Aperçu des semaines

  • Quantitative methods setup and materials

  • 1) Organization, types of arguments, concepts vs. theories

  • Concepts and Measures

    • You will work with 14 Czech administrative units - "kraje".

      It is often argued that socio-economic frustration leads to vote for far-left or far-right parties. Lets asume that economic frustration (at a regional level) has three components - Unemployment, Average Salary, Low Education (education lower than high school "Matura").

      Provide at least three alternative operationalizations agregating the above mentioned components. Explain briefly the logic behind their construction and try to ilustrate the diference in output classification (measurement) produced by alternative operationalizations. Try to briefly defend which operationalization is the best.

      In the next step try to conceptualize and operationalize "far-(right/lef)" political party. In this case the eventual classification must be binary (0; 1).

      Then, take the data from 2021 and measure "kraje" according to the economic frustration, and measure/classify political parties participating in the elections (with more than 1,5 %).

      Plot the putative cause (econ. frustration) against the putative effect (share of far-left/right parties).

      Provide a brief evaluation of the putative relationship.

      Hand in as a PDF (title page and the text with inserted tables adn charts) plus Excel (tables with calculations and charts + you may want to use the fisrt list as a codebook).

  • Building a database

    • Database: COVID impact

      Your task is to build a dataset containing (i) GDP growth predicitions of the European Commision (EC) for the EU countries (including the UK) for 2020 and 2021 as issued just before the COVID (Feb ´20), (ii) current estimates of the actual GDP growth in 2020 and predictions for 2021 (EC - Feb ´21). Add to this dataset the pre-covid values for three marco-economic indicators: (a) GDP per capita (constant), (b) (total) GDP, (c) Gross Governmental Debt (for 2019). Find the data for the macroeconomic indicators in IMF WEO database (use April 2020 edition).

      Calculate the Covid related GDP loss - as % (2019 = 100) for 2020 and 2021. For this task lets assume that the pre-COVID predictions would have been perfect (if there was not COVID). 

      Try to plot all the variables, It is possible that some have extremely skewed distribution - in that case, use log or any other appropriate transformation.

      Hand in as a single excel document (dont forget to include a fine codebook sheet) 

  • Analytical description

    • Our research often starts with a descriptive inquiry that may identify some interesting patterns for further research.

      Your (GROUP) task will be to prepare a short report on the COVID-impact on the EU (+UK) economies.

      Tasks/structure:

      1) Introduce briefly the state of the EU members economies just before the COVID (end of 2019/early2020. Support your argument with a table and a chart (might containg more than one data serie).

      2) Describe briefly the COVID-induced GDP loss over 2020 for EU economies (UK included). DOnt describe all the countries, rather focus on the general trend (within-EU variance might be interesting here) + the extreme (or otherwise interesting) cases. Support your description with a table and a chart.

      3) This leads us to ask a logical question - what is behind the differences in COVID induced GDP loss. SO in the third section look at the possible relationship among the GDP-loss (dependent variable) and (i) the Debt (Gross govern. debt), (ii) GDP per capita, Covid-Deaths for 2020 (see WHO for the data), (iv) Democracy. Given the nature of this task, your inquiry here should be centered on charts (simply plot and compare the putative relationships). Provide a brief evaluation of the visual inspection - which factors seems to be most promissing for further research.

      (iv) provide a brief conclusion

      Dont forget the formal requirements.

      Hand in as a word/pdf (text + the tables and charts), and an excel (data from which you will prepare your report)  

  • Causality

    • HW 4:

      Causality - Read ch. 5 and 6 (Gerring Christenson book)

      1) Define a class of events you would like to explain (define your Y)

      For example: why some (otherwise comparable) countries spend more or less on defense?

      2) Try to generate three alternative theories/explanations (use different frameworks - combinations allowed). You need to define the IV´s (conceptualize them and if needed provide a brief operationalization) + define the logic of your explanations.

      So eventually, you will define the independent variable (or variables) and the general logic of a mechanism connecting IV and DV.

      At least one explanation will use two IVs!

      3) In the next step, construct the (simplified) counterfactual outputs for alternative values of your IVs. E.g. low X leads to high Y, low X leads to low Y. A 2x2 table could be useful here.

      4) Try to find at least two actual empirical cases within the scope of your theory, measure (estimate) the IV´s and propose a hypothesis (regarding the value of your DV). 

      Compare "predictions" (hypothesized values) for your two cases with the actual situation (actual values on the DV).

      5) Summarize your findings into a nice (read "J. Schwabishian") table + short text.

      Your theories will have to be quite simple (naive if you want) unless you want to spend the next 30 hours on thorough research. If some data for your cases are missing, you can make "an expert estimate" (in that case, mention it).

  • Quantitative Methods - Introduction

  • QM - Descriptive Statistics

  • QM - Data Transformation with dplyr

  • QM - Visualisation

  • QM - Statistical Inference

  • QM - Linear Regression

  • Mixed methods

    How to select cases from a larger population or from a statistical model.