Mapping the state of the art of creative cluster research: a bibliometric and thematic analysis

ABSTRACT

The notion of creative clusters has become the focus of a growing number of policy initiatives aimed at revitalizing economies by means of creativity. However, despite their prominence in policy discourse, creative clusters are still a ‘fuzzy’ concept, defined and treated differently in different strands of research. To address these disparities, this paper presents a systematic literature review of creative cluster research (CCR), with the aim of: (1) exploring the state of the art in the field, (2) pointing out some important limitations, and (3) outlining a future research agenda. A total of 355 articles published between 1986 and 2019 were analysed, drawing upon a combination of manual coding, bibliometric analysis, and text mining techniques. This multi-method approach allowed us to provide both a meta-analysis of CCR and an exploration of its thematic content. In so doing, our paper contributes to a comprehensive understanding of how creative clusters have been studied over time, both broadly and in relation to different creative sectors and geographical contexts. Moreover, through the identification of research gaps and boundaries of knowledge in the field, it points to key methodological and conceptual development issues to be addressed in future studies.

KEYWORDS:

• Creative clusters
• systematic literature review
• bibliometrics
• text mining

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 After review and consultation, the keyword term ‘creative class’, which has drawn significant attention amongst international scholars over the last two decades, was also added to the search term list.

2 Due to space constraints, most tables and figures were moved to Appendix A.

3 We decided to focus on abstracts as these concisely summarise the entire content of the paper, representing the best type of data for text mining analysis of scholarly work.

4 We used T-Lab Plus 2020, which is a content exploration and text mining package providing statistical tools for text analysis based on a lexicometric approach (Lancia 2020).

5 ‘LUs’ correspond to words, multi-words, and lemmas, ‘elementary contexts’ are defined as portions of text (in this study they refer to sentences ending with punctuation marks and length up to 1000 characters) and ‘documents’ refer to abstracts.

6 The TF-IDF is a measure proposed by Salton (1989) that allows evaluating the weight of a term within a document, according to the following formula:

$w_{i,j}={\displaystyle \frac{t{f}_{i,j}}{max{f}_{ij}}Xlog{\displaystyle \frac{N}{d{f}_i}}}$

Where:

$t{f}_{ij}$ =  Number of occurrences of $i$ (term) in $j$ (document)

$d{f}_i$ =  Number of documents containing $i$

$N$=  Total number of documents

$t{f}_{ij}$(Term frequency value) can be normalised as follows: $t{f}_{ij}=t{f}_{ij}/Max{f}_{ij}$

7 Lexical units with irrelevant content (i.e. stop words such as ‘the’, ‘of’, ‘these’, ‘between’ or other words not relevant to the analysis) were excluded and others were renamed or coded with the same LU according to a content and synonym’s analysis (e.g. firm*, company*, enterprise* were labelled under the word ‘firms’, factor*, component*, element* were labelled under the word ‘factors’).

8 We excluded articles published in 2019 as, at the time of the analysis, their number could still rise and was not representative of the entire research conducted in this year.

9 As is common in the bibliometric literature (e.g. Mingers and Leydesdorff 2015), we categorised as highly cited those articles that received more citations than the top 10% of the articles in the same WoS thematic area and year of publication.

10 We decided to perform thematic analysis based on documents (i.e. abstracts) rather than on elementary contexts (i.e. sentences) to avoid potential bias arising from abstracts including more sentences belonging to the same theme.

11 The chi-square test is applied to all the intersections of the contingency table lexical units x thematic clusters and the ci-square values show the significance of word occurrences within each cluster (see Appendix B for more details).

Additional information

Funding

This work was supported by Arts and Humanities Research Council (AHRC) Policy and Evidence Centre for the Creative Industries: [Grant Number AH/ S001298/1].