The experience of computerized content analysis of articles of the journal "Bulletin of Finance, Industry and Trade" for 1917

Voronkova Dar'ya Sergeevna PhD Candidate, Section of Historical Informatics, History Department, Lomonosov Moscow State University 119192, Russia, g. Moscow, ul. Lomonosovskii Prosp., 27 korp.4, kab. G-432 dasevo@yandex.ru Other publications by this author

The article is devoted to the results of a computerized content analysis of the articles of the journal "Bulletin of Finance, Industry and Trade" for 1917. The journal is in many ways an indispensable source of economic information, including in the field of state financial policy, international trade, taxation, etc.

In the revolutionary year 1917, the pages of the magazine reflected the economic processes taking place in Russia against the background of the ongoing World War.

Content analysis is a well-established and at the same time promising method of historical research (including periodical printing has been studied with its help for a long time and fruitfully) ^[1]. However, the named journal has not previously become the object of content analysis. At the same time, he attracted the attention of historians, and there are new works in which he is a source base ^{[2; 3; 4; 5; 6; 7]}.

"Herald..." as a source

One of the defining source characteristics of the mentioned journal is its official status as a printing body of the Ministry of Finance. By 1917, it had been published for 32 years (since 1885, and before that there was a direct predecessor – the Index of Government Orders for the Ministry of Finance, published in 1865-1884). In some sources, the year of the beginning of the publication of the "Bulletin" is indicated in 1883. Perhaps the discrepancies are related to the fact that in 1885 the journal was reorganized. The editorial board received at least formal independence, having been allocated from the staff of the Ministry of Finance, and its budget. Such changes are so important that sometimes 1885 is taken as the year of the magazine's foundation. However, in the process of reorganization, even the editor–in-chief was not replaced - both in the "Index ..." and in the "Bulletin ..." this position was held by A. K. Veselovsky. Let us stop at 1884 as a compromise date for the beginning of the publication of the "Bulletin ...". There are a number of justifications in favor of this version, including articles in the Brockhaus and Efron Encyclopedic Dictionary (https://www.booksite.ru/fulltext/1/001/007/020/20375 .htm; https://www.booksite.ru/fulltext/1/001/007/104/104187.htm ; https://www.booksite.ru/fulltext/1/001/007/020/20375.htm ), as well as bibliographic descriptions of the journal in the RNB (https://www.booksite.ru/fulltext/1/001/007/020/20375.htm – It seems to us that here, with the exception of the first line, in which "Bulletin..." dates from 1865, the ratio of "Index..." and "Bulletin ..." is correctly described) and in the NEB (https://www.booksite.ru/fulltext/1/001/007/020/20375.htm ). It is noteworthy that in the "white" Omsk, a short-term attempt was made to "reanimate" the "Bulletin ...", publishing it once every two weeks. This, in particular, is evidenced by one of the surviving April issues of the "Bulletin ..." for 1919 with the indication: "The journal is published by the Ministry of Finance of the Russian Government." (https://www.booksite.ru/fulltext/1/001/007/020/20375.htm).

The initiator of the publication of the new magazine was initially the Minister of Finance N. H. Bunge, and in fact this event was sanctioned by Emperor Alexander III himself. Article 10 of the "Rules on the taxation of commercial and industrial enterprises with an additional fee (percentage and layout)", approved on January 15, 1885, regulates the procedure for publication in the "Bulletin ..." extracts from minutes of general meetings of joint-stock companies and partnerships, as well as balance sheets and annual reports of such associations ^[8]. The publication was mandatory, besides it was paid. The publication of the "Bulletin..." became a logical continuation of the tax reforms of the 1880s. Being the mouthpiece of government policy, Vestnik ... provided the audience with materials carefully selected by the editorial board (among the authors there are many, in particular, leading economists of their time).

On the website of the publishing house "Budget", quite a lot of attention is paid to the circumstances of the appearance of the "Bulletin ...": "... a new magazine was created, which became known as the "Bulletin of Finance, Industry and Trade". <...> The widespread use of telegraphic communication led in 1894 to the formation of a special telegraph department in the editorial office of the Bulletin. The quality of work, accuracy and timeliness of receipt of commercial and other telegrams caused an ever-growing demand for the services of the department from other periodicals and individuals. After all, although at that time there was already a "Russian Telegraph Agency", it was not engaged in commercial information. Under these conditions, the telegraph department, constantly developing, in 1902 was transformed into a separate institution – the "Trade and Telegraph Agency". The management and management of the agency continued to be carried out by the editorial team of Vestnik, and its editor was appointed director of this organization. The quality of the agency's work was noticed by the press. And soon many periodicals switched to him for information service. Foreign telegraph agencies concluded agreements with him on the exchange of telegrams. In 1904, it was decided to transform the "Trade and Telegraph Agency" into the "St. Petersburg Telegraph Agency".^[9]. Thus, the editorial board of the "Bulletin ..." based its work on the principles and technologies advanced for its time, contributed to the development of the telegraph network and gave rise to the publication of a number of print media, such as the "Trade and Industrial Newspaper" and "Russian Economic Review".

