The BLDPI Index
BLDPI (Belarus LGBTQ+ Discourse Pressure Index)
This is a research tool developed by LGBTQ+ House to quantitatively measure the discursive pressure on the LGBTQ+ community within Belarus' public information space. Only sources with editorial offices located within Belarus are included in the calculation; sources whose editorial offices are located outside the country and operate ‘on Belarus’ are deliberately excluded to ensure the validity of the measurement of the domestic information environment specifically. Telegram is used as the monitoring platform: within the chosen monitoring framework and given the availability of public posts, it ensures the comparability of content extraction and sufficient comprehensiveness in representing source types for regular monitoring; The unit of observation is a public message from a specific source (including reposts), as we measure discursive pressure as the fact of dissemination and amplification of the signal by that source, rather than the original authorship.
The processing architecture is implemented as a cascading NLP pipeline — searching for and selecting relevant messages, normalisation, semantic classification and feature extraction — followed by mandatory manual analytical verification of every message included in the calculation (100% human-in-the-loop), quality audits and repeated checks of the stability of results on manual subsets; coding scales and categorisation rules are fixed and applied consistently.
A key stage in the process is assigning a pressure intensity rating (Dp) to each message based on the semantic encoding of its content and context: not only explicit wording is taken into account, but also implicit rhetorical signals (including sarcasm, innuendo, discrediting techniques and stigmatising vocabulary), as well as ways in which the topic is instrumentalised to create or intensify pressure. The final coding is validated by mandatory manual analytical verification of the messages included in the calculation, which reduces the risk of systematic errors in interpretation and classification.
Sources within the monitoring framework are classified according to fixed categories used in calculations and visualisations: pro-government civil society organisations and affiliated bodies, the judicial and legal sphere, regional media, national media, personal propaganda channels and influencers, security agencies, state-owned enterprises, executive authorities, as well as ‘tabloid’ online media outlets.
All changes to the reference framework are recorded and applied consistently; where there are significant changes, their impact on the comparability of the series is indicated.
The index comprises two complementary metrics:
PI (Monthly Pressure Index) — a comprehensive assessment of the level of pressure over the month: each message is assigned a pressure intensity score Dp, which is aggregated taking into account the institutional weight of the source category Wk (a formalised assessment of the source’s institutional capacity to exert pressure by virtue of its position and authority within the social/administrative structure) and the audience reach La, where reach is normalised logarithmically: La = log10(1 + subscribersa). For a message: PIp = Dp · Wk · La, for a month: PI = Σ PIp; the interpretation is straightforward: the higher the PI, the greater the discursive pressure.
PDI (Monthly Positive Deviation Index) — an indicator of positive deviations: PDIp = D+p · Wk · La, PDI = Σ PDIp; interpretation is straightforward: the higher the PDI, the greater the number of positive deviations — signals of support and pressure reduction in the media space.
Discursive pressure by source category (PI) and positive deviations (PDI)
The PI column shows the total pressure for the month, whilst the segments show the contribution of aggregated source categories. The PDI for the same period is shown alongside, reflecting the total positive deviations.
Daily trends: posts showing pressure and positive sentiment
The graph shows the number of posts per day exhibiting signs of discursive pressure and supportive signals, in order to identify dates of increased activity and campaign spikes in the information landscape. This is a frequency indicator; only the number of posts is taken into account. Assessment of pressure intensity, source category weighting and logarithmic normalisation of reach are not applied.
Results and summary reports
Results and summary reports
Each month, we publish a brief summary of PI/PDI trends and attach a PDF summary of the reports included in the calculation. The public version contains aggregated results; more detailed breakdowns and insights are available to partners via private access.
In February 2026, 194 negative posts were recorded, and the Discourse Pressure Index (PI) stood at 1102.0. Compared with the previous month, the pressure decreased: the Discourse Pressure Index (PI) decreased by 31.2%, and the number of negative posts by 39.8%. At the same time, the positive trend remained virtually imperceptible: only two positive posts were noted, and the positive deviation index (PDI) stood at 27.0.
