quinta-feira, 15 de novembro de 2018

Análise tamanho efeito e poder da estatística (análise utilizando o SIGMA STAT e GPower)

  1.   Variância explicada pelo efeito especial = somo dos quadrados das diferenças
  2.   Erro da variância = resíduo
  3.   Eta parcial ao quadrado = 1/2

Tamanho do efeito = calculado pelo G Power a partir do cálculo do (3)
IMPORTANTE: Numerador DF = graus de liberdade (2 grupos -1) x (2 momentos -1) = 1
Número de grupos: considerar 2 momentos e 2 grupos. Assim, são 4

Two Way Repeated Measures ANOVA (One Factor Repetition)          

Data source: Data 1 in Jousi.SNB

Balanced Design

Dependent Variable: TBARS

Normality Test:                Failed   (P < 0,050)

Equal Variance Test:       Failed   (P < 0,050)

Source of Variation          DF          SS            MS              F                P         
Grupos                                    1        1,965       1,965         59,994       <0 nbsp="" span="">Group
Sujeitos(Grupos)                   40        1,310       0,0328                                         
Tempos                                   1        1,763       1,763         64,978       <0 nbsp="" span="" style="color: red;">Treatment
Grupos x Tempos                   1        1,734       1,734         63,897       <0 nbsp="" span="" style="color: red;">Interaction
Residual                                40        1,086       0,0271                                         
Total                                      83        7,858       0,0947                                         


The difference in the mean values among the different levels of Grupos is greater than would be expected by chance after allowing for effects of differences in Tempos.  There is a statistically significant difference (P = <0 a="" comparison="" differ="" from="" group="" isolate="" multiple="" nbsp="" o:p="" others="" procedure.="" s="" the="" to="" use="" which="">

The difference in the mean values among the different levels of Tempos is greater than would be expected by chance after allowing for effects of differences in Grupos.  There is a statistically significant difference (P = <0 a="" comparison="" differ="" from="" group="" isolate="" multiple="" nbsp="" o:p="" others="" procedure.="" s="" the="" to="" use="" which="">

The effect of different levels of Grupos depends on what level of Tempos is present.  There is a statistically significant interaction between Grupos and Tempos.  (P = <0 o:p="">

Power of performed test with alpha = 0,0500:  for Grupos : 1,000
Power of performed test with alpha = 0,0500:  for Tempos : 1,000

Power of performed test with alpha = 0,0500:  for Grupos x Tempos : 1,000





terça-feira, 13 de novembro de 2018