Disclaimer: English Kinda Thing

The sole purpose of the "English Kinda Thing" is to document my attempts to correct my own mistakes in standard English usage and to share the resources I find. In no way do I attempt to teach nobody English through these blurbs--just as I intend not to teach nobody to be a neurotic and psychotic handicap in Ratology Reloaded or Down with Meds! :-)

Wednesday, May 16, 2012

Free Resources for Learning a Language Online

Since I am sort of interested in doing some language learning ding... found the following link and would like to share it with ya....

elearn Magazine: 25 Free Resources for Learning a Language Online

Sunday, May 13, 2012

Memory and schizophrenia

Came across this interesting article... essentially, might be an article base on the dual-coding type of theoretical framework... looking at the plausible differences between schizophrenic patients and normal people memory-related performance.  Since many studies have investigated the topic from the perspective of auditory channel, this study also added the component of visual channel into the equation.


http://140.116.207.99/handle/987654321/92741



Regardless... what the abstract of this writing tells me is that... oops... my memory is fxcked-up-- well, ain't like I haven't complained about my bad memories.

An interesting thought that came to my mind, independent of this writing, was that... well... it is sort of like an issue of double-jeopardy...

What exactly might be the source of the perceived differences?

The condition itself or the side effect of the medication?

Would it even be a question worth asking since... well... psychotics need to be medicated anyways?

Not my business anyways... since it's already my job to be psychotic.  Leave some jobs for others to perform.  8-O lol


Wednesday, April 18, 2012

Eta squared for Two-Way Repeated Measures ANOVA (Mixed design)

The rule thumb for eta-squared effect size is .01 for small, .06 for medium and .14 for large according to Cohen (1988)

The calculation of Eta Squared for Mixed design based on SPSS output-- which is sort of different from the general ANOVA models- can be found here.

Essentially:
eta squared for within=(ss within effect)/(ss within total)
eta squared for between=(ss between effect)/(ss between total)

One thing worth mentioning is that-- say you have one between variable called group, you will see on your between-factor output 3 sources: Intercept, Group, Error.  The SS total for between factor= SS Group+SS Error.  Let go of the intercept.





Thursday, March 15, 2012

Convert partial eta squared to eta squared?

I went on a mission trying to figure out how I might be able to convert partial eta squared reported in SPSS to eta squared (thank God that my professors are 12 hour time difference away 8-X lol).

Thank God, the pursuit got cut short after I came across the linked article.

There is no way to convert partial eta squared to eta squared.  You have to go back to calculate the sum of squared total.

eta squared=SSeffect/SStotal

Partial eta squared=SSeffect/(SSeffect+SStotal)

Hope this helps.

http://yatani.jp/HCIstats/ANOVA

Tuesday, March 6, 2012

Training segmental productions for second language intelligibility

The study of Burleson (2007), native Mandarin speakers were  requested to practices minimal pairs of vowel contrasts.  The participants listen to a model pronunciation first before speak to microphone to have a computerized speech recognizer providing feedback on the their utterance.  Results of the evaluation study found significant improvements in participants' pronunciation. 

Interestingly, the author observed a drop in the pronunciation of the [i]-[I] contrast.  It was speculated that such observations might be due to the influences of the preponderance stimuli and participants might have overapplied what they have learned.

http://gradworks.umi.com/32/55/3255507.html

Friday, February 24, 2012

Creating a weighted average effect size

The following link gives you an idea about how to calculate the weighted average effect size.

http://effectsizefaq.com/2010/05/30/can-you-show-me-how-to-do-meta-analysis-in-just-2-minutes

Essentially, you need to have information on effect size r and the sample size

Say following is what your data looks like:
For effect size 1: r1=0.72, n1=21
For effect size 2: r2=0.53, n2=43
For effect size 3: r3=0.67, n3=44

The weighted average effect size
=(r1*n1+r2*n2+r3*n3)/(n1+n2+n3)
=(0.72*21+0.53*43+0.67*44)/(21+43+44)
= 67.39/108
=0.623981 (0.62)




Of course you can go up to effect size n and I can put it all the way to the nth except for...


(A nice song to break your from the tiring pursuit towards  a weighted average effect size)
 
At the same time, if you are like me started from Cohen's d, you would have to convert Cohen's d into r.

http://www.uccs.edu/~faculty/lbecker/

While Becker's effect size calculator might come in handy, if you need to do a whole lot of conversions, you might consider simply put all things into excel everything would be spitted out in split second with one paste...

If this case, you would need to understand a secret code on Becker's page...

This question mark hidden inside of a black diamond... apparently indicates..."squart root." 8-O lol


Bon chance.

Let me know shall you have any question.

(Wouldn't love to make the algorithm etc looking for beautiful except for... a bit too tired for the day.  Sorry.)

A cheatsheet on reporting Stats APA style

Need to find a cheatsheet on reporting stats APA style?  The following are good resources.

In addition, remember, if you provide stats in a table, don't put it in the text or vice versa.

  1. http://my.ilstu.edu/~jhkahn/apastats.html
  2. http://www.facelab.org/debruine/Teaching/Meth_A/files/Reporting_Statistics.pdf 

Thursday, February 23, 2012

Sample size Statistical Rules of Thumb

Been working with this dataset... since t-test is pretty robust and should be good enough to handle n=10, I've working with t-test with not to much hesitation.

Than, there came to the time when the sample size went down to 8 and 9, I am sort of nervous... thus... on a mission in search of the rule of thumb for the sample size...


http://library.athabascau.ca/drr/HADM%20499/Vanvoorhis%202001%20Statistical.pdf

Regardless, after recalculating the data I was working on, results of the Wilcoxon Signed Ranks Test didn't differ too much from those of the paired-sample t-tests.

Monday, February 20, 2012

Repeated measures with two within-subject variables

Getting old here... what my teachers taught me before have been sent to the warehouse somewhere on another planet.... might have been Neptune, I think.

So I was trying to figure out how to conduct a repeated measures test with two within-subject variables.

Finally, I hit the following document and, voila, now I recall.... http://frank.mtsu.edu/~dkfuller/psy629/labs/withinanova.pdf.

Thank God that both the professor and the school are 13-hour time difference away. 8-O lol

So say the Chinese, "天高皇帝遠." lol

Wednesday, February 15, 2012

Pronunciation Teaching - Bibliography

Been working on second language acquisiton kinda work recently with the main focus on  pronunciation.  Only if I had came across this site earlier....

A whole lot of readings targeting L2 pronunciation.

Pronunciation Teaching - Bibliography