IN Statistical Reasoning in Psychology: Assumptions and Limitations of Correlation Coefficients: Explain three assumptions for the evaluation of a correlation coefficient. Explain three limitations for evaluation a correlation coefficient. EXAMPLES OF CM (classmates) RESPONSES are: CM1 answers: There are three assumptions for the evaluation of a correlation coefficient, which are Homoscedasticity, Linearity and Normality. When the variance of data points are dispersed along the regressive line is equal we term that as the homoscedasticity assumption. While when describing the pattern of data using a straight line is linearity and when we assume the data points are normally distributed, is known as the assumption of normality. The limitations of correlation coefficient can be used to interpret a correlation coefficient that do not show cause, where the outliers change the direction and the strength of correlation. Also the generalization of direction demonstrating the strength of correlation outside the range of data measurement. These assumptions therefore are based on numbers between -1 and +1 calculated to represent the liner dependence of two variables or sets of data. CM2 answers: Three assumptions for interpreting a significant correlation coefficient are: 1.) homoscedasticity: is the assumption that the variance of data points dispersed along the regression line is equal. 2.) linearity: the assumption that the best way to describe the pattern of data is using a straight line. 3.) normality: the assumption that data points are normally distributed. Three additional considerations, or limitations, for interpreting a correlation coefficient are that (1) correlations do not demonstrate cause, (2) outliers can change the direction and the strength of a correlation, and (3) never generalize the direction and the strength of a correlation beyond the range of data measured (restriction of range). CM3 answers: The correlation between coefficients are subject to whether or not they have no outliers, the variance needs to be equal and the variables must be normally distributed. likewise the limitations of a correlation is that correlation does not automatically relate a causation, it does no indicate the direction of interaction and it might even be showing a correlation measurement from another third variable rather than the two variables on the graph