| name | Explanation | classification | Conditions of Use |
| Product moment correlation | Linear relationship between the two variables reaction | Pearson | (1) two variables are continuous data obtained by the measurement; (2) two variables were tested generally normal, or near normal, at least symmetric distribution is unimodal; (3) must be paired data, i.e. two variables to be measured on the same sample from an overall or; (4) a linear relationship between two variables; (5) large sample, n≥30. |
| Rank correlation | Non-linear relationship between two variables reaction | Spearman | 1) the correlation between the variable order. Values of the order of gradation value into a variable; 2) applies to the original data to sequential data variable; 3) equidistant raw data or geometric variable, but the overall non-normal distribution, should not use product-moment correlation analysis of data. 4) the sample size required is not necessarily greater than 30. |
| Kendall | Statistics there are n objects, each object has attributes of two coefficients. All statistical properties of the objects by the value 1 are arranged, without loss of generality, the value at this time was 2 are arranged attribute is scrambled. Let P be the magnitude relation between two attribute values are arranged same on the number of statistical target. | ||
| Quality-related | A variable mass, the amount of another variable | Point two related (real dichotomous) | Continuous variables (or equally variable or geometric variables) subject to normal or near normal. Dichotomous variable data regardless of whether the distribution is normal. Between variables is linear . Dichotomous variable is the real binary, such as gender, coin of positive and negative, right or wrong topic. Which demarcation point as close as possible in value. The sample size is greater than 80. SPSS is a special case pearson correlation coefficient. |
| Two related (human dichotomous) | Continuous variables follow a normal distribution, or near normal. Is a linear relationship between the variables. Dichotomous variable is artificial division, according to test scores as divided into qualified and unqualified. Which demarcation point as close as possible in value. The sample size is greater than 80. In SPSS, also pearson correlation coefficient corresponding to the calculated two continuous variables. | ||
| Multi-family related | Both variables are continuous variables, wherein a variable according to certain criteria to be artificially divided into a plurality of categories. Two related multi-family related exceptions. | ||
| Quality-related | Values for these variables are classified into several categories according to the nature | Column with relevant | When two variables or a variable two variables are divided into three or more categories, used to indicate the correlation between two variables. |
| φ related | When two variables are dichotomous variables, whether real or man-half-half, can be used. | ||
| The relevant quarter | Normal two variables are continuous variables, and exhibits a linear relationship, only two variables are artificially into dichotomous variable represents the correlation between the two variables column. |
A small chestnut:
