Wednesday, October 15, 2008
HW due Oct 21
Create a new variable that indicates the mean of teachers’ fall and spring ratings of children’s approaches to learning (T1LEARN and T2LEARN). Now, transform this new variable into a dummy variable that identifies children who are below and above the mean (1 = below-average approaches to learning, 0 = above-average approaches to learning). Then, using t-test and chi-square analyses, examine the relationship between low/high approaches to learning and children’s SES (WKSESL), age at school entry (P1AGEENT), literacy and mathematics skills (C1RSCALE, C1MSCALE), gender (FEMALE), and race/ethnicity (RACE5). Place your results into a single table. As with every assignment, your paper should reflect the very compelling and interesting “story” present in your findings.
Subscribe to:
Post Comments (Atom)
8 comments:
Hi Megan,
This is Chrissy and Jaime. We're working on our homework early, and we just got stuck at the point of transforming our new variable into a dummy variable. We thought we followed the steps and directions, but its just not working where the numbers are coming out in a way that makes sense. Will you take a look and tell us where we went wrong with the syntax?
Thanks!
WEIGHT BY child_wt.
FREQUENCIES VARIABLES=T1LEARN T2LEARN
/FORMAT=LIMIT (10)
/STATISTICS=STDDEV MEAN MEDIAN MODE SKEWNESS
/HISTOGRAM NORMAL
/ORDER=ANALYSIS.
COMPUTE TLEARNAVE=MEAN(T1LEARN,T2LEARN).
VARIABLE LABEL TLEARNAVE 'AVERAGE OF T1LEARN AND T2LEARN'.
FREQUENCIES VARIABLES=TLEARNAVE
/FORMAT=LIMIT(10)
/STATISTICS=STDDEV MEAN MEDIAN MODE SKEWNESS
/HISTOGRAM NORMAL
/ORDER=ANALYSIS.
DESCRIPTIVES VARIABLES=TLEARNAVE
/SAVE
/STATISTICS=MEAN STDDEV MIN MAX SEMEAN.
RECODE ZTLEARNAVE (MISSING=SYSMIS) (LOWEST THRU 0=1) (0.001 THRU HIGHEST=0) INTO TLEARNAB.
VARIABLE LABELS TLEARNAB 'BELOW THE MEAN'.
VALUE LABELS TLEARNAB 1 'BELOW AVERAGE APPROACHES TO LEARNING' 0 'ABOVE AVERAGE APPROACHES TO LEARNING'.
FREQUENCIES VARIABLES=TLEARNAB
/FORMAT=LIMIT(10)
/STATISTICS=STDDEV MEAN MEDIAN MODE MIN MAX
/ORDER=ANALYSIS.
EXECUTE.
Hi again Megan,
Just to make sure I understood what Doug asked us to do with T1LEARN and T2LEARN before I run the analyses:
1) take their average
2) z-score that average
3) turn that new z-scored average into a 0-1 dummy.
This was a bit fast for me at the end of class so I wanted to make sure I understood. Many thanks,
Karen
Chrissy and Jaime - I ran your syntax, and it is working for me.... I'm not totally sure where you're stuck, but maybe your question has do to with the dummy variable?? Because the data are skewed and not quite normal and therefore the mean and median are not aligned - breaking the ZTLEARNAVE variable into above average and below average using the mean doesn't result in 50% of the cases above and below the mean. Don't worry about this problem, go ahead with the using the mean to divide the cases.
Let me know if this doesn't answer your question!
Megan
Karen -
Yes, that's right.
-Megan
It seems a little illogical to me to relate children's learning skills during the kindergarten year to reading and math scores in the fall of kindergarten. Shouldn't we be looking at their scores at the end of the year, or maybe in first grade?
Hi Megan, Should we set missing values for all variables used in our analysis to sysmis? Alex
Hi Alex,
Yes, you should set missing values to sysmis when recoding your variables.
Per your first question, I'm not sure that it is necessarily illogical to look at the relationship between the two. (Off the top of my head, the approaches to learning is a teacher rating, not some objective rating, while the cognitive IRT tests are perhaps more objective - so there might be an interesting story to tell there.) But in any case, if you have other variables that you think would be more interesting than the cognitive tests include those instead.
Alex - I also checked with Doug on your question, and here is his response - might be more helpful than what I wrote.
-M
"Children's approaches to learning are really quite stable over the school year--point is to see how children's outer manifestations of their ability to "do school" are related to their measured academic skills; essentially, it's a "school readiness" question."
Post a Comment