Notesale: Turn your study into money

Document Preview

Extracts from the notes are below, to see the PDF you'll receive please use the links above

---

 
 
 
 
 
 
 
 

PSYC134/PSYC135 
RESEARCH METHODS AND 
STATISTICS 
Revision Summary Notes 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

DESCRIPTIVE STATISTICS 
 
Variables: 
An Independent variable (IV) is manipulated, changed or controlled in a 
scientific test
...
  
 
Data Levels: 
When designing research and considering the variables, the measurement 
of the variables is crucially important
...
g
...
 
● Ordinal (Ordered): Data can be placed in order but the intervals 
between the data measurements are not equal e
...
shoe size (size 6 
is not twice as big as size 3), army ranks (a colonel is not twice as 
important as a captain)
...
g
...
g
...
 
 
Design: 
● Between (also called unrelated or independent): Each participant 
contributes a score in only one of the conditions e
...
you test male 
reaction times and female reaction times
...
g
...
 
 
Descriptive statistics are used to ​describe data​
...
Descriptive 
statistics may help to identify extreme scores or non-normal distribution of 
data
...
It is the only measure of central tendency that 
applies to nominal (categorical) data as well as numerical 
data
...
With an odd number of scores the 
central number is the median (e
...
if you have 11 values 6 is the 
exact middle, as it has five figures on either side
...
g
...
Unlike the mean, extreme scores do not affect it
...
The mean is by far the most useful measure of central 
tendency and almost all inferential statistics are based on it
...
 
○ 5% trimmed mean: useful when data are unduly influenced by 
outliers (extreme scores)
...
The 
mean is then recalculated using the remaining 90% of data
...
It does not take into 
account all the scores, only the extremes, thus a lot of 
information is ignored, and it is heavily influenced by extreme 
scores
...
Thus, the bigger the 
variance the more spread there is in the scores
...
It is simply the square root of the variance and has 
the advantage therefore that it is measured in the same units 
as the original data, as opposed to the square of the 
units
...
However, the 
SD is also useful as approximately two-thirds of data points will 
lie within 1 SD of the mean
...
We will return to this when we discuss data 
distribution in more detail
...
e
...
05 - sample provides enough evidence that you can reject 
the null hypothesis for the population - statistically significant 
- Alpha level - the level at which we accept a result to be significant 
- Read test statistic from SPSS, also report degrees of freedom 
 
 
“A 2x2 chi square test was conducted to assess the association between 
variable 1 and variable 2
...
52, p=
...
” 
 
However, need to report effect size, as need to know how strong the 
association is  
Phi for 2x2 chi square and Cramer’s V for other chi square designs 
 
Effect size
...
3 medium,
...
There was a statistically significant 
association between variable 1 and variable 2, χ
​ 2​ ​ (1, N=60)=8
...
004, 
φ=
...
” 
 

 
If alternative hypothesis is supported, null can be rejected  
 
 
Problem with p values - type 1 and 2 errors 
Type 1 - incorrectly rejects the null - saying there is a difference when there 
isn’t one 
Type 2 - incorrectly accepts the null - saying there isn’t a difference when 
there is one 
 
Overcome type 1 and 2 errors by changing alpha level  
- Divide alpha level by the amount of tests going to conduct and 
make that the significance level - known as the Bonferroni 
correction 
 
 
P-hacking 
- Method of manipulating data to achieve significant results 
 
Preregistration 
- When researchers predefine what analysis they’re going to do before 
conducting the study 
 
 
ETHICS IN PSYCHOLOGY 
 
Ethics: 
1
...
Ethics in psychology - deception, harm, consent, etc
...
 
● 2 principles: 
○ Right or wrong depends only on the results of the act 
○ The more good consequences, the better/more right the act
...
What people do 
is important, not the consequences
...
 
○ Moral rules 
○ Immanuel Kant 
■ While trekking in the Andes you come across a guerilla 
leader who has captured 5 local villagers
...
If you refuse to shoot, he will kill all 5
...
 
 
Ethics in Psychology 
● Ethical approval: 
○ Confidentiality and anonymity 
○ Deception 
○ Protection from harm - both psychological and physical 
○ Informed consent 
○ Right to withdraw and non-participation 
● Confidentiality and anonymity: 
○ If this cannot be guaranteed the participants must be 
informed 

● Deception: 
○ Not without strong scientific/medical justification 
○ Or if likely to object or show unease once debriefed 
○ Sufficient information at the earliest stage
...
 
