Experimental Design Tasks in Cognitive Psychology: Types and Selection Guidelines

Hello and welcome to the course basics of experimental design for cognitive psychology. I am Dr. Arkwarma from the

department of cognitive science at ID Kpur. Uh we are in the week fourth of the course and we are discussing

different types of uh experimental designs and their nuances. You would remember in the last lecture we talked

about Cunningham and Wal Raven's uh you know uh elements of an experimental task where we talked about task stimuli uh

and uh you know other elements that basically go into making a good experimental task. Uh so in the last

experiment we uh basically first talked about the participants sampling and so on. In this one we going to talk about

task. In the next lecture again we'll probably talk about the task and then the stimula and then we'll move on to

other kinds of experimental designs and uh eventually we'll go on to reaction time studies and so on. So uh let us uh

start and we'll see uh what we have to do. So now one of the most important decisions for an experimental to make uh

once uh the research question has been zeroed upon you have a research question in mind uh the hypothesis is worked out

uh is that the task you know is basically deciding upon the task within whose framework uh you know participants

performance will be measured and the hypothesis tested. So the for example you want to you have an hypothesis about

that you know uh we talked about that say for example color of ink may have a influence on memorizability or we have

something like uh you know symmetric uh uh uh shapes are considered more aesthetically beautiful this and that.

So we can have any number of uh you know hypothesis but what kind of task do we give in which task this question is best

answered for example uh let's take this question which type

of uh shapes or let's say even uh let's let's play with this let's say the hypothesis is symmetric shapes are

considered more aesthetically beautiful as opposed to asymmetric or non-ymmetric shapes all right now this is a broad

question that we have we want to test this what kind of task is best suited. Now you'll see interestingly uh the

degree to which we want an answer or what is it with that answer that we want to do will basically also decide or help

us decide uh you know what kind of task you will do. For example, you can ask this in a free range kind of a task

where you can just ask okay uh you can give some uh figure figures and you can give some shapes and you can say which

of the two or which of the five are more beautiful and then you let the participants describe that. You can even

ask the participants to describe them that their uh you know uh judgment of a particular shape or a figure in in some

detail that why do they find it beautiful or attractive and so on and then basically hope to uh you know go

through thematically uh their answers and figure out that okay this is uh considered more beautiful whether it is

symmetric or not is symmetry a factor in that shape being considered beautiful and so on. On the other hand, if you are

interested in for example just knowing uh that whether people can uh classify a particular shape as symmetric

asymmetric, you give a reaction time kind of a set study and then a second task would be that okay uh you know we

basically ask you to whether you are able to detect symmetry or not and if you are able to detect symmetry is it

more beautiful less beautiful than the uh you know other reference shapes or not. So uh depends upon what is it that

you want to do what kind of research question you have in mind what kind of hypothesis that you want to test on the

basis of these factors you will select a given task and there are a variety of tasks. So for example there is uh

obviously a variety of tasks that are available and uh several tasks are typically used to uh probe even one

single cognitive function. For example, uh there's a lot of literature on language switching studies on inhibitory

control on executive functions. And you'll see there are very popular tasks, language switching task, task switching

scenarios where say for example a famous task was uh you know between color and shape discrimination task. Uh you have

the stop signal task, you have the flanger task, you have the ent task, you have the go no-go task. There are so

many of these tasks each of which are sort of uh you know geared towards broadly one uh function which is your

executive functions but within that uh you will know that attention networking task talks about uh you know different

networks of attention orienting flangask talk about something else there is congruence in congruence there are so

many factors that you can figure out through a given task. So in that sense a task is an extremely important uh you

know uh decision to make and then we can sort of worry about okay what are the factors how does a uh you know uh

experimental zero in upon the task that he or she wants to use. Now the first and the most important factor that

probably helps in deciding uh upon which task we use is the research question. Okay. So the more explicit and

unambiguous your research question is going to be, the easier it will be to decide upon the task. So for example,

