Essential Guide to Case Studies in Cognitive Psychology Research

Hello and welcome to the course basics of experimental design for cognitive psychology. As I mentioned in this week

two, I'll be giving us a broad survey of the different kinds of research design so that we are aware of the

considerations that goes into each kind of research design and we'll also uh be able to uh you know make an informed

decision about how and when or what kind of research design to choose uh that basically suits the kind of research

questions that we are answering. Now uh we're talking about descriptive research designs uh since the previous

lecture and one of the most important uh you know on one of the most widely used descriptive research designs is

basically case studies. All right. So case studies include data based on a very small set of individuals sometimes

just one or two individuals uh which are selected on the basis of their unique characteristics. So a case study

approach is most important to uncover what is happening with a specific individual and that specific individual

is chosen because there are certain unique characteristics of these individuals or they're undergoing

certain unique uh let's say uh circumstances or there is something very interesting uh that is uh to be

understood to be had when you are performing or when you are observing this case in extreme detail. For

example, Jean Pia uh the developmental psychologist used observations of his own children to develop the stage theory

of cognitive development. I'm sure all psychology students are aware of Pia's theories of cognitive development which

is basically built on the basis of his observations of his own children and some around uh you know his house. Other

famous examples for example include this study of the patient M. Leourne or more famously known as tan by Paul Broka to

develop detailed descriptions of what happens when uh the uh Broca's area which is you know the inf inferior

temporal uh uh area in the left hemisphere uh is damaged and it leads to corresponding difficulties in uh speech

production. Okay. So uh these kinds of situations are actually fairly unique and they basically are potentially

extremely informative about what happens when a particular uh instance happens. Say for example when a brain injury

happens what are the consequences and you want to study all of these consequences is in detail so that you

can basically uh develop a particular theory about the processes involved. Let's say when uh Paul Broka came across

this particular patient, he was only be being able to say tan tan t 10 and a few exploitives here and there but was not

able to produce uh coherent grammatically correct speech. Uh once this person died uh you know uh this was

uh you know uh his uh skull was postumously exumed you know his postmodern was carried out

histologology was carried out on his brain which revealed that this was the area of his lesion. So that basically is

is an interesting aspect in terms of case studies. Now uh while case studies are typically conduct so as I said case

studies are typically conducted with individuals who have unusual or abnormal experiences and possess characteristics

uh or are going through unique situations. The idea as I said is uh that the individuals will provide a

detailed ins insight into statistically rare instances uh and be extremely informative about these process or

phenomena. One of the most famous proponents of the case study approach has been uh you know Siban Freud who I

always call the rockstar of psychology because he was a master of using uh the psychological difficulties of his

patients his individuals to draw conclusions about uh basic psychological processes. Say for example his uh

detailed descriptions about the subconscious and the unconscious mind. these detailed treaties on

interpretation of dreams on basically uh you know also the the nature of personality the id ego and the super ego

kind of things most of these things were basically developed most of these insights were developed through a series

of case studies uh which is basically uh individuals who are coming to him with particular kinds of psychological

problems particular kind of psychological difficulties and Freud was take extensive interviews with these

individuals and basically make these observations later connect these this observation with other cases form a

pattern and that is basically what led to some of his most famous theories. An example uh could be the study of little

hands uh you know Freud treated this uh little kid uh who was afraid of horses. Now uh when Freud goes deep into uh you

know how the child behaves and other uh basic details also about uh the family and stuff he basically finds out that

this fear of horses could be interpreted in terms of rep repressed sexual impulses towards the mother uh fear of

the father and so on. So uh again uh we'll not discuss the merits of what Freud was investigating and later

interpreting as uh but the idea is that case studies are extremely informative. They tell us something interesting,

something unique about these statistically rare occurrences and sometimes can also be connected across

similar cases of of a kind and be developed into theories. Other instances and we'll take some

examples just to sort of you know drive home the point. Other cases of uh you know utilizing the case study approach

is that of neurologists who uh use this approach to understand and investigate disorders of the brain and some

injuries. Now for example Roger Perry uh studied uh who are called split brain patients. These patients are those

individuals who have had their two hemisphere surgically disconnected to prevent severe epileptic seizures. Now

when this phenomena was first observed the people would survive lead a relatively healthy life so to speak uh

but very interesting behavioral patterns emerged. For example, if you look at and there are a bunch on YouTube, there are

a bunch on uh cognitive neuroscience textbooks that you can find uh you know the behavioral outcomes of split brain