○ Exceptions:  
■ Observations in a public place 
■ Information in the public domain, e
...
internet 
● Deception: 
○ Debrief as soon as possible 
● Debriefing: 
○ purpose/results 
○ deception/discomfort 
○ Procedural problems 
● Withdrawal and non-participation: 
○ Participants can ask to withdraw their data 
○ Withdraw retrospectively 
○ Data destroyed 
○ Gowers et al, (2007) 
■ Property of experiment? 
■ Self selecting sample? 
 
Internet research 
● Informed consent: 
○ children/vulnerable adults? 
○ Participant identity? 
○ Chat rooms? 
● Public or private? 
○ Pittenger (2003) 
■ Widespread - anonymous - open/public forums 

■ confidential/limited access forums 
● Deception and debriefing: 
○ Public: debriefing not required 
● If deception is involved, participants must be debriefed - 
electronic/telephone debrief? 
 
Controversial Experiments 
● Zimbardo - Stanford prison experiment 
● Milgram (1963) - obedience to authority 
● Rosenhan (1973) - to challenge validity of psychiatric diagnoses 
● Watson (1920) - Little Albert - conditioning a phobia 
 
Animal research 
● 26 million animals each year in US 
● Largely concerned with rats and mice
...
 
○ Excluded from US animal welfare acts 
● Arguments for: 
○ Finding new drugs and treatments 
○ Improving human health 
○ Ensures safety of drugs 
○ Similarity to humans - can generalise 
● Arguments against: 
○ Cruel and inhumane 
○ Drugs not necessarily safe 
○ Expensive 
○ Animals and humans are never exactly the same 
● For or against? 
○ Simon Chaitowitz - animal rights activist 
■ ‘Throughout the past six years, I have felt terribly guilty 
about the drugs and procedures I've undergone because 
I know that so many animals have suffered in their 
development’
...
g
...
36, ​p < ​
...
​” 
​
 
● Type of correlation 
● Was it significant? 
● Positive or negative figure? 
● Variables 
● Inferential stats in APA format 
 
0
...
3 = medium, 0
...
05, so p values below
...
025, so e
...
p value
...
004 
 
Pearson’s correlation 
 
“A Pearson’s correlation analysis showed there was a significant positive 
correlation between variable 1 and variable 2, r(48) =
...
001
...
1 r = small effect size, 0
...
5 r = large effect size 
 
Visualising correlations - scatter plots 
● Makes it easy to see strength of correlations 
● Should show line of best fit, which represents pattern of data 
● Correlational analysis essentially test how well this line fits the data 
 
Visualising correlations - correlation matrix 
● If you have lots of variables and you correlate them you could 
present this info in a table known as a correlation matrix 
● Should be in APA format 
 
SUMMARY 
★ Correlational analyses are used to assess relationships (associations) 
between variables 
★ Correlations can be positive or negative 

★ Using these analyses a correlation coefficient is calculated telling us 
the strength of the association and the direction 
★ Correlation does not always imply causation 
★ Spearman’s correlation is used when data is ordinal level or is 
skewed 
★ Pearson’s is used for interval or ratio data that is normally distributed 
 
 
 
SURVEY METHODS 
 
Designing a questionnaire 
● Topic 
● Type 
● Draft 
● re-examine/revise 
● Pilot study 
● Edit and procedures for use 
 
Questionnaire items 
● Don’t use 
○ Double barrelled questions 
○ Loaded questions 
○ Negative wording 
● Use of 
○ Open-ended 
■ Advantages: detailed answers, quick to design, 
participant led 
■ Disadvantage: long time to analyse, subjective 
interpretation 
○ Partially open-ended 
■ MCQ plus others… 
○ Closed-ended 
■ true/false 
■ MCQ 
■ Likert scale 
■ Disadvantages: guessing (50:50 guess), unsubtle, 
complex to design, theory led 
 
Questionnaire fallacy 

● Mistaken belief that a questionnaire provides a true picture of what 
people do and think 
 
Questionnaire bias 
● Order effects - question can affect responses 
● Priming - thinking about the answer to one question whilst 
answering the subsequent question 
● Overcoming order effects/priming 
○ Counter-balance question 
○ 2 versions 
■ One order then another - compare results - split-half 
analysis 
○ Randomise 
● Demand characteristics - ppts believe they know what the 
experiment is about - change behaviour, can be overcome by using 
disguised questionnaires 
● Social desirability - ppts try to make themselves look as good as 
possible 
● Acquiescence - balance of negative and position questions needed 
● Cultural bias - language that could be misunderstood, various 
interpretations of words, etc
...
unobtrusive - generally, data which exist before research 
begins, e
...
physical traces, archival data 
2
...
g
...
Focus groups 
b
...
g
...
g
...
g
...
g
...
g
...
 
