StayinDE

Method and sources

How this quiz calculates your result

We coded where the seven parties stand on 29 statements, checking each position sentence by sentence against the party’s manifesto for the 2025 federal election. We compare your answers with each party to get your “match”, then turn the matches into a “spectrum” that adds up to 100%. Every step is explained below. Data as of 27 September 2026.

1. How we chose the statements

There are 29 statements, covering 11 topics. We chose them according to four principles:

Political science research has pointed out again and again that the choice of statements affects the matching result (Walgrave et al. 2009; Lefevere and Walgrave 2014). That is why the full result page shows every statement, every coded party position and its original text.

2. How we coded the party positions

The source is the manifesto each party published for the federal election on 23 February 2025. Each cell (one party on one statement) is coded on a five-point scale:

Where a manifesto says nothing on a statement, we first look for other official positions the party took in 2025: press releases from its parliamentary group, interviews with the lead candidates on the Bundestag’s youth website, and speeches in Bundestag debates. We use a source like this for 6 of the 203 cells, and the result page names it each time. Where we still find nothing, the cell is marked “not in the manifesto”. That applies to 14 of the 203 cells. The party with the most: BSW (5 cells).

Quote check: the result page quotes 187 sentences from the manifestos in the original German. A script compares each one with the text layer of the manifesto PDF, confirms that it matches word for word and records the PDF page number automatically.

Since the election: the CDU/CSU and the SPD have acted differently from their manifestos on some points. For example, in March 2025, before the new government was formed, the two parties amended the constitution together with the Greens to loosen the debt brake (Schuldenbremse); after taking office in May, they kept dual citizenship. Scoring still follows the manifestos; the result page marks changes since the election separately under each statement concerned.

We did not use Wahl-O-Mat data. The terms for the dataset from the Federal Agency for Civic Education (bpb) forbid using it to build similar matching tools, and we follow them. In the Wahl-O-Mat, the parties fill in their own answers; our coding comes from the original manifesto text. The two methods differ, and each has its own statements.

Links to the seven manifestos

3. How the match is calculated

This follows the scoring principles the bpb has published for the Wahl-O-Mat: 2 points for the same view, 1 point if one side is neutral, 0 points for opposite views; statements marked as important count double, and skipped ones are left out. The Wahl-O-Mat uses a three-point scale and we use a five-point one, so we extend the same principle to “a quarter off for each step apart”. Researchers call this calculation city-block distance, and it is also the method the Wahl-O-Mat uses (Gemenis 2024). Studies that compared several matching methods found that the statement-by-statement methods mostly give the same advice (Louwerse and Rosema 2014; Mendez 2012).

4. How the spectrum adds up to 100%

Spectrum share = eβ × match ÷ the same value summed over all seven parties, with β = 10.

This is a multinomial logit (softmax) transformation. It comes from the random utility choice model of the economist McFadden (1974). The spatial theory of voting in political science (Downs 1957; Enelow and Hinich 1984) uses it to describe how voters choose between several parties. Paul W. Thurner, a political scientist at LMU Munich, and his colleagues have used models of this kind to study past German federal elections (Thurner 2000; Mauerer, Thurner and Debus 2015; Mauerer, Pößnecker, Thurner and Tutz 2015).

β = 10 means that a party whose match with you is 10 percentage points higher gets about 2.7 times the share. We set this value ourselves; we did not estimate it from real voting data. So the spectrum shows “relative closeness” and cannot be read as a probability of voting for a party. The ranking of the parties depends only on the match; β only affects how far apart the shares spread.

5. The political map and a check against expert ratings

The horizontal axis (economy) uses 8 statements; the vertical axis (society, migration, climate) uses 15. Each statement is converted according to the direction “agree” stands for, then the results are averaged and scaled from 0 to 10. The 6 statements on foreign and defence policy, China policy, the EU and the climate dividend (Klimageld) are not on the map, but all of them count in the spectrum. The vertical axis is based on GAL–TAN, a concept widely used in political science (green, alternative, libertarian vs. traditional, authoritarian, nationalist; Hooghe, Marks and Wilson 2002).

Check: we compared the positions of the seven parties, calculated from their manifestos, with the Chapel Hill Expert Survey 2024 (CHES; Rovny et al. 2025), a survey of experts on party positions that is widely used in European political science. The Pearson correlation is 0.98 on the economic axis (lrecon) and 0.91 on the social axis (galtan); for the CDU/CSU we used the CDU’s values. The two datasets are independent of each other and agree on the broad direction. There are two main differences. Our positions sit closer to the ends of the axes than the expert ratings, because manifestos are usually worded more sharply than the experts’ overall ratings. And on the social axis, the order of the CDU/CSU and the BSW, and of the SPD and the FDP, is the other way round in CHES; in CHES these two pairs are only 0.4 to 0.5 points apart to begin with.

The map only shows the broad direction. Research has found that the “closest party” you get by placing people in a one- or two-dimensional space often differs from the result of a statement-by-statement comparison (Louwerse and Rosema 2014). So your spectrum is always calculated from the statement-by-statement match.

