Today’s focus is on practicing things in Jamovi and in Excel.
You will turn something in from today’s lab on Brightspace (here). You should plan to turn it in at the end of the lab, but it may be turned in up next week. Make a document in whatever program you would like, write your name at the top, and label it Lab 2. You can just paste the results into the textbox in Brightspace, or upload a PDF/word doc. (Pages files don’t load well in Brightspace, so please save as a PDF if you are using Pages.) I do not need your spreadsheet or data. Just these answers.
As you go through this lab, feel free to discuss with your classmates.
Import data. Find the pizzaplace.csv dataset on Brightspace or here. Load it into Jamovi. Read about it more here.
Start by going through the variables, either by clicking on the Variables menu at the top or by going to Data and double-clicking on a variable name (e.g., ID). What are they? Do they make sense compared to the brief description you just read?
Open the same dataset up in Excel (if you have access to it) or in Google Sheets. (If you have it, use Excel over Sheets, as it has more functionality. You can probably do this in Numbers, if you have that and would prefer to use it.) In either program, you can go to the File menu and then click Open, and find the file on your computer. In Google Sheets, you’ll need to click “Upload” once you access the Open menu. (In Excel, you can probably just drag the file onto the program icon in your dock [Mac] or taskbar [PC].) I’ll sometimes refer to Sheets/Excel as S/E below.
What is different in Excel/Sheets compared to Jamovi? What’s the same? (Hopefully the data looks identical!)
Use both Jamovi and Sheets/Excel to find the average price of a pizza. Use the =AVERAGE() function in S/E; use Analyses/Exploration/Descriptives in Jamovi. What difference do you see between the scores on Jamovi vs. S/E? Add the answers here as item #1 in your answers to turn in.
Identify the median price for all of the pizzas in S/E, using the =MEDIAN() function. Does this one look the same as the one in Jamovi? Why or why not? Add this median price to the answer sheet as item #2. What does it signify that the median is so close/so far from the mean?
Imagine that you’re the owner of this pizza place, and you want to figure out which pizzas are different from the normal prices. How would you do that? Well, you might look for outliers! With your Exploration/Descriptives tab up, and
pricein the Variables: find the plots tab (below) and turn on a Boxplot. Ask it to label outliers. Recall from our class on the median absolute deviation (MAD) that this is a method for identifying outliers… how can we make that happen? (You do not need to add this to your answer sheet, since you haven’t done it yet. This is the next question.)We’re going to calculate the MAD ourselves. Not the most fun—perhaps!… but also potentially useful, and conceptually worth doing. For this, I’m assuming you have price in column G of your spreadsheet. Let column H be blank (clear it if needed), and make sure the median is in a cell you can refer to (not in H) or written down. I’ve calculated the median of all the data; it is in cell J5. Label column H, row 1 (cell H1) “abs.deviation” or something like that. In row 2, write the following formula:
=ABS(G2-$J$5). (If you want to use your actual median value, replace$J$5with the median value.) What’s that doing? From the outside in: the=sign tells S/E that this is a formula.ABS()asks for the absolute value (i.e., removes negatives).G2is the first price (a Hawaiian medium classic pizza for $13.25). The term$J$5refers to the cell with our median in it. The $ signs tell E/S not to change this cell when we do what we’re doing next (fill), below.Select cells H2:H7 using your mouse or the Shift & arrow keys. Then either press CMD+D on a Mac (CTRL+D on a PC), or (in Excel) go to Edit: Fill: Down. (In Sheets, you will need to use the key sequence.) What happens? Your first three cells should now read 3.25, 0.5, 0.5.
Use your keyboard or mouse to select all the way to the bottom of the data in row H. Fill all the way down in H. You may see a loading bar at the top right for a second or two in Sheets. In an empty cell to the right, find the median of cells H2:H49575. This is the median absolute deviation (MAD)! If we want to find values that are 3 MAD above the median of the data, and 3 MAD below the median of the data, what would the range be? Add this range to the answer sheet as answer #3.
There’s a neat function in S/E called
=COUNTIF(). If you’re feeling like you’re getting things thus far, try to see if you can figure out how to use this=COUNTIF()function to count how many of the prices in column G are more than 3 MADs over the median, and how many are smaller than 3 MADs under the median. Once you start typing it, S/E will give you a bit of an idea what should go first and second in the parentheses. If you do this question, it is answer #4. It is not required, and I will discuss it in class. Even if you don’t do this question, look over the first 20–30 prices. Do any of these seem like outliers to you? What about a price of $30 for a pizza from this shop?In Jamovi, turn on the Histogram under plots for these data. Look at the histogram. Does it seem normally distributed (i.e., does it look like a normal distribution)? Answer why / why not in your answer sheet as answer #5.
In Jamovi, turn on the Q-Q plot under plots. In a Q-Q plot, datapoints are plotted along a diagonal line. The closer they are to falling on the line, the more normally-distributed the data. Does this confirm your decision on #10 (answer 5)? Feel free to add to your answer.
In Jamovi, uncheck all of the descriptives. Also, uncheck all of the plots except for boxplot + outliers. At the top of the Descriptives menu, add the variable/column size into the “Split by” section. They’ll show up in alphabetical order in the plot. Does price align with size?
Go to the Data tab in Jamovi, and double click on the size variable. Change
Measure typeto Ordinal instead of Nominal. Put them in the correct order. Then go back to the Analyses -> Exploration -> Descriptives, and remake the plot from question #12. Does it make more sense now?In Split By for the Descriptives menu, replace size with type. This isn’t an ordinal variable (except maybe by your preference). Which type of pizza has outliers in the boxplot? This is answer #6 on your answer sheet. We’re done with Jamovi for today. You can close it if you’d like.
Back to S/E. We’re going to calculate z-scores for some of our values. Make sure column I (I as in IDAHO) is blank—you can insert a blank column if you’ve been typing in it. In row 1 of column I, label it
z-scores. In another empty cell (probably off to the right, maybe in column J or K), calculate the standard deviation of all of the prices in column G using the population SD function:=STDEV.P().Using the mean (average) which we calculated in #5, and the standard deviation from #15, let’s calculate a z-score. As you’ll recall from class, \(z=\frac{(X-M)}{SD}\) where \(X\) is an individual value, \(M\) is the mean of the population, and \(SD\) is the standard deviation of the population. (For the moment, the pizza prices are our population.) Can you figure out how to calculate that by yourself, in Excel? It will be quite similar to what we did in question 8. Don’t forget to use parentheses—you can understand what’s going on without them on paper, but you will need them in S/E. Write the equation you use in cell I2 to do this as answer #7. Your answer for cell I2 should be, rounded, -0.90 (to three decimals, it’s -0.897).
Also note that you can use a “relative formula”, which refers to a specific cell, or just enter the values for SD and M. Your choice.
Fill down for the first 20 or so values. Are any of them more then 2 z-scores away from the mean? This is your final answer #8.
Reuse
Citation
@online{dainer-best2024,
author = {Dainer-Best, Justin},
title = {Finding {Your} {Jamovi} {Footing} {(Lab} 2)},
date = {2024-02-08},
url = {https://faculty.bard.edu/jdainerbest/stats/labs//posts/02-finding-your-jamovi-feet},
langid = {en}
}