Lecture 3,4,5

Question 1

Do our assumptions need to be true to interpret our results?

Answer 1

Yes if they are not true we cannot trust our results

Question 2

What are parametric tests focused on?

Answer 2

Continuous data
The one overarching assumption is you have atleast 1 continuous variable

Question 3

What does the central limit theorem say?

Answer 3

That if N>30, the sampling distribution is normal anyway

Question 4

What are the graphical displays we use?

Answer 4

Histogram and probability plot or quantile plot

Question 5

What does the KS test test?

Answer 5

Tests if data differs from a normal distribution
This test gives us our p value

Question 6

What is the null and alternative for normality?

Answer 6

Null- there is no difference between our variables distribution and a normal distribution
Alternate- there is a difference between our variables distribution and a normal distribution

Question 7

True or false: Statistical tests provide the probability based on the null?

Answer 7

True

Question 8

Do we reject or fail to reject the null based on the p value?

Answer 8

Yes it is based on the probability value

Question 9

What does reject and failing to reject mean? (Ks test)

Answer 9

If we reject there is a difference between our variable and a normal distribution

If we fail to reject there is no difference between our variable and a normal distribution

Question 10

When do we not want to reject the null?

Answer 10

We don’t want to reject the null when it comes to assumptions. We want to meet our assumptions so it is normally distributed

Question 11

What is a type 1 error?

Answer 11

Occurs when we believe that there is a genuine effect in our population when in fact there isn’t.
It is a-level (usually 0.05)
This shouldn’t exist

Question 12

What is a type 2 error?

Answer 12

Occurs when we believe that there is no effect in the population when there is

Question 13

What does independence mean?

Answer 13

It means one persons score is not related to another
It is implemented at the design stage

Question 14

What is the homogeneity of variance when you have several groups?

Answer 14

It is when the variance is equal in both groups

Question 15

What is the homogeneity of variance when you have continuous variables

Answer 15

It is when the variance is spread out equally at every value

Question 16

When comparing two means which variable will be categorical which will be continuous

Answer 16

The independent variable will be categorical

The dependent variable will be continuous

Question 17

What does a t-test measure?

Answer 17

It measures the mean between two groups

Question 18

What is a dependent t test

Answer 18

Both groups are the same. They go through the same thing.

Question 19

What is a independent t test

Answer 19

Both groups are different from each other

Question 20

What is a one sample t test

Answer 20

We only have 1 sample we collected data from and we are comparing our same to the larger population (known value)

Question 21

What is the null and alternate hypothesis in a one sample t test

Answer 21

Null- there is no significant difference between the sample mean and the population mean

Alternate- there is a significant difference between the sample mean and population mean

Question 22

What is the null and alternate hypothesis in a independent t test

Answer 22

Null- there is no significant difference in means between the experimental and control group

Alternate- there is a significant difference in means between the experimental group and the control group

Question 23

What is the null and alternate hypothesis in a dependent t test (paired samples t test)

Answer 23

Null- there is no significant difference between the means before and after

Alternate- there is a significant difference between the means before and after

Question 24

What are the 4 things t tests assume?

Answer 24

1. Independence
2. At least one continuous variable (it’s the dependent variable)
3. Variables are normally distributed
4. Homogeneity of variance

Question 25

What is the null and alternative for ks test of normality?

Answer 25

Null- there is no difference between our variables distribution and a normal distribution Alternate- there is a difference between our variables distribution and a normal distribution

Question 26

How can we identify from a table of data that we did a t-test?

Answer 26

There will be a Levene’s test of homogeneity of variance

Question 27

What is the null and alternative of Levene’s test of homogeneity of variance?

Answer 27

Null- there is no difference in variance between our two groups Alternate- there is a difference in variance between our two groups

Question 28

The p value is what when we met our assumption (our sample is not significantly different from equal variances)

Answer 28

P > 0.05

Question 29

The p value is what when we violated our assumption (our sample is significantly different from equal variances)

Answer 29

P < 0.05

Question 30

If we violate our assumption in a t test, which row do we interpret in the set of data

Answer 30

If we violate (p < 0.05) we interpret the BOTTOM row

Question 31

In an independent t test, what is our probability value on the data table?

