# Statistics 2

HideShow resource information
• Created by: Shannon
• Created on: 13-02-16 13:18
What does the research design affect?
Which statistical test you use
1 of 92
When will the research design change?
If you are finding a difference or a relationship
2 of 92
What is correlational design?
Investigates whether a relationship exists between two variables
3 of 92
Which correlational design is used if data is normally distributed?
Pearsons' product moment correlation (r)
4 of 92
Which correlational design is used when data is NOT normally distributed?
Spearman's Rho Correlation Coefficient
5 of 92
What is an experimental design?
Used to investigate whether a difference exists between two variables
6 of 92
What is an independent variable?
Manipulated
7 of 92
What is a dependent variable?
Measured
8 of 92
What is an extraneous variable?
May have an impact on our other variables
9 of 92
What three criteria's does an Experiment have to satisfy in order for it to be a TRUE EXPERIMENTAL DESIGN?
- Experimental manipulation - Standardisation of procedures - Random allocation of participants to IV conditions
10 of 92
What is random allocaiton?
Each possible outcome has an equal chance of being picked
11 of 92
What is Quasi-Experimental?
When all 3 criteria's can't be met
12 of 92
What is a directional hypothesis also known as?
One tailed
13 of 92
What is a non-directional hypothesis also known as?
Two-tailed test
14 of 92
What is a one-tailed test?
Specifies the direction of the relationship between the variables
15 of 92
What is a two-tailed test?
No direction of the relationship or difference is predicted
16 of 92
What is 0.05? (1 in 20 chance)
The probability that something will happen by chance - statistically significatn
17 of 92
If we raised the probability to 0.1 what type of error would we find?
Type 1 Error - Find effects that are not there
18 of 92
If we lowered the level to 0.01 what type of error would we find?
Type 2 Error - We would miss genuine effects
19 of 92
What does it mean if the bars DO NOT overlap in error bars?
A real difference between the population means
20 of 92
What does it mean if the bars DO overlap in error bars?
Suggest no real difference between the population means
21 of 92
What is a WITHIN PARTICIPANT DESIGN?
If participants take part in more than one condition (related sample)
22 of 92
What is a BETWEEN PARTICIPANTS DESIGN?
If participants only take place in one condition (Independent samples)
23 of 92
What is a MIXED DESIGN?
A combination of within and between factors - More than 1 IV is needed
24 of 92
What are the 2 advantages of WITHIN PPT DESIGN?
- Fewer ppts required - greater control over confounding variables e.g. individual differences
25 of 92
What are the 2 advantages of BETWEEN PPT design?
Removes the chance of practice or fatigue effects - Harder for ppts to guess the purpose of the study
26 of 92
What are the 2 disadvantages of WITHIN PPT DESIGN?
Increased chance of practice or fatigure - incerased chance of ppt guessing the study = demand characteristics
27 of 92
What are the 2 disadvantages of BETWEEN PPT DESIGN?
More ppts required - Not much control over confounding variables - individual differences
28 of 92
What do T-test show?
If there is a significant difference between the means of 2 conditions
29 of 92
What is the independent (unrelated) t-test?
Calculates whether the mean scores from 2 different groups are significantly different
30 of 92
When does independent t-test work best?
When score is normally distributed
31 of 92
What does the independent t-test use?
Means from two different samples of ppts
32 of 92
When do you NOT use an unrelated T-test?
When you have 3 or more groups of ppt
33 of 92
What test should you use
ANOVA
34 of 92
When do you use a RELATED t-test
If you have 2 sets of scores from one group of ppts
35 of 92
What is a paired t-test?
Paired
36 of 92
What is the formula for T-test
Difference between group means (M1-M2) / Variability of groups (Measure of spread within groups) - See sheet
37 of 92
The higher the T- value....
The bigger the difference - less likely to accept null hypothesis
38 of 92
What does the T- value tell us?
About the size of the difference (magnitude)
39 of 92
What is the issue with selecting a sample from a population?
We can't be certain how representative it is of the population
40 of 92
We don't use all members of the population therefore isn't entirely accurate
41 of 92
What is sampling errors?
The degree to which the population parameter differs from the sample statistics
42 of 92
What does a larger sample mean?
Smaller sampling error
43 of 92
What does a smaller sample mean?
Larger sampling error
44 of 92
What is the null hypothesis for a t-test?
No significant difference between two means
45 of 92
If the population means are the same what is the sampling distribution
