|
Size: 716
Comment:
|
← Revision 14 as of 2019-11-19 11:55:48 ⇥
Size: 1490
Comment:
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 3: | Line 3: |
| Suppose we have two variables, X and Y we wish to correlate. To obtain the 95% Confidence Interval for their correlation firstly standardise the two variables by subtracting their mean and dividing by their standard deviation. This can also be done using CROSSTABS in the EXPLORE menu. | Weaver and Koopman (2014) use SPSS macros available from [[https://sites.google.com/a/lakeheadu.ca/bweaver/Home/statistics/spss/my-spss-page/rhoci | here]] to obtain confidence intervals for each element of a matrix of Pearson correlations showing associations between variables. |
| Line 5: | Line 5: |
| Perform a linear regression using one of the standardised variables as the predictor of the other after selecting 95% confidence intervals for the regression estimate in the window obtained by clicking on the statistics button in the regression window. | This method is also suggested by Howell (2002) and has advantages over using 95% confidence intervals based upon regression coefficients of standardised variables in giving asymmetric intervals which are contained within the range [-1,1] and is, therefore, to be preferred. |
| Line 7: | Line 7: |
| There is also an[http://faculty.vassar.edu/lowry/rho.html: on-line calculator.] |
A 95% confidence interval using the backtransformed Fisher transformation for a ''single'' Pearson correlation may also be computed using this [[http://vassarstats.net/rho.html|on-line calculator.]] __References__ Bonett DG (2019) Point-biserial correlation: Interval estimation,hypothesis testing, meta-analysis, and sample size determination. To appear ''British Journal of Mathematical and Statistical Psychology''. Confidence intervals for different types of bi-serial correlations used as alternative two group effect sizes to Cohen's d. Howell DC (2002) Statistical Methods for Psychologists. Fifth Edition. Wadsworth:Pacific Grove, CA. Weaver B and Koopman R (2014) An SPSS Macro to Compute Confidence Intervals for Pearson’s Correlation. To appear in ''The Quantitative Methods for Psychology'' in Spring 2014. |
How do I obtain 95% Confidence Intervals for a (Pearson) correlation in SPSS?
Weaver and Koopman (2014) use SPSS macros available from here to obtain confidence intervals for each element of a matrix of Pearson correlations showing associations between variables.
This method is also suggested by Howell (2002) and has advantages over using 95% confidence intervals based upon regression coefficients of standardised variables in giving asymmetric intervals which are contained within the range [-1,1] and is, therefore, to be preferred.
A 95% confidence interval using the backtransformed Fisher transformation for a single Pearson correlation may also be computed using this on-line calculator.
References
Bonett DG (2019) Point-biserial correlation: Interval estimation,hypothesis testing, meta-analysis, and sample size determination. To appear British Journal of Mathematical and Statistical Psychology. Confidence intervals for different types of bi-serial correlations used as alternative two group effect sizes to Cohen's d.
Howell DC (2002) Statistical Methods for Psychologists. Fifth Edition. Wadsworth:Pacific Grove, CA.
Weaver B and Koopman R (2014) An SPSS Macro to Compute Confidence Intervals for Pearson’s Correlation. To appear in The Quantitative Methods for Psychology in Spring 2014.