Mastering Reaction Time Studies in Cognitive Psychology Experimental Design

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

department of cognitive science at IIT Kpur. This is the week seven of the course and we are now uh at the level

where we are talking about methodologies. In this week I will talk to you about behavioral uh

methodologies. Uh I will start with reaction time studies today. Now the study of time relations. So

let's let's basically get a little bit of a background of what reaction time studies are are about because they are

the most fundamental experimental method that are used in cognitive or experimental psychology. So whatever

method you're using if you're using uh you know reaction times you can basically add reaction times to all

kinds of studies. So even if you're doing an fMRI study or an eye tracking study or any EEG study, you will find

that typically the first and the most basic task is the reaction time task. The reaction time task uh is the most uh

reliable and one of the most oldest tasks to basically figure out what kind of mental processes go in when we are

performing a mental function. So we'll give you uh some background to this and we'll take some examples. Then I'll move

to the history of reaction time tasks and then finally we'll give you some practical suggestions on how to

construct reaction time studies. So these will be the lectures 31, 32 and 33 uh for this week. All right. So let's

first start with a bit of a background about what reaction time studies are and what are they basically about. All

right. So uh this is a quotation from Joseph Chalro. Uh let's look at it. The study of the time relations of mental

phenomena is important from several points of view. All right. Uh first it serves as an index of mental complexity

giving the sanction of objective demonstration to the results of subjective experience. So whatever a

person is going through whatever mental process for example you show somebody a picture of a house versus you show

somebody a picture of a face whatever mental processes uh you know start from the time that stimulus is shown till the

time you ask the person to make a response. Let us say uh you've shown a face and you are asking the response for

whether the face is a male face or a female face. Or if you've shown a house and you're asking the question that you

know press this key if this is a man-made object, press this key if it is a you know natural object. So uh

starting from the uh exposure to a given stimulus and obviously uh we have seen in experimental psychology we make

stimuli uh vary on particular critical uh criteria that is typically our independent variable manipulation.

uh reaction time studies are basically uh you know designed to uh what they want to do is they want to uh understand

the chronology and the time course of what goes on when a particular stimulus is exposed to a participant. So that is

basically what it says. It says it serves as an index of mental complexity giving the sanction of objective

demonstration to the results of subjective experience. So given how complex the stimulus is. Say for example

if it is just a single word then the complexity is less uh it is quickly sort of uh you know converted to sound to

meaning and the person is able to respond to whether this is a word or an on word in a lexical decision task or

the person is ready to name it or the person is ready to tell you whether it is an inanimate object, animate object,

concrete object, abstract object any kind of things. The more complicated the stimulus is, the more complicated the

corresponding mental processes will be. And in that sense, the corollary is that the more complicated the mental

processes are pertaining to a given stimulus, the more time it will take. So the reaction time studies uh

philosophically are designed to tap into the mental processes that will start once the stimulus is shown and depending

upon how complex that those mental processes is the time will increase. That is the broad assumption.

of the second part is the sanction of objective demonstration of subjective experience. So whatever you are feeling

whatever competitions are going on in your head are obviously uh subjective and they are sort of happening at your

own pace but an objective demonstration of where you are at with that specific mental process or with those specific

mental processes can be tapped in by asking the participant to respond within a given time span. All right. Also it

indicates a mode of analysis of the simpler mental acts as well as the relation of these laboratory products to

those uh of daily life. So whatever uh you know you are asking your participant to reproduce in the laboratory. Say for

example you've uh shown them a you know uh a list of words and you're asking them whether they are uh you know real

words or not real words or you've shown them faces. you were asking there whether it is you know uh a male face,

female face, happy face, sad face any kind of those decisions. Uh these simpler mental acts of sort of

calculating these things they basically uh are the mental uh you know they are the laboratory products that are

happening in the head once the participant gives you the reaction time it basically becomes uh relatively

similar to how they would perform that task in their daily lives in in outside the lab. Okay. uh also reaction time

studies demonstrate the close interreation of psychological with physiological facts and uh and an

