16.14
Microsoft Excel: corrélation de Pearson
Microsoft Excel est un outil puissant pour l'analyse statistique, notamment pour calculer le coefficient de corrélation de Pearson, qui mesure la forc…
La corrélation est une relation entre deux variables. Considérons les données hypothétiques suivantes organisées dans Microsoft Excel.
Pour effectuer une corrélation univariée, la variable X doit être disposée dans une colonne et la variable Y dans une colonne à sa droite.
Pour tracer ces deux variables, sélectionnez d’abord les deux colonnes dans l’onglet Insertion, recherchez des graphiques et sélectionnez un nuage de points.
D’autres éléments tels que les lignes de tendance peuvent également être ajoutés aux conceptions de graphiques.
En supposant que les données proviennent de populations normalement distribuées, le coefficient de corrélation de Pearson — r — est calculé à l'aide de la fonction CORREL sélectionnée pour les données.
Une autre fonction, PEARSON, renvoie également la même valeur.
Ces fonctions mesurent l’intensité et la direction de la relation linéaire entre deux variables. Dans ce cas, il est de 0,985, ce qui suggère une corrélation forte et positive.
La fonction RSQ renvoie la valeur au carré de r. Cette valeur est le coefficient de détermination qui mesure la proportion de la variance dans la variable dépendante qui est prévisible à partir de la variable indépendante.
View the full transcript and gain access to JoVE Core videos
Q1: What does Pearson's correlation coefficient measure?
Pearson's correlation coefficient, denoted as r, measures the strength and direction of the linear relationship between two continuous variables. The coefficient ranges from -1 to 1, where values close to 1 indicate strong positive correlation, values near -1 indicate strong negative correlation, and values around 0 suggest no linear relationship between the variables.
Q2: How do you calculate Pearson's correlation in Microsoft Excel?
In Microsoft Excel, use the CORREL function to calculate Pearson's correlation coefficient. Enter the formula =CORREL(array1, array2), where array1 and array2 represent your two data columns. For example, =CORREL(A1:A10, B1:B10) calculates correlation between variables in cells A1:A10 and B1:B10. The PEARSON function returns the same result.
Q3: How should data be organized in Excel before calculating correlation?
To perform correlation analysis in Excel, arrange the X variable in one column and the Y variable in an adjacent column to its right. Select both columns together, then use the CORREL function or create a scatter plot through the Insert tab to visualize the relationship between your two variables.
Q4: What does the coefficient of determination (R²) tell you?
The coefficient of determination, calculated using the RSQ function in Excel, represents the squared value of the correlation coefficient r. This value measures the proportion of variance in the dependent variable that is predictable from the independent variable, indicating how well the linear model explains the data variation.
Q5: Can you visualize correlation relationships in Excel?
Yes, you can create a scatter plot in Excel to visualize the correlation between two variables. Select both data columns, go to the Insert tab, choose Charts, and select a scatter plot. You can enhance the visualization by adding trendlines and displaying the equation, which helps you observe how closely data points fit a straight line.
Q6: Why is correlation different from causation?
Correlation measures the strength of a linear relationship between two variables but does not imply that one variable causes changes in the other. Two variables can have high correlation without a causal relationship. Additionally, Pearson's correlation captures only linear relationships and may miss complex, non-linear associations between variables.
Q7: What advanced correlation analysis options does Excel provide?
Excel offers the Data Analysis Toolpak add-on, which provides correlation matrices for analyzing relationships across multiple variables simultaneously. This tool makes it easier to compare correlations across datasets and conduct more robust statistical analysis beyond simple pairwise correlation calculations between two variables.