6. Limitations

7. References

  1. Downs, A. (1957). An Economic Theory of Democracy. New York: Harper. books.google.com/books/about/An_Economic_Theory_of_Democracy.html?id=kLEGAAAAMAAJ
  2. Enelow, J. M., & Hinich, M. J. (1984). The Spatial Theory of Voting: An Introduction. Cambridge University Press. books.google.com/books/about/The_Spatial_Theory_of_Voting.html?id=lXY6AAAAIAAJ
  3. McFadden, D. (1974). Conditional logit analysis of qualitative choice behavior. In P. Zarembka (Ed.), Frontiers in Econometrics (pp. 105–142). Academic Press. eml.berkeley.edu/reprints/mcfadden/zarembka.pdf
  4. Thurner, P. W. (2000). The empirical application of the spatial theory of voting in multiparty systems with random utility models. Electoral Studies, 19(4), 493–517. (the author was then at the University of Mannheim; professor at LMU since 2008) doi.org/10.1016/S0261-3794(99)00025-6
  5. Mauerer, I., Thurner, P. W., & Debus, M. (2015). Under which conditions do parties attract voters’ reactions to issues? Party-varying issue voting in German elections 1987–2009. West European Politics, 38(6), 1251–1273. doi.org/10.1080/01402382.2015.1026562
  6. Mauerer, I., Pößnecker, W., Thurner, P. W., & Tutz, G. (2015). Modeling electoral choices in multiparty systems with high-dimensional data: A regularized selection of parameters using the lasso approach. Journal of Choice Modelling, 16, 23–42. (all four authors at LMU’s departments of political science and statistics) doi.org/10.1016/j.jocm.2015.09.004
  7. Hooghe, L., Marks, G., & Wilson, C. J. (2002). Does left/right structure party positions on European integration? Comparative Political Studies, 35(8), 965–989. doi.org/10.1177/001041402236310
  8. Walgrave, S., Nuytemans, M., & Pepermans, K. (2009). Voting aid applications and the effect of statement selection. West European Politics, 32(6), 1161–1180. doi.org/10.1080/01402380903230637
  9. Garzia, D., & Marschall, S. (2012). Voting Advice Applications under review: The state of research. International Journal of Electronic Governance, 5(3/4), 203–222. doi.org/10.1504/IJEG.2012.051309
  10. Marschall, S., & Schultze, M. (2012). Voting Advice Applications and their effect on voter turnout: The case of the German Wahl-O-Mat. International Journal of Electronic Governance, 5(3/4), 349–366. doi.org/10.1504/IJEG.2012.051314
  11. Mendez, F. (2012). Matching voters with political parties and candidates: An empirical test of four algorithms. International Journal of Electronic Governance, 5(3/4), 264–278. doi.org/10.1504/IJEG.2012.051316
  12. Wagner, M., & Ruusuvirta, O. (2012). Matching voters to parties: Voting advice applications and models of party choice. Acta Politica, 47(4), 400–422. doi.org/10.1057/ap.2011.29
  13. Gemenis, K. (2013). Estimating parties’ policy positions through voting advice applications: Some methodological considerations. Acta Politica, 48(3), 268–295. doi.org/10.1057/ap.2012.36
  14. Lefevere, J., & Walgrave, S. (2014). A perfect match? The impact of statement selection on voting advice applications’ ability to match voters and parties. Electoral Studies, 36, 252–262. doi.org/10.1016/j.electstud.2014.04.002
  15. Louwerse, T., & Rosema, M. (2014). The design effects of voting advice applications: Comparing methods of calculating matches. Acta Politica, 49(3), 286–312. doi.org/10.1057/ap.2013.30
  16. Rosema, M., & Louwerse, T. (2016). Response scales in voting advice applications: Do different designs produce different outcomes? Policy & Internet, 8(4), 431–456. doi.org/10.1002/poi3.139
  17. Garzia, D., & Marschall, S. (2019). Voting Advice Applications. Oxford Research Encyclopedia of Politics. Oxford University Press. doi.org/10.1093/acrefore/9780190228637.013.620
  18. Gemenis, K. (2024). Artificial intelligence and voting advice applications. Frontiers in Political Science, 6, 1286893. doi.org/10.3389/fpos.2024.1286893
  19. Rovny, J., Polk, J., Bakker, R., Hooghe, L., Jolly, S., Marks, G., Steenbergen, M., & Vachudova, M. A. (2025). The 2024 Chapel Hill Expert Survey on political party positioning in Europe: Twenty-five years of party positional data. Electoral Studies, 97, 102981. doi.org/10.1016/j.electstud.2025.102981

Wahl-O-Mat scoring principles: Federal Agency for Civic Education (bpb), “Häufig gestellte Fragen zum Wahl-O-Mat” (6 February 2025) bpb.de. None of the authors or institutions above took part in, reviewed or endorsed this quiz.

Sources for the statement backgrounds

Update log

See the full update log

Start the quiz (free)