Week 4 - Statisitical Models Correlation & T-tests Flashcards
It is critical in psychological research to operationalise key terms because…
The reader needs to know what is actually being studied
A normal distribution as a…
Bell shape
“It was predicted that individuals who completed more years of education would have lower cardiovascular disease burden”. What is the correct characterisation of this statement?
It is a directional hypothesis.
We use the terms independent variable and dependent variable for….
Experimental and quasi-experiential designs
We use the terms predictor variable and outcome variable in…
Observational
We are running a study manipulating bedroom light levels to investigate effects on sleep efficacy. What is the (1) IV and what is the (2) DV?
(1) bedroom light, (2) sleep efficacy
What is not a measure of central tendency?
a.
Skew
b.
Mean
c.
Median
d.
Mode
Skew
Leptokurtic means…
Positive kurtosis
Negative kurtosis is also …
Platykurtic
Normal distribution is also
Mesokurtic
when do we run a Spearman Correlation?
When the data is non parametric
When do we run a Wilcoxin test (as opposed to a t-tes)
when the DV is non paramtertic
We traditionally use cohens d to measure effec size for t-tests. True or False
True
a d-value of 0.2 is considered…
small effect
a d-value of 0.5 is considered…
moderate effect
a d-value of 0.8 is considered…
large effect
Testing skew or kurtosis:
W statistic - maximum value of 1 = data looks perfectly normal
True or False
True
Testing skew or kurtosis:
The smaller the value of W the less normal the data are?
True or false
True
p value of W statistic
<.05 = non normal data
True or false
True
p value of W statistic
>.05 = normal data
True or false
True
p value of W statistic
>.05 = non normal data
True or false
the higher the df = more statistical power
True or False
True
df = participants - 1
True or false
True
What are the assumptions of Pearson correlation?
Linear relationship (straight line).
Homogeneity of variance (homoscedasticity).
Parametric data/normality.
Independence.
At least one variable needs to be continuous.
The other variable can be continuous or dichotomous.
What are assumptions of t-tests?
Parametric data/normality.
Independence.
Homogeneity of variance (homoscedasticity).
DV needs to be continuous.
IV needs to be dichotomous (groups or time points
Pearson correlation
r= 0 no relationship
r=1 perfect positive relationship
R -1 perfect negative relationship
True or False
True
Pearson correlation also measures effect size
-1.0 to- 0.9 is a very strong negative correlation
-0.9 to -0.7 strong neg
-0.7 to -0.4 moderate neg
-0.4 to -0.2 is a weak negative correlation
-0.7 to -0.4 is moderate negative
0 - 0.2 is a negligible positive
0.2 to 0.4 is a weak positive correlation
0.4 to 0.7 Moderate pos
0.7 to 0.9 is a strong pos
0.9 to 1.0 is a very strong positive correlation
True
Why do we use t-test
When you want to compare two means
When do we use between independent t-test?
If those two means are from different groups - assigned to 1 condition only
i.e., group 1 vs group 2
When do we use within groups dependent/paired t-test?
if those two means are from the same people - assigned to both conditions
i.e., time 1 vs time 2 using the same groups to compare
Parametric data (normal)
Bell curve
Not too skewed (sway left or right)
Not too kurtotic (flat or peaky)
No outliers (extreme values)
T or F
True
