Annotated Tweets from German politicians.

A dataset containing 1,000 hand-coded Tweets, drawn from [tw_data].

Usage

tw_annot

Format

A data frame with 20838 rows and 8 variables:

status_id

Tweet ID

ane

Binary hand-coding of anti-elitism

ppc

Binary hand-coding of people-centrism

pop

Binary hand-coding of populism. 1 if either ane or ppc is 1

coder

Coder identifier. A or B. AB if Tweet was parallel code. In this case, the coding is 1 if at least one coder decided to code 1.

user_id

Twitter user ID

twitter_handle

Twitter handle

party

Political party

followers_count

The number of followers in Aug 2022

created_at

Date and time when the Tweet was created

full_text

Uncleaned Tweet

rel_test

Indicates if Tweet was parallel-coded for reliability test

ane_A

Binary hand-coding of anti-elitism from coder A

ane_B

Binary hand-coding of anti-elitism from coder B

ppc_A

Binary hand-coding of people-centrism from coder A

ppc_B

Binary hand-coding of people-centrism from coder B

Details

The dataset contains 1,000 Tweets which were drawn as stratified random sample from the population data [tw_data]. Each political party is represented with (at least) 120 Tweets. The populist parties AfD (250 Tweets) and Die Linke (150 Tweets) were oversampled, as we anticipated more populist content from these parties.

TWo expert coders, one the author of this package, hand-coded the Tweets along two binary categories for populist communication: Anti-elitism and people-centrism.

The coding followed the instructions documented in the online supplementary files of Thiele (2022).

90 Tweets were parallel-coded for reliability testing, resulting in Krippendorff's Alphas of .86 for anti-elitism, and .71 for people-centrism, as documented by the variables ane_A, ane_B, ppc_A, and ppc_B.

References

Thiele, D. (2022). Pandemic Populism? How Covid-19 Triggered Populist Facebook User Comments in Germany and Austria. Politics and Governance, 10(1), 185–196. https://doi.org/10.17645/pag.v10i1.4712