Instructions:

Return to the EVE character dataset. Review the week $1$ assignment for details on the data.

CCP remains interested in your analysis of the relationship between character creation choices $($particularly character Gender$)$ and duration of gameplay. In digging into the setting and data more you noticed that some of the players $($userIDs$)$ create multiple characters $($charlDs$).$

Background:

The attached dataset $($EVE$\_$char in R format, or EVE$\_$char.csv in text$)$ is an anonymized random sample of proprietary data for users of the video game EVE Online. Eve is an online role$-$playing game hosted by the CCP Games. In EVE players create characters and then compete and interact with others in a fictional universe. See EVE Online and EVE Online $-$ Celebrating $15$ Years of EVE $($below$)$ for more details on the game.

Dataset:

The dataset contains the following variables for a random sample of $311742$ EVE characters from six real$-$world nations:

userID: an ID code for each unique user playing EVE

UGender: the self$-$reported user gender

UCountryIP: user nation, based on IP address

Uage$\_$decile: user self$-$reported age

charID: an ID code for each unique character

CGender: the character gender chosen by the player. This choice has no effect on the capabilities or resources of the character.

CCareer: the character "career path" chosen by the player. This choice has no effect on the capabilities or resources of the character.

CViolence: $1=$ the character engages in violence during the first week of game play

CCorp: $1=$ the character joins a player run corporations

CDMY: the month$/$year of character creation

LogonHours: the number of hours of cumulative game play

Do the following:

Run and report two regressions for LogonHours with CGender as one of your explanatory variables:

for model $1$ exclude userID fixed effects

for model $2$ include userID fixed effects $($see the Ife package in R$)$

Why does the coefficient on CGender change from model $1$ to model $2,$ and how do you interpret that coefficient in each model?

Why does the R$-$squared increase, and the degrees of freedom drop, in model $2?$

What does a "fixed effect" mean?

Should model $2$ include all cases? Or there are some we should drop here?

Dataset Files:

EVE$\_$char

EVE$\_$char.csV

Submission Format Guidelines:

Treat this assignment as if you are writing a professional memo. The report must consist of prose $($PDF or MS Word format$)$ and contain neatly formatted tables and figures. You may use bullet points to summarize facts and findings but must fully explain those points. You may not exceed one page of prose and one page of tables or figures. Submit your code as a separate appendix.