For my final blog post, I decided to take a trial run at online data collection and analysis. My broader research interests are related to the flow of Tajik migrant labor to Russia and its resilience to issues such as economic hardship, bureaucratic hurdles, and xenophobia.
As discussed in my last post, data from the World Bank evidently show that during past times of economic hardship in Russia, Tajik remittances drastically fell. Since the early 2000s, there have been two major economic disasters which have impacted Tajik remittances: The Great Recession from 2007-2009, and Russia’s oil industry crash beginning in 2014. Yet, these past two scenarios were purely economic, thus it proves difficult to compare them with the world’s current crisis, Covid-19, a pandemic which has had terrible public health, social, and economic impacts.
Overall, Covid-19 has sickened and killed millions of people across the globe. The pandemic has been a public health and economic catastrophe, and socioeconomic problems which may have existed prior to the pandemic have merely been exacerbated. Due to Tajikistan’s economic interdependence on Russia, Tajik migrants have faced a severe amount of hardship this year. Tajikistan is merely one of many countries in the world, but it serves as an interesting case when analyzing the impact of catastrophes on the lives of labor migrants, particularly in the case of Covid-19. Yet, because the Covid-19 pandemic is ongoing, it proves difficult to study and fully understand the scope of a crisis which is evolving. As a result, I have decided to take to the web to try and better understand the lives of Tajik labor migrants and how the current Covid-19 crisis may or may not be affecting their livelihoods.
Before getting into the meat of my research question, attempted methodology, data collection, and analysis, I’d like to begin with an image of a Facebook post.
This is a post on the Facebook group, “Таджики в Москве”, “Tajiks in Moscow.” As you can see, not much information is given about the person who posted, and I’ve purposefully omitted their name and photograph from the image. However, based on my own observations, it is clear that the person who posted is female and is likely past the age of thirty. This person took to the web to seek advice. It appears that she is currently in Moscow and is seeking advice for how to return home to Tajikistan. Her request led to 77 comments and 1 share. Within the comments section, there is all sorts of advice about where to browse for tickets and what to expect given the current atmosphere of travel to and from Russia.
This is just one post within a large Facebook group, but it highlights how Tajik migrants may be using online networks to seek out and share information. This one post also demonstrates how fruitful yet ambiguous gathering data from online spaces can be. While this post, and thousands like it, are merely sitting on an online group page, totally open and accessible, it is impossible to know much about the people posting and interacting, but I will discuss these limitations and give more examples of other interactions later in my blog post.
Here is a link to the Facebook group: https://www.facebook.com/groups/2830409070518217
And here is an image of the Facebook group’s banner and title.
Research Question and Methodology
As mentioned earlier, my overall research question is: How has the Covid-19 pandemic impacted Tajik migrant labor to Russia? As it is rather impossible to fully answer this question by using data acquired from a social network, my research question for the purpose of this blog post is: How have Tajik migrants utilized the online network “Таджики в Москве”?
Moreover, for my broader thesis project, I originally planned to pull four separate months of data from this online Facebook group. The months I had planned to collect data for were: September 2019, December 2019, March 2020, and June 2020. When I actually began collecting data from this Facebook group, I quickly discovered that the group was created in July 2019. Thus, in December 2019 the group was still quite new and there is no way to know how many members the group had at that time. This is a definite limitation as it makes longitudinally comparing and contrasting group activity much more complex. I merely have access to the current number of group members which, on today’s date (December 14, 2020) is at 1,992 individuals. Moreover, I’ll add that this group page is public and visible to those who search for it, but only members may create posts.
I have found that it is difficult and very tedious to access past data on this Facebook group page. Facebook unfortunately does not have a mechanism to enable me to only look at a certain time frame within a group page. Rather, I merely need to scroll, and scroll, and scroll some more in order to find what I am looking for. Also, I am not entirely sure whether data collected from this group page will actually be of use to my thesis project. Therefore, I decided to make data collection as simple as possible for the purpose of this blog post. I began with what was the newest, November 2020, and I pulled every post from that time frame. I then decided that I could use November 2020 as a trial in order to see:
1. How I would analyze data
2. How I would develop a coding mechanism
3. How I would separate the group page usage of admins versus group members
In this blog post, I will not be discussing the third task on the list but I am hoping to look further into the concept of group admins versus regular group members in the coming weeks.