Answer 31

It is the two-sided p value

Question 32

In a t test do we want to reject the null from our two sided p value

Answer 32

Yes we do because we want to find something from our intervention. We want our probability to be smaller then 0.05 to conclude there is a meaningful difference

Question 33

What is the second last thing we report?

Answer 33

95% confidence intervals

Question 34

What does a 95% confidence interval mean?

Answer 34

It means if you did the test 100 times 95/100 times the estimate will fall in this boundary

Question 35

Should 0 be in your range in a 95% confidence interval

Answer 35

No 0 should not be in the range is results are significant That would mean your possible effect are 0 or no effect

Question 36

What is the last thing we want to report

Answer 36

The effect size

Question 37

What are the cohen’s d effect sizes

Answer 37

Small effect is .20 Moderate effect is .50 Large effect is .80 (This is plus or minus. So -.30 would still be a small effect)

Question 38

Can a cohens d effect size be larger then 1?

Answer 38

Yes that just means it’s a very big effect size

Question 39

Where do we find the effect size (cohens d) on a graph table

Answer 39

It is under point estimate and cohen’s d

Question 40

If our t test has a p value of 0.044 do we have an effect?

Answer 40

Yes it is less than 0.05 so we have an effect and our intervention worked

Question 41

Which of the following terms best describe the sentence: “Children in the sports program will have different mental wellbeing than children not in the sports program” A. An ordinal value B. An operational definition C. A null hypothesis D. An alternate hypothesis

Answer 41

D (this answer because there IS an effect)

Question 42

Why would you run a t test over a correlations test

Answer 42

A t test compares the means of 2 groups and a correlations test is used to understand the relationship between two variables (measures strength and direction)

Question 43

What is the difference between a chi squared test and a Pearson correlation test

Answer 43

Chi squared is used with categorical data and tests the association between categorical variables to expected data Pearson correlation is 2 continuous variables and assess the strength and direction of a linear relationship

Question 44

What does a cohens d allow us to do

Answer 44

It allows us to start comparing the size of the mean difference that we found across different studies

Question 45

What is a Pearson correlation

Answer 45

Is it a way of measuring the extent to which two variables are related Trying to look at the average pattern between these two continuous variables

Question 46

What is linearity?

Answer 46

It is visual. It can be observed through a scatter plot. It puts one variable on the x and one on the y axis and we want to see how strongly they are related.

Question 47

What does it mean if we meet the assumption of linearity?

Answer 47

It means the dots are going up in a positive direction

Question 48

For the assumption of linearity, what can the values range between

Answer 48

-1 and +1

Question 49

What is the null and alternate hypothesis for correlations

Answer 49

Null- there is no significant relationship/association/correlation between…. Alternate- there is a significant relationship/association/correlation between….

Question 50

What letter represents the correlation correct

Answer 50

r

Question 51

What happens if there is a correlation of 0?

Answer 51

Then there is no relationship

Question 52

What is correlation and causality?

Answer 52

Correlation coefficients say nothing about which variables causes the other to change. So if exam anxiety reduces exam performance, better exam performance can reduce anxiety. We don’t actually know which one is correct

Question 53

What is a partial correlation

Answer 53

It measures the relationship between two variables adjusting for the effect that a third variable has on them both

Question 54

What is a semi partial variable

Answer 54

A measure of the relationship between two variables while adjusting for the effect that a third variable has on one of those variables

Question 55

What is an assumption in a chi square test?

Answer 55

Expected frequencies are greater than 5

Question 56

What a standardized residuals?

Answer 56

It is data given to us in a chi square test The standardized are z scores (if the value lies outside of +/- 1.96 then it is significant because p < 0.05) (If falls between +/- 1.96 then they are not different then chance)