At the end of the quarter, a "Systematic list of articles" was published, but in 1917 only two were published – for the first and second quarters.

Research methodology

As noted above, content analysis is successfully used as a method of analyzing text sources and, in particular, the historical press.

The software that made it possible to carry out a computerized content analysis of the materials of an array of articles from the issues of the "Bulletin ..." of 1917 was the MAXQDA program ^[10]. The work uses the MAXQDA version – MAXQDA 2022.

Development of a system of categories and analysis of their occurrence frequencies

After repeatedly reading the journal articles, converting the source file into recognized electronic text and importing this text into the MAXQDA program, a frequency dictionary was created in the MAXDictio module. Based on the most common words, indicators were identified, which were consolidated into a system of categories. The text is then indexed using the auto-coding option with a dictionary.  Indexing in content analysis is the markup in the text of segments where indicators of the corresponding categories meet. Indexing was based on indicators for which only the root bases of words were set. In total, 14 categories were constructed (all of them, with the exception of the "countries" category, are single-level, and the latter includes opposing and neutral states, and the composition of the corresponding blocks was formed taking into account the realities of 1917, therefore, for example, Italy was on the side of the Entente). A number of the author's articles highlight the most important problems for the journal and analyze their topics ^{[11; 12; 13]}.

The results are presented in the following table (Table 1):

Table 1. Semantic categories of texts and indicator words

The unit of account is the offer, and the unit of analysis is the identified category. Tables of word frequencies have been compiled for the journal as a whole (words that do not carry a semantic load are excluded using a stop list - a list of service words that are not subject to analysis). 

In MAXQDA, it is possible to create a word cloud (visualization of a frequency dictionary), but a cloud of category codes will be more informative for us (Fig. 1).

Fig. 1. Code cloud

It can be seen that the most common category is "finance", followed by "trade", "industry", "war" and "political sphere". This is also confirmed in dynamics if we present the categories by frequency of occurrence for each quarter of 1917 and display the result in the form of a so-called heat map (with this graphical representation, the warmest shades correspond to the most common categories, and the lower the frequency, the colder the color; Fig. 2):

Fig. 2. Heat map of the frequency of occurrence of categories

The analysis of quarterly indicators is chosen for two reasons. Firstly, if you do not combine the 43 issues of the Bulletin ... into any groups, the final tables will be too cumbersome and difficult to understand. Secondly, if the numbers are combined monthly, there will be a question with a double number 19-20: it is dated May 14 (27) and May 21 (June 3), 1917, which means that, being a single one, it would formally have to fall into different months, and it is impossible and incorrect to separate its materials. Thus, the only option is to combine the numbers into groups of documents by quarter. The first quarter – numbers inclusive from 1 to 13-14 (double), the second – from 15 to 26, the third – from 27 to 38 and the fourth (incomplete) – numbers 39-43. We see that the "finance" category tops the list throughout the year. The "trade" category has been among the main ones for three quarters, and only in the fourth interest in it is noticeably falling, giving second place to the "war" category. For the first quarter, the three categories are closed by "industry", in the second – "trade", in the third – "political sphere" ("industry" at the same time comes in second place), and in the fourth, the last, in which the publication of the magazine ceased, the third place is shared with the political sphere by the category "war". Despite such a wide range of issues discussed on the pages of the magazine, it is possible to outline something in common about why they received such attention. In our opinion, the unifying factor is the search for additional sources of state income, ways to solve the problems of supplying the population, getting out of the gradually escalating transport and fuel crisis and analyzing foreign experience in these areas.

In order to more accurately identify the main categories and work with them in the future, we used a useful tool MAXQDA – "code coverage". This indicator for groups of documents (corresponding to the four quarters of 1917) was 87.6-91.7%. Such high values indicate the completeness of all stages of encoding the source text data. Let's look at the distribution of the frequency of occurrence of categories, provided that the values are rounded to tenths (the "percentage of total text" setting is selected, Fig. 3).

