# Blog SOCY601- Quantitative Data Analysis

A causal relation between two variables exists if the occurrence of the first variable (Independent variable) causes the other (Dependent variable). The first variable is the cause and the second is the effect. A correlation between two variables does not necessarily indicate causality. In causality, one variable is partly responsible for causing another variable. On the contrary, if there is a causal relationship between two variables, they must be correlated. In case of some variability of one variable can be accounted by the other variable, two variables are associated.

Source: The Wall Street Journal

Causal relationship occurs when change in one variable can be attributed to the change in another variable. The primary step of establishing the causality is demonstrate the association between two variables in a bivariate analysis. Also, time order of occurrence of independent variable and dependent variable can help in proving the causality. Logically, causality can be proved if the independent variable occurs before dependent variable in time order as cause comes before the effect. In order to imply that a causal relationship exists both independent and dependent variables must be tested by controlling extraneous variables. Inclusion of the extraneous variables can change the interpretation of the importance of the independent variable. The best way to establish causality is through controlled experiment with subjects assigned to control group and experimental group. In the experiments, all the variables should be constant except the one, which is being tested to determine the effect of that specific variable. Then, the outcome of these two groups can be compared to establish causality between two variables.

Often, the cause and effect relationship in the variables is not obvious. That’s where intervening variables can be useful to explain the relationship between two variables. Extraneous variables can influence the relationship between two variables and can also influence the outcome of an experiment. Extraneous variables can threaten the internal validity of the outcomes causing spuriousness. To establish a non-spurious relationship between variables in is important to rule out any extraneous variables. By designing efficient study, taking an experimental approach whenever possible, carefully collecting the data and using statistical control, non-spurious relationship can be proved.

For the peer-reviewed article analysis, I found an article titled “Need for Growth, Achievement, Power and Affiliation: Determinants of Psychological Empowerment” by Sumi Jha. This article was published in the Global Business Review Journal in 2010. The objective of this article is to study the influence of motivational needs on psychological empowerment. Author Jha explains that independent variable in this study is motivational needs (need for growth, achievement, power and affiliation) and dependent variable is psychological empowerment in terms of intrinsic task motivation (meaning, competence, self-determination and impact) (Jha, 2010). Author also states that statistical techniques like correlation, multiple regression and canonical correlation are used to measure the significance and strength of relationship between identified independent and dependent variables (Jha, 2010). The textbook chapter states that the multiple regression analysis can be used to test how multiple independent variables are related to dependent variables. Author has done a great job of explaining how multiple independent variables such as – need for growth, achievement, power and affiliation are related to dependent variable – psychological empowerment. Author has not introduced any control variables in this study. Also, variables have not been identified as extraneous or intervening. In my opinion, extraneous variables such as educational attainment and team leadership can be introduced to strengthen the results. Author does not make any claims regarding causal relationship between the variables. Significant positive relationship was found between independent variables need for growth, need for power and need for achievement with dependent variable psychological empowerment (Intrinsic task motivation). Results also state that there was no significant relationship found between need for affiliation and psychological empowerment (Intrinsic task motivation).

Author of this article does not mention any limitations of the study. I would suggest that one limitation would be generalizability of the study. This study involves 319 participants from few five star restaurants in the Mumbai area limiting its generalizability. The study design involves closed ended questions, which can result in limited information from respondents. Author strengthened the study by analyzing various independent variables with dependent variable. Also author states by using canonical correlation, it provided  “a sound basis for assessing the overall and interactive impact of variables on each other designated as first and second set or dependent and independent set of variables”.

References,

Jha, S. (2010). Need for Growth, Achievement, Power and Affiliation: Determinants of Psychological Empowerment. Global Business Review, 11(3), 379-393. Global Business Review-2010-Jha-379-93

Schutt, R. K. (2015). Investigating the social world: The process and practice of research. Thousand Oaks, CA: Sage Publications.

## One thought on “Blog SOCY601- Quantitative Data Analysis”

1. Julie Honnold

Great cartoon! Also, thanks for pointing out the strength of experimental design in examining causal assertions. One minor thing that I’d like to mention is that you say that a carefully conducted experiment can “prove” non-spurious (hence, causal) relationship. Technically, examining statistical significance involves testing the hypothesis of no relationship (null hypothesis), right? (I’m sure you know that!) The analysis will tell the researcher whether or not the null can be rejected at a certain level of confidence, say 95%. That’s a good reason we don’t say that our favorite hypothesis – the opposite of the null – is proven. Rather, it has escaped disproof at a certain confidence level. Typically, you’ll see some language like “results provide evidence to support our hypothesis,” but not to “prove” it. Also, remember the confidence level; there’s a 5% chance that our rejection of the null is incorrect.