reliability analysis

[Johnny Amora]

Dear all,

I am conducting reliability analysis (part only of the
entire analysis) of a newly developed instrument with
50-items, which aims to measure the teacher
performance.  The instrument was administered to a
total of 300 students to rate the teaching performance
of their respective teacher.  There were 30 teachers
rated; thus, on average, there were 100 students who
rated for each teacher.

I need your expert opinion because I dont know which
one is correct:

(1) Do I analyze the 300 sample to conduct the
reliability analysis?

Or

(2) Do I analyze the aggregate data-teacher level?
That is, for each item I compute first the mean
ratings of the students within each teacher, then use
the 30 teachers as sample for analysis?

Thank you.

Johnny


      ____________________________________________________________________________________
Be a better friend, newshound, and
know-it-all with Yahoo! Mobile.  Try it now.  http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Sara House]

John said:

I am conducting reliability analysis (part only of the
entire analysis) of a newly developed instrument with
50-items, which aims to measure the teacher
performance.  The instrument was administered to a
total of 300 students to rate the teaching performance
of their respective teacher.  There were 30 teachers
rated; thus, on average, there were 100 students who
rated for each teacher.

John -

I'm a little confused about your sample size.  Did you have 300 or 3000?  If there were in fact 30 teachers being rated, then for a sample of 300 that would mean 10 students rated each teacher (30 * 10 = 300).  And if there were in fact 100 students rating each teacher (on average), then your sample would have to be 3000 (30 * 100 = 3000).  To someone who knows more about reliability analysis than myself, would sample size (300 versus 3000) make a difference in how you would analyze?

Sara House
[hidden email]

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[KATHY GREEN]

Hi John,

It sounds to me like you might consider a generalizability approach to
this--multiple items, multiple teachers, and multiple students. There is
an article by Mushquash and O'Connor (2006) in Behaviora Research Methods
that has the link to a downloadable (free) spss syntax file.

Kathy

On Sun, 27 Jan 2008, John Amora wrote:

> Dear all,
>
> I am conducting reliability analysis (part only of the
> entire analysis) of a newly developed instrument with
> 50-items, which aims to measure the teacher
> performance.  The instrument was administered to a
> total of 300 students to rate the teaching performance
> of their respective teacher.  There were 30 teachers
> rated; thus, on average, there were 100 students who
> rated for each teacher.
>
> I need your expert opinion because I dont know which
> one is correct:
>
> (1) Do I analyze the 300 sample to conduct the
> reliability analysis?
>
> Or
>
> (2) Do I analyze the aggregate data-teacher level?
> That is, for each item I compute first the mean
> ratings of the students within each teacher, then use
> the 30 teachers as sample for analysis?
>
> Thank you.
>
> Johnny
>
>
>       ____________________________________________________________________________________
> Be a better friend, newshound, and
> know-it-all with Yahoo! Mobile.  Try it now.  http://mobile.yahoo.com/;_ylt=Ahu06i62sR8HDtDypao8Wcj9tAcJ
>
> =====================
> To manage your subscription to SPSSX-L, send a message to
> [hidden email] (not to SPSSX-L), with no body text except the
> command. To leave the list, send the command
> SIGNOFF SPSSX-L
> For a list of commands to manage subscriptions, send the command
> INFO REFCARD
>

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[denielnill]

Reliability analysis provides complete picture about asset performance management and optimization. Maximo reporting is the analytical tool that triggers the data driven decisions such as spare parts cost, maintenance schedules, work planning and scheduling etc. Asset tracking is foremost important, in order to decide the optimum spare part quantity to be ordered. It facilitates in optimizing spare part cost, which is an important factor that contributes in reducing total cost of ownership for owing the assets.