Fig. 3. The frequency of occurrence of categories is calculated using the "code coverage" function

So, we can make the following list of the most common categories for 1917 as a whole:

· "finance" (46.5%);

· "industry" (19.6%);

· "trade" (19.1%);

· "political sphere" (18.0%);

· "war" (17.0%);

· "socio-economic sphere" (14.6%);

· "taxes" (11.8%);

· "legal field" (11.2%).

It is for them that we will calculate and visualize the frequencies of joint occurrence below.

Analysis of the frequency of joint occurrence of categories

 The MAXMaps module, which allows you to build informative graphs, is well suited for visualizing the frequencies of joint occurrence of categories. We are interested in the proximity of the codes (the user himself can set the necessary distance for the text of the document – for example, the fixed codes can be located in neighboring or one or two paragraphs from each other) and the intersection of the codes in the encoded segments. In the constructed graph, the maximum distance between codes in one paragraph is set by the proximity of categories. The result is shown in Fig. 4. The fatness of the lines is an indicator of the high frequencies of the common occurrence of categories, the frequencies themselves are indicated on these lines. The core of the graph consists of those categories between which the strongest connections are.

Fig. 4. Graph of frequencies of joint occurrence of categories by their proximity

All the main categories shown in the graph are closest to the "finance" category. The paired values are as follows:

· "finance" - "political sphere": 6803;

· "finance" - "war": 6152;

· "finance" - trade": 5993;

· "finance" - "industry": 5938;

· "finance" - "socio-economic sphere": 4781;

· "finance" - "legal field": 4662;

· "finance" - "taxes": 3475.

In addition, the strong connection between the categories "trade" and "industry" is noteworthy (4136). Thus, financial issues clearly form the core of the entire content of the "Bulletin ...". They are most strongly associated with the political sphere, as well as the categories "war", "trade" and "industry", while the latter two are expected to be mutually close. MAXQDA helps to identify these hidden relationships. Let's focus on the structure of categories in more detail, analyzing each of the following graphs in Fig. 5-12 (these graphs are constructed based on the intersection of categories; in order not to overload the images, 10 categories intersecting with the central one are shown).

Fig. 5. Graph of co-occurrence frequencies for the category "finance"

Finance for the authors of the "Bulletin ..." – mainly issues of money and check circulation, the budget, the State Bank, the foreign exchange market, etc. This explains the prevalence of links with the categories "political sphere", "trade", "war", "industry". Here is what B. Remez writes about the reasons for the preparation of monetary reform: "During the war under way, as in the previous historical epochs of shocks to the monetary system, quite a lot of articles have already appeared on the issue of the upcoming monetary reform. The original feature of modern projects is that they are mostly aimed not at returning to metal circulation, but at streamlining the existing paper and money economy" ^[14]. Wartime also left its mark on the financial sphere. A. N. Guryev notes on this occasion: "Modern war, with all the complex complex of phenomena generated by it, has already been certified by the whole world as unprecedented, unheard of, without any precedent in history. In accordance with this attestation, all previous views based on historical precedents have been subjected to a decisive revision. Only in one area we still have a historical stencil. This area is the paper and money economy generated by the war" ^[15].

Fig. 6. Graph of co-occurrence frequencies for the category "trade"

Figure 6 shows that the strongest relationship between the category "trade" and the category "finance". There are also noticeable links with the categories "industry", "Entente and allies", "war" and "taxes". The latter two are also not strange in this context: we are talking about the fact that, on the one hand, military spending needs to be replenished with proceeds from trade, and on the other, the question "where to get additional funds?" the authors of the "Bulletin ..." are trying to solve "at the junction" of the trade and tax spheres.

Fig. 7. Graph of co-occurrence frequencies for the category "industry"

The category "industry" (Fig. 7) is closely related to the categories "finance", "trade", "war", somewhat weaker – with socio-economic and political spheres. In general, this is the whole complex of the most important issues that are most closely considered on the pages of the magazine.

Fig. 8. Graph of co-occurrence frequencies for the category "war"

In the case of the frequencies of the joint occurrence of the category "war" with other main categories (Fig. 8), an unobvious result can be noticed. The connection with the category "finance" is also the strongest here (war is always expensive!), followed by the categories "industry", "political sphere", "trade" by a large margin. But the links with the subcategories of countries (which, we recall, are divided into three blocks), although quite naturally bring the Entente and allies to the fore of them, are still not as strong as one might expect. This result reveals hidden information that is difficult to obtain by traditional methods.

It should be noted that, despite the above, the experience of the allies was covered in sufficient detail in the magazine. For example, in the 43rd issue of Vestnik ..., which became the last, there are also reflections on possible measures to rest