Biostatistics Question: Analyzing these groups - USMLE Forums
 USMLE Forums         Your Reliable USMLE Online Community     Members     Posts
 Home USMLE Articles USMLE News USMLE Polls USMLE Books USMLE Apps
 USMLE Forums Biostatistics Question: Analyzing these groups
 Register FAQs Members List Search Today's Posts Mark Forums Read

 USMLE Step 1 Forum USMLE Step 1 Discussion Forum: Let's talk about anything related to USMLE Step 1 exam

#1
04-21-2011
 USMLE Forums Guru Steps History: 1+CK+CS Posts: 390 Threads: 40 Thanked 365 Times in 144 Posts Reputation: 375
Biostatistics Question: Analyzing these groups

Researchers wish to determine the efficacy of a new weight loss regimen. To do this, they compare the amount of weight loss by 40 patients on the regimen (Group A) with 40 people selected at random from a commercial weight loss program (Group B) and 30 volunteers who were trying to lose weight on their own (Group C). Results of weight loss study are shown below :

Group A
Mean : 17.3 lb, SE : 3.4, N : 40

Group B
Mean : 12.7 lb, SE : 4.3, N : 40

Group C
Mean : 5.4 lb, SE : 5.3, N : 30

SE = standard error, N : sample size

Based on the results above and using a criterion level of p<0.05, the researchers are most likely to conclude that :

A. Group A lost more weight than Group B and Group B lost more weight than Group C.
B. Group A showed more weight loss than Group C, but not more than Group B.
C. Group A showed significantly more weight loss than either of the other two groups.
D. No answer is possible without the results of statistical test being given.
E. No significant differences were found among the three groups.
__________________
Time is immortal when you believe on something...
 The above post was thanked by: abeaveli (08-23-2015), bebix (04-22-2011)

#2
04-21-2011
 USMLE Forums Master Steps History: Not yet Posts: 557 Threads: 69 Thanked 1,038 Times in 351 Posts Reputation: 1058

I would say B: Group A showed more weight loss than Group C, but not more than Group B.

The SEs overlap in Group A and B (meaning group A has a range of 13.9-20.7 whereas Group B has a range of 8.4-17.0), so no difference there, but Group A doesn't overlap with the values in C.

I always mix up SE and SD; I know you can use the SD to reason through this like i just did, but is that correct to use the SE that way?
 The above post was thanked by: harlesmd (04-21-2011)
#3
04-21-2011
 USMLE Forums Guru Steps History: Not yet Posts: 298 Threads: 45 Thanked 713 Times in 210 Posts Reputation: 723

I was thinking B, but also have the same question about SE vs SD
 The above post was thanked by: harlesmd (04-21-2011), Prepping123 (04-21-2011)
#4
04-21-2011
 USMLE Forums Guru Steps History: 1+CK+CS Posts: 390 Threads: 40 Thanked 365 Times in 144 Posts Reputation: 375

This is a good question to test your concept on Standar Deviation, Standard Error and Confidence Interval, that's why I post it here. It can be very tricky. I will wait for more responses.
__________________
Time is immortal when you believe on something...
#5
04-21-2011
 USMLE Forums Guru Steps History: CK Only Posts: 475 Threads: 39 Thanked 351 Times in 169 Posts Reputation: 374

I'm inclined to say D
__________________
Skill+Hardwork+Preparation=Success
To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

 The above post was thanked by: harlesmd (04-21-2011)
#6
04-21-2011
 USMLE Forums Guru Steps History: Not yet Posts: 298 Threads: 45 Thanked 713 Times in 210 Posts Reputation: 723

Quote:
 Originally Posted by 1TA2B I'm inclined to say D
D sounds enticing, but it's probably a cop out answer choice like the answer choices in bioethics that say "refer to ethics committee" or "go to court"
 The above post was thanked by: harlesmd (04-22-2011)
#7
04-21-2011
 USMLE Forums Master Steps History: Not yet Posts: 557 Threads: 69 Thanked 1,038 Times in 351 Posts Reputation: 1058

Quote:
 Originally Posted by apx85 D sounds enticing, but it's probably a cop out answer choice like the answer choices in bioethics that say "refer to ethics committee" or "go to court"
lol I agree! Plus I wouldn't say NO answer is possible; you can't conclude statistical significance if you don't have the results of a statistical test, but you can conclude something if you want to.
 The above post was thanked by: harlesmd (04-22-2011)
#8
04-21-2011
 USMLE Forums Veteran Steps History: Not yet Posts: 236 Threads: 4 Thanked 123 Times in 95 Posts Reputation: 133

Iam going with C
 The above post was thanked by: harlesmd (04-22-2011)
#9
04-22-2011
 USMLE Forums Master Steps History: 1+CK+CS+3 Posts: 1,357 Threads: 194 Thanked 3,259 Times in 881 Posts Reputation: 3269

I would say D

Strictly speaking, we should do an ANOVA (calculated SSB & SSW) and then a post hoc analysis.
With the ANOVA, the only conclusion would be "at least one mean group is different", but then, with a post hoc (i.e. bonferroni), we would have to compare A&B, B&C and A&C (the type one error should be correct it)
 The above post was thanked by: harlesmd (04-22-2011)
#10
04-22-2011
 USMLE Forums Guru Steps History: 1+CK+CS Posts: 390 Threads: 40 Thanked 365 Times in 144 Posts Reputation: 375

We have various answers here, unfortunately no one hit the correct one.