The main structural change of the month relates to the ‘Media’ category: its share in the overall discourse pressure index increased by 11.7 percentage points compared to the previous month. This means that even against the backdrop of an overall decline in negative posts and aggregate pressure, the media segment’s contribution to shaping the negative agenda became noticeably greater. In other words, in February the pressure weakened in quantitative terms, but within the structure of the negative discourse, the role of the ‘Media’ segment strengthened.
In January 2026, public pressure intensified sharply: 322 negative posts were recorded, and the Discursive Pressure Index (PI) reached 1601.0. Compared to the previous month, the number of negative posts increased by 47.0%, and the Discourse Pressure Index (PI) by 39.7%. This is not merely an increase, but an absolute record for the entire observation period in terms of the number of negative posts.
There was virtually no positive sentiment during the reporting month: only one positive post was recorded. The most pronounced surge in pressure occurred on 19 January 2026, when 27 negative posts were published – this constituted an anomalous peak within the month. Thus, January is characterised by a maximum concentration of negative posts alongside an almost complete absence of positive ones.
In December 2025, 219 negative posts were recorded; the Discursive Pressure Index (PI) stood at 1146.0. Compared with the previous month, the number of negative posts decreased by 1.8%, yet the Discursive Pressure Index (PI) increased by 1.0% at the same time, indicating that the high level of negative sentiment persisted, with a slightly greater concentration of pressure in individual posts.
There were no positive posts during the reporting month. No pronounced structural shifts or anomalies were noted in December; therefore, the overall picture for the month can be characterised as consistently unfavourable: the negative narrative persisted, and no compensatory positive activity was observed.
In November 2025, 223 negative posts were recorded, and the Discursive Pressure Index (PI) stood at 1135.0. Compared with the previous month, the pressure decreased significantly: the Discourse Pressure Index (PI) decreased by 32.3%, and the number of negative posts by 23.6%. At the same time, the positive narrative remained minimal: only 3 positive posts were recorded, and the Positive Deviation Index (PDI) was 43.0.
Against the backdrop of an overall decline in pressure, the ‘Propaganda’ category gained significance in the structure of negative content: its share in the overall pressure index increased by 13.1 percentage points compared to the previous month. This indicates not a uniform weakening of pressure, but its redistribution in favour of a more pronounced propaganda component. 10 November 2025 stands out in particular, when an abnormal surge was recorded – 19 negative posts in a single day, which was the highest daily figure of the month.
In October 2025, 292 negative posts were recorded, with a Discourse Pressure Index (PI) of 1676.0. Compared with the previous month, the Discourse Pressure Index (PI) decreased by 11.9%, whilst the number of negative posts, conversely, increased by 0.7%. This makes October the month with the highest number of negative posts recorded since monitoring began: such a volume — 292 — had not been recorded previously.
The pressure structure has shifted significantly: the share of the ‘Media’ category in the overall pressure index decreased by 20.8 percentage points compared to the previous month, whilst the contribution of the ‘Propaganda’ category increased by 12.1 percentage points and that of the ‘State Apparatus’ category by 11.1 percentage points. Against this backdrop, positive posts remained few and far between: five such posts were recorded in October, and the Positive Deviation Index (PDI) stood at 72.0, indicating an extremely weak positive counterbalance within the overall body of publications.
In September 2025, 290 negative posts were recorded, and the Discursive Pressure Index (PI) stood at 1902.0. Against this backdrop, the positive narrative was minimal: only 4 positive posts were identified, with the Positive Deviation Index (PDI) at 71.0.
The main event of the month was an abnormal surge in pressure on 17 September 2025, when 55 negative posts were published – the highest daily figure for the reporting period. Against a backdrop of generally stable monthly trends, it was this peak that marked the most tense episode of September and confirmed the continuing imbalance between negative and positive posts.