patients. Uh sometimes you'll come across an individual who's uh you know seeing something here but not being able

to name it, seeing something in the right visual field, able to draw but not name it and so on. uh there are there

are interesting anecdotal uh you know evidences there are uh detailed documentations of what happens when

let's say this you know bunch of uh fibers the corpus colosum are disconnected uh surgically and the left

and the right hemispheres are left to be on their own they are not communicating with each other all right so this has

also been an interesting sort of uh you know uh coalition of a bunch of cases with you know with patients who had uh

their brains split because that was the only surgical alternative to prevent epileptic seizures around the time. Now

uh another very important case uh you know in in psychology cogary psychology in particular was that of Phineas Gage.

Uh I'm sure you would remember if not you can go and you know look at some of my previous lectures. Uh now the Phineas

Gage was basically a railroad worker and he was basically pumping dynamite into the uh track and they're basically

trying to uh you know uh clear the stones and this and that. Uh once what happens is that he uh you know inserts

an iron uh rod but he fails to sort of uh you know take adequate precautions and as the dynamite burst the iron rod

basically enters through his uh you know through his cheek and sort of uh goes right through it taking away a portion

of the frontal cortex. Now before this accident happened, uh Fenus Gage was supposed to be a likable, sociable

worker, was very happy, very regular, uh had extremely nice uh you know uh relationships with people around. Uh but

once this accident had happened, uh his behavior changed uh drastically. Okay. A person who was once popular and sociable

and was liked by all his colleagues, all his co-workers was now uh you know uttering profanities, abuses, was not

regular to work, was getting into fights a lot and his behavior considerably changed. Now uh neurologists basically

found that this is interesting that you know damage in a specific area of the brain is leading to changes in in

personality and they made you know detailed attempts to study the behavior of fine cage and uh learn about the role

of the prefrontal cortex in managing social behavior. This is again a few cases that I'm going to talk about.

Another if you remember is the case of HM a patient uh who was studied extensively by professor Glenn Humphre

the late professor Glenn Humphre where what happened was similar to the split brain patients you know a surgical

alternative to prevent his epileptic seizures was followed now uh what happens in the case of H was that a part

of his you know the hippocampus which is uh area in the brain that is responsible for forming new memories uh hippocampus

in both in the left and the right hemisphere was surgically removed because that was probably near to the

source of the epileptic seizures that this guy was having. Interestingly once this uh you know uh surgery was

performed the epileptic seizures stopped but something very interesting happened. The ability of HM to form new memories

completely vanished. he could remember everything in some uh you know to some degree uh prior to the operation was not

but was not able to form new memories after the operation. So till the day he died he could not really make memories

and remember anything uh beyond a very small period of time. Uh and therefore he became a source of curiosity for a

lot of neuro neurologists. He was studied extensively by professor Glenn Humphre but several others also who

basically you know went to this uh guy tested him across different range of scenarios across different tasks in

terms of different cognitive functions to study what is the um you know in uh a separable role of the hippoc campus not

only in memory formation but in other uh you know uh cognitive functions as well. So in that sense you can I think you can

appreciate that case studies are uh extremely useful in in getting us some idea of how uh you know things are

especially when uh you know there are instances which are statistically rare or unique. All right. Now while this is

uh this is uh good it is extremely important and it yields uh you know very interesting knowledge there are certain

disadvantages to the case study approach as well. For example, case studies are based on unique and often irreplicable

uh experiences or the or situations of a very small set of individuals. Say for example, if neurologists

were to develop uh theories or were to develop uh you know broader theories based on specific cases like that of

gage or like that of uh TAN or like that of HM. It will be very difficult to generalize

this across patients. The first reason would be uh it is almost statistically you know impossible to have another

individual have exactly same uh you know injury. Say for example uh injury to the exact same coordinates of the brain.