analysis of the former uh is shown dis indispensable to the right compensation of the latter. All right. So unless you

are able to analyze uh you know the psychological processes you will not be able to relate them to the physiological

reaction. See giving a response in a reaction time study is a physiological phenomena. So you your brain should make

a decision. That decision is uh to be uh you know conveyed to the afrant muscles. The muscles will decide when to respond

or which keys to press. There is a relationship between what is psychologically happening when you're

evaluating a given stimulus versus what is physiologically transpiring in the way you're giving the response. It is a

fast response, slow response, it is a physiological response, it is a reaction time response, all of that. So if you

look at this quote, it basically says there are three things. First is uh reaction time studies or these

demonstrations serve as an index of mental complexity. They are a mode of analysis of what kind of steps are going

on when a particular process is taking place and it inter it basically demonstrates the interreation of the

psychological processes that go on with the physiological processes that basically are responsible for giving

responses. Also uh it suggests the it suggests means of lightening and shortening

mental operations how fast and how quickly certain mental operations can take place and thus offers sometimes

mode of improving educational methods. Also it promises in various directions to deepen and widen our knowledge of

those processes by the complication and elaboration of which our life is wonderfully built up. So basically

remember in the beginning I've talked to you about atomization of uh mental processes. What are the steps that go in

to make a decision? What are the steps that underly a particular behavioral response? So all of that basically is

very nicely demonstrated in this course by Joseph Jastro in his paper the time relations of mental phenomena. basically

says that these time relations basically mirror a lot of what goes on in the brain when the brain is evaluating a

particular stimula complexity uh you know the chronology chronometry and also in that sense it basically tells us it

gives us an estimate of what kind of building blocks are there behind a given response and how can we sort of increase

the steps decrease the steps how can we basically and to what level we can describe and deconstruct the processes

that go on uh between a stimulus and a response. Now let's talk a bit generally about

reaction time studies, reaction time research. Now time is is a very important variable when one is talking

about uh you know measurement of mental processes. Remember uh experimental psychology is all about measurement of

mental processes. It is different to that of qualitative psychological methods where you let the person sort of

be and decide and they describe whatever they are feeling at their own pace. In experimental psychology research, you

want these processes to transpire within a given moment of time and you want to sort of know how much time each aspect

of that decision has taken. Physiologists and experimental psychologists have long been interested

in how the mind works and they have examined mental processes by measuring how fast people were able to respond to

a stimulus or uh to perform a task. So the idea is as I've been saying you give a stimulus you basically let the person

uh respond to one aspect of the stimulus and you calculate how much time the person took in responding to that of

aspect of the stimulus or say for example you give a task such as a task switching uh scenario you basically take

measure how much time that the person takes to switch from let's say task A to task B this research is therefore uh you

know a lot of times it has been referred to as mental chronometry basically measuring or timing the mental mental

processes that go on. Now, reaction time research refers to any empirical study in which a reaction

uh in which a research question is answered through the measurement and the analysis of the time individuals take in

responding to a stimulus or performing a task. So, basically reaction time studies is all about measuring the time

a person takes to perform a given task or to react to an aspect of a stimulus. So for example, you have uh words that

are rhyming with each other and words that are not rhyming with each other. How much time does a person take to uh

respond to two consecutive rhyming words versus two consecutive non- rhyming words? And if you subtract that and I'll

come to subtraction and other things later. It'll give you an estimate of how much time does a brain uh take up to

compute the effect of rhyming. Okay. So that's broadly what we are doing. A task for this a task for me measuring

this mental uh you know timing uh is typically it is very simple as I was just saying. So for example in a lexical

decision task which is one of the uh you know very commonly used tasks individuals are asked to decide if a

letter string is a word or not. So they present we typically present people with a string of letters and we ask them

whether it is an actual word or not. participants would uh usually respond by pressing two buttons. One uh for if it

were a non-word and the other for if it were a word. Okay. Now this is one kind of task but you can have any number of

tasks. You can basically ask if it is rhyming with the previous word or not. If it is a word indicating an animate

object or an inanimate object. You can ask any kind of questions. The person basically is expected to evaluate the

