Correlation is the process of establishing a relationship between two...

Correlation is the process of establishing a relationship between two or more factors. Correlation is an important concept that can be misused. One misuse is saying that factor A is caused by factor B just because correlation is found. Cause cannot be implied simply from correlation. Find two examples in scholarly articles within the last 10 years that use correlation analysis. One of the articles must use correlation to imply causation correctly and one article should not have justification to imply cause.

Summarize both articles in at least 500 words.

Explain why cause was appropriate in one article and not in the other.

What would be needed for the second article to justify a statement of cause?

Article Summaries

The study by Liu et al., 2023 titled ‘Correlation Analysis of Health Factors of Elderly People in Traditional Miao Dwellings in Western Hunan’ shows that correlation implies causation. One of the most common problems worldwide is the aging population. China is faced with the same problem with a total population of 1411.78 Million People and 260 million of the entire population is aged 60 years and above. This accounts for 18.70% and 190 million people are aged 65 years and above. As the population is advancing in age, it is experiencing difficulties to live in a suitable living environment characterized by urban-rural disparity. More specifically, in the rural areas, the population of the elderly aged 60 and 65 years and above accounts for 23.81% and 17.72% respectively. On the other hand, for the population aged 60 and 65 years and above, 7.99% and 6.61% respectively live in urban areas. The issue of aging is more serious in Xiangxi Tujia and Miao Autonomous Prefecture of Hunan Province in China and the health and living environment issue arouses more interest and attention in research. Physical health problems such as strokes and respiratory infections and mental health issues such as anxiety and depression arise from the elderly’s living environment. Since the COVID-19 pandemic, human health has become a major global concern more specifically the elderly’s health status.

In this research study, the study took place in Chinese Traditional Villages in Miao Autonomous Prefecture and Xiangxi Tujia. The Miao traditional villages have maintained their traditional culture and architecture was the major focus of the study. The villages selected were Liangdeng Village in Phoenix County, Laojiazhai Village, Zhaogang Village, and Zhushan Village. The sample of the study was selected by first observing the living habits and daily production conducted by native residents. Data from the sample was collected through interviews and questionnaires on health status and other related factors. The total number of respondents selected for the study was 97 elderly people aged 60 years and above. SPSS software was used to correlate the factors of the elderly and health status in the traditional Miao dwellings in Western Hunan.

Correlation analysis was conducted by converting the score values as the total indicators of the health assessment and those of the dependent variables. The independent values assessed are the annual per capita income, number of permanent household members, level of education, gender, and age, and the 23 assessment factors grouped into the four independent variables of natural environment, infrastructure, traditional dwellings, and daily behavior. Correlation analysis of the elderly's health status was conducted using SPSS. Pearson's correlation values of between -1 and +1 and values greater than 0 exhibit positive correlation and the opposite of that exhibited negative correlation.

In the research study, the correlation was enhanced by increasing the absolute value where * shows a correlation at 0.05 level and ** represents a correlation at 0.01. The results of SPSS exhibited both positive and negative correlations. The Pearson correlation values of the number of permanent members of the household, daily behavior, traditional living, and infrastructure were 0.250, 0.213, 0.209, and 0.228 showed a positive correlation at 0.01 level. For the age, the Pearson value to age was -0.847 and this showed a negative correlation at 0.01 level and for the income, the Pearson value was 0.278 and this exhibited a positive correlation of 0.01.

The second article by Wilscher et al., 2021 titled ‘Genetic correlation and causal relationships between cardio-metabolic traits and lung function impairment’ is a perfect example that correlation does not imply causation. Cardio-metabolic traits and obesity are public health problems in most world parts. It is predicted that, by 2025, 21% of women and 18% of men will be obese. Obesity is attributed to non-communicable diseases such as various types of cancers, cardiovascular diseases, and Type of Diabetes. Over time, there has been a growing pool of research on the relationship between obesity and chronic lung disease and lung function. The research by Wilscher et al., 2021 assesses the relationship between the cardio-metabolic traits of coronary artery disease (CAD), systolic blood pressure (SBP), diastolic blood pressure (DBP), pulse pressure (PP), triglycerides (TG), total cholesterol (TC), low-density lipoprotein (LDL), high-density lipoprotein (HDL), C-reactive protein (CRP), type 2 diabetes (T2D) and body mass index (BMI) and three lung functions such as ratio of forced vital capacity (FVC), first second forced expiratory volume and FEV1/FVC.