Also uh we know that functional and structural conductivity uh of the brain differs across individuals. So if one

were to develop theory about uh let's say the role of uh the prefrontal cortex in social behavior, it cannot be made

just on one case alone. All right. So whatever we know about the prefrontal cortex and its role in social behavior

is actually derived from a large number of uh uh you know case studies, large number of experimental investigations,

neuromaging investigations and so on. Okay. So uh while a single cases are or let's say one or two or three cases are

extremely important they may help inform us about say for example they can lead to certain hypothesis you know the way

people were performing experiments with phenas gauge or with HM can actually be a source of certain hypothesis which

could then be tested in a range of normal individuals or tested with other patients who may have injuries similar

to those areas but not exactly those areas because that's extremely rare. All right. So, uh these instances, these

cases, they cannot tell us much about whether the same experiences would manifest in other individuals as well in

similar situations or uh even it'll be difficult for the case studies to tell us why uh the particular situation the

specific unique situation uh of a given individual is manifested is manifesting in the way it is. Okay? because there is

there is a lot of uh you know details about the individual uh that has probably happened prior to the accident

or prior to the surgery and we don't really know we know once this event has happened these are the consequences we

don't know everything else about the individual from the past and so on also there is no comparison group so

remember when we want to make causal inferences we want to have a comparison group we want to have a control group we

want to have an experimental group we want certain kinds of controls uh all of That obviously is not possible with case

studies. Okay. So there is no comparison group in these situations that did not experience the unique situations that

these cases are experiencing and therefore researchers cannot really determine how these individuals would be

like if they were not having this unique experience. So again it's it's very difficult to match uh you know two

individuals on all situations. Okay. Now on top of that this one individual has had this particular injury. You cannot

really compare the two because there are too many variables in play. All right. Now, given that there are some of some

advantages and some disadvantages to case studies and let's say observational research designs that we've talked about

in the previous lecture, uh we can talk a little bit about how do people collect data, how do people organize data in

these uh you know uh methods. So even though observational research designs and case studies provide a detailed uh

insight into ongoing ongoing behavior, they may not often be as objective as one might like or they might also be

colored with say for example experimental bias to be uh precise. So as the observer has chosen which

individuals to study which so this is extremely selective the sampling is extremely selective. The observer has

chosen which individuals to study, which behaviors to record or which to ignore. uh and also the method of how how they

will interpret these behaviors. Say for example uh Sigman Freud interpreted most of the problems of his patients in terms

of unconscious repressed impulses. Uh they could be uh sexual in some cases. They could be uh you know uh uh violent

in some cases. everything was in some sense broadly connected back to problems uh or traumatic experiences which were

rooted in the past experiences and in in some sense repressed and logged back into the unconscious or subconscious

mind. So uh in in that sense what happens is that the uh generalizability the scientific validity of whatever you

are deriving from these uh you know observational research uh designs or case studies uh becomes extremely

difficult. There's also a very interesting tendency. We've talked about papar and we've talked about

verification and falsification. A lot of times what you'll see is the way people uh describe case studies is basically

that they say okay I had this hypothesis when I was studying this case and see this this this confirms my hypothesis.

There is hardly scope to falsify anything in that sense and then that also as you might be aware by now makes

this whole thing a little bit weak does not stand on proper scientific grounds. Again uh another thing is that the data

that is observed that is basically being collected uh from these uh forms of research methods can sometimes be

extremely sketchy you know uh it's it's unstructured it's unorganized it can just be in a form of field notes you

know things that you are listening to making notes hazardly sometimes and they're not in that sense aminable to

detailed analysis and assessment although uh you know very organized case studies have also uh you know been

organized we'll talk come to them in a bit so In a lot of cases, these problems can be overcome by using systematic

observation and to create quantitative like measured variables. What are the things that you would want

to sort of uh use to maximize the potential of these designs? Uh one is systematic observation. For example, the

researcher needs to specify ahead of time exactly which observations will be made on which people and in what times

and places. So in that sense you already know what you're looking for. You're not there and collecting everything that is

available. You're looking for something. You're doing some research and in that research you know exactly what it is

you're looking for. Rest of that is just junk data and will not be useful. So these decisions by the way also are

based on theoretical expectations. Say for example if a uh you know if in today's time uh you know a patient

presents himself or herself to a neurologist door with a frontal cortex injury uh the researcher the neur

neurologist will have a ears of theory to inform him what to expect and what not in that patient. Now in that sense

the hypothesis that this neurologist or psychologist would form when they are testing this new patient will be

informed by theory that has been uh you know accumulated over several uh years uh for that matter. All right. So these

decisions about what to observe, who to observe, what specific things to observe, what times, what task and so on

are based on theoretical expectations about the type of events that are going to be of interest. specifically about

the uh you know this specificity about the behaviors of interest has the advantage of both focusing the

observer's attention on these specific behaviors and also reducing the mountains of data that might be

collected if the observers observers were just randomly recording everything that they find. Another thing uh in

cases for example pomerans and colleagues who were basically interested in uh assessing how and when young