stimulus and you know uh bring all his knowledge on to sort of evaluating that stimulus and given that this is

typically time bound. So uh the first impression is the quickest judgment is typically tapped into. Okay. So whatever

task you're using the researcher is basically interested in measuring how fast a person is going to respond to a

given stimulus. Now this timing that a person takes is referred to as reaction time. And reaction time is typically

measured in milliseconds from the onset of the stimulus to the point of time when a response is given. So basically

yeah so the time a person takes to respond to a stimulus is typically referred to as reaction time. It is

usually measured in milliseconds and this mill second measurement starts from once the stimulus is projected on the

screen till the time the person presses the key. So the time from the onset of the stimulus till the response that is

basically what we are measuring that is what is referred to as reaction time. Now reaction times basically our primary

data for exploring the research question in reaction times research. Uh although a lot of studies also use error rates or

accuracy information as well. Now the use of reaction time data is typically based on the premise that cognitive

processes take time and by observing how long it would take individuals to respond to different simile or perform a

different task in uh perform a given task in different situations that is the information that reaction time tasks

that the reaction time data will provide us. We can basically create our questions. We can sort of adapt our

research questions to you know trying to answer how is the mind working. How is the mind evaluating a given stimulus?

What aspects of uh you know the stimulus are participating in the decision-m process and on the basis of that we can

sort of decipher or infer mechanisms that would have gone into evaluating a given stimulus. Say for example in a

lexical decision task the understanding is that a person would convert the printed form to a sound and then the

sound will be mapped onto the meaning which is probably already there if it is a known word in the mental lexicon that

is your mental dictionary and this process of consulting the mental dictionary to check whether a presented

letter string is an actual word or not that is the process which is called lexical access. So when you're doing

lexical decision task, the researcher is interested in knowing the time it has taken you to perform lexical access.

That is the time that we are interested in measuring. And uh the time of lexical access depends on a variety of

parameters. It may depend on uh frequency of the word. It may depend on the length of the word. It may depend on

the uh task of the word. Any any of those things. So for example you uh get uh you know uh

participants uh to your lab and you ask them to uh you know produce a set of first language words. Okay. So for

example my first language is Hindi. I go to a lab and I am asked to produce let's say uh within 1 minute produce as many

first language words as you can. So I go on saying amapita and so on. uh or you can ask me that okay you produce uh uh

second language words you produce uh you know words in your second language for me that is English. So I start saying

tablechair uh desk uh uh this that and so on. Now the thing is what I will do as a researcher is I will note the time

that I that the participant has taken for producing first language words. I will note the time that the participant

has taken to produce second language words. I will subtract them in some sense and I'll basically get an estimate

of how much more or how much less time it has taken me to produce the words in a given language and then there can be

other things I can look at which topic the words were on what kind of prompts were given to produce the word and so on

but the idea here is uh if I am the researcher I'm interested in calculating uh what took me so long what basically

contributed to whatever time I have taken to perform this task So on comparison of naming times in the

example, our intuition will basically tell us that the second list should take longer because that is the second

language. I'm you know as most people I'm probably poorer in my second language than I am on my first language.

Does not mean I'm uh you know how poor am I on second language but at least compared to the L1 I should be

relatively less uh you know good in the second language. So why is this happening? Is there a proficiency

difference? Are the words of the second language more difficult to pronounce? What is the case? So basically once you

get this time difference, you want to know why this has happened and that is what you are as a researchers interested

in. Okay. And that is basically the second part of the question. What is going on in our mind during that extra

time? So let's say I take 400 milliseconds to produce words of the first language and I took 450

milliseconds to produce words of the second language. As a researcher, I would be interested in knowing where did

that extra 50 milliseconds go. What really happened? Was it difficult to uh recruit the words from the second

language? Uh were the second language words of higher length, had more syllables, etc. Were the words of the

second language uh not being you know I was not able to produce because I have a smaller vocabulary in my second

language. Basically you create situations to create these time differences and then you create other

situations to understand where these time differences have arisen from. Now a longer reaction time so that's

what we will continue a longer reaction time in one condition than in the another may reflect the involvement of

more mental operations. Okay. So if you've basically taken more time in task uh two as compared to task one, it

basically tells the researcher that there were more steps in task two as compared to task one. Okay. Also a

higher level of complexity of operations or a higher degree of difficulty encountered by the language processor.

