any experts here willing to assist me with some practice problems for fg ?

Got a sample problem?

Rotten Tomatoes is an American review-aggregation website for film and television. The company was launched in August 1998 by three undergraduate students at the University of California. This company is currently interested in three films that were released in theaters two year ago (2019) – pre-Covid times: (1) Detective Pikachu, (2) Avengers: Endgame, and (3) Parasite and how well they performed during the first week of their release. They gathered 30 avid movie watchers to watch one of the movies (10 people in each group) and asked each person to rate the movie on a 10-point scale: 1 = movie was bad, do not recommend, 10 = movie was great, highly recommend. The data is presented below. (30 points)

Det. Pikachu Endgame Parasite
10 8 7
10 8 5
9 8 6
9 9 4
8 8 3
8 9 5
9 10 5
8 9 6
10 8 5
10 9 4

A. What is the independent variable?
The dependent variable here would be continuous, the movie ratings here is the dependent variable.
B. What is the dependent variable?
The independent variable here is categorical one, which consists of the three groups (Pikachu, Endgame, Parasite).
C. What statistical analysis should be conducted?
ANOVA can be computed to check for the difference in means between the three groups.
D. Is your analysis significant?
The analysis is significant since p-value less than 0.05 in the ANOVA table. Significant p-value indicates that the null hypothesis of equal variance across the group is rejected and there lies difference in means between the three groups significantly at 5% level of significance. yes, the analysis is significant.
E. Is the variance equal across the groups?
The variance is not same across the group.


LMK if this looks right

Guess you have to be carefull here.
Dependent variable (ratings) was measured on a 1-10 point scale, so the variable can take a finit set of values (1 till 10) and therefor has to be considered discrete instead of continous.

Also stats-purists will argue, that rating-scales like this are only ordinal-scale level, since it is very unlikely that the distance between 9 and 10 is indead equal to the one between 4 and 5.
They also will argue, that no method assuming normal-distribution will be appropriate for this scale, since it cant be normaly distributed (because of its fixed end points and narrow range).

Be careful with anova-assumtions (Homoscedasticity) in regards to D and E and fall back to robust methods if in doubt.

This post was edited by Wirbelschwein on Oct 10 2021 12:16pm