StatisticsStatistics

Hypothesis TestingHypothesis TestingHypothesis is a ‘testable statement’Types = alternate, research,

experimental (H1), null (H0)They are 1 or 2 tailed (directional or non

directional (but not the Null)They include an IV and a DVThey are operationalised (precisely

defined in terms of how the IV and DV will be manipulated or measured)

Aim of research is to accept/reject H1 or H0

Complete the exercise

Inferential StatisticsInferential Statistics1. Analyse using descriptive statistics (tells

you whether a difference exists or not)2. Analyse using inferential statistics (tells you

whether the differences are significant)3. If the sample has yielded significant results,

we can infer the same is true of the population

4. The statistical tests stringently process the data to tell you whether or not chance has caused the outcome

5. Therefore part of the whole process involves having a null hypothesis (HO) and levels of significance or probability e.g. 5%

‘‘Proof’Proof’In science it is only possible to

prove something is not the casee.g. ”all swans are white”Can’t be provedBut can be disproved – HOW?

Significance levelsSignificance levels

Likelihood or probabilityExpressed as a % and as a

decimale.g. heads or tails 50% or 0.5Picking the ace of hearts from 4

aces 25% or 0.25Likelihood of having schizophrenia

1% or 0.01

Ruling out ChanceRuling out Chance

• Standard level of significance in Psychology =

5% (0.05)• Accept H1 – then p < 0.05• Accept H0 – then p > 0.05• BUT a significant result might still

be wrong 5 times in 100 – in other words it happened due to chance – we live with this risk

Type 1 and type 2 errorsType 1 and type 2 errors

Accept H0 Reject H0H0 is actually true

OK Type 1You conclude there is an effect when there isn’t

H0 is actually false

Type 2You conclude there isn’t an effect when there is

OK

Type 1 and type 2Type 1 and type 2Type 1 is more likely when we

have a high significance level e.g. 10%

Type 2 is more likely when we have a low significance level e.g. 1%

At 5% both are equally likely

What you need to be able What you need to be able to doto doIdentify an appropriate statistical

testExplain your choiceState a conclusion based on a

stats testWrite a null hypothesisExplain why a particular stats

test was used

Descriptive StatisticsDescriptive StatisticsDescriptive statistics give us a

way to summarise and describe our data but do not allow us to make a conclusion related to our hypothesis. For example, measure of central tendency such as ____________, ___________ or ____________. It also includes graphs and charts, and measures of dispersion such as _________________ or ______________ _____________.

TasksTasksMeasure of central tendency

◦Match the definitions to the terms, and the strengths and weaknesses.

Measure of dispersion◦What is a range?◦Read about standard deviation

What are bar charts and scatter graphs? Define and draw an example.

Complete the memory experiment tasks

What is the point?What is the point?Why do we bother to use

inferential statistical tests?◦Inferential statistics allow us to draw

conclusions from findings.◦They allow us to see whether our

results are the result of something happening, or are just down to chance.

Cross out the words on the sheet and fill in the table

Pg 20Pg 20Read the example about the chip

bins and female drivers.Read “Using Statistical Tests” on

pg 22-23

Answer the questions on the sheet

Levels of significanceLevels of significanceThis refers to the minimum probability

we will accept that our results are due to chance. If it is too lenient, then our results may appear to be significant when in fact they are not. If it is too stringent, then our results may appear to be insignificant when they actually they are.

In psychology, we generally aim for a significance level of _______%. This means that we can be _________% certain that our results are not due to chance.

This is written as P≤_______

ExampleExampleRead the example about biscuits

On the sheet, cross out the right words and fill in the table

Levels of measurementLevels of measurementNominalOrdinalIntervalRatio

NOIRRead the descriptions on the

sheet give your own examples

Choosing the right testChoosing the right testDesign Nominal data Ordinal data Interval data

Repeated measures

Sign test Wilcoxen signed ranks

Related t test*

Matched pairs

Sign test Wilcoxen Related t test*

Independent measures

Chi square Mann Whitney U

Unrelated t test*

Correlation Chi square Spearmans rho

Pearsons product moment*

* Refers to Parametric tests (see handout)