29/2016
2016-07-29
Methodology Note about Random Samples
At CBOS we reach our conclusions about the population through the use of statistical methods, which allow us to evaluate in a representative manner such things as support for political parties. Making use of statistical methods means that we have to use random samples. Thanks to these we can both reach conclusions about the population and determine the scale of statistical error in our evaluations. For more about our methodology (in Polish only) please see the CBOS website, in the BADANIA tab.
Results of the July survey of political party preferences* with additional information about statistical error.
In July support for Law and Justice is estimated to be 39.3%, while the statistical error for the sample is +/-3.9% (with a confidence level of 95%). This means that it is 95% certain that the interval 35.5% to 43.2% includes the actual value of support for Law and Justice in the population. Accuracy of the estimated support for the other parties is greater. Thus, support for Civic Platform is estimated at 15.0% with +/-2.8% accuracy, for .Modern 13.9% with +/-2.7% accuracy and for Kukiz’15 7.9% with +/-2.2% accuracy.
*See the ‘Political party preferences in July’ CBOS report (by K. Pankowski). The ‘Current Events and Problems’ (314) survey was conducted using computer assisted (CAPI) face-to-face interviews on a random representative sample of 983 adult residents of Poland.