 
Preparing for qual data analysis 
● Analysis process depends on the specific methodology you are using 
● Generally, all involve coding and theme development 
● Focus on depth, detail, context, complexity 
 
Reflexivity 
● About acknowledging how aspects of our identities, histories etc
...
(1996) 
 
● Facilitated by researcher 
● Getting people to think about, discuss/debate an issue (or set of 
issues) related to the research question 
● Key is group interaction 
 
Why/when use them? 
● To access attitudes, feeling, beliefs and experiences of a group rather 
than just individuals (interviews) 
● To generate discussion/debate about your research question/area of 
research 
● When people may not feel comfortable talking in individual 
interviews 
● Epistemology 
○ Interpretivism (understanding) not positivism (explanation) 
 
Type of research questions 
 
Benefits of focus groups 
● Allows for interaction and discussion 
● Ideas can be shared 
● Provides window into culture 
● Allows for people to change their mind 
● Many forms of communication; jokes, anecdotes, teasing, arguing 
● Conflicting data 
● Allows ppts to interpret each others’ responses 
● Provides large amount of info in short space of time 
● Can be empowering to ppts 
● Facilitates discussion of taboo topics 
 
Limitations of focus groups 
● Research has less control over data produced 
● Time-consuming 
● Less confidentiality 
● Determining conflicts in opinion 
● Power dynamics 
● Social desirability bias 

● Maybe less detailed or in-depth than some interviews 
● Particularly difficult to arrange 
● Generalisability and reliability 
 
Alternatively, can conduct a focus group study 
● 6-10 per group 
● Sample: heterogenous or homogenous? 
● Incentives 
● Advertising 
● Semi-structured 
● 1-2 hours 
● Recording equipment 
● Moderator 
● Observer 
 
Photovoice 
 
= (definitions) 
 
● Encouraging individuals to reflect about their community 
 
1
...
Taking the photographs 
3
...
Second focus group discussion 
5
...
Disseminate the findings (photo-exhibition) 
 
Why issues/phenomena have been explored through photovoice? 
 
Photovoice methods: impact 
● Last phase: dissemination of the findings and advocating for change 
● Photographs and accompanying captions presented through an 
exhibition used as tool to bring ppts’ voices to the attention of key 
stakeholders, raise their awareness about the issues identified, so 
that they can stimulate change 
 
Benefits of using photovoice 
● Encouraged engagement between ppts and stakeholders 
● Invited ppts to successfully engage with the photographs, with the 
collective discussions offering an opportunity to critically think as a 
group about strengths and issues of their community 

● Provide deeper level of knowledge - combination of different 
methods allow us to gain greater comprehensiveness of 
understanding of the topic from the ppts’ POV 
● Creative tool to reach key stakeholders 
 
Challenges of using photovoice methods 
● Ethical aspects related to taking photographs 
○ Important to ask written permission before taking photos of 
individuals 
○ People may feel protective of their communities 
○ Issues of photo ownership 
● Factors that can prevent ppts taking photographs that they want to 
take 
○ Photographing negative social concepts 
● Time frame for taking photographs 
 
Ethnography 
 
= involves ppt observation with researcher participating/observing the 
daily lives and events being studied over extended period of time 
● Other characteristics 
○ Using multiple methods of data collection 
○ Placing emphasis on context 
○ Focusing on what people do as well as what they say they do 
● Approach is underlined by assumption that accessing beliefs and 
behaviours in the context in which they occur will aid understanding 
and provide a holistic perspective 
 
When use ethnography? 
● Useful when developing understanding of complex relationships 
from multiple perspectives 
 
Method 
● Aims, not hyp - flexible design 
● Reliant on  
○ Gatekeepers 
○ Snowballing 
○ Networking 
 
Multi-sited ethnography 
● The ‘object’ being followed can consist of 
○ The people 

