# Class Exercise on Theory in Program Evaluation

Document Sample

```					                              Class Exercise on Statistical Analysis
PM 536
September 14, 2012

This exercise is review for choosing the appropriate statistical procedures and tests. For each
question below write the appropriate statistical test and any transformations or other procedures
needed to conduct the analysis. Then write a sentence describing the result.

1) The association between gender (1=male 2=female) and MPH program track (1=Bio/Epi,
2=HC, 3=HP, & 4=Nutrition).

2) The association between gender (1=male 2=female) and MPH program track (midterm exam
score). Which variable is independent and which dependent?

3) The association between gender (1=male 2=female) and interest in football rated as low,
medium and high.

4) Whether attitude toward contraception, measured as a 7 item likert scale (alpha=.75) increased
significantly between baseline and follow up among a two-wave panel sample of 1,000 randomly
selected respondents?

5) In (4) above, how would you test whether it increased differently for men and women?

6) In (5) above, how would you test whether a media campaign was associated with changes in
campaign?

7) In (6) above, how would you test whether the media campaign was more effective among
women than men?
Variable |      1           2    Uniqueness
-------------+--------------------------------
q10a |   0.58102   -0.35716     0.53485
q10b |   0.66227   -0.45192     0.35717
q10c |   0.59314   -0.08275     0.64134
q10d |   0.08049     0.77817    0.38797
q10e |   0.31485     0.56020    0.58704
q10f |   0.51183     0.09332    0.72932
q10g |   0.49089     0.30261    0.66745
q10h |   0.56381   -0.32552     0.57615
q10i |   0.34078     0.37732    0.74150
q10j |   0.50707   -0.16623     0.71525
q10k |   0.49777     0.17450    0.72177
q10l |   0.30303     0.57073    0.58244

. egen bf_1=rmean(q10a q10b q10c q10f q10g q10h q10j q10k)
(85 missing values generated)

. egen bf_2=rmean(q10d q10e q10l)
(88 missing values generated)

reg bf_1 q1gen q3age q4b, b

Source |       SS       df       MS              Number of obs   =     1773
-------------+------------------------------           F( 3, 1769)     =     2.90
Model | 1.33829501      3 .446098336            Prob > F        =   0.0339
Residual | 272.293069 1769 .153924855              R-squared       =   0.0049
Total | 273.631364 1772 .154419506              Root MSE        =   .39233

------------------------------------------------------------------------------
bf_1 |      Coef.   Std. Err.      t    P>|t|                     Beta
-------------+----------------------------------------------------------------
q1gen |   .0379335   .0215399     1.76   0.078                 .0429328
q3age |   .0016037   .0007975     2.01   0.044                  .048862
q4b |   -.013156    .008224    -1.60   0.110                -.0383389
_cons |   3.401938   .0623882    54.53   0.000                        .
------------------------------------------------------------------------------

. reg bf_2 q1gen q3age q4b, b

Source |       SS       df       MS              Number of obs   =     1770
-------------+------------------------------           F( 3, 1766)     =     8.74
Model | 10.3783104      3 3.45943679            Prob > F        =   0.0000
Residual | 699.324817 1766 .395993668              R-squared       =   0.0146
Total | 709.703128 1769 .401188879              Root MSE        =   .62928

------------------------------------------------------------------------------
bf_2 |      Coef.   Std. Err.      t    P>|t|                     Beta
-------------+----------------------------------------------------------------
q1gen | -.1024162    .0346114    -2.96   0.003                -.0718579
q3age | -.0026681    .0012803    -2.08   0.037                -.0504362
q4b | -.0531411    .0132106    -4.02   0.000                -.0960347
_cons |    2.29904   .1002629    22.93   0.000                        .
------------------------------------------------------------------------------
Association between breast cancer attitudes and independent variables.

R2

```
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
 views: 0 posted: 9/15/2012 language: Latin pages: 3
How are you planning on using Docstoc?