It is usually not enough to use only descriptive statistics to investigate the deep trends prevalent in a population. Inferential statistics, including regression and correlation analyses, are used for this purpose. This is especially important in those studies for which establishing causal relationships is essential. For example, the web article chosen discussed a scientific study looking for a link between electronic device use and mental health. Regression, including multiple regression, was used to identify the strength of the causal relationship between increased device use and psychological well-being disorders. The material showed that the use of regression models could be confusing and even distant if significant limitations were not taken into account. For example, “while most of these models showed some small positive correlation, one in particular — the one that accounted for the most factors — turned up a negative correlation between electronics usage and anxiety scores” (Spalding, 2021). Indeed, regression is a valuable statistical tool for determining the strength and nature of the relationship between variables. However, it is not enough to take multiple data sets, run regression analyses on them, and be confident that cause and effect have been found. On the contrary, using this tool means that more serious consideration of the data, including the source of the data collection and possible deviations from trends, is necessary.
While it was generally clear from the material that the use of regression does not provide universal answers for the relationship between devices and mental well-being, the overall narrative could have been improved by the use of additional tools. For example, any regression models could have been explained with appropriate graphs and equations, making it easier to understand. In addition, the piece reported the use of primary data but did not provide a descriptive summary of the nature of the sample. Too high an average age of the respondents, for example, could have been a reliable limitation for this work since the use of electronic devices obviously differs between generations. Thus, the use of a fuller range of statistical models would have made this paper more readable.
Reference
Spalding, K. (2021). Your computer is (probably) not making you depressed. IFLS.