are you interested in reaction times? Are you interested in the phenomenal experience of uh you know something

being considered attractive or not? Are you interested in comparisons? What is it that you really want to check? And on

the basis of that you will choose your task. There's a very interesting uh you know texonomy of experimental tasks

which is uh you know offered by Cunningham and Wal Raymond in their book uh which if presented here on the right

side for you to see is that say for example depending upon whether your research question is general or very

specific and depending upon what kind of answers you seeking. Are you seeking rather vague answers or very concrete

answers? You basically have this whole range of task on a on a continuum almost. Okay. So for example you can

start from free description task or rating task or a four choice two alternative four choice or three

alternative four choice kind of task uh to sometimes nonverbal physiological tasks which are eyetracking imaging and

those kinds of things. Uh and again it sort of has to do with the kind of question that you have. All right. So uh

if the experimentter say for example wants to know the value of a specific parameter in a new visualization

technique that provides the fastest and the most accurate identification of brain tumors then a task will be needed

that can basically uh measure the response uh speed as well as also uh you know uh count uh identification

accuracy. If that happens then is where you will be able to decide about how good the task is working. On the other

end, as I was saying earlier, if the experimental wants to know whether people notice some form of facial

expression in the motion of collected dots, there are a lot of these random dot stereoggram studies and so on. So,

first is whether people notice some form of uh facial expressions in this collection uh you know in this

collective movement of dots and if so then which emotion are they seeing. Okay. So this basically if you have this

kind of a question it allows the participants as much leeway as possible without influencing the nature of

answers and it'll provide you all the answers need. So you you probably tend to give a more freeranging task. Say for

example uh a task that allows for different kinds of responses to emerge and not constrain them between yes no

this that and you know that kind of responses. So the distinguish uh you know the

distinction between specific and general you can see is basically you know represented here in this taxonomy by

Cunningham and Bal Raven with be the research question. how uh general your research question is versus how uh

specific uh it is or say for example how concrete answers you want versus how vague answers can do that is something

uh you know which is uh interesting and this is sort of a very interesting uh you know way to look at uh the kind of

tasks that are available used in both experimental studies as well as some other uh you know qualitative studies as

well. So at one end of the continue were the tasks that answer broadway vague questions which are very difficult to

interpret very you know uniquely and quickly because the phenomenal experience of the participant sort of

you know comes into play. At the other end are tasks that easily support unique representations but focus on very

specific questions and thus generate very specific answers. We'll talk about those things. All right. Now uh

accordingly uh if you look at the kind of tasks that are available these guys divide these tasks into meta tasks,

direct tasks and physiological tasks. Okay, let's let's uh dive slightly deeper and understand these. So the

metatasks basically uh here the participants are essentially asked how they think uh or believe they would act

in a given situation. Oh do you think this is beautiful? Do you think it is not? what would you do in you know in

terms of categorizing these stimula those kinds of things. Okay. So a simple experimental sort of an example could be

one could present a number of images of facial expressions uh rendered in different manners say smileys schematic

faces real faces uh you know caricaturish rendering renderings and so on. And then you ask the participant

which style he thinks or she thinks is the best at communicating the expressions. In which of these three or

four or five different formats the do the expressions come out best. Okay. Now these kinds of tasks can be called a

meta task because it asks for a form of prediction or introspection into what one is feeling when one is looking at

the stimula. This does not measure how well one can actually recognize the expression. So

this is not a reaction time kind of a task. It is a task where you are simply being asked a question and you can give

a relatively general answer here. Okay. Uh and these are also very important tasks. So metadata tasks are useful in

instances where the experimenters are interested in the participants beliefs. What are they feeling? you know what is

their opinion and where the participants cannot be constrained in terms of response choices and basically that we

are not constraining them in response choices can help us obtain an unbiased insight into the natural boundaries of

the phenomena. So for example, you can present a range of pictures uh you know let's say aesthetics is the uh thing