Research methodology involves performing analyses of cross-trait linkage disequilibrium score regression (LDSC), cardio-metabolic health to lung function association tests, and Mendelian Randomisation (MR) to assess the causal effect of disease on lung function and cardio-metabolic traits. The results indicated a negative genetic correlation between the functions of the lungs and cardiometabolic diseases and traits. Mendelian Randomisation analysis (MR) established that there is a relationship between FEV1/FVC, FVC, and Type 2 Diabetes. The instruments for measuring BMI were associated with C-reactive protein instruments and lung function traits. Overall, the genetic associations exhibited a causal effect of cardio-metabolic traits of the lung function. The lung function analysis revealed a causal effect of FVC/ FEV1 on blood pressure.

Explanation of Cause

Causation in correlation indicates that the occurrence of one event is due to the occurrence of the other. This is referred to as cause and effect. Statistical correlation can be established by the use of the Correlation Coefficient that is represented by the symbol r. The numerical value showing the correlation coefficient ranges from -1 to +1 and this indicates the direction and strength of the relationship. If the correlation coefficient indicates negative values of below 0, a negative relationship between variables is indicated. This means that variables move in opposite directions such as when one of the variables increases, the other decreases, and vice versa. On the other hand, the correlation coefficient with a positive value above 0 shows a positive relationship between the variables. This means that, as one variable decreases the other also decreases and if one variable increases, the other also increases. A coefficient relationship of 0 indicates no relationship between the variables when one of the variables in research decreases where others decrease or increase. In the first research study, both positive and negative correlations have been established as well as a cause-and-effect relationship between the health status of the elderly and infrastructure, dwellings, daily acts, income, and family.

In the second study, causation has not been established. Of worth noting, most times, in statistics, a positive correlation is usually mistaken for causation. A negative correlation has been identified in the second study and there is no causation between the variables in the research. The causation in the study was not established as the study was not controlled by splitting the population or sample in two where both groups can be compared after receiving different types of treatment is assessed. The first research study revealed the daily behaviors and living environment of the elderly and the psychological and physical needs established. Further study should be taken on sustainable development models for the elderly's health in Chinese villages.

Statement of Cause

For the second article to justify the statement of cause experimental research is necessary. Experimental research plays a crucial role in testing the causal hypotheses by allowing the three criteria for time order, association, and nonspuriousness. There are three features of true experiments and they are; two comparison groups (control and experimental), the difference in the independent variables before the change in the dependent variables to establish the time order, and the random assignment of the comparison groups. In true experimental design, the relationship between dependent and independent variables as two or more groups differs due to their value (Rogers & Revesz, 2019). One group in an experimental study should receive treatment which is manipulation of the independent variables and this is referred to as an experimental group. In the experiment, the other group does not receive any manipulation and this is referred to as a control group. In the case where a true experiment cannot be conducted due to high cost and time consumed, a quasi-experiment can be conducted. In quasi-experimental design, the comparison group that should be compared to the treatment group is predetermined. The respondents are not randomly assigned to the control and experimental group. There are three types of quasi-experimental design that should be conducted and they are; before-and-after designs, nonequivalent control group designs, and ex post facto control group. Nonequivalent control group designs have designated comparison and experiment groups before the treatment is assigned and not created in the random assignment (Rogers & Revesz, 2019). The before-and-after designs have post post-test and pretest and no comparison group. This means that the respondents of the study are earlier exposed to the treatment just like in the control group. On the other hand, the ex-post factor control group designs utilize nonrandomized control groups.

References

Liu, Z., Li, Z., Zhang, F., Yang, G., & Xie, L. (2023). Correlation Analysis of Health Factors of Elderly People in Traditional Miao Dwellings in Western Hunan. Buildings, 13(6), 1459. https://www.mdpi.com/2075-5309/13/6/1459

Rogers, J., & Revesz, A. (2019). Experimental and quasi-experimental designs. In The Routledge handbook of research methods in applied linguistics (pp. 133-143). Routledge. https://www.taylorfrancis.com/chapters/edit/10.4324/9780367824471-12/experimental-quasi-experimental-designs-john-rogers-andrea-r%C3%A9v%C3%A9sz

Wielscher, M., Amaral, A. F., van der Plaat, D., Wain, L. V., Sebert, S., Mosen-Ansorena, D., ... & Jarvelin, M. R. (2021). Genetic correlation and causal relationships between cardio-metabolic traits and lung function impairment. Genome Medicine, 13(1), 104. https://genomemedicine.biomedcentral.com/articles/10.1186/s13073-021-00914-x