ANOVA CODING

1) When do you reject Null Hypothesis?

• Fcal < Fcrit

• Fcal > Fcrit

• p value> 0.05

• Never reject

 2) Why do you perform Tukey's HSD test

• To know variance in the group

• No reason

• Must perform after Anova

• To know which two groups means are different

3) Why do you make use C in the following block of code:

model = ols('height ~ C(water) + C(sun) + C(water):C(sun)', data=df).fit()

• To make sure it is a dependent feature

• To make sure that machine knows that these are separate features

• To indicate that these variables should be treated as categorical.

• To find the sum of each feature

sm.stats._____(model, typ=2)

4) Fill in the blanks above code.(We want Anova table as output)

• anova_lm

• twoway_anova

• f_oneway

• anova_table

import scipy.stats

scipy.stats.f(3,24).ppf(0.95)

5) This is the code for finding critical value for one way anova. By looking at degrees of freedom how many treatments are we considering for the test?

• 24

• 4

• 1

• 5

6) melt syntax (For example: df.melt) is used to____

• To melt the features

• To make datasize smaller

• Transform this long format into wide format

• Transform this wide format into long format

7) Which of the following test where ANOVA assumptions cannot be checked?

• Bartlett’s

• Shapiro-Wilk

• Tukey's HSD

• Levene’s test

8) Points on the QQ plot fall along a straight line. What does it suggest?

• Data is straight forward

• Data is approximately normally distributed.

• Data is good

• Nothing

9) Which of the following test is used to check the Homogeneity of variances?

• Post hoc test

• Ad hoc test

• Tukey's HSD

• Bartlett’s test

10) You are performing F test for 95% Confidence Interval. Your p value of F test is 0.5. What is your conclusion regarding null hypothesis?

• Reject

• Accept

• No idea