that you want to study. You can present them with a range of pictures uh from aesthetically very beautiful to

aesthetically not so appealing and you can basically keep on asking the participants you know broader questions

and you'll see that there develops a pattern that will be able to probably you know categorize these pictures into

aesthetic aesthetically positive and aesthetically slightly negative pictures. So that is where these kinds

of tasks probably you know uh they they work best. Now there are other kinds of task as

well in the middle. These are the direct task. What are the direct task? They basically ask the participant to

actually perform the task. Okay. Say for example given two images judge which of the two is most beautiful. We are

measuring your reaction time and accuracy. So two pictures come and you present uh you press uh X for left one

being more more beautiful and zed for the right one being more more beautiful and you're basically asking the

participants to give you a timed response. All right. So some forms of rating and force choice task and we'll

talk about them in some detail later uh especially uh the specialized force choice task and a non-verbal real world

task basically are referred to as direct tasks. Okay. And these are also very important because they provide a direct

window into how people will actually respond in different situations. However, because these are direct uh

direct uh you know uh choice tasks, they're also difficult to use sometimes because it is not possible to recreate

that same scenario all the time in the lab. Okay. uh if you really want to say for example uh check whether and how

people respond to say for example APS pictures you know you cannot create those kinds of scenarios in the lab

again and again and to that extent the task sort of becomes ecologically less valid and so on. Okay. Finally, the uh

third kinds of task, you know, broad category of tasks that is considered important are the physiological tasks.

Okay. Uh these basically include measuring the body's uh you know, reactions such as heart rate, body

temperature, neural firings and so on. And they are extremely uh useful because they provide a very direct unbiased view

of what elements of the stimulus the participants really saw and how they're responding to that, you know,

physiologically at the physiological level. But again, so the idea is that these physiological responses are able

to provide us a direct window into what the participants actually felt. How did their body react, how did their mind

react and so on. But again these are relatively uh difficult to implement because uh you know most uh research

questions involve real world behavior or subjective experiences and then at that uh and for that reason making solid

definitive connections between physiology and uh you know what a person will do in that real situation is is not

very precise. So there is obviously a debate about say for example when you are showing a picture of a uh you know

let's say you know an unpleasant picture versus if you are actually putting the person in that scenario what kind of

reactions will happen is that a direct one to one correspondence or is that say for example the difference between

stimulation and actually doing something. So uh again these are very important tasks and they are widely used

in experimental studies but they have their uh you know own issues and we'll talk about them at some point.

Now, so which method one uses, which task one uses, let's take this uh slightly forward. Typically, uh if you

see, I very briefly surveyed the three kinds of tasks that are available. Uh we also looked at this texonomy that

Cunningham and Walraven offer. Uh but if you see in practice, it boils down to the researchers own preferred methods

and their experience and their expertise in what kind of methods they do. Say for example, some people like doing uh

experimental uh studies with reaction time experiments. Uh these are called behavioral uh measures, reaction time

experiments or eyetracking experiments. Some people are slightly more comfortable with let's say uh

electrphysiology. So EEG and ERP kind of experiments. Others are more uh you know uh interested in doing say for example

things like neuromaging you know fMRI and uh you know FTCD and PET and uh you know those those kinds of things. uh

similarly uh you know there are other kind of uh tasks that people would do. So uh it eventually boils down to what

kind of research question you have what are what is the cognitive process that you are probing and also to a certain

level uh you know the degree of experience the degree of preference and expertise that the researcher carries a

lot of people I I know would like to do only behavioral tasks and will not really bother too much about neuro

imaging and so on and while some others will probably you know they'll have their hypothesis and they would love to

go to the scanner and directly you know perform neuroiming experiments to basically look at how does the brain

respond to uh you know these uh uh the kind of stimula that they want to sort of play with. So we'll talk about these

tasks in in much detail towards the later parts of the course. But uh again the point that we want to sort of uh

consider today is that what is the question that you're asking and how does that question constrain the choice of