Data Collection
Because data collected from this group page can range from a funny video of a goat damaging a car (posted on November 16th by a group admin) to a desperate family member posting about a missing brother who was last seen in Moscow, it would be difficult to analyze and compare the actual contents of each post. I decided instead that I could analyze the perceived intended purpose of each post and create qualitative codes. This would allow me to analyze how this Facebook group and network is functioning within a given month.
After collecting every post from the month of November – in both Russian (or Tajiki) and English – as well as the name of the person who posted, I designated the perceived purpose of each post. Through this process, I took more of an inductive approach, allowing the data to shape the codes I created.
Please feel free to check out a draft of my data here:
https://docs.google.com/spreadsheets/d/1Us3fG6hC8UAWLDI-1c7e0_vPE0H8Vtp6u5E88Q1liio/edit?usp=sharing
I left out member names from the spreadsheet shared on this blog. Also, the English translations I pulled were generated by Facebook’s translation mechanism, thus are far from perfect. There are a handful of posts in Tajiki. I can read and understand Tajiki well enough to code for intended purposes, but I have not yet taken the time to perfect translations, either in Tajiki or in Russian. However, please feel free to scroll through my draft of November data.
Coding Mechanism
After brainstorming the intended purpose of each post, I took my first shot at developing a qualitative code book. When developing a code book, it is necessary to have a code name, a code definition, inclusion criteria, exclusion criteria, and examples of “close but no’s”. As an undergraduate at Arizona State University, I worked as a research assistant in an anthropology lab and was trained in qualitative text analysis. With each code book, the goal was to create sets of definitions that were so clear and concise that any two researchers could read a code book and analyze a set of data in almost the same exact way. In that sense, code books enabled us to take qualitative data and make them quantifiable. After my first attempt at creating a code book based on data from the month of November, I realize that my definitions are incomplete and need to be more thoroughly defined. I also recognize that some codes may split apart, some may converge, and there may be some additional codes added.
Analysis
In total, there were 97 posts for the month of November 2020. Only 6 of those posts were directly related to sharing information about Covid-19. However, there may have been some posts that were indirectly related to impacts of the Covid-19 pandemic.
The most common type of posts in November were related to jobs. Thus, perhaps the greatest purpose of this online network is for job sharing. There were only a couple of posts from people who were in search of work. Instead, there were many job advertisements. There were several repeat posts from a woman who was looking for a cleaner, thus I may need to decide what to do with repeat posts later. Moreover, there were several posts from individuals who were looking for entire teams for construction or installation workers. By far, the most common jobs that were advertised were construction jobs, cleaning jobs, store jobs, and courier jobs. There was also an interesting link to an article about clinics in search of migrant labor which was shared. Based on this data, it would be interesting to compare the month of November to earlier months of the pandemic to see whether there are more job postings. Perhaps the economy is beginning to rebuild and industries are in need of labor.
‘Cultural connectedness’ was another common code. There was a very fine line between the ‘cultural connection’ code and the ‘light-heartedness’ code, and I will need to sharpen my definitions. Many of these posts were made by one of the group admins, but there were other people who posted music videos, jokes, historical videos about Tajikistan, and there was even one video about a Plov restaurant.
Examples of Various Posts
Next I’d like to go through a handful of data which were collected for the month of November. Data will appear as images of original posts. Again, I’ve left out personal information such as names and profile images.
Here is an example of one of the posts I easily coded as “Job Posting”. The Facebook translation of the Russian is a bit strange, but overall, it is easy to distinguish this post as a job advertisement.
The next post is an advertisement for an apartment for rent. Interestingly, the advertisement includes a comment that those of Central Asian nationality are welcome as tenants.
Next is another job posting. The reason why I found this post interesting is due to its mention of a required patent for Uzbek and Tajik citizens. We have been reading about work permits (or patents) throughout the semester, so it has been interesting to see patents being mentioned.
