Correlation, Regression, and T-Tests
Linear regression analyses are conducted after correlation has been established (Cohen, West & Aiken, 2014). Correlation shows how two variables are related, that is how one variable behaves when the other changes. Regression analyses somehow quantify this relationship -- we use it to predict the value of the dependent variable at different values of the independent variable. Since regression analyses yield numerical values, they can only be performed on quantitative, measurable values. RQ1 in the week one assignment -- how does engagement in social media relate to loneliness -- fits the qualification for simple regression analyses -- the independent variable (engagement in social media is measured in terms of the number of hours one spends on the internet) whereas the dependent variable is measured in terms of quantified categories ranging from 'very low' to 'very high', where quantification would involve assigning numbers to categories, for instance, 'very low' is assigned a low value of 1, 'low' is assigned a relatively higher value of 2, and 'very high' is assigned the highest value, say 5.
Regression analysis in this case will help the researcher predict i) how one's degree of loneliness would change with changes in the number of hours spent on social media; and ii) the relative strength of social media as a predictor of loneliness. For this research question, we could formulate the null and alternative...
A simple regression analysis will most often yield three different tables - the model summary, the ANOVA table, and the table of coefficients. The table of coefficients will often give the unstandardized regression coefficients, the standard error, the t-values, and the significance levels. The unstandardized coefficients are then used to formulae the regression equation. For this particular case, we could expect something like:
Loneliness = 23.971 -- 3.016 SME (where SME represents the number of hours spent engaging in social media).
From this hypothetical equation, we can predict the degree of loneliness at different levels of social media engagement, say 5 hours and 10 hours as shown below:
a) Loneliness = 23.971 -- 3.016(5)
=8.891
b) Loneliness = 23.971- 3.016(10)
= -6.189
Part Two: Research Article Critique
Makhbul and Hasun (2011) conducted an exploratory study to assess how such entrepreneurial factors as support from others, style of leadership and access to…
