Science can be everywhere. Scientific questions, and, hopefully, scientific answers. Regularly at the university's cantine, the salad bar offers nice options, but – in my opinion – ruins them with raisins. When I shared this taste of mine with my colleagues, only a few people answered that they like them. We therefore questioned the abundance of raisins. It started as a joke, but as practical scientists, we now want evidence.
There are several key questions to address. First, of course, what is the proportion of the global human population who likes raisins? Then, to dig a bit more in the data, is there geographical/cultural pattern in the response? Is there any correlation with sex and age?
To collect data, we created an online form, available here: https://goo.gl/forms/4SgSP8Lfh2TtZlNi2. Please take the survey if you haven't yet. It's completely anonymous and we don't track respondents in any way.
The data are then processed in R, first tidied with the
tidyverse package, then plotted with the
ggplot2 package (actually part of
tidyverse now). A first bar plot investigates where do the respondents come from; a second one investigates the age and sex of the respondents. Then come the statistics. The data will be analysed with relevant tests for each particular question: chi-square test, ANOVA, linear mixed-effect model.
If the dataset grows big enough, I'm already considering using it as an ice-breaker during my "Introduction to statistics with R" lectures. It should demonstrate how a scientific process must be conceived: 1. starting with a question, 2. designing the collection of the data, and 3. planning how to analyse them to answer the question, even before having them.