Research Methods: inferential Statistics

Question 1

What are the differences between descriptive and inferential stats

Answer 1

1. Descriptive
   > measures of central tendency and dispersion
   - mean
   - median
   - mode
   - range
   - standard deviation
2. Inferential
   > tells of significance
   - 8 tests

Question 2

When do we use sign tests? And how do we carry them out?

Answer 2

• when to use
- test of difference (experiment)
- related data (data comes from same source/repeated measures)
- nominal data (in categories, eg. 3 porridge and 7 toast)

• STEPS..
1. Take one column from the other (subtract)
> doesn’t matter wch but be consistent

2. Find what kind of difference,
   By counting positive and negative signs
   > eg 18 negative, 1 positive
3. Identify smaller number (eg. 1)
   So the calculated critical value in this example = 1
   > in sign tests, critical value is S (eg. S=1)
   >
   > N = number of ppl in study
   MINUS those whose scores didn’t change (eg. N=19)

Question 3

How to tell if results are significant or not with significance level

Answer 3

The significance level is 0.05, wch means result is significant
- means 95% due to significance
- means 5% probability results due to chance

..
There are two types of critical values found
calculated (eg, S=1)
table (eg, S=4)
> always comparing calculated to table value

Calculated must be EQUAL or LESS THAN to be statistically significant
Not due to chance

Question 4

What are one and two tailed hypotheses

Answer 4

One tailed = directional hypothesis

Two tailed = non directional hypothesis
> no previous research / open minded
> or previous research contradicts

Question 5

When do we accept and reject null hypotheses

Answer 5

If statistically significant,
We reject null and accept alternative hypothesis

If by chance,
We accept null and reject alternative hypothesis

Question 6

How do we know when to use/identify correct inferential statistics tests

Answer 6

diagram 1
Find with
- experiment, association or correlation?
- related/unrelated?
- level of measurement

..

Is test looking for difference (experiment),
association or relationship (correlation - one IV )?

1. If experimental study, will have key words
   - experiment
   - difference
   - cause (IV)
   - effect/affect (DV)
   - causality
2. If a test of association,
   The word association will be there
3. If correlation, will have words
   - correlation
   - relationship

Question 7

Difference between related and unrelated data

Answer 7

Unrelated data
> diff ppl in each condition/each time

Related data
> data from same ppl in all conditions (repeated measures)

Question 8

What are levels of measurements

Answer 8

Nominal
- data in categories
- eg. how many in each category? 12 dogs, 7 cats, 1 rat

Ordinal
- data put in rank order (lowest to highest)
- not equal units of measurement; subjective and opinion based
- eg. How many got each category? 1 love it, 10 hate it

Internal
- data in rank order
- with equal units of measurement; objective
- How many in each category? 1 got 1 min, 3 got 5 mins, etc

Question 9

do questions in this

Answer 9

like right now

Question 10

Example of a statement of significance

Answer 10

When N=12 and a one tailed test is used at the 0.05 level of significance,
The table value is 2.
The calculated value is 5, so is not equal/less than to
So result isn’t significant; null hypothesis is accepted

Question 11

What are errors

Answer 11

Type 1 and 2 errors refer to osassions when we have results
That are in small proportion that are due to chance
(Error = Inunderstanding of results)

• type 1 error
- when null is rejected but shdve been accepted (false positive)
- more likely to make type 1 error if use a loose
Lvl of significance (eg 0.1) as probability due to chance is higher

• type 2 error
- when null is accepted but shdve been rejected (false neg)
- more likely to happen if we use a tight lvl of significance (eg. 0.01)