Tutorials

by (2014)
31 Flashcards & Notes
20 Students

Study smarter with eFaqt summaries

  • Available on desktop, tablet, mobile & print
  • Questions with answers about the material
  • Access to 300 000 online summaries
  • Smart study features & timers for more results

Get this summary

Samenvatting - Tutorials

  • 1 Tutorial 1

  • Do we have to measure performance at all levels of analysis?

    No. However, scholars, reviewers and editors should carefully examine the level of analysis used in the theoretical development of hypotheses before empirical generalization
  • a.Popularity bias: selecting measures that are popular in the research community due to external sources, even if they are not the best measures to use;
    b.Convenience bias: selecting measures based on the availability of them;
    c.Data source bias: overlooking measures because of lesser quality/quantity; and
    d.Resource bias: selecting the less costly and time consuming measures
  • selection bias
    a.Popularity bias: selecting measures that are popular in the research community due to external sources, even if they are not the best measures to use;
    b.Convenience bias: selecting measures based on the availability of them;
    c.Data source bias: overlooking measures because of lesser quality/quantity; and
    d.Resource bias: selecting the less costly and time consuming measures.
  • 2 Tutorial 2

  • what is the purpose of geographic distribution?
    It shows the spread of the sample. This helps the reader (and obviously the researcher) to get an idea of the type, size and spread of sample he or she is dealing with. These observed country differences in acquisition patterns suggest for example that it may be important to take into account country and time effects.
  • purpose of descriptive statistics?
    1.By examining the data before the application of any technique for analysis, the researcher gains several critical insights into the characteristics of the data:
    a.First and foremost, the researcher attains a basic understanding of the data and relationships between variables; and
    b.       Second, the researcher ensures that the data underlying the analysis meets all of the requirements for any type of analysis.
  • columns of Table 3 show some stars (*). What does this mean?
    Whether the difference, between the mean and median between family and non-family targets, is significant statistic-wise at pre-specified significance levels (5% and 1%).
  • 3 Tutorial 3

  • Ordinary Least Square (OLS) regression
    is a mathematical technique to best represent the linear relationship between X and Y. It minimizes the sum of squared deviations between the line and the observations.
  • Coefficient of Determination 
    It is the portion of explained variation by the regression line.
    (R2 does not have to be significant to confirm hypotheses (i.e. significant coefficients)!)
  • F-test
    can be performed to test the overall model. It tests the possibility whether all parameters are jointly zero. The test value needs to be compared to critical values of the F-distribution or p-values.
  • ·Five assumptions of a classical linear regression model
    1.Y is a linear function of X:
                      No curvilinear relationship;
    If needed, use (logarithmic) transformations (e.g. log X, 1 / X);


    2.Expected value of the error term is zero;


    3.Error terms are expected to be uncorrelated and have the same variance:
    No multicollinearity and heteroscedasticity;
    Include normality;


    4.X is fixed in repeated samples, thus not stochastic (i.e. not random); 


    5.The number of observation should be bigger than the number of regressors and there should be no exact linear relationship between the regressors.
  • Define 'control' variables
    • A control variable (explanatory, independent, regressor variable) is a variable introduced
    • to help interpret the statistically significant relationship between variables. 
    • That is, you want to exclude its influence on the dependent variable from the model
  • Explain all elements in the second column of the table. In particular, how do they relate to the first column? And which are the independent variables?
    The upper section contains the specific hypothesized relationships. The lower part contains the control variables. The squared value is needed to estimate the curvilinear relationship.
  • In column (2), the first entry reads ‘0.320*** (0.058)’. Explain the numbers and the stars. 
    0.320 indicates that for every ‘1’ competitive pressure, the quality of vertical linkages is 0.320 (i.e. the estimated regression coefficient of local competitive pressure), all other things being equal. The stars indicate that this relationship is significant at p < 0.01 (99% confidence). 0.058 is the robust standard error.
  • 4 Tutorial 4

  • Heteroscedasticity
    : the variability of the dependent variable (Y) is unequal across the range of values of a variable that predicts it (X). Uneven distribution of errors in the scatterplot. 
  • reverse causality: solution
    1.Gather (measure) the data corresponding with X before the data corresponding with Y.
  • correlation coefficient of -1.000 
    (extreme collinearity) for every white, there is one black less.
Lees volledige samenvatting
Maak nu je eigen eFaqt account aan voor toegang tot deze en duizenden andere hoge kwaliteit samenvattingen.

Example questions in this summary

Do we have to measure performance at all levels of analysis?
2
selection bias
2
what is the purpose of geographic distribution?
2
purpose of descriptive statistics?
2
Page 1 of 8