So wherever that extra time has gone, it could either be because it involved more mental operations, uh they were more

complex or they were more difficult. So that is basically what you would want to sort of now uh piggyback on and that is

what you would want to analyze but why did this person take more time in this task as compared to the earlier task.

This kind of reaction time data when it is collected under adequately designed experimental conditions can actually

shed light on what kind of linguistic knowledge knowledge individuals have and how is such knowledge put to use. So for

example, what are the processes that went in? Why did somebody take say for example, we can also come across

participants who performed the second list of second language words faster as compared to first language words. And

then we can analyze that also. Why did that happen? Uh are these perform are these participants systematically

different from the earlier participants who were taking more time in the second language list. So that is basically the

kind of comparisons you would be expected to do and you would be expected to find out why these kind of time

differences appear in the first place given that the experimental list was well controlled the task was done in

well-controlled conditions. There were no other confounding or extraneous variables that were playing a part here.

So in most reaction time studies the focus is not only to measure the absolute speed or rapidity of a person

that you know in performing a task. The focus basically is on how fast an individual performs the same task in

different conditions. So basically the task was naming words. Uh but say for example how fast was I to name words in

my first language as compared to second language or how fast was I to name the words in my second language as compared

to my first language. So two or more conditions are created in an RT experiment by varying the stimula or the

task or the participant systematically. So if you were interested in a between participant design, you could have uh

people who are uh good in second language versus people who are not so good in second language and you

basically have both of these groups uh produce words in L1 and L2 uh and then compare that uh based on the difference

in proficiency what kind of time differences in naming L2 words does emerge. So by examining these kinds of

variations and how these variations affect the reaction times, we can basically make inferences about the

cognitive processes or operations that are involved in language processing for this particular example, but it could be

very simply a perceptual task. It could be an attention based task like a visual search task. It it could be anything.

the difference in the time that the participant takes in different conditions of the task uh in responding

to different stimuli or if there are systematic differences between participants. These three will give us

clues about why that extra time was taken and as researchers you would be interested in knowing why that extra

time was actually needed. Now let's talk a little bit about important uh you know characteristics of

time research. The most important is as the name suggests it is reaction time research. So timing is obviously most

important accurate timing of both presentation of the stimulus and measurement of the responses that is

absolutely paramount. Okay. So the most basic and essential requirement of any reaction time research is accuracy in

timing. Millisecond accuracy is required typically uh you know in both the display of the stimulus. So when does

the stimulus onset and uh when does the response measured and accurate timing in stimulus display has has broadly around

two components. First is duration. How long the stimulus is there. Say for example a lot of studies want the

stimula to be presented for a fixed duration. So when we are conducting laterality studies uh we know that the

time it takes uh you know the eye to make a sakad from the center of the screen to the left or to the right is

around 185 you know 90 uh milliseconds. Walker and Maxoli in their study in 2006 says it's around 200 millconds. So

you'll typically see that uh you know uh studies that are testing for laterality present their stimula under that 200

millcond. I did a lot of experiments where where I presented my stimula for around 175 milliseconds. Uh somebody did

180 millconds. Somebody can also do 150 millconds. When you want to study uh laterality and you want the participants

to keep looking at the center of the screen and not move their eyes either side, uh it is uh uh actually pertinent

that the stimulus goes away before the eyes can actually move. All right. So the accurate timing in stimulus display

is absolutely important for how long the stimulus is there and also when does the stimulus onset. So for example in a

typical uh trial there is a blank screen there's a fixation cross and there is the stimulus how quickly did the

stimulus come did it exactly come at the time you intended it to or say for example there were some uh properties of

the display that led it to come a little bit later because then it will sort of snowball into uh errors and

miscalculation in the overall timing of the instrument of the experiment as well. Okay. So uh in lexical decision