Finally I’ve included a post that I coded as a “Request for Help”. There were quite a few posts in November similar to this. The Facebook group served as a place to share information about missing persons. Therefore, the group may be used as a connection between Tajiks in Tajikistan and Tajiks working abroad, and I’m really hoping that this man is safe and eventually found.
Final Discussion
There is, however, one major disappointment I have from gathering all of the November posts from this particular Facebook group: there was not very much conversation centered on Covid-19. In this sense, I learned a lot more about how this particular group functions online than I did about how the Covid-19 pandemic has impacted Tajik migrants. Although, this can perhaps be perceived as an additional conclusion. Maybe the month of November marks a period of less chaos and movement between Tajikistan and Russia. Perhaps those in Moscow are trying to work hard, as usual, while those who left Russia as a result of the pandemic are still waiting until it is possible to return. I have no way of currently knowing whether that theory is accurate, but as I collect more data from this online group I may be able to notice more monthly trends since the start of the pandemic.
Moreover, one of the broader purposes of collecting data from November was to get my data collection system worked out. Now that I feel like this is doable, I will definitely go back to posts in March and April 2020 and see how they differ from last month’s activity.
Another important aspect to this data collection is that this Facebook group of 2,000 people is not representative of the Tajik labor migrant population as a whole. Instead, it serves as a mere snapshot of the individuals who chose to use this particular site (Facebook) and this particular group page. I am honestly not sure if this is enough information for me to work into a full project. I will need to talk everything over with my thesis advisor and wait to hear comments from Dr. Kamp and whoever else happens to be reading my blog post. Therefore, please give your honest and critical feedback of my research trial run! I know my writing style for this blog post was not the most formal, and I did not include any theoretic sources or additional information. Instead, this post reflects my thoughts and methodology as it is still developing.
Going forward, I will need to do more research on the science of social networks. There has been much published on the methodology of pulling online data from blogs and Twitter. For example, pulling data from blogs often utilizes search engines such as Google and DuckDuckGo as well as key search themes. Often attributed to data collection from Twitter is the analysis of specific Hashtag usages. I am sure that there is also much that has been done on analyzing group interaction within an online social network, but that is precisely what I need to look into.
Additionally, reflecting on this whole process, I think that learning how to systematically gather data from an online source is valuable for a number of reasons. First, when gathering data from open-source sites and virtual communities, researchers are able to capture people’s thoughts, words, and interactions in an unfiltered way. Anyone who has collected surveys or interviews know that it can be difficult to get subjects to open up about certain topics. It can be difficult to build rapport and sometimes no matter how hard a researcher tries, subjects may be uninspired to speak at length about a certain topic. Collecting data online offers a very different scope and dynamic. Also, I personally think that, although there is so much “squishiness” behind the methodology of online data collection, this avenue of research can be especially useful for students. Gathering data from online sources is free, and it can be a very fruitful tool during times such as global pandemics when travelling to conduct research is not feasible.
As mentioned earlier in my blog post, I received a background in anthropology as a research assistant at ASU. I first began dabbling with the idea of collecting data from online communities when I was working for a researcher named Dr. Sarah Trainer who studies fat stigma and cultural differences between expectations of the “ideal” body type. For a whole year, I collected hundreds and hundreds of narratives from online weight-loss blogs. In order to collect the narratives, my boss and I came up with a system of key words and phrases which we plugged into search engines. We then had a set number of ‘hits’ we would sift through each time we searched a key phrase through three different search engines. This project made me stretch intellectually, and it was so challenging, but by the end, I learned much about the methodology behind gathering data from online blogs. I also read many weight loss, fitness, body acceptance, and “fatosphere” narratives. Some of these blogs got particularly creative. There was a site titled, “Runs for Cookies,” and also a site titled, “Bear-iatric.” “Bear-iatric” was centered around a man who was a member of the LGBTQ+ community who identified as a gay “bear”, and also received bariatric surgery.
If you are at all interested, here is our abstract for “The Fat Self in Virtual Communities.”
And with that, I hope you have found this post interesting! Please be sure to comment with any thoughts, questions, or criticisms.