The adult Taking Part survey is a household survey of those aged 16+ in England and collects data on cultural engagement, including public libraries. As part of a MSc in Social Research and Evaluation I undertook a study using logistic regression to analyse the Taking Part data from 2017/18 and 2018/19 to understand:
- Who in England is more likely to use public libraries’ digital collections (e-books, e-audio, e-magazines, and online reference resources) in terms of socio-demographic characteristics?
- How, if at all, is this different from in-person public library users?
Logistic regression is a statistical analysis method used to estimate the probability of ‘an event’ occurring. It is a suitable method to use when the dependent variable has two values (e.g. library user or non-library user) and independent variables that can be either continuous or categorical (e.g. socio-demographic characteristics).
This blog summarises the main findings of the study.
The study finds that the following are strong predictors for using public libraries’ digital collections:
- Visiting a library during childhood.
- High education level.
- Young age (16-29).
Females, those in the higher managerial/professional socio-economic classification (SEC) and those without children in their household are also more likely to use them. Those in the intermediate SEC/ethnically minoritised group and aged 60+ (particularly females) are less likely to use them.
How is this different from in-person public library users? Visiting a library during childhood and a high education level are also predictors for visiting a library, but become stronger for using digital collections. In contrast to digital use, those aged 60+ are, and those with children in their household (particularly females) are more likely to visit libraries. Also, unlike digital use, SEC is not a predictor for visiting libraries.
The study demonstrates that in terms of SEC, education level, age, and ethnicity that those using digital collections are not as representative as those who visit in person. Therefore, library services need to think carefully about developing their digital services to reach a broader range of people.
The study also demonstrates the importance of visiting a library during childhood on adult library use. If more research were carried out to understand this in greater depth, it could help inform the development of library services. Is the best way to develop future adult audiences to invest more in children’s library services or target families?!
Finally, this study shows the importance of going beyond descriptive statistics and using inferential statistical analyses to understand library use. Taking a multiplicative approach can provide more insight, depth and, most importantly, highlight inequalities in provision. For example, most previous studies have found that ethnically minoritised individuals are more likely to visit libraries. The additive model in this study confirmed this, but the multiplicative model highlighted an inequality for older ethnically minoritised individuals who are less likely to use libraries. If the library sector wants to understand who uses libraries, it needs to look more closely at interactions. In turn, the findings from this research can then be used to inform the development of services, attract new users and reduce inequalities.
One of the limitations of the study is the data predates the Covid-19 pandemic and is already out of date. The most recent Taking Part datasets had not been released at the time the research was undertaken. However, it would be interesting to use the same methodology on the datasets for 2020-2021 and 2021-2022. We know from other studies that, during the Covid-19 pandemic, there is some evidence that digital services did not reach everyone despite their use increasing.
The full version of the dissertation can be downloaded here.