task for example a word may be displayed for 500 milliseconds in one experiment and in priming studies typically it's

around 50 milliseconds in other ones. Okay. Now this duration how long you want to present your stimulus is

determined by several factors. Under circum some circumstances accurate display duration is extremely crucial

for the interpretation of results. So for example in a laterality study if you have presented a stimulus more than 200

milliseconds you will not be able to be sure unless you've added uh let's say an eye tracker measurement and so on you

will not be able to show if the participants had moved their eyes or not. If your participants have moved

their eyes then you will not be able to measure whether it was basically uh you know uh

processing of the stimulus in uh in paraphobial vision. That is what that is basically what you want to measure in

your latality studies. Also another example for example in a mass priming study the prime has to be presented long

enough so that it can be processed but it should be brief enough so that it is not consciously processed. When you're

doing priming studies it is very important that the stimula is presented for just the right amount of time.

This is just an example. So fixation cross 200 2,500 millconds. Then you have a forward mass. Then you have the prime,

you have a backward mask, then you have the prime target interval and then there's the target. Now basically what

you have to do is you have to decide on the target and you are hoping that there will be some effect of the prime on the

target. Now for how long the prime is present basically will determine the success or the failure of your priming

uh manipulation. All right. Another element in uh accurate stimulus display is onset timing which refers to

when is the stimulus presented at what point in the trial the stimulus appears. This applies to experiments uh in which

a trial consists of more than one stimula. Say for example if a trial has multiple stimula coming. So typical

trials will be blank screen fixation cross stimulus comes you give a response. Some kinds of trials just like

we saw the masking trial uh uh you know the mask priming trial uh there is uh blank screen fixation cross prime then

again uh a you know a gap and then again the stimulus. So when in this uh you know sequence of uh events in a trial

does your actual critical stimulus pop up. For example a very good example here is the picture word interference

paradigm. Now here a trial consists of the presentation of a picture and then there is a word. Now the exact onset

time when does the picture come and when does the word appear basically uh moderates the interaction between the

picture and the word. If the word is presented too early or the word is presented too late it will have

consequences for how does that interference go. So a strong interference effect is shown to occur if

a semantically related word is presented 150 milliseconds before the onset of the picture uh but not when it is presented

150 milliseconds after the onset of the picture. So uh for semantic interference to occur even before the picture comes

the uh you know the uh interfering uh word should have been presented. uh if the picture comes and the interfering

word has come after then that semantic interference might not take place. Look at this. So for semantic interference

the uh basic uh you know the word should have been presented a little bit earlier so that it can react with the meaning

calculation of the picture. For phological thing the word has to come much later uh than after the picture. So

in this case if the word is coming later and the stimulus on synchrony is positive then you can affect phological

interactions. If the word comes earlier then you can expect semantic interaction. So here you can see very

easily that how does the timing sort of uh you know of presentation of the second stimulus actually have a bearing

on the kind of mental competitions that are going on. Also, accurate timing is required uh in

the measurement of a participant's responses as they are the primary data for answering the research questions

that you're asking. A participant's reaction time is typically measured from the onset of the stimulus to the point

of time when the response is provided. And this is usually done through a computer clock. So every time you sort

of press a key, the computer is anyways, you know, it's it's working its time and that is a time gets response. So the

timer turns on with the onset of a stimulus and stops when the response is received. Accurate reaction time

measurement depends on knowing when to start timing and also the method that you're using for uh reaction uh for

recording reaction time. So a lot of times you'll see when we are doing reaction time research, we would like to

have instruments measuring reaction time which are which have a millisecond level resolution. Sometimes keyboards actually

present a lag. So a lot of times you'll see people use response boxes, they sometimes use joysticks and other ways

to collect responses because they don't want the lag because sometimes and uh you know uh I do a lot of literality

studies and literality studies uh time lag of 5 milliseconds is is actually a lot. All right. So you want the reaction

time calculation to be on point and you want that that there should be no lag between the participant pressing the a

key and it getting recorded by your computer. So according to the testing done by

