deletions | additions       diff --git a/Further Bias Corrections (Not implemented).tex b/Further Bias Corrections (Not implemented).tex  deleted file mode 100644  index c0732db..0000000  --- a/Further Bias Corrections (Not implemented).tex  +++ /dev/null 
...
\subsection{2.2 Further Bias Corrections (Were not implemented)}   The main bias correction mechanism that we employed was described in section 2.1.  The multiplication of the transition matrix with the 2013 results vector can be viewed as giving different respondents different weights. If Ri respondents replied that they voted for party i in 2013, then in the normalized transition matrix, each such respondent has weight 1/Ri. When we right-multiply the matrix with the actual 2013 results vector A, each respondent ends up with the weight Ai/Ri. With this weighting scheme, the total weight of respondents that claim to have voted for party i in 2013 is the same as the actual number of voters for i in 2013. Our sample now 'agrees' with the actual 2013 election results.  To recap, we inserted a weighting scheme that \textbf{controlled for} the publicly known 2013 election results.  In very much the same way, we could have controlled for additional known-in-advance distributions to correct for suspected biases. This would require us to ask the respondents some control questions (age, sex, ethnicity), to identify discrepancies between our sample and the real world distribution (publicly available from CBS), and to produce a weighting scheme that would bring our sample into agreement with the real world distributions.  Eventually, the only controls implemented were the ones mentioned in section 2.1.  diff --git a/layout.md b/layout.md  index ffbec42..06b4c1b 100644  --- a/layout.md  +++ b/layout.md 
...
Introduction.md  Methods.md  Seats distribution forecasting.md  Further Bias Corrections (Not implemented).tex  Results.md  Discussion1.md  Discussion.md