How to do statistical comparisons on unequal size groups?

oem7110 · 2014-05-29 03:17:28

In 2013, there are 100 students taking EXAM
In 2014, there are 150 students taking EXAM

In term of percentage of students within 2014 year, their scores are getting improvement comparing with 2013 year. The main issue is not knowing on how to define the benchmark on 2013.

For example, John's score is 64 on 2013, Mary's score is 73 on 2014, then Mary in 2014 is getting better than John in 2013. Should I use average score on 2013 be benchmark in this case?

Does anyone have any suggestions on how to do statistical comparisons on unequal size groups? so I can compare between 2013 and 2014, in order to determine which year students are doing better.

Does anyone have any suggestions?
Thanks in advance for any suggestions

Download file:
http://1drv.ms/1nvIxOA

Last edited by oem7110 (2014-05-29 03:18:03)

bobbym · 2014-05-29 03:43:48

A 2 sample T test might be what you need. Are the two samples independent?

oem7110 · 2014-05-29 03:51:23

bobbym wrote:

A 2 sample T test might be what you need. Are the two samples independent?

Two samples are independent, could you please show me on how to apply T test in this case? in order to determine which years students are doing better.

Thanks you very much for any suggestions

bobbym · 2014-05-29 03:56:25

Hi;

Take a good look at this. I am going to use it for your problem.

https://onlinecourses.science.psu.edu/stat500/node/50

The method does not seem to care about the different sizes. What are the variances of both populations?

oem7110 · 2014-05-30 00:02:32

bobbym wrote:

Hi;
Take a good look at this. I am going to use it for your problem.
https://onlinecourses.science.psu.edu/stat500/node/50
The method does not seem to care about the different sizes. What are the variances of both populations?

I get no idea on variances of both populations, can I still apply T Test in my case?

Thanks you very much for any suggestions

bobbym · 2014-05-30 00:12:49

There are 2 T tests. The variances are calculated form the data. Can you provide the data in text form?

oem7110 · 2014-05-30 00:27:02

bobbym wrote:

There are 2 T tests. The variances are calculated form the data. Can you provide the data in text form?

Do you have any suggestions?
Thank you very much for any suggestions :>

Text file for download:
http://1drv.ms/1o5pHJU

bobbym · 2014-05-30 00:30:17

I can not download that file. Can you cut and paste the data here.

oem7110 · 2014-05-30 01:34:09

bobbym wrote:

I can not download that file. Can you cut and paste the data here.

Here you go ...

Thanks you very much for any suggestions

2013 2014
4% 46%
83% 26%
30% 92%
89% 74%
38% 21%
54% 13%
51% 64%
47% 87%
7% 13%
77% 95%
57% 19%
24% 20%
3% 1%
73% 93%
85% 28%
4% 14%
39% 19%
74% 71%
59% 16%
91% 51%
51% 64%
45% 54%
67% 7%
87% 81%
9% 39%
30% 41%
15% 31%
50% 89%
61% 25%
46% 86%
70% 69%
77% 16%
25% 20%
81% 10%
7% 95%
67% 95%
23% 51%
56% 52%
88% 4%
43% 23%
22% 69%
36% 60%
43% 78%
26% 81%
65% 29%
55% 70%
53% 12%
80% 96%
68% 96%
7% 72%
33% 14%
20% 4%
54% 52%
68% 87%
24% 95%
31% 72%
53% 22%
45% 91%
34% 37%
3% 16%
23% 5%
87% 38%
33% 29%
71% 20%
62% 98%
22% 60%
61% 48%
30% 59%
56% 83%
32% 40%
80% 59%
36% 29%
2% 63%
55% 1%
72% 90%
69% 49%
20% 64%
9% 67%
50% 24%
6% 70%
67% 44%
36% 45%
76% 62%
34% 64%
27% 95%
47% 24%
86% 43%
74% 40%
18% 18%
61% 30%
67% 2%
29% 57%
46% 77%
65% 81%
63% 29%
28% 5%
89% 96%
68% 8%
53% 26%
27%
97%
34%
44%
59%
50%
21%
14%
41%
32%
71%
40%
2%
48%
36%
19%
27%
22%
84%
8%
47%
41%
92%
69%
67%
25%
54%
82%
77%
27%
81%
50%
10%
79%
40%
92%
34%
45%
90%
39%
21%
40%
98%
96%
14%
5%
33%
56%
25%
79%

bobbym · 2014-05-30 02:16:35

That does not look like real data, a person scored 1, another 4, a couple scored 7?

Agnishom · 2014-05-30 03:59:13

bobbym · 2014-05-31 09:23:33

Hi;

The variances are not close so we will go with a different test. I am computing it now.

According to what I got there is no reason to assume there is any differences between those two means and we are forced to reject the alternative hypothesis that students are doing better in 2014 as opposed to 2013.

Math Is Fun Forum

#1 2014-05-29 03:17:28

How to do statistical comparisons on unequal size groups?

#2 2014-05-29 03:43:48

Re: How to do statistical comparisons on unequal size groups?

#3 2014-05-29 03:51:23

Re: How to do statistical comparisons on unequal size groups?

#4 2014-05-29 03:56:25

Re: How to do statistical comparisons on unequal size groups?

#5 2014-05-30 00:02:32

Re: How to do statistical comparisons on unequal size groups?

#6 2014-05-30 00:12:49

Re: How to do statistical comparisons on unequal size groups?

#7 2014-05-30 00:27:02

Re: How to do statistical comparisons on unequal size groups?

#8 2014-05-30 00:30:17

Re: How to do statistical comparisons on unequal size groups?

#9 2014-05-30 01:34:09

Re: How to do statistical comparisons on unequal size groups?

#10 2014-05-30 02:16:35

Re: How to do statistical comparisons on unequal size groups?

#11 2014-05-30 03:59:13

Re: How to do statistical comparisons on unequal size groups?

#12 2014-05-31 09:23:33

Re: How to do statistical comparisons on unequal size groups?

Board footer