Foster and Foster uh input devices such as a game pad, a joystick and a mouse offer much better accuracy than

recording uh responses that are gained on a keyboard. Now another thing that we have been

talking about so far and in much detail is the rigorous variable manipulation and control. We've talked about how it

is extremely important that whatever uh you know we are sort of tapping in our responses is only due to the

manipulation in the independent variable and there are no extraneous or confound variables in play. So in reaction time

research participant experimenters are extremely careful in isolating the phenomena under investigation from all

other related uh factors. For example, one may have a feeling that the number of translations in the learner's first

language than an L2 word has that an L2 word has would affect the time it takes to translate the L2 word. Say for

example, you have a word in your second language which has many words in your first language. So the number of

translation this L2 word has in your L1 uh when you have to translate it, it will have to say for example, if there

is just one translation, you'll pick that and deliver it. If there are three possible translations, you'll pick

amongst the three and it'll take you more time to come up with an accurate translation. So this should basically,

you know, have some consequence in the time you take to translate a given L2 word. Now you can believe that more L1

translations the L2 word has, the longer it will take uh to translate this word because multiple L1 translations will

compete with each other. In such a scenario, careful uh steps would be taken uh to ensure that the single

translation words only have been selected and words in multiple translation condition should be

separately uh you know presented in research design terms. What you would want to do is that the process of

identifying L2 words with one or multiple L1 translations and you then divide them in two steps is basically

referred to as variable manipulation. So you basically let's say we are interested in uh the time it takes to

produce a translation uh depending upon whether there is multiple translations or single translations. So you would

basically block it. You'll separate the L2 words which have a single translation. And you'll separate L2

words which have multiple translations and you will basically get the reactions to them in separate blocks. And then

you'll compare how much time it took for single translation words to appear and multiple translation and you'll

basically subtract saying okay because this word had three translations it was three times as slower or faster as

compared to this word which had a single translation. That's basically the you know arithmetic mental arithmetic that

one would carry out. So rigorous variable manipulation and to be actually sure of what are the entailed processes

for a given stimulus and for a experimentter to be able to foresee them is extremely important. So rigorous

variable manipulation and control is the key to relating a set of reaction time data to the phenomena such that the

former can actually tell you about mental process. If you are if you have no idea about what are the mental

processes that are going to take place when you present a word when you present a translation word then you will not be

able to estimate what really went wrong and what is it that your arties are composed of. So as researchers we do a

lot of literature search we basically go through several studies and we try and understand what is it that the reaction

time is composed of for a given experimental task. also time-sensitive assessment of

behavior. It is also very important. An important feature of reaction time research is that the data are collected

within the narrow window of time. Uh when the cognitive processes is are still going on, it's not like after they

have finished and then you are measuring because after that a lot of other mental processes will have occupied the uh

mental space and the time that you are measuring is will not be of that specific mental operation. Okay. So

standard part of the instructions of the reaction time is to emphasize that participants have to respond as quickly

as possible because you don't want over uh deliberation on the stimulus. You don't want the participants to think of

oh this option that option and so on the first option that comes and the participant has to press the key. Also

transient display of stimulus and feedback on speed uh often help reinforce their rapidity requirement. So

typically you will see at least in practice uh blocks uh experimental psychologists provide feedback. Oh your

response was so fast or so slow or correct or incorrect so that the participants uh who are in the

experiment can also moderate uh what kind of strategies they are using to respond to your stimulus. This emphasis

on immediate and quick response is actually very important and the use of tasks that make such quick responses

possible and then the collection of timesensitive data. These are some of the uh you know elements of reaction

time research that work to maximize the chance for the observed data to reflect moment by moment unfolding of cognitive

process. Say for example I give you uh you know a stimulus that is just an object and then I gave you a stimulus

that is hand manipulable object. So this is object nonobject and this is tool non-tool. Uh the difference between an

object and a tool is that the tool can be hand manipulated. Say for example a hammer, it can be used to create effect.

