# Research Methods - Stats exam

?
• Created by: miSha91
• Created on: 09-08-16 18:45
What is meant by 'power effiency' in reference to parametric tests?
Power of test refers to its relative ability (1) to detect signifcant effect when null hypothesis is false (1). Or you can say it avoid making a type II error (1)
1 of 22
What is roboustnes in reference to parametric testing?
Fairly accurate results can still be produced even if there has been some violation of the underlying assumptions.
2 of 22
Three assumptions underlying parametric testing?
(1) Data is measured on interval or ration scale. (2) Population variances should be similar (homogeneity of variance). (3) Normal distribution of sample data from parent population.
3 of 22
What is a type II error? how can it be dealt with?
It is an error of pessimism (1) Resulting from retaining a null hypothesis when it should have been rejected (1). This error can be made less likely by choosing a less stringent significance level (1) and increasing the power (1).
4 of 22
What is a type I error? How can it be prevented?
it is an error of optimism (1) that is a result of rejecting a null hypothesis when it should be retained (1). It can be prevented by choosing a more stringent significance level (1).
5 of 22
Why is it conventional to use .05 sig. level?
.05 sig. level is thought to strike a balance between probability of making a type I or type II error (1). It is often used in research that is not deemed 'life or death' (1)
6 of 22
What does the p value tell you?
The p figure that is quoted states the probability of obtaining the observed test result (1) if the null hypothesis is true (1).
7 of 22
What does df?
Degrees of freedom (1).
8 of 22
What is partial Correlation? Give an example.
Its when the correlation between two variables are tested (1) whilst controlling the third variable (1). Example: test correlation between intoxication and alcholic intake (1) whilst controlling bodymass (1).
9 of 22
Is correlation of -.8 weaker than +.8?
Correlation values are the same strength (1) due to numerical values being the same (1). They only differ as they show that the relationships are in different directions i.ie postive or negative (1).
10 of 22
Between which two values can correlation coeffient vary?
They can vary between +1.00 and -1.00
11 of 22
How do standard errors and confidence interval relate?
Standard erros can be used to predict/calculate confidence intervals (1) for the given statistic (1).
12 of 22
What is meant by 95% confidence interval?
There a range of values (1) for any given stat (1) within it we have confidence that the true value of that statistic lies (1) in 95% of samples (1).
13 of 22
What info does standard error of the mean give?
It tells you how much variation there is in the mean (1) across the samples from the same population (1).
14 of 22
What is standard error of the mean?
Its the standard deviation (1) of sampling distribution of the mean (1).
15 of 22
What is a Z-score?
is the number of standard deviations a score is away from its sample mean.
16 of 22
Name three levels of dispersion?
1) Standard deviation (2) range (3) variation ratio.
17 of 22
What are measures of dispersion used for?
Used to indicate the variabilty or spread in a sample scores.
18 of 22
When would you use the median rather than the mean?
Preferred when distribution scores are skewed. As a result of the being a small number of atypical scores.This mean is distorted by this score whereas the median is not.
19 of 22
Define three measures of central tendency.
1) The mean is a sum of score dived by the number of scores (2) the mode frequently occurring value in set of scores and (3) the median is the central value in a of scores.
20 of 22
What is central tendency used for?
It is used to sumamarising up a large amounts of data into one typical or average value.
21 of 22
Name four levels of measurement.
nominal, ordinal, ratio and interval
22 of 22

## Other cards in this set

### Card 2

#### Front

What is roboustnes in reference to parametric testing?

#### Back

Fairly accurate results can still be produced even if there has been some violation of the underlying assumptions.

### Card 3

#### Front

Three assumptions underlying parametric testing?

#### Back ### Card 4

#### Front

What is a type II error? how can it be dealt with?

#### Back ### Card 5

#### Front

What is a type I error? How can it be prevented?

#### Back 