Comparison of methods to account for autocorrelation in correlation analyses of fish data

Abstract

Autocorrelation in fish recruitment and environmental data can complicate statistical inference in correlation analyses. To address this problem, researchers often either adjust hypothesis testing procedures (e.g., adjust degrees of freedom) to account for autocorrelation or remove the autocorrelation using prewhitening or first-differencing before analysis. However, the effectiveness of methods that adjust hypothesis testing procedures has not yet been fully explored quantitatively. We therefore compared several adjustment methods via Monte Carlo simulation and found that a modified version of these methods kept Type I error rates near . In contrast, methods that remove autocorrelation control Type I error rates well but may in some circumstances increase Type II error rates (probability of failing to detect some environmental effect) and hence reduce statistical power, in comparison with adjusting the test procedure. Specifically, our Monte Carlo simulations show that prewhitening and especially first-differencing decrease power in the common situations where low-frequency (slowly changing) processes are important sources of covariation in fish recruitment or in environmental variables. Conversely, removing autocorrelation can increase power when low-frequency processes account for only some of the covariation. We therefore recommend that researchers carefully consider the importance of different time scales of variability when analyzing autocorrelated data.

Mechanisms for concurrent low-latitude circulation anomalies responsible for persistent extreme precipitation in the Yangtze River Valley

“…Both the effective number of degrees of freedom and the significance threshold for the correlation coefficient between filtered series, to be shown in Sect. 5, are evaluated following the method introduced by Pyper and Peterman (1998).…”

Synchronous marine pelagic regime shifts in the Northern Hemisphere

“…Relationships between the first three BioPCs and some key large-scale hydro-climatic indices, NHT anomalies, the AO index, the AMO index and the PDO index (figures 4 and 5; electronic supplementary material, table C), were investigated by means of a linear correlation analysis. The correlation probabilities were corrected to account for temporal autocorrelation by adjusting the degrees of freedom [72]. However, as this technique can be overly conservative, we also examined the uncorrected probabilities.…”

Dynamics of brain activity underlying working memory for music in a naturalistic condition

“…These were mostly weak (r range=0.09-0.21), but their significance had to be corrected due to serial correlation, as the data points are not independent from each other. In order to estimate the significance of the r coefficients, the effective degrees of freedom were calculated between each AC and the WM regressor following Pyper and Peterman's (1998) However we needed to test also whether a linear combination of the AC, rather than individual components, could account for our WM variable. Thus we performed multiple regression whereby the WM variable was regressed against the whole set of AC, solving the general linear model…”

“…To correct the Fisher z-scores for significance, we first estimated the effective DoF by following a non-parametric procedure described in Pyper and Peterman (1998) based on a Monte-Carlo simulation. The effective degrees of freedom are given by Eq.…”