If there is systematically more time taken in responding to tools than to other objects which are also man-made

objects and basically resembling in every other feature to these tools, then we know that the extra time that these

tools have taken is the computation or simulation of manipulability and the simulation of uh action with those

objects. That is if if that is the level at which we design our task and that is the level to which we are very clear

about what our stimulus entails only then we will be able to know exactly what kind of time uh or mental

operations have gone on in a given task. So as Tyler points out, fast responses fast response tasks tap the listener's

representation of the input as a specific at a specific moment in time. Given the input available to the

listener when he or she makes the response, it is then possible to infer what types of analysis must have been

performed upon this input to produce the effects rendered in the response. And that is basically what we are interested

in. Now many reaction time tasks are referred to as online tasks. So when a participant is actually performing the

processing and then you are getting the response these are said to provide online measures of cognitive processes

because they help reveal what is going on in the participant's mind while a cognitive process is unfolding. So what

are the steps? What kind of computations are going on that is why these are called online tasks. Okay. All tasks

that require fast responses and produce reaction time data can be considered online task because you're uh computing

and responding at that same time and the response is is in some way tapping on your online computations that you're

performing in response to that stimulus. So with the understanding that different tasks differ in the degree to which

reaction time data reflect what analysis is being done and what say for example what linguistic knowledge is being

involved or what mental representation is being built at a given or a particular point in time.

Also something that is extremely important in buildup of reaction time task is the stepby-step progression. It

is very important to basically know exactly what steps go on in the analysis of your stimulus. Say for example a

particular stimulus might require steps 1 2 and three to give a response. Another stimulus which is slightly more

complicated might require steps 1 2 3 and four. So you would basically need to know what are the steps that are

involved in computing or in coming up with a response to a stimulus and that is what as a researcher you are

interested in. So any scientific inquiry represents a continuation of an effort to further one's understanding of a

phenomena one step at a time. So a lot of times you'll see experiments are designed in such a way that they can tap

into each specific step of evaluating a given stimulus. So this progressive nature of research of reaction time

research is basically very very important. It is not rare for example for a study to include a large number of

experiment. For example there were 10 experiments in Trabaso and colleagues's study 11 experiments in Topolinsky and

Stra's study and so on. There's typically why a lot of studies use many experiments is because reaction time

research aims to have rigorous variable control so that a target variable under investigation can be isolated and then

linked to the observed reaction time data. A lot of times you will see say for example feature search versus

conjunction search. The stimula are designed such that the procedure is designed such that you actually know

exactly what aspect of stimulus evaluation has led to extra time being taken. So this often means that only a

single or a very small number of variables are examined in any specific experiment. In other words, typically

reaction time experiments are very uh narrow and they're very focused. They are very unlike uh you know qualitative

experiments where the a large number of variable are allowed to play a part in how the participant is responding how

the participant say for example introspectively is telling you about what they have done. Now in order to

understand or in order to basically uh work on the generalizability of a finding in an reaction time experiment

say for example to understand whether a finding is restricted to the exact stimula that have been used exact task

or exact participant. It becomes necessary under several circumstances to create follow-up experiments to see uh

you know to what level does this really uh you know generalize uh what levels the same effects are actually

reproduced. So the follow-up experiments usually incorporate variables that were not initially considered in earlier

experiments and sometimes they involve a different type of material, another kind of participants or a different type of

task altogether because you'll see the number of tasks that will yield similar finding about a given stimulus is going

to tell you what aspects of a stimulus are tapped in a given task as well. Now the second reason is that uh studying

complex phenomena such as human language or human behavior perfect variable control is not always possible and is

not realistic under several uh you know experimental conditions. So in one set of experiments you will have control for

something in another set of experiments you're controlling for a different set of variables and after you've sort of

created uh several follow-up experiments each controlling one very specific part of that variable you will be able to

deduce the effect of the actual uh independent variables. So it is therefore it becomes difficult to

predict or anticipate all the relevant uh variables in any given experiment and how these variables may have affected a

result. Falloff experiments therefore uh have to be done to explore an unexpected finding or to test alternative

explanations. Remember the judgment of causality is also important because you want to rule out any possible

explanations. uh alternative explanations. You really want to be extremely sure that whatever results

you've got, they've basically there because of your uh very specific independent variable manipulation.

