Why national polls do not straight translate the number of seats to the Sejm. Why can you have a percent of support in a nationwide survey and get less tickets?

1. Elections are not ‘national’ only district

Mandates are allocated to individual constituenciesnot on the basis of 1 national score.

I mean:

• Each territory has a circumstantial number of mandates (e.g. 10, 12, 20),

• You vote for the list of candidates in this district,

• The national score is only the sum of the results from these districts — but this not linear dependence.

2. Regional diversity support

The parties have different levels of support in different parts of the country.

Example:

• The A organization may have 40% in the country, but for example 60% in the west and 20% in the east.

• Party B can have even 40% anywhere.

As a result:

• Party A “spent” votes where the majority already has a majority (because more than 50% in the territory does not give additional mandates),

• Or he could lose tickets where he doesn't have adequate votes.

Therefore geographical distribution of support is crucial.

3. Method for converting votes into mandates (D’Hondta)

In Poland it is utilized D’Hondt methodwhich:

• It favors larger parties,

• Operates separately in each district,

• It depends on the number of mandates in a given territory (the more mandates, the more proportional the result).

Therefore, in order to calculate a real consequence in tickets from a national survey, it is essential to:

1. Estimate the score in each territory (based on erstwhile elections, demographic data, etc.),

2. Use the D’Hondt method in each of them,

3. Summing up tickets from all the districts.

That's what it is. algorithm for conversion of the poll to mandates.

4. Example

Let's say:

• Party A: 40% national

• Party B: 35%

• Party C: 10%

• Party D: 8%

• The rest: 7%

The national thing is, A has a clear lead.

But after being counted as districts:

• In districts where A has 60-70%, its “surplus” votes do not give additional mandates,

• In districts where he is 25-30%, he may not get any mandate,

• As a result, organization B, despite a smaller overall score, can score more mandatesif he has more equal support.

5. Why It Needs algorithm or simulation model

Bo:

• Each territory has a different size and past of voting,

• The D’Hondt method is discreet (it is not easy to convert percentages into mandates by model),

• Data are needed on the geographical distribution of support that national surveys do not themselves show.

Therefore, it is utilized conversion modelswhich simulate:

"How would votes spread in individual districts in the current poll, and what mandates result?"