children compared their own performance with other children uh they wanted to do this study. So what they did was in this

study one or two adult observers sat on chairs uh and they basically adjacent to work areas or adjacent to uh you know

next to the classrooms of elementary school children and they were basically just recording everything on the laptops

and uh you know about the behavior of the children. Now see what happens here is that uh even prior to starting this

data collection uh these researchers went in with specified behavioral categories uh which would later be used

by the observer. So for example these guys know what they are observing for number of instances uh uh or or things

like that. So when these categories are formed before data collection based on theoretical predictions and based on uh

you know previous theory uh the researchers would have an idea about uh you know what do they observe in these

children what kind of behaviors they take note about and what are the specific uh instances of those behaviors

occurring. Also the method of recording observations is important. So what to

observe and how to observe? How to observe is equally important. Say for example, even before the researchers

begin to regard the behavior, the uh observers uh you know or the uh uh researchers would have spent three or

four days in the classroom learning, practicing and revising the coding methods and also letting the children

get used to their presence so that the data collection exercise is uncorrupted. Okay, there is practice. There is

something that you've sort of said, okay, uh when I was spending time with these children, I saw this, this, this,

this, this. Uh and then the researcher will say, okay, out of the 10 things you've listed, four are important. All

right, as the coding categories were well defined in this colleague study, there was a very good interrator

reliability. So across these number of observers they were basically uh reliably talking about the same

phenomena and also to ensure that this interrator reliability remains uh the experimenters frequently carried out the

reliability analysis on the codings that were being obtained over time again and again uh with every observation period

so that they know that the data is being collected is clean and it's correct. Now this is extremely important because uh

you know even in uh children's social behavior for example there are some behaviors that would occur more

frequently some would occur infrequently and it is important to be sure that uh you know the behaviors are being uh

coded in a reliable in a systematic manner over the course of each observation

period. For example, several types of data could be collected. Uh for one the observers encoded what is called event

frequencies. For instance, the number of verbal statements that indicated social comparison. For example, uh statements

like my picture is the best or how many did you get wrong or how many did you get right? Things like that. So where

where comparison is there. So you can tally them. You can say oh these are the instances at these times these two kids

were comparing. And again depends on what kind of uh you know uh theoretical uh what framework you've gone in with.

Also the observers could code for let's say event duration. Not only frequencies but also duration. For instance, the

amount of time that the child was attending to their work versus the amount of time the child is attending to

the work of others. So finally all the children were interviewed after the observation had ended and this sort of

closed the overall loop and it provided rather systematic data uh to the observers uh recorded well for uh you

know subsequent uh analysis and interpretation. Another slightly equally important thing

actually in terms of collecting uh you know observational data is choosing the kind of uh you know sample that you're

targeting. What is your target population? What is the sample that you would go with? All right. So for example

uh permanent and colleagues use three basic sampling strategies to reduce the amount of data that they needed to

record. First they utilized event sampling. They focused on specific behaviors that were theoretically

related to social comparison. Second, they employed individual sampling. That is the observers were randomly selected.

Observers typically selected one child to be observed during one observational period. Okay. So, which individual is

being sampled for observation session number one or two or 20? What are the events within this individual that I'm

going to observe? And lastly, time sampling. Each child was observed only for a period of up to 4 minutes. So in

this 4-minute period, this is what the behavior of this particular child looked like and then maybe his turn will come

later. All right. Now these data you can see now the data that is going to come out will be extremely systematic. So

these data were coded as they were being observed and the observed events could now be more precisely recorded as it was

guided by a theory. It was guided by particular sampling strategies. Finally, Pomance and colleagues actually

found that the older children were using uh were basically using more subtle uh social comparison strategies and they

basically thought of social comparison or comparisons typically as uh you know boastful behavior or unfair behavior. So

you can see initially children are more innocent and they get into comparisons and this and that but as they grow older

they get a sense of that this is unfair, this is boastful and probably not socially accepted behavior. So you can

see uh using a a systematic research design even in uh observational uh studies uh you can actually come up with

interesting insights useful insights uh with high reliability and validity. Okay. So uh in that sense uh you know uh

what we have is a very good idea of how observational research or case studies can actually you know inform us and what

kind of insights can be gained from this. How can we what are the basic concerns in in in coding data uh in

organizing it interpreting it later and analyzing it and interpreting it later all right so thank you I'll talk to you

about other uh research designs in the next lectures