Zero
46 of 92
If there is no difference between our means what would we expect the t-value to be?
Closer to the middle region of the sampling distribution
47 of 92
What do t-tests help us calculate?
How far from zero the difference between the two obtained means is
48 of 92
The higher the t-value
The more likely it is that any difference between the group means is not the result of a sampling error
49 of 92
What is the degrees of freedom in an Independent t-test?
n-1 for both groups
50 of 92
What does a related t-test test?
Whether one mean is significantly different from another
51 of 92
What are the two sources of variation?
Variation due to the difference between the means (IV) and Variation due to uncontrolled factors (Error Variance)
52 of 92
How can we reduce Error variance?
By measuring the same thing twice as individual differences are removed
53 of 92
When do you use related t-test?
Single group of ppts that are measured twice or are matched into pairs
54 of 92
When don't you use related t-test?
When you have two sets of scores from a single group of participants on different measures
55 of 92
Give an example
If you took extraversion and intelligence
56 of 92
What is a One-Way ANOVA?
An extention of the t-test
57 of 92
When is a One-Way ANOVA used?
When we have three or more groups of participant (Still 1 IV and 1 DV)
58 of 92
How many levels does the IV have?
2 (groups of participants)
59 of 92
What does an ANOVA do?
Comapres the between group variance with the within groups variance
60 of 92
What is the F-ratio
Between groups variance (variance due to IV) / Within groups variance (Error variance)
61 of 92
Within groups...
Variance is minimal
62 of 92
Between groups...
Variance is larger
63 of 92
The larger the between groups variance in relation to the within groups variance...
The larger the F ratio
64 of 92
What does it mean if the F ratio is small?
The within-groups variance is larger than between groups
65 of 92
What does it mean if the F ratio is large?
The within-groups variance is smaller than between groups
66 of 92
What does the null hypothesis state?
That there is no difference among the means (between groups)
67 of 92
What will the F value be if the null hypothesis is true?
1
68 of 92
If the between groups variance is larger than the within groups variance what does this mean?
The IV has an effect therefore the ratio will be greater than 1
69 of 92
What is the significance level?
0.05
70 of 92
What is a Type 1 error?
1 in 0
71 of 92
What happens if you increase the number of tests on the same data?
The chance of finding a significant result by chance is increased
72 of 92
When can't you use a one way anova?
When there is more than 1 IV
73 of 92
What ANOVA do you use instead?
A factorial ANOVA
74 of 92
IVS = Factors
Conditions = LEVELS
75 of 92
What are the four sources of variance
1. Main effect of factor 1 2. Main effect of factor 2 3. The interaction between Factor 1 and Factor 2 4. Error variance
76 of 92
What is the main effect of factor 1/2?
The effect that one factor has on the DV, irrespective of the other factor
77 of 92
What is the interaction between factor 1 and 2
The ways in which factor 1 interacts with factor 2 - behaves differently
78 of 92
What are the three significant test?
3 Hypothesis = 2 main effects and 1 interaction
79 of 92
How does two way anova interpretation begin?
With examination of the interaction
80 of 92
ANOVA terminology
Check revision sheet
81 of 92
What does the number of numbers show?
The number of factors (IVs)
82 of 92
What do the numbers themselves represent?
The levels in each of the factors
83 of 92
When is anova USED?
When we want to analyse all the possible sources of variance in a study
84 of 92
Parallel line in a graph shows
No interactio
85 of 92
No parallel lines on a graph
Possible interaction
86 of 92
Lines cross on a group
Interection
87 of 92
When is a Post Hoc used?
After insepcting the ANOVA results
88 of 92
What does a Post Hoc do?
Compare every mean with every other mean
89 of 92
What does every significant outcome mean?
The means are different and tells us where the differences in the ANOVA actually are
90 of 92
When do you perform a post hoc
If results of your ANOVA were significant and if you have more than 2 levels of factors
91 of 92
What does TUKEY'S HONESTLY SIGNIFICANCE DIFFERENCE DO?
Controls the type 1 error rate well
92 of 92

## Other cards in this set

### Card 2

#### Front

When will the research design change?

#### Back

If you are finding a difference or a relationship

### Card 3

#### Front

What is correlational design?

### Card 4

#### Front

Which correlational design is used if data is normally distributed?

### Card 5

#### Front

Which correlational design is used when data is NOT normally distributed?