task. Okay. So while some methods can be more appropriate for certain types of questions say for example when probing

high level versus low-level cognitive processes those kinds of things are possible. An important uh consideration

here also by the way uh depending on say for example or something that sort of feeds into your choice of method is the

possibility for controlling all sorts of external noise and variance. Okay, it's very important because uh whichever

research question that you want to approach and you want to answer will require quite a few experiments to be

performed. You are sometimes using different kinds of task as well and the experiments also say for example need to

be thought out uh you know much in advance planned very well before they are carried out. So the thing is uh a

lot of times uh the choice of task or choice of method gets governed by which is the most logistically feasible. Say

for example uh in u neuro imaging experiments movement is is not uh you know considered movement sort of creates

artifacts and it's not done in eg also probably movement creates artifacts and it's discouraged in behavioral task for

example uh you know these are much more easily tested much uh uh more uh logistically uh nicer to set up you just

need a you know a laptop uh uh you know and a screen a laptop say for example or sometimes a desktop and a keyboard that

is enough. So uh these uh things sort of factor in uh depend you know and they help us decide which kind of task you

want to use also uh the kind of data that will come out. So uh what kind of data comes out what how will you deal

with that data how will you analyze that data uh interpret that data to sort of give some conclusions about the behavior

that you wanted to study those things also uh play a very very important part okay another issue and we we you'll see

that when we go deeper into each of the kinds of tasks so be task ietracking eg fmri you will see the these things that

we're discussing now sort of they'll start making much more sense or you'll come back and revisit this lecture and

say okay these are the considerations that we were talking about. Now another important thing a sort of a meta uh

factor will you know uh in descri in choosing the method or in choosing the task is basically how much do your

participants know when they are going into the experimental lab and uh you know how much should they know. Uh now

in in this uh something interesting is that say for example with the information processing metaphor it is

possible to know and it is possible to categorize you know particular cognitive functions as lowle or high level or

early or late cognitive processes and that is something very interesting because say for example uh see the broad

thing is you show a stimulus uh the stimulus basically is sensed you know the person sees that stimulus once the

person sees that stimulus the process of sensation converts this external physical stimulus whatever if you are

showing a picture or playing a sound we'll convert this into uh you know nerve impulses so transduction happens

at the very first step in the pipeline so if something is just about detection if if your hypothesis is just about

detection of particular stimuli or the timeline of detection for example then you're talking about early level

processes on the other hand there are also high level or slightly later processes which uh say for example like

aesthetic judgment or decision-m which are generally considered to requiring intention and are happening slightly

later in that pipeline. Okay. So where do you want to place your task? What is your hypothesis? Do you want to test

these early level process or do you want to test these later level process, lowle, high level? Whichever dichotomy

you are comfortable with. Okay. Now a question that needs to be taken into account while choosing a task with

respect to this differentiation of high versus low-level task is whether the participant can intentionally influence

their performance. Okay. Now the kind of task that you're using and when you get the participant in the experimental lab

is there a possibility that the participant can understand the demand characteristics and the participant can

sort of play with the task and try and give you the you know the expected responses. Okay. So a factor here is

this cognitive impenetrability. So when you're talking about low-level sensation perception kinds of task, attention kind

of task there uh these tasks are relatively cognitively impenetrable because you cannot intentionally uh you

know get inside the task understand and influence your performance there. Okay. Uh so in these cases you can sort of uh

you know use uh these over kinds of tasks uh you know detection and visual search and so on because even if the

participant decides they cannot uh influence much of their performance okay although there is obviously always a

gray area and the participants can you know work out a strategy work out a pattern in your stimulus presentation

work out say for example uh I don't know some kind of low-level uh variances in your sim and use that so one has to be

in that sense also very careful what kind of task what kind of stimula what kind of stimula presentation uh etc am I

going to use so that the participant cannot judge uh the kind of task I'm using and cannot start uh catering to

the demand characteristics of this scenario. So demand characteristics uh is is very important that creates a

