Residual Reading Time for Self-Paced Reading Experiments

[Michiel Spape]

Hi,
I see, it sounds quite useful! Anyway, so what you need is
A) reading times for various words; I don't know how you obtain these exactly, but I guess from E-Prime, before the rest of the experiment starts or something
B) average reading times for each word-length
... such that you can calculate by means of simple linear regression the 'line', and therefore, later on, the residual.

What exactly is the problem? You just make something like an array (or a dozen of variables: sum2, sum3.. sum14) where you keep the sum of each reading time for each wordlength, then after all reading times are known, you divide them by the amount of reading times encountered, and then you calculate a linear regression. It's a bit of work, I grant you that, and if you're not brilliant with statistics, you will have to look up how these work exactly (I think there are decent articles with examples on Wikipedia), but I think you're right in saying that it makes sense to do this within E-Prime rather than afterwards. How far do you get, though, and where do you get stuck?

Best,
Michiel

 
Michiel Spapé
Research Fellow
Perception & Action group
University of Nottingham
School of Psychology


-----Original Message-----
From: e-prime at googlegroups.com [mailto:e-prime at googlegroups.com] On Behalf Of Dwivedi Lab
Sent: 29 June 2010 16:00
To: E-Prime
Subject: Re: Residual Reading Time for Self-Paced Reading Experiments


Residual reading time (RRT) is a way to correct for sentence length,
word length, and individual differences between participants' reading
speeds. For example, a sentence with five words is read faster than
one with 10; a sentence with 10 long words is read faster than one
with 10 short words; etc. Since we're using a self-paced design,
naturally some people will read and progress through the sentences
faster than others. By calculating RRT, we can eliminate this bias.
First thing we do is plot all of the raw reading times against the
number of characters per word for each participant and create a line
of best fit. This line represents the average speed that that
particular person reads depending on the number of characters (ie word
length). The general trend is that as the number of characters
increase, so does reading time. We do this for every participant,
resulting in average reading times that are specific to every
participant. Using the formula of the line of best fit, we can
determine the average reading time for words that have 2 characters, 3
characters, 14 characters, etc. From there we compare the actual raw
reading time to the average value. The amount by which the raw value
deviates from our calculated value (ie the line) is known as the
residual reading time. These values appear as plus or minus values
(+/-).
This way, two participants may read the same word at completely
different speeds, but now we compare that value to their own reading
pace in order to determine whether they're taking longer to read that
word or not instead of comparing it to a group average which is less
accurate.

Our method for calculating RRT was adopted from the following article,
explained in the appendix: "Semantic Influences on Parsing: Use of
Thematic Role Information in Syntactic Ambiguity Resolution" Trueswell
and Tanenhaus, 1994

Hopefully this makes sense


On Jun 29, 10:24 am, Michiel Spape <Michiel.Sp... at nottingham.ac.uk>
wrote:
> Perhaps you can start by saying what residual reading times are?
> Sorry, I'm more used to cognitive fields than psycholinguistics (I assume residual reading, as a concept, is commonly known there?)
> Best,
> Mich
>
> Michiel Spapé
> Research Fellow
> Perception & Action group
> University of Nottingham
> School of Psychology
>
> -----Original Message-----
> From: e-prime at googlegroups.com [mailto:e-prime at googlegroups.com] On Behalf Of Dwivedi Lab
> Sent: 29 June 2010 15:00
> To: E-Prime
> Subject: Residual Reading Time for Self-Paced Reading Experiments
>
> We are doing a series of self-paced reading experiments and need to
> calculate the residual reading times for each individual participant
> then combine all the participant data into one group file for
> analysis. So far we've been importing the data into excel and
> manipulating it there in order to get residual reading time values. We
> have created a series of excel files to streamline this process,
> however these files are prone to error and end up taking more time to
> fix than they actually save. Is there a way to calculate residual
> reading time data from within E-Prime? Or is there a faster, fool-
> proof way to calculate residual reading time in general?
>
> Any help is appreciated!
>
> --
> You received this message because you are subscribed to the Google Groups "E-Prime" group.
> To post to this group, send email to e-prime at googlegroups.com.
> To unsubscribe from this group, send email to e-prime+unsubscribe at googlegroups.com.
> For more options, visit this group athttp://groups.google.com/group/e-prime?hl=en.
>
> This message has been checked for viruses but the contents of an attachment
> may still contain software viruses which could damage your computer system:
> you are advised to perform your own checks. Email communications with the
> University of Nottingham may be monitored as permitted by UK legislation.

-- 
You received this message because you are subscribed to the Google Groups "E-Prime" group.
To post to this group, send email to e-prime at googlegroups.com.
To unsubscribe from this group, send email to e-prime+unsubscribe at googlegroups.com.
For more options, visit this group at http://groups.google.com/group/e-prime?hl=en.

-- 
You received this message because you are subscribed to the Google Groups "E-Prime" group.
To post to this group, send email to e-prime at googlegroups.com.
To unsubscribe from this group, send email to e-prime+unsubscribe at googlegroups.com.
For more options, visit this group at http://groups.google.com/group/e-prime?hl=en.