Você está na página 1de 9

opened on: 26 Jan 2017, 17:07:02

1. DESCRIPTION OF VARIABLE.
a) Gender of the student is nominal data.
b) Course work is continuous data.
c) Exam is continuous data.
d) Weighted score is continuous data.
e) Residence is ordinal data.
5(a)

pie chart showing the g-total

below 59 80++
60-79

5(b) Majority of the student scored 60-79, followed by 80++ and the rest 59
and below
6(a)
RECODE of total Freq. Percent Cum.
below 59 10 25.00 25.00
60-79 21 52.50 77.50
80++ 9 22.50 100.00
Total 40 100.00
52.5% of the male students scored 60-79, followed by 25% who scored 59 and
below and the rest scored 80++
6(b)
RECODE of total Freq. Percent Cum.
below 59 4 21.05 21.05
60-79 10 52.63 73.68
80++ 5 26.32 100.00
Total 19 100.00

52.63% of the male resident students scored 60-79, followed by 26.32% who scored
80++ and the rest scored 59 and below.
6(c) RECODE of total Freq. Percent Cum.
Below 59 6 28.57 28.57
60-79 11 52.38 80.95
80++ 4 19.05 100.00
Total 21 100.00
52.38% of the male resident students scored 60-79, followed by 28.57% who scored
59 and below and the rest scored 80++.
7(a) COURSE WORK
Percentiles Smallest
1% 14 14
5% 18 18
10% 20 18 Obs 63
25% 26 18 Sum of Wgt. 63
50% 32 Mean 30.25397
75% 36 37 Largest Std. Dev.
6.377805
90% 37 37 Variance 40.6764
95% 37 39 Skeweness -.7598593
99% 39 39 Kurtosis 2.521428
The median of the score out of 40 obtained in course work assessment was 32 with
minimum 14 and maximum 39. The high skeweness tends to pull down the mean.
SCORE OUT OF 60
Percentiles Smallest
1% 10 10
5% 17 12
10% 26 16 Obs 63
25% 34 17 Sum of Wgt. 63
50% 40 Mean 39.60317
75% 47 56 Largest Std. Dev.
10.8517
90% 53 58 Variance 117.7593
95% 56 59 Skewness -.5948934
99% 60 60 Kurtosis 3.518714
The median of the score out of 60 obtained in the final examination was 40 with
minimum 10 and maximum 60.
WEIGHTED SCORE
Percentiles Smallest
1% 18 18
5% 18 18
10% 18 18 Obs 63
25% 24 18 Sum of Wgt. 63
50% 28 Mean 26.09524
75% 28 30 Largest Std. Dev.
3.635463
90% 30 30 Variance 13.21659
95% 30 30 Skewness -1.175589
99% 30 30 Kurtosis 3.526213
The median of the weighted score adopted on admission to program was 28 with
minimum 18 and maximum 30.

8(a)

A scatterplot showing the relationship between cw and exam scores


60
50 40
exam
3020
10

15 20 25 30 35 40
cw

Fitted values SCORE OUT OF 60

There is a moderate positive relationship between course work and exam scores.
catterplot showing the relationship between total and wht scores

8(b)

8(b)
100
total
80

15 20 25 30
wht
60

total Fitted values


40

There is a positive moderate relationship between the total and weighted scores.
9(a) Kruskal-Wallis equality-of-populations rank test total, by(sex)
chi-squared = 0.651 with 1 d.f. probability = 0.4199
There is no significant relationship between total score and sex significant at
5%level since (p>0.05)
Kruskal-Wallis equality-of-populations rank test total, by(resd)
chi-squared = 2.581 with 1 d.f. probability = 0.1082
There is no significant relationship between total score and resd significant at 5%
level since (p>0.05)
spearman total cw, stats(rho p) star(0.05)
Number of obs = 63 Spearman's rho = 0.7732
Test of Ho: total and cw are independent
Prob > |t| = 0.0000
There is a strong positive relationship between total and course work significant at
5% level.
20
spearman total exam, stats(rho p) star(0.05)
Number of obs = 63 Spearman's rho = 0.8844
Test of Ho: total and exam are independent
Prob > |t| = 0.0000
There is a strong positive relationship between total and exam significant at 5%
level.
spearman total wht, stats(rho p) star(0.05)
Number of obs = 63 Spearman's rho = 0.8175
Test of Ho: total and wht are independent Prob > |t| = 0.0000
There is a strong positive relationship between total and wht significant at 5% level
9(b) g_total sex, chi2
Pearson chi2(2) = 1.4288 Pr = 0.489
There is no significant relationship between categorized total score and sex since
(p>0.05)
g_total resd, chi2
Pearson chi2(2) = 2.2658 Pr = 0.322
There is no significant relationship between categorized total score and residence
since (p>0.05)
Kruskal-Wallis equality-of-populations rank test g_total, by(cw)
chi-squared = 32.509 with 18 d.f.
probability = 0.0191
There is a relationship between categorized total score and course work significant
at 5% level.
Kruskal-Wallis equality-of-populations rank test g_total, by(exam)
chi-squared = 43.498 with 32 d.f.
probability = 0.0845
There is no significant relationship between categorized total score and exam at 5%
level.
Kruskal-Wallis equality-of-populations rank test g_total, by(wht)
chi-squared = 35.738 with 4 d.f.
probability = 0.0001
There is a relationship between categorized total score and weighted score
significant at 5% level.
10 (a)
spearman exam cw, stats(rho p) print(0.10)
Spearman's rho = 0.4600
Test of Ho: exam and cw are independent Prob > |t| = 0.0001
There is a moderate positive relationship between course work and exam score
significant at 1% level.
10 (b) spearman total wht, stats(rho p) print(0.10)
Spearman's rho = 0.8175
Test of Ho: total and wht are independent Prob > |t| = 0.0000
There is a strong positive relationship between total and weighted scores significant
at 1% level

