p-verdier, Met 3431, flipped, etc

Njål Foldnes
Bergen. Juni. 2015.

Tema

  • Met 3431
  • Endringer
    • Software
    • p-verdi kontroverser
  • Flipped classroom

Studentstemmer

  • Whats expected: 4.2
  • Course literature: 3.0
  • Acquired knowledge: 4.2
  • Presentation: 4.4
  • Total average: 4.2 (2010: 2.3, 2014: 4.0)

Total average by campus

  • Bergen: 4.2
  • Oslo: 4.0, 4.1, 3.9
  • Stavanger: 4.2
  • Trondheim: 3.6

Karakterer totalt

plot of chunk unnamed-chunk-2

Oppmøte og eksamenspoeng

plot of chunk unnamed-chunk-4


Call:
lm(formula = Poeng.eksamen ~ totalt.oppmøte + gpa + Interesse, 
    data = resJoin)

Residuals:
    Min      1Q  Median      3Q     Max 
-31.419  -9.690   1.701   8.531  30.609 

Coefficients:
               Estimate Std. Error t value Pr(>|t|)    
(Intercept)    10.44864    4.84227   2.158  0.03243 *  
totalt.oppmøte  2.26355    0.36375   6.223 4.06e-09 ***
gpa             0.10334    0.03635   2.843  0.00505 ** 
Interesse       1.12143    0.42547   2.636  0.00922 ** 
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 13.08 on 161 degrees of freedom
Multiple R-squared:  0.2789,    Adjusted R-squared:  0.2654 
F-statistic: 20.75 on 3 and 161 DF,  p-value: 2.027e-11

NHST kontroverser

  • Journalen Basic and applied social psychology forviser NHST
  • Mindless statistics. Gigerenzer
  • Gigerenzer, Gerd, Stefan Krauss, and Oliver Vitouch. “The null ritual.” The Sage handbook of quantitative methodology for the social sciences (2004): 391.
  • Problems with NHST (http://www.andrews.edu/~rbailey/Chapter%20two/7217331.pdf)
  • Ioannidis, John PA. “Why most published research findings are false.” PLoS medicine 2.8 (2005): e124.