response bias that basically creates a tendency for the participant to respond in the way they

think that we need. Okay. A lot of times the participants are out there to you know they're motivated to produce the

results that the experimental is wanting in or sometimes the opposite of that is also possible. Okay. So what kind of

demand characteristics the demand critics should not be easily discernible. Uh the experimental must be

uh aware of the fact that what kind of strategies the participants are uh you know uh adopting and uh how are these

response strategies affecting the response criteria. Okay. Now if you put all of these together, all of these

things together, you will be able to decipher whether there is a response bias in your participants responses or

not. And it is important in that sense to work out that response bias and take steps to either at the level of the

experiment and data collection to avoid this or later uh you know using statistics say for example signal

detection theory and so on work that out that okay what kind of uh you know response biases are present in my data

and how can I sort of work those things out okay so the issue of response biases becomes extremely difficult to handle

when participants know what is the topic what is the goal of our experimental studies so if you're studying uh you

know processes that require participants to intentionally make choices to perform the task and those kind of tasks if free

desript description tasks provide considerable leeway in how the participants proceed proceed in

providing us answers. This is where the the issue of response bias motivation strategies and demand characteristics

becomes extremely tricky. Now how do you how does one address response bias in task? Uh first is

concealing expectations. It's very important that the participant is unaware of the true purpose of the

experiment and expected responses from the experiment. You can obviously de debrief the participants later. We'll

talk talk about this when we are talking about the ethics of research. Uh but concealing expectations is important for

a lot of studies. Uh preserving anonymity. So participants should be categorically assured that the data that

we are going to use is being coded in an anonymized fashion and it is going to be used in an anonymized fashion and there

are no identifiers that can lead back to the participant for whatever way there's no judgment evaluation etc. So anonymity

sometimes if you can if you can assure the anonymity of the entire enterprise to their participants it helps uh you

know them to basically give you good responses correct respon good is probably not the best word but it helps

them to uh respond naturally uh to the kind of experimental task that you're working with uh also yes as I just

mentioned uh using statistical techniques to detect response biases and to uh you know uh address them and to

sort of weird them out is is also extremely important. Uh sometimes altering the response

criteria for example uh you know you can provide feedback uh reward for correct answers or uh higher rate of accuracy

and punishment for lower rate of accuracy. Uh that sometimes requires uh you know giving participant feedback.

I've seen a lot of behavioral experiments do provide feedback but there's also a bit of a trade-off here

because sometimes the participants basically uh understand uh and they sort of will become extremely slow and

they'll get into this response you know uh uh speed accuracy trade-off. So that is also something that you would want to

uh you know be slightly careful about. Uh adding cache trials for example uh adding cash trials is an explicit way to

catch specific response biases. For example, if the participant is responding yes, yes, yes, yes, yes to

all trials, you could add some filler trials, you could add some cash trials whose response is not yes. And if the

participant responds to those uh trials also in you know that the participant is not sincerely responding and in that

sense you can basically ideally you know work out the data work out the participant and say okay these kinds of

responses are not acceptable or you at least mark that and know that this uh you know the data from this particular

participant is going to be unreliable. All right. Sometimes also using a covert task. So a lot of times what you uh what

people do is that they uh have something else in mind but they use a completely covert task. For example, if you if the

if you are interested in examining how people respond uh to upwards you know in in the direction of uh you know up

versus down. How do people do that? Uh they can typically just get the participants to be performing a lexical

decision task. And in the lexical decision task the idea is that you just ask participant whether is the you know

whether the word presented on the screen is a actual word or not. So you ask that and you include some upwards and some

filler words there and you basically let the participants tell you just uh you know if this is a word or not. Now if

the participants actually respond to up words faster then you'll see that their reaction time with the up words are

actually faster as compared to the filler words and the other kind of words. Okay. So upwards are basically

positive uh uh words. So in that sense it basically ensures that participants are paying attention. They are

