, (2005). The study used the age of the respondent and respondent's sex rates of prevalence by considering the dynamic changes in the demography of the population in different years. A total of 33,609 interviews were conducted in the year 2001 with a standard deviation of 16.35 recording the lowest deviation from its mean, 29,561 were conducted in the year 2002 having a standard deviation of 16.69, 29,802 in the year 2005 with a standard deviation of 17.90 and 28,923 interviews were conducted in the year 2008 having the highest deviation from its mean(SD=21.08). The adults with the age of 15 years and above recorded a total of 103,367. The population displayed some disparities in the demographic features for the respondents, thus implying there was a change in the structure of the population. Read More