11(a) multiple linear regression regress total sex wht resd


Prob > F = 0.0000 R-squared = 0.5354
Adj R-squared = 0.5118
total Coef. Std. Err. t P>|t|
[95% Conf. Interval]
sex 4.523627 2.845699 1.59 0.117
-1.170603 10.21786
wht 3.030859 .3746159 8.09 0.000
2.281255 3.780464
resd 2.410133 2.727994 0.88 0.381
-3.04857 7.868836
_cons -16.70967 11.01093 -1.52 0.134
-38.74249 5.323152

The model is a good fit since (p<0.005)


51.18% of total score is explained by sex, weighted score and residence.

Total score = -16.70967 + 4.523627 sex + 3.030859 weighted score +


2.410133 residence

Sex and residence are not significant predictors of total score while weighted score
is a significant predictor of total score.

11 (b) Logistic regression logit d_total i.sex wht resd


LR chi2(3) = 22.74 Prob > chi2 = 0.0000
Pseudo R2 = 0.3095
d_total Coef. Std. Err. z P>|z|
[95% Conf. Interval]
2.sex 1.976994 1.040048 1.90 0.057
-.0614622 4.015449
wht .4287776 .1226188 3.50 0.000 .1884491
.6691061
resd .725453 .7216722 1.01 0.315
-.6889984 2.139904
_cons -10.77803 3.255958 -3.31 0.001
-17.15959 -4.39647

The model is a good fit since (p<0.005)


Females are most likely to excel than the male students.
Resident students are more likely to excel than nonresident students.
Sex and residence do not significantly predict excelling while weighted score
significantly predicts excelling
log close
log: C:\Users\TONNY\Desktop\jane\JANE.smcl
closed on: 26 Jan 2017, 20:20:41

12. LOG FILE


log using "C:\Users\TONNY\Desktop\jane\JANE.smcl"

label variable sex "GENDER OF STUDENT"

label variable cw "COURSE WORK"

label variable exam "SCORE OUT OF 60"

label variable wht "WEIGHTED SCORE"

label variable resd "RESIDENCE"

label define SEX 1 "MALE" 2 "FEMALE"

label define resd 1 "yes" 0 "no"

generate total = cw+exam

recode total (min/59 = 1 "below 59") (60/79 = 2 "60-79") (80/max = 3 "80++"),


prefix(g_)

graph pie, over(g_total) sort title(pie chart showing the g-total)

tabulate g_total if sex ==1

tabulate g_total if sex ==1&resd==1

tabulate g_total if sex ==1&resd==0

summarize cw exam wht, detail

twoway (lfit exam cw) (scatter exam cw), ytitle(exam) xtitle(cw) title(A scatterplot
showing the relationship between cw and exam scores)

twoway (scatter total wht) (lfit total wht), ytitle(total) xtitle(wht) title(A scatterplot
showing the relationship between tota and wht scores)

kwallis total, by(sex)


kwallis total, by(resd)

spearman total cw, stats(rho p) star(0.05)

spearman total exam, stats(rho p) star(0.05)

spearman total wht, stats(rho p) star(0.05)

tabulate g_total sex, chi2

tabulate g_total resd, chi2

kwallis g_total, by(cw)

kwallis g_total, by(exam)

kwallis g_total, by(wht)

spearman exam cw, stats(rho p) print(0.10)

spearman total wht, stats(rho p) print(0.10)

regress total sex wht resd

recode total (min/64 = 0 "failled") (65/max = 1 "excelledl"), prefix(d_)

logit d_total i.sex wht resd


log close

MAKERERE UNIVERSITY
COLLEGE OF BUSINESS AND MANAGEMENT SCIENCES
INSTITUTE OF STATICS AND APPLIED ECONOMICS

DATA ANALYSIS 111 COURSE WORK


NAMES REGISTRATION STUDENTS` SIGNATU
NO. NUMBER RE
ASIIMWE SAUL 14/U/5444/EVE 214008760

NANZIRA JANE 14/U/12832/EV 214011501


E

NALWOGA WINNIE 14/U/11820/EV 214006437


E

Você também pode gostar