Let's let's take one study as the uh you know uh parting example. Now, Costa and Santa Savan they basically uh you know

created a study they where they wanted to uh you know explore a particular phenomena called asymmetry and switching

tasks and switching cost and this uh was first discovered by Muter and Alport in 1999. Now in the 1999 study bilingual

speakers were asked to name the pictures in two languages with the background color of the pictures telling which

language they have to name a given picture into. Okay. So sometimes this background will change. Sometimes they

have to name in L1. Sometimes they have to name in L2. L1 is your native language. L2 is the second language. Two

important findings came out in that study. First one was intuitive. The naming times were longer. When

participants had to switch from one language to the other, say for example on trial one, if I'm naming in English

and trial two if I'm naming in Hindi, the time from trial one to trial two will increase. So that is okay. That is

understandable. When I'm switching, I'll take some time. The second was slightly odd. It say it showed that switching

cost was bigger when a participant switched from a weaker L2 to a stronger L1 than from stronger L1 to weaker L2.

So when say for example I'm naming one one trial in uh English then the next trial in uh Hindi then I was taking more

time. On the other hand when I was naming in Hindi and then going to naming in English I was taking less time.

people would assume or let's say intuitive explanation would be that naming an L2 should be harder and

therefore going from L1 to L2 should take more time but here what is happening is going from L2 to L1 is

taking more time and people were puzzled and they were worried about oh what is happening here and uh and a lot of

experiments around that time in the 1990s early 2000s basically focused on this phenomena they were trying to

understand what is going on here so an explanation was rendered by uh green's inhibitory control and explanation and

basically what they said was that to overcome the suppression that they had accorded to the L1 overcoming that

separation and then naming in L1 was basically taking that time. So Santan reasoned that if this explanation was

correct the asymmetry should be smaller and eventually disappeared as the bilingual speaker becomes more

proficient in L2. when the see the explanation that green gave was because they are extremely uh proficient in L1

and less proficient in L2. When they suppress the L1 they suppress it with a lot of uh you know internal force

internal resources and to overcome that suppression takes time because there is a asymmetry between the proficiency in

L1 and L2. Now when the proficiency between L1 and L2 becomes almost equivalent then that suppression and

coming out of suppression should take less time and that is basically what these guys said. They said that when the

participant will become equally proficient in their L2 then the switching cost should sort of go away.

2004 they conduct a study to test this proficiency based explanation and uh it had six experiments. One was intended to

replicate the findings of asymmetry uh you know in switch cost among adult L2 listeners as they that was replicated.

Second was the main experiment in which they tested Spanish speakers who were highly proficient in Catalan as well

consistent with the prediction of proficiency based explanation. Here you see that the asymmetry did in fact

disappear. All right. So when the participants were equally proficient in their L1 and L2 Spanish and Catalin in

this case the asymmetry of switching costs did actually disappear. Third experiment was to uh address a

specific design issue of the earlier experiment uh that is using a small number of pictures repeatedly. So thus

instead of using 10 pictures they used 40 pictures and which were used as a critical stimulus. Now results

replicated those of experiment two suggesting that the earlier finding was not due to uh the you know more number

of features and so on. The proficiency based explanation was further examined in experiment four which involved naming

pictures in L1 and a much weaker L3 where the uh you know no asymmetry was found. In this series of experiments if

you see 1 2 3 4 5 6 a new finding emerged. Bilingual seem to name uh pictures faster in L2 uh than in L1 in a

mixed language condition. So in mixed language condition they were actually able to perform better in L2 as well.

This basically you know this example and how uh systematically the experiments are constructed to uh basically observe

what mental computations have uh taken place. It basically illustrates several things that are extremely important in

reaction time research. One often needs to uh replicate a finding if the purpose is to explain a finding or to test an

explanation. Okay. So how critically how comprehensively we build experiments is is actually very very important. All

right. So I'll stop here. I will uh uh continue this discussion in the next lecture. Thank you.