PSYA01H3 Lecture Notes - Lecture 10: Standard Deviation, Descriptive Statistics, Big Bang
This preview shows half of the first page. to view the full 2 pages of the document.
Psychology Lecture 10 (finishing chapter 2)
Study the textbook for the midterm. The lectures are a way to help you through understanding the
content in the book.
Active studying- you take notes while reading the book, or write a M/C question (which is a good
method to maximize your study). You can also rely on the external links on portal to get some extra
Mode: what there is the most of in a sample.
You use all three (mean, median, mode). How closer to the data points to each other are the numbers?
How typical are they?
Slide 30: the first set in the example has values that are close together, but the second set has a wider
range. If you considered them as an age range for example, and one set represents the age in a
classroom, it would be more difficult to teach the second group because they have such a large age
range (how do you teach a 5 year old what you teach a 20 year old at the same time?)
M.A.D.: mean absolute deviation- you take each deviation point, subtract the mean from it, and
takes an absolute value of that to see how far it deviates from the mean.
Ex: first set from the ex. 18, 19, 20, 21, 22 mean=20.
Deviation: -2, -1, 0, +1, +2 mean=0
If you just ask how far you are from the mean, it’s always zero. That’s why there needs more
mathematical concepts involved in variability. You remove the sign and just add up the deviation
(now called absolute deviation) deviation mean=1.2
Variance does not use the absolute deviation, but gets rid of signs by squaring the numbers.
Ex: first set’s square deviation: 4,1,0,1,4 mean square deviation=2 what does that mean?
Take the square root (standard deviation)= 1.41. Now it’s a little closer to M.A.D.
Variance and standard deviation are preferred over M.A.D. among statistical methods.
Slide 27-30 are considered descriptive statistics. However, scientists use a different method: Inferential
Slide 31: person of interest is suspected for cheating on you. What do you do if you’re not sure, but
think you need to take action? You select data, and you become convinced. You’re convinced because
your null-hypothesis (that they aren’t seeing each other) is rejected. The other hypothesis, that they are
cheating, is increasing their chances the more incidences occur (they don’t respond that often, they
have a different scents). That is supported by data that don’t support the null-hypothesis. When you
confront your partner, you argue by stating the low-probability data (no response, different smell etc. ),
and their response can decide whether the partner’s cheating or not.
You're Reading a Preview
Unlock to view full version