Corrigendum to: Semantic Memory Impairment Across the Schizophrenia Continuum: A Meta-Analysis of Category Fluency Performance.

[This corrects the article DOI: 10.1093/schizbullopen/sgaa054.].

Corrigendum to “Semantic Memory Impairment Across the Schizophrenia Continuum: A Meta-Analysis of Category Fluency Performance” by Tan et al. Schizophrenia Bulletin Open, 2020. doi:10.1093/schizbullopen/sgaa054.

In the meta-analysis article by Tan et al.[1], unintentional manual entry errors in clustering variable definitions and participant groups occurred during the data screening stage and led to some inaccuracies in the reported clustering and switching meta-analyses for the chronic schizophrenia (CSZ) group.

The authors have rechecked all the data used in the manuscript and re-run the meta-analyses of clustering and switching performance within the CSZ group using the corrected data. The interpretation of the data does not change and remains correct. The incorrect and correct text and corrections to Tables 1 and 2 and Figure 4 are presented here:

Table 1: The clustering variables for three studies[2-4] should be described as ‘number of clustered words’. The clustering variable for one study[5] should be labelled as ‘words per cluster’.

Table 2: Changes made to values associated with ‘Number of clustered words’ and Switching’.

Figure 4: Changes made to Fig 4c and 4d.

Pg 4, Characteristics of included studies: ‘10 studies (8 CSZ, 2 ROP) reported 2 types of clustering measures: mean cluster size (number of items reported ÷ number of clusters) and number of clusters. Only mean cluster size had a sufficient number of included studies for meta-analysis’ should read ‘5 CSZ studies reported 3 types of clustering measures: number of clustered words, mean cluster size (number of clustered words ÷ number of clusters) and number of clusters. Only number of clustered words had a sufficient number of included studies for meta-analysis’.

Pg 4, Characteristics of included studies: ‘5 studies reported number of switches made (4 CSZ, 1 ROP)’ should read ‘5 CSZ studies reported number of switches made’.

Pg 4, Psychosis Spectrum Groups vs Healthy Controls: ‘Mean cluster sizes were smaller in the CSZ group compared to HCs (d = .65), and HCs also made more category switches compared to CSZ (d = .97)’ should read ‘The CSZ group had smaller numbers of clustered words compared to HCs (d = .84), and HCs also made more category switches compared to CSZ (d = .82)’.

Pg 9, The Breadth of Measures Within the Category Fluency Task: ‘The scope of the current study was limited as we were only able to analyze productivity as well as errors, mean cluster size, and switching performance in the CSZ group’ should read ‘The scope of the current study was limited as we were only able to analyze productivity as well as errors, number of clustered words, and switching performance in the CSZ group’.

The authors would like to thank Dr Petar Gabrić for bringing these issues to our attention and sincerely regret the errors.