responding correctly and they are doing this sincerely. And these filler trials basically you don't even a lot of times

you don't even analyze them or just sometimes you analyze them to make sure that the you know phenomena that you're

measuring in your critical trials is actually being uh you know uh followed. So these are some of the methods in

which you can uh address uh you know response bias in experimental task and uh going forward we'll discuss some

different kinds of tasks that are available. So uh we'll broadly discuss these four kinds of task freedescription

task, rating skills, force choice task and specialized multiple choice task. So in this lecture I thought that we'll

very briefly touch upon these two kinds of task the free description and the rating skills. Let's talk about them.

Now, free description task as I was just saying you know these are qualitative tasks that are basically designed uh to

test the belief and the opinion of your participant. Okay, they're basically asking them to provide explicit

word-based answers to your questions. The questions are typically presented in a written form and participants normally

perform this experiment in the absence of the experimental so that there is a slightly higher feeling of anonymity and

assurance and the participants are not reactive but they are naturally giving your giving you the responses. Okay. Uh

the responses can be written uh mostly they are written sometimes audio or video tapes can also be made and can be

analyzed obviously after taking enough uh you know uh the considered uh permissions and assuring the

confidentiality of uh uh you know this whole enterprise. What kind of questions can we ask

through these tasks? So typically they these tasks are best at uh answering broad general questions that seek broad

general sometimes vague answers. Okay. Newer experiments, exploratory experiments typically would use these

kinds of tasks. Let's take what kind of questions they could be. So, uh if you have a question like what words would

people use to describe this uh you know a given facial expression. So you can give them three four five facial

expressions and ask them okay give me a word for this. And here you're not restricting the range of responses. So

somebody can say word one word two say somebody say I'm showing a smiley picture. Somebody can say happy somebody

can say ecstatic. Somebody can say joyful. you're not restricting the kind of responses that are coming from the uh

participants. Uh or you can ask for example what do people first notice in this display and again you're not

leading uh or presenting options already. So you are basically allowing your participant to tell you whatever he

or she might feel. In that sense you sort of get a broader answer. You get a obviously vague but you get a more

general answer which will tell you what the participant thinks or feels about this given stimulus.

Okay. So what kind of tasks typically are considered in the free description task uh quality? We've talked about

interviews, questionnaires, we talked about long answer, short answer and partial report kind of uh things. I'll

not discuss them again in detail because I think we've spent enough time in interview structured and semiructured

and unstructured and so on. We've talked about questionnaires also. So we've talked and uh long answer, short answer

and partial report are basically types of questionnaires where uh wherein there is a difference in the kind of answers

you're telling uh you know you're expecting from your participants. So in some case you can just ask them to write

whatever they want. This is a long answer format. Sometimes you ask them to write within any amount of words. So

within 100 words answer this within 50 words answer this and so on. And partial reporters say for example if you

remember uh you know grid kind of presentation. So uh you know the experimentter presents a 3 + 3 or a 4 +

4 grid and shows a few numbers and then ask the participant to partially report. So just tell me the first row, just tell

me the third row, just tell me the second column, something like that. So these are some of the ways in which you

can get responses. Again, you're not restricting the kind of responses the participants can give. And these are

useful tasks and they'll they'll tell you broadly about how the participant is feeling or having an opinion about a

particular phenomena. Now other kinds of important task is the rating scale task. For example in a

rating scale task typically what happens is that the participants will place a numerical value on some aspect of each

stimulus. So you you can give the participants a different stimula. You can say on a scale of 1 to five uh you

know uh how you know how beautiful uh these 3 four five pictures are and then you can basically ask them to you know

assign a number and then uh you know accordingly you get your responses. You can also do something like okay uh let's

say if uh picture A is uh you know uh let's say in in terms of some measurement is bright is uh you know 20x

bright let's say that's a unit how bright is the second one or the third one or the fourth one. So somebody could

say oh uh second one is twice as bright as first or half as bright as the first one and so on. So these basically uh you

