Anova and Standard Error Flashcards

1
Q

What do you need to know to calculate 95% confidence interval?

A

Use when data is normally distributed and either we know the sample size is large OR we know the true standard deviation of the population

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

What does the 95% confidence interval show you?

A

The true values of the data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

What is sampling error?

A

The difference between a metric you make from a group and their true value.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Where does the effect of sampling error on obtaining an estimate come from?

A

Standard Deviation

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

What does a small standard error suggest?

A

A lower influence of sampling error whe estimating mean

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

What would happen to the standard error number when the population becomes more vairable?

A

It increases

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

What would happen to the sample size when standered error increases?

A

it decreases

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

What does standard deviation describe?

A

The variance of measurement and data

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

How do you calculate 95% confidence interval?

A

1) 1.96 x standard error = answer
2) Add answer to mean
3) subtract answer from mean

creates interval start and end points

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

How would you calculate the 95% confidence interval if you didn’t know the standard deviation?

A

1) t 0.05 (2 tailed), d.f. multiplied by SE = answer

Will need to look up t table for this information: t 0.05 (2 tailed), d.f.

2) Add answer to mean
3) Subtract answer from mean

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

How do you calculate standard error?

A

Standard deviation divided by sample number

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

How do you calculate standard deviation?

A

1) Calculate mean of data
2) Subtract mean from all data points
3) Square all these numbers individually
4) Add all the now squared numbers
5) Divide this answer by total number of original data points.
6) Remove square root

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Why use 1.96 to multiply for 95% confidence interval?

A

To get 95% of all values beneath main part of linear curve you need to move the curve by 1.96 standard deviations.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

If the 95% confidence interval larger when using a T distribution or smaller?

A

Larger

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

What do the error bars seen in confidence interval graphs tell you?

A

How precisely the CI was calculated

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

If you increased the SD what would happen to the 95% CI?

A

It would increase

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
17
Q

What would happen to the 95% CI if you were to increase the sample size?

A

The CI decreases

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
18
Q

When would you use a 1-way ANOVA and why is it good it only generate 1 P value?

A

When testing for differences between more than 2 groups. It generates a single P-value for easy comparison and less chance of a type 1 error.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
19
Q

What does a 1-Way ANOVA tell you about the results?

A

It confirms differences between groups

20
Q

What are the assumptions for a 1-way ANOVA?

A

1) Random Sampling
2) Independance
3)Homogeniety of variance
4) Normality

21
Q

What are the two ways you can randomly sample something?

A

1) Randomly selecting individuals from a population
2) Placing people into groups based on characteristics and randomly selecting one member from each group to get different treatments.

22
Q

What do we mean by independance?

A

Every value generated if independant from one another e.g. you cant get three values from one person and pretend their from different people, instead calculate a mean from these values.

23
Q

How big does the sample size of a 1-way ANOVA need to be?

A

At least 2 independant measurements per treatment.

24
Q

What happens to power when there is more treatments but a small sample size?

A

Power decreases

25
Q

What is the difference between ANOVA and a GLM?

A

Nothing really apart from GLM’s can be used for more things than an ANOVA

26
Q

Do GLM’s do the same things as ANOVA?

A

Yes

27
Q

Do GLM’s have the same assumptions as an ANOVA?

A

Yes

28
Q

What is degrees of freedom?

A

Number of independant peices of information included in an analysis

29
Q

What happens to degrees of freedom when calculating a mean?

A

The degrees of freedom decreases (n-1).

30
Q

What are some things which can naturally cause variation?

A

Genetics
Environment
Measurement error

31
Q

In variation when would you reject the null hypothesis?

A

When variation is above 1

32
Q

What information do you get when generating a 1 way ANOVA?

A

Degrees of freedom
Sum of squares
Mean Square
F value
P value

33
Q

What are residuals?

A

The distance from a data point to the fitteed value.

34
Q

What does homogeniety assume with residuals?

A

that they are all similar

35
Q

What are some Post Hoc tests of ANOVA and what do they do?

A

P-value, effect sizes

These help compare between treatments.

36
Q

In graph form how do you tell that the assumption of normality is met?

A

All data points fall near the QQ line

37
Q

When should you check assumptions?

A

Before statistical test

38
Q

What do good residuals look like on a graph?

A

Same amount of data below and above line, no pattern and a straight line.

39
Q

What do bad residuals look like on a graph?

A

Different amount of data below and above line, pattern and an uneven line.

40
Q

How do you quantify within variation - the total sum of square calculation?

A

1) Calculate mean of all sample measurements (the grand mean)
2) Subtract every individual point from this grand mean
3) Square all answers from (2)
4) Add all squared values

41
Q

How do you calculate residual sums of squares (this is the sum of square for each individual treatment not as a whole e.g. if you have 3 treatments you’ll get three answers).

A

1) Calculate mean of all sample measurements from one treatment
2) Subtract every individual point from this mean
3) Square all answers from (2)
4) Add all squared values

Repeat for all treatments.

42
Q

How do you calculate variation among groups?

A

1) Calculate the grand mean of all data points
2) Calculate individual treatement means
3) Square grand mean
4)Take treatment mean from grand mean
5) Square then add to other treatment means

43
Q

How to work out variance analysis?

A

Variation among and within sample
Divide by variation within sample

44
Q

What number shows there is no vaiance in a variance analysis?

A

less than 1

45
Q

How do you change variation to variance/ mean square

A

sum of square divided by degrees of freedom

46
Q

How do you calculate F?

A

Calculate the mean square for treatment 1 divide it from the mean square of treatment 2

47
Q

When would you use the F value?

A

In an ANOVA