PSYC 200W Lecture 4: Conceptual Review
28 Jun 2018
School
Department
Course
Professor
Levels of Measurement - Assign numbers to responses to summarize results and compare them
Nominal - Numbers assigned to responses are just labels such as gender (1,2). Nothing to do
with the values of the number. Can't carry out mathematical functions
Ordinal - they have names and magnitude. The numbers have value and they indicate the order
but not the distance such as ranking (GPA, races)
Interval scales - name, magnitude, equal intervals. It tells us about the ranking and also the
distance between two values and the distance is always equal. For eg - IQ differences between
100 and 110 is the same as 140 and 150. there are no zero points - latitude, there is 0 but it is not
an absolute zero because the 0 here is a value and not the absence of latitude
Ratio - name, magnitude, equal intervals, and absolute zero. Eg - height, weight where there is a
zero point.
Nominal, Ordinal - qualitative, categorical
Interval, ratio - quantitative, continuous
Rating Scales:
Less than 5 options- ordinal
More than 5 - interval
More detail is better to understand so using a scale that gives a more fine answers. More powerful tests
require more detailed data.
Population - entire group about whom research question is asked
Sample - subset of pop
Hypothesis testing- inferential stats
null hypothesis statistical testing is used to determine whether differences between the means of the
experimental conditions are greater than expected on the basis of error variance alone
H0 - null hypothesis - statement of no effect
HA/1 - alternate hypothesis - something is happening
Type 1 - Falsely reject - set by researcher, rule of thumb = 0.05
Type 2 - Falsely fail to reject - not set by researcher, too few participants , rule of thumb = .20
H1: "Somethings happening"
Non-directional
oThere's a relation between x and y
oGroup 1 and group 2's scores will differ
Directional
There's a positive relation between x and y
Group 1 will have higher scores than group 2
Directional results = more powerful research
P-value - expresses probability that obtained difference in conditions is due to error variance
Power - finding an effect if its there to find - ability to reject the null hypothesis when it is false -
opposite of beta --- (1 - b)
We shoot for power of .80 or higher
find more resources at oneclass.com
find more resources at oneclass.com
