Introduction
In this assignment, we will discuss and describe the application of probability theory and statistics to a real world situation. The goal of this assignment is to perform statistical inference for two research questions. According to the instructions, Dr. Kava performed a case study to examine the relationship between alcohol consumption and attractiveness at night clubs. The experiment was designed to test the so-called “beer-goggle” effect. Our task is to test any interesting hypotheses in this paper, using the dataset and variables collected by Dr. Kava.
Methods
Subjects or Participants
We are operating with the data set of 200 observations – 200 participants were recruited by Dr. Kava for her study. The participants were randomly assigned between two groups. One group received alcohol-free beer and the other group received 4 portions of alcohol beer. Next, groups of 20 participants were taken to a local night club, where there were 50 individuals pretending to be patrons of the club. The participants of the study were not aware about that those people were a part of the experiment. No specific instructions were provided to the participants, except that before the end of the night they should rate the attractiveness of all the people at the club on a scale from 1 to 100. The participants rated all the 50 individuals and each that rating was averaged and recorded for each participant. Three months later, participants were brought back to the Dr. Kava, where they were shown the pictures of those 50 individuals. The researcher asked them to rate their attractiveness once again, on a scale from 1 to 100. The new ratings were averaged and saved for each participant.
The following variables were included in this research:
age = age of participant
attract1 = attractiveness rating from the club
attract2 = attractiveness rating from 3 months later
female = coding scheme for gender variable
gender = gender of participant
alcohol = coding scheme for alcohol_grp variable
alcohol_grp = an alcohol group in which the participant belonged
In this study, we will examine the following two research questions:
Is there any difference in the average rating scores of “alcohol” group (those, who received 4 beers) between the rating from the club and rating from 3 months later?
Is there any difference in the average rating scores of “non-alcohol” group (those, who received 4 beers) between the rating from the club and rating from 3 months later?
Procedure
Two research questions are tested. For the first research question, the following steps will be performed:
Set the level of significance alpha at 0.05
Choose the appropriate statistical test. Since the test is performed two times for the groups with the same individuals, paired t-test should be performed.
Calculate appropriate statistics and report the probability of type 1 error.
Make a decision about the null hypothesis
Interpret the results
For the second research question, the following steps will be performed:
Set the level of significance alpha at 0.05
Choose the appropriate statistical test. Since the test is performed two times for the groups with the same individuals, paired t-test should be performed.
Calculate appropriate statistics and report the probability of type 1 error.
Make a decision about the null hypothesis
Interpret the results
Results
For the first research question, paired t-test reported t(122)=39.671, p=0.001, indicating that the null hypothesis is rejected. There is enough evidence to support the claim that there is a significant difference in the average rating scores of “alcohol” group between the rating from the club and rating from 3 months later (at the 5% level of significance)..
For the second research question, paired t-test reported t(76)=28.364, p=0.001, indicating that the null hypothesis is rejected. There is enough evidence to support the claim that there is a significant difference in the average rating scores of “non-alcohol” group between the rating from the club and rating from 3 months later (at the 5% level of significance).
Discussion
In this paper, we examined the relationship between alcohol consumption and attractiveness at night clubs. We have tested how the rating scores of alcohol and non-alcohol groups are different between rating in a club and rating in the lab 3 months later. It is appeared that both groups indicated significantly different results in the average rating scores in the two rating procedures. This suggests that the change in the level of attractiveness of perception may be caused not only alcohol, but also on the overall situation in general (being in a club or in the laboratory).
Appendix (R code)
//Convert file to .csv
d <- file.choose();
//Choose your .csv file
mydata<- as.data.frame(read.csv(d),header=T);
newdata=mydata[which(mydata$alcohol_grp=='4 beers'),]
t.test(newdata$attract1,newdata$attract2,paired=T)
newdata2=mydata[which(mydata$alcohol_grp=='none'),]
t.test(newdata2$attract1,newdata2$attract2,paired=T)