○ An artefact 
○ A metaphor 
○ A story 
○ A biography 
○ A conflict 
● It is this conceptual link that differentiates MSE from comparative 
studies 
 
Types of data collection 
● observation/field notes 
● Photos 
● Docs 
● Informal interviews 
● Formal interviews 
 
Benefits of ethnography 
● Possible to see interlinking between diff aspects of the culture under 
study; looking at the wider processes, relationships, connections of 
the actors in the setting rather than looking at one person’s POV in 
isolation 
● Researcher can establish rapport and trust with the ppts, which in 
turn can facilitate honest, uninhibited discussion 
 
Challenges 
● Time-consuming 
○ Observations 
○ Field notes 
○ Quantity of data 
● More subjective than other qual methods 
● Greater potential to invite ethical issues associated with invasion of 
privacy and informed consent 
● Depicts the people it studies as being in a void, uninfluenced by time 
and broader social constructs 
 
Types of qual analysis 
● Thematic analysis 
● Grounded theory 
● IPA 
● Framework analysis 
● Narrative analysis 
● Content analysis 
● Discourse analysis 

● Conversation analysis 
 
 
 
 
T-TESTS 
 
Stroop test (1935) 
● Demonstrates cognitive interferences 
● Reaction time delayed due to mismatching stimuli 
● Quicker when there’s congruence than incongruence 
 
T-tests 
● Two types - used for within subjects and between subjects designs 
● Within subjects - paired/related/dependent 
● Between subjects - unpaired/unrelated/independent 
● Used when we have two diff conditions or groups 
● Parametric test 
● Data must be interval or ratio and normally distributed 
● Between-subject t-tests have an additional assumption: 
homogeneity of variance 
 
Effect sizes 
● ‘Cohens D’ 
● D value 0
...
5 medium and 0
...
There 
was a significant difference between congruent and incongruent 
conditions, t (5)= 1
...
02, d
​ =
​ 0
...
17±4
...
83±7
...
 
 
Direction of effect is stated - gained by looking at means of each condition 
 
Stroop experiment 2 
● There may be practice effects, so can do between-subjects design 
● One group completes congruent task, another group completes 
incongruent task 
 
Independent samples t-test 

● Homogeneity of variance - each of our conditions should have 
similar variances 
● Error bars = SDs, therefore if they’re similar, there is homogeneity of 
variance 
● If not, heterogeneity of variance 
● Use stats test ‘Levene’s test’ to know if we have homogeneity of 
variance - tests null hyp that the population variances are equal 
● If Levene’s test is significant → no homogeneity of variance → 
assumption is violated 
● Leads to increased likelihood we will find significant effect when 
there isn’t one (Type I error) 

 
Welch Test 
● Used when you do not have homogeneity of variances when ‘equal 
variances are not assumed’ 
 
Independent samples t-test write-up 
 
An independent samples t-test was conducted to assess the difference 
in reaction time between congruent and incongruent conditions
...
88, p
​ ​=
...
92, with those within the congruent 
condition having significantly lower reaction times (17
...
80) than 
those within the incongruent condition (20
...
10)
...
But it is based on 
ranking the data rather than just the sign of the difference
...
etc
...
of values 
Effect sizes 
● P values only tell us if effect is 
significant 
● Tells us nothing about the size of difference 
● Wilcoxon - calc in Excel 
● r=
...
30 medium, r=
...
64, p
​ ​<
...
47, with 
improved performance when mnemonics were used (​Mdn​= 8, range=6) 
than when they were not used (​Mdn​=5, range=6)
...
50​,​ p
​ ​=
...
43, with cats (​Mdn​= 7, range​= 
18) performing better than the Trump administration (​Mdn​= 5, 
range=6)
...
1 small, 0
...
8 large 
Write-up:​ There was a significant effect of drink type on subjective 
intoxication, X​2​(2)= 60​
...
001, ​W=
​ 1
...
 
