Practical issues in design and implementation

AO1

• Practical issues impact on choices that researchers make when selecting measures in their research to meet the aims of a study 

• Ethical issues such as gaining fully informed consent would need to be considered against potential for causing demand characteristics that impact upon the findings (validity vs ethics) 

• Deciding on what apparatus or materials to use in a study could be an issue that depends on the time or money available to a researcher  

• Gathering a sample, such as random, volunteer or opportunity, would depend on the access a researcher has to target population 

Baddeley

Validity

• Independent groups design, no order effects, no practice/fatigue  

• Recalling word order rather than words themselves  

• 15-minute interference task before surprise recall task to ensure no STM being used 

• Low ecological validity due to artificial stimuli  

• Participant variables between groups 

Reliability

• Same list of 10 words for everyone, standardised procedure 

Generalisability

• 72 pps of mixed men and women, representative 

• Split into smaller groups, less than 20 pps in key groups  

• All participants from Applied Psychology Research Unit, and all pps volunteers – similar characteristics  

Sherif

Validity

• Field experiment so natural settings, high ecological validity 

• Range of different research methods used (observing, tape recording, tests, quantitative as well as qualitative data collected) 

• No control group to compare findings 

• Risk of researcher bias and exaggeration of findings from Sherif, subjective interpretation of observations 

• (Deceived boys to ensure high internal validity, allowed 2 boys to drop out due to homesickness caused an imbalance in number between the Eagles and Rattlers)

Reliability

• Multiple observers, inter-rater reliability 

• Used a numbered scoring system for boy’s friendship patterns – quantitative data (objective) 

• The observers could only observe the boys for 12 hours a day, not able to see or hear everything 

Generalisability

• Sherif screened the boys beforehand, removing any from troubled backgrounds or with antisocial behaviour  

• Removed confounding variables that might have affected prejudice

• Sample of 22, only boys (gender bias), 11-12 white protestant American boys (cultural bias) 

Raine

Validity

• Matched pairs design was used, matched on sex, age and ethnicity and schizophrenia (n=6)

• All participants were taken off medicine 2 weeks before their brain scan. This made sure that all NGRI’s brain activity was not altered by drugs, making the results more accurate 

• Canthomeatal line is the distance between ear to ear, and this varies between people 

• Only looks at brain activity, this could be a very reductionist view of human behaviour 

Reliability

• Each participant underwent a PET scan of their brain, injected with radioactive tracer and left for 32 minutes. Each participant completed the same continuous performance task to allow the radioactive tracer to be up taken by the brain 

• Subjectivity in interpretations of PET scans 

Generalisability

• Sample of 41 NGRIs (39m and 2f) and then a control sample – ratio of m:f was representative to wider society as men commit more violent crimes than women 

• NGRIs are not representative to wider society, cannot be generalised  (although raine used this sample for ethical reasons as they had already received the PET scan as a part of their court trial and did not have to be given it by Raine)

Watson and Rayner

Validity

• Hid the bar behind a curtain to ensure Albert didn’t associate the loud sound with the bar, and instead with only the white object/rat in front of him 

• Lacks ecological validity – findings cannot be generalised to other settings outside the laboratory situation – artificial stimuli and unnatural method (although this allowed for a more standardised procedure)

• Vagueness over what constituted a fear response in Little Albert = subjectivity

• Throughout Little Albert was tested with wooden blocks to establish whether he was becoming more fearful/less playful on the whole, or whether his response was specifically to the white rat

Reliability

• Well standardised procedure, stages planned out and sections of classical conditioning identified .Gathered data upon Albert’s baseline phobic rates through pre-conditioning, finding that he had no phobic response to a white rat, a monkey and a rabbit

• Weak reliability as the study cannot be replicated – goes against the BPS ethical guidelines 

• Albert was too young to consent, and his mother wasn’t asked to give informed consent  

Generalisability

• Poor generalisability, only one participant, white American boy. Results may have been affected by participant variables – Little Albert may have had hydrocephalus 

Rosenhan

Validity

• Good ecological validity, a natural environment – nurses didn’t know they were being observed so behaviour not affected by demand characteristics  

• Lack of internal validity as pseudo patients were faking a mental condition so doesn’t tell us about real mental health issues  

• Rosenhan took part in his own study – researcher bias 

Reliability

• All claimed to hear the same 3 words ‘empty’ ‘hollow and ‘thud’ and then as soon as they were admitted to the hospital they requested to be discharged  

• Not all followed the procedure – one had a romance with a nurse.  

• Field experiment so limited controls over experiences  

Generalisability

• Range of different psychiatric hospitals were assessed – 12 different hospitals from all over the state, well-funded, under-funded, over-staffed, under-staffed etc.  