know are tasks that will be that you know you assign a value to some aspect of each stimulus and this value can be

used to determine how does that stimulus compare with along the same dimension to other kinds of stimula. If you want to

conduct a rating scale task that aspect of the stimula that is to be rated must be explained to the participant. So if

you want to talk about symmetry for example, you have to explain what symmetry is and then basically ask the

participants to tell you how symmetrical do they think pictures 1 2 3 4 5 are okay or bright or beautiful or how fair

this judgment is something like that. You'll have to tell them exactly what dimension of these different stimula you

want your participant to uh you know consider and rate accordingly. Okay. So rating tasks are designed to give

insight into how the elements of a class of stimula paintings uh you know cities expressions etc vary along this given

dimension. So sincerity aesthetic value size and so on. Okay. What kind of questions can be asked here? Say uh what

are people's preferences among the following participants? Uh among the following paintings or how similar are

these expressions on a scale of 1 to five? How similar do you find painting five to be to painting one? How sincere

do people think these expressions are? On a scale of 1 to 7, tell me how sincere these expressions are where one

is uh not sincere at all and seven is extremely sincere. Now here you are basically asking them uh to choose a

value between uh you know on this given continuum and that is what the participant will tell you. So there are

these three four variants of this which I wanted to very briefly discuss is ordered ranking. It's a simple rating

task. In essence participants are simply asked to list the simile uh in order along a particular dimension. So you

have seven stimuli, rate them in the order of aesthetic uh pleasantness. You have five simul rate them in the order

of size. Uh things like that. Okay. So once the dimension that you want your participant to attend is clearly

defined, then the stimula are presented to the participant typically all at once and then the participant is asked to you

know place the stimula in order. So ordered ranking uh rank 1 2 3 4 5 that you know like that. Uh sometimes you can

ask people to uh you know estimate the magnitude. For example, uh you know rather than being given a fixed number

of labels, participants are asked to assign any number that they want to the stimula. Uh these numbers should

represent a more or less intuitive uh indication of precisely how much of the dimension that particular thing has. Say

for example, how beautiful is this? So somebody says, okay, 100 units. All right. So if this is 100 units

beautiful, how beautiful this second uh picture is or how beautiful this third picture is. And then somebody will say

okay this is 100 and this is 200 and this is 50 and this is 40 and something like that. All right. So the going

forward the value first stimulus is typically assigned a value and the second third fourth are all judged in

reference to this first stimulus whatever value it has been given. This third one is the liyker scales that I

was saying. So liker scales typically are 1 to 5 5 1 to 7 1 to 11 uh kinds of scale where one disco 1 and five are two

ends of the continuum. So uh on the basis of say for example we did you know a scale uh in in terms of uh you know on

the scale of 1 to five how easy it is to uh you know imagine uh you know a given stimulus. So for example imageability is

what we were trying to see. So how easily you can bring to mind the image of this given stimula. So one is not at

all five is extremely easily or seven is extremely easily or say for example you can say on a scale of 1 to five how

difficult is this expression or easy where one is extremely easy and seven is extremely difficult. So this is similar

to magnitude estimation except now experimental already defines the scale. It gives you a fivepoint scale or a

sevenpoint scale and on this scale you are asking your participants to make responses. All right. Then you have

semantic differentials. For example, it's it's just a variation of the liquor scale paradigm. Uh where the end points

of the fixed range of numbers are assigned to bipolar opposite terms. So this is basically uh you know a variant

of the liquor scale which I saying. So good, one is good, five is uh bad and on and and the three say for example and

that is why typically you use odd numbers oddnumbered scales for this where you have a natural midpoint where

the participants say okay it is three it's not good not bad something like that all right

so this is just the two kinds of task that I wanted to talk about today I will continue this uh uh you know uh

discussion in the next lecture where we talk about the force choice and the specialized force choice uh kind of task

before we move on to stimula and other discussions. Thank you.