 
Omnibus tests 
● Friedman and Kruskall-Wallis are both Omnibus tests - tests which 
look to see if there are differences between conditions overall (i
...
 
between more than 2 groups)
...
00, p
​ ​<
...
00
...
001) more intoxicated following alcohol (Mdn= 8, 
range=6) relative to placebo (Mdn=3, range=4)
...
001) following 
alcohol relative to the control drink (Mdn= 0, range= 3)
...
001) higher levels of intoxication 
following placebo relative to the control drink
...
01 /31= 0
...
33 
 
Eta squared 
● tells us something very specific about our results 
● Tells us how much variance (in the form of a percentage) in our 
results is accounted for by out IV 
● Suggested cut offs for eta squared:
...
06 medium,
...
16, p
​ =
​
...
33
...
002​),​moderate drinkers (​p​=
...
036)​
...
48), light drinkers and non-drinkers 
(​p=
​
...
21)
...
 
 
 
 

 
SINGLE CASE RESEARCH 
 
Overview 
● Nomothetic vs idiographic 
● Case studies 
● Single-case experiments 
○ Experimental designs 
 
Nomothetic research 
● Groups 
● Universal laws 
● Predict average behaviour 
● Developmental trajectories 
Idiographic research 
● The individual 
● Unique 
● Rehabilitation 
● Risk assessment 
Conflict 
● Mental health conditions 
● Cannabis and treatment of pain 
 
“If you want to know something about a person, why not first ask him?” - 
Allport (1953) 
The clinical’s goal “
...
” 
 
Single case research 
● Developing broad psychological theories 
● E
...
Amnesia 
● General understanding of hippocampus function 
● Diff parts of the brain - diff kinds of memories 
 
Amnesia 
● Diff kinds of memory: STM, LTM, procedural, episodic, semantic 
● Two main types of amnesia:  
○ Anterograde 
■ Impairment of memory after trauma 
■ Only LTM affected 
○ Retrograde 

■ Impairment of memory prior to trauma 
■ Episodic memory affected 
 
HM - case study 
● Anterograde amnesia 
○ Intact STM 
○ LTM impairment - no new memories 
○ Memory prior to surgery normal 
○ Retained old skills and acquired new skills - but is unaware he 
has them 
 
Clive Wearing 
● musician/conductor 
● Total amnesia - anterograde and retrograde  
● Some memories from before illness 
● Unable to form new memories 
● Memory only lasts 7-30 seconds, music ability (procedural memory) 
remains intact 
 
Single case research 
● Case studies: e
...
Freud 
● Single case experiments: e
...
Skinner 
● Case studies: 
○ Intensive study of a unit/system 
○ Richly detailed info from variety of sources 
○ Exploratory 
○ rare/unusual conditions 
 
Freud - case studies - interpretation of dreams 
 
Rat man 
● Obsessive thoughts 
○ Cutting his own throat 
○ Father and fiancee tortured by rats 
● Sexual experiences in infancy 
 
Dora 
● anxiety/hysteria 
● Repressed desire 
● Manifestation of jealousy toward the relationship between Frau K 
and her father 
 

Wolf man 
● Severe depression 
● Nightmare about white wolves 
● Caused by seeing parents having sex 
 
Advantages of case studies 
● Good source of ideas for research 
● Opportunity for innovation especially in clinical samples 
● Rare phenomena 
○ E
...
Luria (1968) - the mind of a mnemonist 
 
Issues with case studies 
● Objectivity - confirmation bias 
● Cause and effect 
○ covariation/correlation 
○ Time-order relationship 
○ Eliminate possible alternative causes 
● Generalisation 
○ Representativeness? 
○ Is the phenomena the same in other people? 
 
Single case experiments 
● Skinner - School of experimental psychological research 
● Behaviourist - operant conditioning 
● Behaviourism evolved from single case experimental research 
designs 
● Training animals using operant reinforcement 
 
Applied behaviour analysis 
● ABAB design (reversal): 
○ Baseline - intervention - baseline - intervention 
○ Clear change when intervention is added or removed = 
effective treatment 
○ No ethical or practical issues 
○ Continuous assessment 
○ Baseline assessment - standard for improvement 
○ Stability of performance/behaviour - evaluation of change from 
baseline 
○ Different phases 
■ High internal validity 
■ Testing diff interventions 
○ Data evaluation 

■ Mean - compare mean values across consecutive 
observation points which constitute a ‘phase’ 
■ Level analysis - diff between last observation in one 
phase and first observation in the next 
■ Trend analysis - rate at which changes appear within the 
diff phases 
■ Latency - how quickly changes between diff phases 
become apparent 
● Some interventions can’t be reversed: e
...
practical or ethical reasons 
 
 
 
 
 
 



---