• Small sample size of only 8 pseudopatients and only 12 hospitals. All hospitals were US and therefore culturally biased – only a reflection of western diagnostic systems  

Van Izjendoorn and Kroonenberg

Validity

• High ecological validity in the Strange Situations Test for western countries using a daycare system, as the procedure aims to replicate a nursery environment. High internal validity of this procedure as tightly documented – no extraneous variables such as other people entering the room  

• SST developed by an American psychiatrist Ainsworth, therefore the procedure could be viewed as having Eurocentric bias – imposed etic/cultural bias – conclusions about attachment styles in other cultures may be invalid  

Reliability

• Used the SSP which is a standardised procedure – 3-minute sampling and intervals that can be replicated easily  

• Excluded individuals with disabilities for example Down’s Syndrome for standardisation – ensure that these do not act as confounding variables 

• Meta-analysis therefore reliant upon the methodology of others, unsure of standardisation within the procedures of the studies and VI + K did not conduct this themselves  

Generalisability

• Large sample size used of 32 studies including 1990 mother-infant pairs – results should be generalisable to a range of different cultures 

• However, not all countries were studied – for example no countries from Africa or South America, and the majority of the samples (18) were from the USA 

• Only studied mother-infant pairs so may not be representative of relationships with the father 

Clinical practical

Validity

• Subjective interpretations of what was considered positive/neutral which would reduce reliability. However attempt to standardise this - created a list of words as a collective deemed to be neutral and negative and used this to compare to words found in the article

Reliability

• 20 minutes on each article to scan and record positive/negative descriptor words for standardisation

• Interrater reliability

Generalisability

• Use of three different articles from each newspaper (The Guardian or The Daily Mail) specifically reporting on schizophrenia is quite a suitable sample size as this is a very specific/niche topic. The three articles were chosen unbiasedly by selecting the first three that came up on a google engine search

• However only used one newspaper (The Guardian) to show broadsheet and one (The Daily Mail) to show tabloid - this is not representative

Social practical

Validity

• Questionnaire as a form of self-reporting data on a socially sensitive topic like obedience runs risk of social desirability bias. Inclusion of red herring questions to try and reduce this risk of demand characteristics, also did not inform participants specifically what the nature of the study was about (slight deception - breach of ethical guidelines deemed necessary, inclusion of a debrief)

• Inclusion of both quantitative and qualitative data - qual subject to researcher bias/subjectivity?

• Piloted study to check for understanding

Reliability

• standardisation over all participants receiving the same questionnaire with the same questions in the same order - replicability

• however no standardisation over what environment the questionnaire was completed in

Generalisability

• Small sample size, limited to only students aged 16-18 in the local area. Opportunity sampling technique may be biased. Not representative of others cultures (England as an individualistic culture may implicate upon obedience levels in comparison to collectivist cultures)
  Removed age as a confounding variable

Cognitive practical

Validity

• Procedure lacks ecological validity as nature of the task is artificial and not representative of daily life

• objective interpretation of results as quantitative data

Reliability

• Independent researchers = lack of standardisation over methodology - however all told to wear neutral colours, speak in a neutral tone and conduct experiment in a silent room

• Standardisation over the rate at which digits were read aloud

Generalisability

Biological practical

Validity

• Ethics brief and debrief - fully informed consent, informing participants of the aggression questionnaire. This could have decreased the validity by increasing the risk of social desirability bias in the participants’ answers to the questions 

• Objective quantitative data provided

Reliability

• Standardised procedure as all participants completed the same Buss-Perry questionnaire, answering the same questions

• However independent researchers measuring finger quotient - asked to measure fingers on dominant hand, however this may have differed between participants left vs right. Attempt to standardise the way in which finger was measured

Generalisability

• opportunity sample of 16-18 year olds, half male half female. Lacks generalisability as a small sample of only 28 participants - not representative of wider society, results cannot be generalised beyond study

Learning practical

Validity

• High ecological validity as a naturalistic observation, participants did not know of their involvement in the study (ethical breach - hwvr near public security camera) so less risk of social desirability bias or demand characteristics

• only completed at one time of day - not representative perhaps of behaviours at different times (afterschool - people may have been in more of a rush, motivations differ)

Reliability

• set length of time observing, stayed in the same location, use of tally method to record results - standardisation

• however observations carried out by independent researchers - unlikely that exact same methodology and procedure

Generalisability

• opportunity sampling - sample not representative, particular time of day migth mean individuals with similar characteristics

Child practical

Validity

• High internal validity as controlled extraneous variables by completing in a silent room, all pps completed individually so they could not confer - reduces conformity bias

• we checked understanding for all the different descriptions of emotions by providing participants with list of definitions before beginning the experiment

• included a practice test before beginning to ensure participants understood what they were completing

Reliability

• All participants viewed the same photos and received the same options of emotions to choose from. Instructions were standardised and each photograph was shown for 8 seconds to standardise the procedure for all participants.

  
  

Generalisability