Research methods 2

What is the baseline significance level in psychology research and why?

0.05- balance between making a type 1 and type 2 error

What does a higher range indicate?

Less consistency

What does p ≤ 0.05 mean?

Probability that the results are due to chance is less than or equal to 5

What are stringent levels?

Highly significant e.g. p ≤ 0.01

If we need to be more confident than accepting a 5% probability that the results are due to chance

What are non-stringent levels?

When are they used?

Low significance e.g. p ≤ 0.1

If we need to demonstrate an effect/difference/relationship

What is a type 1 error?

False positive

Wrongly accepting experimental, when null should be accepted

Happens with lenient significance levels

What is a type 2 error?

How to reduce it happening?

False negative

Wrongly accepting null, experimental should be accepted

Happens with stringent significant levels

2) use 5% level, use less stringent levels or increase sample size

What to do if it is unclear whether data is ordinal or interval?

Regard as ordinal

How to tell whether the result is significant based off critical and calculated values?

If there is an R in the test, critical has to be higher than calculated for the result to be significant

What is the critical value and what is the calculated value?

Critical- value from statistical test table (have to find from table)

Calculated- value produced after statistical test done (have to work out)

How to calculate the calculated value in sign test?

Sign that occurs less frequently

How to calculate df for chi squared?

(no. of rows-1) x (no. of columns-1)

What is content analysis?

What is the purpose of content analysis?

A technique for analysing data according to themes

Make analysis more objective and identify trends

How to carry out content analysis?

1) Researchers come up with aim/hypothesis to investigate

2) Select sample

3) Create coding units

4) Carefully go thru sample + record any instances of the coding unit that occurs

5) Produces quantitative data

Strengths of content analysis? (3)

No ethical issues- no consent needed as content is published

Flexible method- used to suit any aim

Statistical analysis is possible- identify patterns in data- easier to compare findings from similar studies- acts as reliability check- scientific technique

Weaknesses of content analysis?

Reductionist- detailed, in depth sample reduced to numbers- subtle message in commentary is lost and loses its richness and complexity as it is taken out of context

Subjective- categorising and coding is biased as it is based on what the researchers believe is important- not valid or trustworthy

Key features of thematic analysis?

Coding + identifying patterns and themes

Specific e.g. quotes to support or challenge themes

Original qualitative form

How to carry out thematic analysis?

1) Familiarise with data + transcribe

2) Coding to initially analyse transcript

3) Look for emerging themes

4) Define and name each separate themes

5) Write up report

Strength of thematic analysis?

Less reductionist than content analysis- data still in depth + contains original message

Weaknesses of thematic analysis?

No statistical comparison- reliability harder to establish

More subjective- bias from the researchers gender, culture, background are highly likely when interpretations are made

How to define interval data?

The units are of fixed intervals/differences between the scores are fixed

Why are correlations a useful preliminary tool for further research?

Relatively quick + economical to conduct- uses forms of established (secondary) data to assess patterns of variables before researcher commits to more lengthy and time consuming research methods

Forms basis of starting point for further experimental research

Why do correlations allow studies of variables which can’t be manipulated in experiments?

Don’t require manipulation of behaviour and are used when it may be unethical/impractical to manipulate variables artificially in experiments

Doesn’t break ethical guidelines

Weakness of correlations?

Variables aren’t being manipulated so we can’t state whether one variable has caused the effect on the other variable as there could be other extraneous variables that are having an effect

Can’t establish cause and effect

What does a negative skew look like and what does it mean?

Long tail end at left side (negative = left)

Mean at left side of median and mode

Large amount of data about mean score i.e very easy test so very high scores

What does a positive skew look like and what does it mean?

Long tail end at right side (positive = right)

Mean at right side of median and mode

Large amount of data below mean score i.e. very hard test so very low scores

When are bar charts used?

With discreet categorical data e.g. hair colour

When are histograms used?

Used to illustrate distribution/frequency of data items

With continuous data e.g. weight

When are scattergrams used?

To show a relationship between 2 variables i.e. correlations

Strength and weakness of mean?

Strength- most sensitive measure of CT- - includes all values in data set- most representative

Weakness- easily distorted by extreme values- could be unrepresentative

Strength and weakness of median?

Strength- not affected by outliers- useful when mean isn’t appropriate + easier to calculate

Weakness- not as sensitive as mean- doesn’t include all data values

Strength and weakness of mode?

Strength- easiest measure to calculate + unaffected by outliers

Weakness- crude measure- unrepresentative in small data sets- less useful when several modes

Strength and weakness of range?

Strength- easiest measure of dispersion

Weakness- only takes two most extreme values into account- unrepresentative

Strength and weakness of SD

Strength- sensitive and precise measure of dispersion- all values taken into account for calculation

Weakness- doesn’t say full range of data + can be affected by outliers

What does a large SD indicate?

A large variation from the mean- values are spread out around mean score

Data points out over wider range of scores

What does a small SD indicate?

Not a lot of variation from the mean- values are clustered together around the mean

Data points close to mean of the set

What is primary data?

Data which is collected by the researcher first hand and is gathered directly from the p’s themselves for the purposes of the research and to test the hypothesis put forward by the researcher

Strength of primary data?

Authentic- collected first hand from the p’s themselves- specifically targeted to meet the researcher’s needs + focused on purpose of research

Weakness of primary data?

Time consuming to collect- investigations require planning and preparation

What is secondary data?

Data which has been collected from p’s by other ppl (not the researcher themselves) and already exists

Researcher makes use of this data for their own research purposes and to test their hypothesis

Strength of secondary data?

Easily accessible- requires minimal effort to collect as it already exists

Weakness of secondary data?

Content may not meet the researcher’s needs and it could be lacking in valuable info. or be outdated

What is a meta analysis?

Researcher combines findings from a number of previously published studies, all dealing with the same research question, and provides a statistic to represent an average and the common overall effect

Strength of meta analysis?

Greater statistical power- more ability to generalise findings to wider population

Evidence based

Weakness of meta analysis?

Difficult + time consuming in searching for appropriate studies to examine

Complex statistical skills and techniques required

Strength of quantitative data?

Objective- free from bias

Capable of being analysed statistically- allowing comparisons to be made

Weakness of quantitative data?

Fails to consider participant emotion or feeling

Lacks insight into reasons behind human behaviour

Strength of qualitative data?

Allows respondents to develop thoughts, feelings and emotions

Gives meaningful insight- high in external validity

Weakness of qualitative data?

Difficult to analyse statistically- comparisons difficult to make

Conclusions based on subjective interpretations

How use of inferential stats would improve investigation

• allow the researcher to establish whether the relationship is significant/due to chance

• allows the (alternative) hypothesis to be accepted/null to be rejected

• can more accurately show the strength of the relationship

• to increase the scientific credibility/validity/objectivity of the research.