Facebook Attraction

In this WPA, you will analyze data from a study on attraction. In the study, 1000 heterosexual University students viewed the Facebook profile of another student (the “target”) of the opposite sex. Based on a target’s profile, each participant made three judgments about the target - intelligence, attractiveness, and dateability. The primary judgement was a dateability rating indicating how dateable the person was on a scale of 0 to 100.

The data are located in a tab-delimited text file at http://nathanieldphillips.com/wp-content/uploads/2016/04/facebook.txt

Here is how the first few rows of the data should look:

##   session sex age haircolor university    education shirtless intelligence
## 1       1   m  23     brown   3.Geneva    3.Masters     2.Yes        1.low
## 2       1   m  19    blonde   2.Zurich 1.HighSchool      1.No     2.medium
## 3       1   f  22     brown   2.Zurich  2.Bachelors     2.Yes        1.low
## 4       1   f  22       red   2.Zurich  2.Bachelors      1.No     2.medium
## 5       1   m  23     brown   3.Geneva  2.Bachelors      1.No     2.medium
## 6       1   m  26    blonde   2.Zurich    3.Masters     2.Yes       3.high
##   attractiveness dateability
## 1         3.high          15
## 2       2.medium          44
## 3       2.medium         100
## 4         3.high         100
## 5       2.medium          63
## 6         3.high          76

Datafile description

The data file has 1000 rows and 10 columns. Here are the columns

Data loading and preparation

A. Open your WPA.RProject and open a new script. Save the script with the name WPA7.R.

B. Using read.table(), load the tab-delimited text file containing the data into R from http://nathanieldphillips.com/wp-content/uploads/2016/04/facebook.txt and assign it to a new object called facebook. Make sure to specify that the file is tab-delimited with the argument sep = \t and contains a header with the argument header = T.

C. Using write.table(), save the data as a text file called facebook.txt into the data folder in your working directory. That way you’ll always have access to the data even if it’s deleted from the website you downloaded it from.

Understand the data

D. Look at the first few rows of the dataframe with the head() function to make sure it looks ok.

E. Using the summary() function, look at summary statistics for each column in the dataframe. Make sure everything looks ok.

Answer guidelines

For example, here is how I would analyze and answer the question: “Was there an effect of diets on Chicken Weights?”"

Answer: There was a significant main effect of diets on chicken weights (F(3, 574) = 10.81, p < .01). Pairwise Tukey HSD tests showed significant differences between diets 1 and 3 (diff = 40.30, p < .01) and diets 1 and 4 (diff = 32.62, p < .01). All other pairwise differences were not significant at the 0.05 significance threshold.

One-way ANOVAS

  1. Was there a main effect of the university on dateability? Conduct a one-way ANOVA. If the result is significant (p < .05), conduct post-hoc tests

  2. Was there a main effect of intelligence on dateability? Conduct a one-way ANOVA. If the result is significant (p < .05), conduct post-hoc tests

  3. Was there a main effect of haircolor on dateability? Conduct a one-way ANOVA. If the result is significant (p < .05), conduct post-hoc tests

Multi-independent ANOVAs

  1. Conduct a two-way ANOVA on dateability with both intelligence and university as IVs

    • Do your results for each variable change compared to your previous one-way ANOVAs?

  2. Conduct a multi-way anova including ALL independent variables predicting dateability. Conduct post-hoc tests on the conditions that differ.

    • Add a new column to the dataframe called all.aov.dateability that has the predicted dateability for each person according to the multi-way ANOVA you just ran

    • Create a scatterplot showing the relationship between the actual dateability and predicted dateability. Add appropriate labels to the plot

ANOVAs on subsets of data

  1. It turns out that the experimenter who ran sessions 1 through 30 (a man) was trying to score a date and slipped in his own profile picture into the study. We can’t trust these data. Repeat your multi anova from question XXX ONLY for sessions 31 through 50. Do your conclusions change compared to when you analyzed the data from all sessions?

Interactions

  1. Create a plot (e.g.; pirateplot(), barplot(), boxplot()) showing the distribution of dateability based on two independent variables: sex and shirtless

    • Based on what you see in the plot, do you expect there to be an interaction between sex and shirtless? Why or why not?

    • Test your prediction with the appropriate ANOVA

CHECKPOINT!

Understanding how one-way ANOVAs make predictions

  1. Conduct a one-way ANOVA on the effect of attractiveness on dateability:

    • Add the fitted values from your previous ANOVA back to the dataframe as a new vector called attractiveness.aov.dateability

    • Round attractiveness.aov.dateability to the nearest 3rd decimal place using the round() function. (Hint: Assign the variable in the dataframe to a rounded version of itself).

    • Look at all the unique values of attractiveness.aov.dateability with table(). How many different values does the ANOVA predict?

    • Calculate the actual mean dateability for each level of attractiveness with aggregate() or dplyr()

    • Based on what you’ve found, how does an ANOVA fit specific values to data? In other words, if you conduct a one-way ANOVA, and make predictions for groups based on that model, what will the ANOVA predict for each observation in each group?

One-way vs. multi-way ANOVAS

  1. Let’s study the relationship between university and attractiveness on dateability.

    • Conduct a one-way ANOVA on dateability with attractiveness as the IV (I know you just did it, but do it again)

    • Conduct a one-way ANOVA on dateability with university as the IV

    • Conduct a single multi-way ANVOA with both variables (use formula = attractiveness + university). What is your conclusion?

    • Did something change? If so, explain your findings!

Using lm() coefficients to understand group differences

  1. Conduct a multi-way ANOVA on dateability with sex and education as independent variables. No matter if they are significant or not, conduct post-hoc tests.

    • Repeat your analysis using regression instead of ANOVA to get regression coefficients.

    • What are the default values of sex and education in your regression analysis?

    • What dateability does the regression predict for a female with a high-school education?

    • What dateability does the regression predict for a male with a PhD?

    • Are the significance levels for group differences the same in your ANOVA post-hoc tests and your regression analysis?

More interactions

  1. Create a plot (e.g.; pirateplot(), barplot(), boxplot()) showing the distribution of dateability based on two independent variables: sex and shirtless

    • Based on what you see in the plot, do you expect there to be an interaction between sex and shirtless? Why or why not?

    • Test your prediction with the appropriate ANOVA

  2. Create a plot (e.g.; pirateplot(), barplot(), boxplot()) showing the distribution of dateability based on two independent variables: university and haircolor

    • Based on what you see in the plot, do you expect there to be an interaction between university and intelligence? Why or why not?

    • Test your prediction with the appropriate ANOVA

Predicting new data

Here are data for 3 of your friends:

sex age university intelligence shirtless attractiveness
m 22 1.Basel 3.high 1.No 3.high
f 23 1.Basel 1.low 2.Yes 3.high
m 26 2.Zurich 1.low 2.Yes 1.low
  1. Create a new ANOVA object on the original data set containing ONLY variables in the data set for your 3 friends. Then, using this ANOVA, predict the ratings of your 3 friends