Preamble

This month, we return to the theme of the spatial distribution of engagement with the K-Wave. We have already looked at K-Wave fans in Europe in the 2010s and early 2020s using survey data, and this time we turn our attention to the USA in the year 2020 using digital behavioural data. More specifically, we will be considering whether the presence of proportionately larger Korean-identifying populations is related to spatial patterns of engagement with the K-Wave, rather than trying to characterise patterns of engagement with the K-Wave over time as we did before. To this end, for each US state we will look at the proportion of the population that identifies as Korean and online engagement with the K-Wave, at least, as it is measured in terms of relative volumes of search traffic related to timely K-Wave concepts.

Context

To understand what we might get out of doing this, a few bits of context will be helpful. Before getting into the nitty-gritty of data and analysis, though, we’ll discuss why the US is a good place upon which to focus for this, and why 2020 is a good year for it.

The US is home to the second largest population of Korean people outside of the Korean Peninsula. In case you were wondering, the largest is in China, with an especially large concentration of people in the Yanbian Korean Autonomous Prefecture, which borders North Korea. The contemporary Korean diaspora in the US is conventionally said to stretch back to the early twentieth century and the arrival of a few thousand people to work on the plantations in Hawai’i. Since then, the population has grown to in excess of two million people, with large communities found in Los Angeles as well as in and around New York City and its neighbouring state of New Jersey. Outside of these historic centres, there are Korean-identifying people living in every state. The common, albeit increasingly fragmented media ecosystem of the US, along with the variation in proportion of Korean-identifying population in each state makes it an ideal case for exploring any possible relationship between the presence of Korean-identifying people and the spatial distribution of online engagement with the K-Wave.

Having maximally dependable figures for the total and Korean-identifying population of each state was part of the motivation for focusing on 2020 as that was a year in which US conducted a decennial census. On top of that, 2020 was quite the year for the K-Wave in the US. In terms of K-Pop, BLACKPINK released The Album, and BTS were prolific. The boyband released the album Map of the Soul: 7 in February of that year followed by Be in November along with the Japanese-language release Map of the Soul: 7 - The Journey in July. Notably, all of these enjoyed the commercial success to reach the heights of the US Billboard 200 chart with Map of the Soul: 7 and Be topping the chart while The Album and Map of the Soul: 7 - The Journey peaked at numbers 2 and 14, respectively. On-screen, Parasite won the Academy Award for Best Picture, and there were a couple of high profile streaming successes in the form of Crash Landing on You (albeit released at the very end of 2019) and Itaewon Class. We make the assumption that the contemporaneous use of search terms relating to these developments in the K-Wave would reflect public engagement with the K-Wave at the time to at least some extent.

It’s worth being explicit that by investigating this we are in no way making the claim that the K-Wave is simply a reflection of engagement with Korean popular culture on the the part of Korean people living outside of South Korea. We recognise the potential for complete disjuncture between the various ‘scapes’ of global transcultural flows. People, products, language, ideas, and many other things have been able to metaphorically move independently from one another for quite some time now. For example, we do not suppose that Korean media necessarily ‘travels with’ Korean people (i.e., with regard to the ethnoscape), Korean capital (i.e., with regard to the financescape), or the key concepts of Korean culture (i.e., with regard to the ideoscape). We emphasise, though, that what we recognise is the potential, rather than the necessity, of complete disjuncture. That is, the complete and total separation between the different scapes of global transcultural flow is possible, but not automatically the actual state of affairs in the world as a result of that possibility. Such disjuncture is not a given. Thus, our investigation here is carried out to determine whether there are any signs of a relationship between the presence of proportionately larger Korean-identifying populations and local engagement with the K-Wave, and if so to propose plausible mechanisms for this.

Data

Data on the total and Korean-identifying populations of each state was sourced from the US Census Bureau. Search traffic data was obtained from Google Trends. As Google is the most widely used search engine in the US by an enormous margin, we consider this data adequately representative of the internet search behaviour of the populations of each state. We nevertheless acknowledge that this does not capture the entirety of search traffic in the US.

From the raw population data, the proportion of Korean-identifying residents of each state could be calculated. This is visualised in the map below.

The search traffic data consists of scaled scores for each US State between 0 and 100 for the year 2020. These scores are calculated as explained on this website, reproduced here:

Google Trends normalizes search data to make comparisons between terms easier. Search results are normalized to the time and location of a query by the following process:

  • Each data point is divided by the total searches of the geography and time range it represents to compare relative popularity. Otherwise, places with the most search volume would always be ranked highest.

  • The resulting numbers are then scaled on a range of 0 to 100 based on a topic’s proportion to all searches on all topics.

