MIRC October 9, 2007 – Soc 497 Fall 07

 

 

<jerry> First, do we all remember the format of the test..

<jerry> There are three parts..

<jerry> 1) objective (MC and TF)

<Judy> yes

<jerry> 2) definitions

<jerry> 3) short answers

<jerry> There are 50 points.

<jerry> No scantrons or bluebooks needed..

<jerry> Do we have that down?

<jerry> Judy, do we understand that much?

<Judy> yes

<jerry> okay, second issue then..

<jerry> this test deals only with the material covered since the last test..

<jerry> that means three things:

<jerry> 1) measurement control

<jerry> 2) experimental control

<jerry> 3) statistical control

<jerry> taken together, #1 is the first part of an operational definition..

<jerry> that is, a measurement unit..

<jerry> while the last two have to do with the "procedure" in an operational definition

<jerry> Does that resonate, Judy?

<Judy> yes

<jerry> okay, then it remains to structure your knowledge of these three components to have a successful test.

<jerry> first, measurement control..

<jerry> there are three issues I would like you to remember..

<jerry> the first is the levels of measurement..

<jerry> can you name all four?

<jerry> Judy?

<Judy> Nominal, ordinal, interval and ratio..right

<jerry> good

<jerry> please know what they mean and what operations are implied by each..

<jerry> The second issue is the process of imposing operational definitions..

<jerry> that is, nominal concept, indicator, operational definition and variable..

<jerry> to answer your question, an indicator is simply an intervening conept more precise than...

<jerry> a nominal concept, but less operational than an operational definition..

<jerry> The third issue has to do with the reliability and validity of a concept..

<jerry> remember reliability means consistency and validity means accuracy..

<jerry> what are the two types of reliability?

<jerry> Judy?

<Judy> test/retest and split half

<jerry> good.

<jerry> and what are the four types of validity

<jerry> Judy?

<Judy> sorface, construct, criteron, predictive

<Judy> i ment face on the first one

<jerry> good

<jerry> study these three issues having to do with measurement...

<jerry> and you should be okay.

<jerry> The second major topic of this test..

<jerry> is experimental control..

<jerry> please know four things..

<jerry> 1) the definition of experimental design..

<jerry> 2) simple designs.

<jerry> 3) complex designs

<jerry> 4) quasi designs.

<jerry> what three things do you need for an experiment?

<jerry> Judy?

<jerry> they are:

<jerry> 1) at least two groups

<jerry> 2) random assignment

<jerry> 3) measurement of an outcome

<jerry> do you remember these?

<Judy> forgot the lat one

<jerry> okay, please reflect on what they mena.

<Judy> ok

<jerry> make that "..they mean".

<jerry> secondly, remember that simply designs mean one IV...

<jerry> and complex designs mean two or more IVs.

<jerry> that complex designs are done precisely so that we can interpret "interaction" effects.

<jerry> remember, for example, what a 3 x 2 CRD means:

<jerry> 1) two IVs

<jerry> 2) three attribute on one variable and 2 on the other;

<jerry> 3) a total of six conditions

<Judy> got it

<jerry> also, remember that when designs get too big, ...

<jerry> we try to conserve on subjects...

<jerry> either by taking some of the conditions out of the design...

<jerry> or reusing subjects..

<Judy> ok

<jerry> finally, remember that a quasi-design is missing one or more conditions of a true experiment

<jerry> typically lack of random assignment..

<jerry> finally, the Third critical topic is statistical control..

<Judy> ok

<jerry> meaning that we need to sample our populations to include those with the attributes of...

<jerry> variables we deem to be potential noise or exogenous variables.

<jerry> the difference between the attempt to engage internal validation through experimental control...

<jerry> and external validation through sampling...

<jerry> is precisely the difference between taking charge of the variables and holding them constant..

<jerry> vs. measuring them in their naturally occurring setting and statistically holding them constant..

<jerry> I suggested to you that there are two types of samples..

<Judy> probability

<Judy> and non probability

<jerry> and?

<jerry> good, what is the difference between the two re: a sampling frame?

<Judy> probability uses simple random sample

<jerry> well, it uses more than that..

<jerry> but which has a sampling frame, probability or non-probability samples.

<Judy> well it uses random smples while non-probability uses samples such as convenience

<jerry> but what of my last question, which uses a sampling frame?

<Judy> non

<jerry> nope..

<jerry> remember a sampling frame is a listing of the population elements..

<Judy> by sampling frame do you meen a system to get randm samples

<jerry> nope, see above; a listing of the population elements

<Judy> ok

<jerry> without such a list, we cannot calculate the "probability" of an element occurring.

<jerry> hence the term, non-probability

<jerry> please study the difference..

<jerry> also, please be able to articulate different types of probability samples.

<jerry> and non-probabiliy samples..

<Judy> theres 4 types that i need to know

<jerry> yes..

<Judy> and 3 for non..correct

<jerry> yes

<jerry> finally, you need to know about the size of a sample

<Judy> the formula

<jerry> please understand that there are three issues having to do with size

<jerry> can you name them?

<Judy> Error in prediction

<jerry> yes

<Judy> confidence interval

<jerry> yes

<Judy> variability in population

<jerry> good

<jerry> these are the essential concepts for our test..

<jerry> now, do you have any questions?

<Judy> not @ the time

<jerry> okay, then I will see you on Thursday

<Judy> thank you so much

<jerry> you are welcome