Research Methods - Lecture 3: Introduction to analysing data in psychology – Inferential Statistics

Question 1

How can you determine the strength of the evidence?

Answer 1

The strength of the evidence depends on how likely it is that we might have got the same result just by chance

Question 2

What is a significance level?

Answer 2

The criterion probability, called a significance level, that we normally use is 0.05, or 5%, or 1/20

Question 3

How can we decide if something has a real effect?

Answer 3

If the probability of our data occurring just by chance is less than 1 in 20, we’ll believe that it’s a real effect

Question 4

2 rival hypotheses about the data

Answer 4

• H0 – the null hypothesis
• There is no real effect, results obtained just by chance or sampling error
• H1 – the experimental or alternative hypothesis
• There is a real effect, results not obtained just by chance or sampling error

Question 5

Calculations for finding out if there is a real effect

Answer 5

• Work out the probability of our results if H0 were true
• If that probability is less than the significance level (usually .05) reject the null hypothesis and conclude that there is a real effect

Question 6

We have two independent samples of scores and we want to know if they’re significantly different

Answer 6

• We’re really asking – how likely is it that these two samples were drawn from the same population, and the apparent difference between them is just sampling error?
• If that’s not likely, then they come from different populations and represent a real difference
• So we need a test for the between-groups design…

Question 7

Frequency distribution

Answer 7

If we draw a graph of the number of times each score occurs in the sample (frequency) we get a picture of the sample as a whole

Question 8

How can frequency distributions differ?

Answer 8

In terms of their location or average value and variability

Question 9

Bell curve shape of frequency distribution

Answer 9

Most ppl score in the middle and very few on the extremes either side

Question 10

What needs to be considered when deciding if there is a statistical difference?

Answer 10

Location and spread (average and variability) of frequency distribution

Question 11

How do we measure location or average?

Answer 11

• Usually by calculating the mean
• Add up all the scores and divide by the number of scores
• Equation for mean 𝑋ത 𝑋ത = (∑ 𝑋)/𝑁
• X means “score”
• Ʃ means “add these things up”
• N means “number of scores

Question 12

How do we measure variability or spread?

Answer 12

• Usually by calculating the standard deviation
• Subtract the mean from each score, square the differences, add them up, divide by N – 1, and take the square root
𝑠 = (square root of): ∑(𝑋 − 𝑋ത)^2/(𝑁 − 1)

Question 13

So whether two samples are significantly different or not depends on two things:

Answer 13

• How far apart are their means?

* How variable are the samples?

Question 14

Bigger difference between means…

Answer 14

is more likely to be significant

Question 15

More variability implies…

Answer 15

less likely to be significant

Question 16

If samples are quite variable…

Answer 16

you’re going to need a big difference between their means to be significant

Question 17

A test for the between-groups design

Answer 17

• It’s called Student’s t and it compares the difference between means with the variability or standard error:
𝑡 = (Mean 1 - Mean 2)/standard error

Question 18

What is standard error?

Answer 18

Combined calculation of standard deviations

Question 19

What does 𝑡 value give?

Answer 19

Information on how far samples are relative to how spread out they are