  • Different regions that show the same search interest for a term don’t always have the same total search volumes.

The meaning of spatial and temporal variation in interest scores is also explained elsewhere, higher interest scores are found in places or times:

…where your term has a higher probability of being searched.

Given all this, we will not be able to conclude definitively that people in states with higher interest scores carry out a greater number of searches for K-Wave terms than those in states with lower interest scores. Rather, higher scores may be taken as indicators of proportionately greater interest in the K-Wave in a general sense.

Moving on, the following K-Wave terms were selected for inclusion in the data set on the grounds of the then-contemporary developments in the K-Wave described in the preceding section:

The interest scores for each search term are visualised by state as below.

From the above visualisations we can see that not only is the Korean-Identifying population spread far from evenly over the US, engagement with the K-Wave as it is measured in terms of search traffic varies enormously between states. Furthermore, different search terms garnered different relative levels of interest between states.

Analysis

As stated at the outset, we are not just trying to establish spatial patterns in engagement with the K-Wave, but to determine whether there is a relationship between the presence of a Korean-Identifying population and engagement with the K-Wave. Taking all of these keywords together, we can get a mean index score for interest in the K-Wave in each state and determine whether there is a positive correlation between that and the proportion of each state’s Korean-identifying population.

For determining the presence and strength of a correlation between a larger proportion of Korean-identifying residents and a higher interest score, we selected the Kendall Rank Correlation measure. This choice was made as our data violates key assumptions of the Pearson Correlation test, specifically our data are not normally distributed and there is an outlier present. That is, while the mean index scores for most states are in the twenties or thirties, that of Hawai’i is in the nineties. Given the presence of this outlier Kendall Rank Correlation is preferred to the also non-parametric Spearman Rank Correlation. Based upon the results of this test, then, we can confirm that states ranked more highly in terms of their proportion of Korean-identifying residents were likely to rank more highly in terms of their mean interest scores (τ = .703, p < .001). In other words, larger populations of Korean-identifying people appear to correlate with greater online engagement with the K-Wave.

As noted above, we do not uncritically accept the idea that this finding simply reflects engagement with the K-Wave on the part of Korean people. This leaves us with the question, then, of precisely how the presence of a proportionately larger Korean population leads to greater engagement with the K-Wave over the entire population. A close examination of each state would no doubt reveal specific, concrete material or cultural factors which could lead to greater engagement with particular pieces of media or the K-Wave in general there. An example of a concrete factor that may have a role in reducing engagement with the K-Wave is the relative scarcity of access to wireline broadband and internet infrastructure in some areas of the US. Conversely, a possible cultural factor influencing amount or intensity of engagement with the K-Wave by state could be the differential valorisation of local and traditional values over cosmopolitan consumption.

A more general mechanism explaining this specific correlation, though, could be found in the the mere-exposure effect (for the origins of the idea in a paper from the 1960s see here, for the Wikipedia overview see here). In brief, the theory behind this psychological phenomenon suggests that ‘mere’ (that is, not positive, complex, intense, etc.) repeated exposure of an individual to a given stimulus enhances one’s attitude towards that stimulus. It has, significantly, been put forward as a mechanism in reducing racial prejudice, albeit under experimental conditions (click through here for the paper). The theory tentatively proposed here is that, as it is highly likely that the non-Korean populations of US states with larger proportionate populations of Korean-identifying people have a higher degree of ‘exposure’ to Korean-identifying people and culture, the positive attitudes this ‘mere exposure’ engenders may be a factor that contributes to greater engagement with the K-Wave throughout the population of the state.

Conclusion

Correlation is by no means definitive proof of causation. This is amply demonstrated by the internet-famous spurious correlation generator. Nevertheless, it is suggestive. While we only ran one correlation test, one of the variables was derived from the mean of a number of observations. This adds weight to the argument that the presence of a proportionately larger Korean-identifying population is a causal factor in engagement with the K-Wave on the state level, especially if one adopts a stance which recognises the regularity approach to establishing causation. What’s more, the mere-exposure effect provides a plausible mechanism by which non-Korean identifying people who more frequently encounter Korean people or culture could come to be more positively disposed towards, and consequently interested in, various media, products, or personalities related to the K-Wave that those who do not.

In short, it is worth entertaining, or at least looking further into, the claim that the presence of larger proportionate populations of Korean-identifying people is a factor which leads to greater engagement with the K-Wave at the local scale.

Acknowledgement
This work was supported by the Core University Program for Korean Studies of the Ministry of Education of the Republic of Korea and Korean Studies Promotion Service at the Academy of Korean Studies (AKS-2021-OLU-2250004)