Warning: Many online articles and other sources only provide aggregated tables and not the individual-level data used to create those tables. Think about what your sampling units are (what does a row in your data represent?) and be sure that the variables you will be analyzing are measured at the individual sampling unit level. Data Collection Considerations Consider the following, as whether or not you meet these criteria will impact your grade: . Measure numeric variables so that you can expect to get at least 10 different values in your sample. For example, if you ask people how many caffeinated drinks they had yesterday, you will likely only get responses of 0, 1, 2, or 3. To get a wider range of values, ask them how many they drank in the past week. . For categorical variables, do not have more than 3 possible outcomes and make every attempt to get at least 15 observations in each category. If both of your explanatory variables are categorical, attempt to get at least five responses in every possible combination of categories. . You will be asked to cite at least one source (academic or non-academic) that provides a basis the expected relationships between your explanatory variables and response variable. Your project can attempt to confirm, refute, replicate, or expand on a previous study