Although p value is not mentioned above, we can still conclude the statistical significance of this study by computing confidence interval from each group.

The question set the the criterion level of p <0.05, so we must compute 95% confidence interval.

95% CI = mean +/- (z x SE)
z score for 95% CI is 2

For group A :
95% CI = 17.3 +/- (2 x 3.4)
= 17.3 +/- 6.8
= 10.5 to 24.1

Using the same formula
For group B :
95% CI = 4.1 to 21.3

For group C :
95% CI = -5.2 to 16

The confidence interval of each group is overlapping to each other, thus we can conclude that there were no significant difference among three groups (E)
__________________
Time is immortal when you believe on something...
 The above post was thanked by: doc_study (04-22-2011), heights (04-22-2011), noothan (04-22-2011)
#11
04-22-2011
 USMLE Forums Guru Steps History: CK Only Posts: 475 Threads: 39 Thanked 351 Times in 169 Posts Reputation: 374

Quote:
 Originally Posted by harlesmd We have various answers here, unfortunately no one hit the correct one. Although p value is not mentioned above, we can still conclude the statistical significance of this study by computing confidence interval from each group. The question set the the criterion level of p <0.05, so we must compute 95% confidence interval. 95% CI = mean +/- (z x SE) z score for 95% CI is 2 For group A : 95% CI = 17.3 +/- (2 x 3.4) = 17.3 +/- 6.8 = 10.5 to 24.1 Using the same formula For group B : 95% CI = 4.1 to 21.3 For group C : 95% CI = -5.2 to 16 The confidence interval of each group is overlapping to each other, thus we can conclude that there were no significant difference among three groups (E)
Good Job! Thanks.

A couple of queries!

1. How do you get "2"?

2. What is the relationship between standard deviation and standard error?" It looks like SD equals -muliplies "z" of SE.
__________________
Skill+Hardwork+Preparation=Success
To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts.

#12
04-22-2011
 USMLE Forums Master Steps History: 1+CK+CS+3 Posts: 1,357 Threads: 194 Thanked 3,259 Times in 881 Posts Reputation: 3269

Quote:
 Originally Posted by harlesmd We have various answers here, unfortunately no one hit the correct one. Although p value is not mentioned above, we can still conclude the statistical significance of this study by computing confidence interval from each group. The question set the the criterion level of p <0.05, so we must compute 95% confidence interval. 95% CI = mean +/- (z x SE) z score for 95% CI is 2 For group A : 95% CI = 17.3 +/- (2 x 3.4) = 17.3 +/- 6.8 = 10.5 to 24.1 Using the same formula For group B : 95% CI = 4.1 to 21.3 For group C : 95% CI = -5.2 to 16 The confidence interval of each group is overlapping to each other, thus we can conclude that there were no significant difference among three groups (E)
Actually this is not quite correct. You can calculate each 95CI%, but since u are doing 2 comparisons, you increased your alpha error (this error follows a binomial distribution). I know that this may sound a little bit technical...but this is way we use ANOVA and we don´t do multiple t test or multiple 95CI%....

the number 2 is just an aprox. of Z alpha (normal dist) two sided tail (2.5%each side =1.97).
 The above post was thanked by: harlesmd (04-22-2011)
#13
04-22-2011
 USMLE Forums Guru Steps History: 1+CK+CS Posts: 390 Threads: 40 Thanked 365 Times in 144 Posts Reputation: 375

Definitely you can confirm this result with ANOVA test, and calculating the f value (for ANOVA test, you don't need to do it manually, you can use the software) and again the p value will be > 0.05 (not statistically significant). Thus choosing option D won't be the best answer.
__________________
Time is immortal when you believe on something...
#14
04-22-2011
 USMLE Forums Master Steps History: 1+CK+CS+3 Posts: 1,357 Threads: 194 Thanked 3,259 Times in 881 Posts Reputation: 3269

Quote:
 Originally Posted by harlesmd Definitely you can confirm this result with ANOVA test, and calculating the f value (for ANOVA test, you don't need to do it manually, you can use the software) and again the p value will be > 0.05 (not statistically significant). Thus choosing option D won't be the best answer.
hahaha...I know that...and u also can do multiple t test with corrections or permutations. The all point is the concept....In this case was pretty obvious the answer, but the concept is wrong ...and u can do it manually, trust me, haha
#15
04-22-2011
 USMLE Forums Guru Steps History: 1+CK+CS Posts: 390 Threads: 40 Thanked 365 Times in 144 Posts Reputation: 375

Quote:
 Originally Posted by bebix hahaha...I know that...and u also can do multiple t test with corrections or permutations. The all point is the concept....In this case was pretty obvious the answer, but the concept is wrong
Sorry but I didn't really get your point. Which concept is wrong ?
__________________
Time is immortal when you believe on something...

#16
04-22-2011
 USMLE Forums Master Steps History: 1+CK+CS+3 Posts: 1,357 Threads: 194 Thanked 3,259 Times in 881 Posts Reputation: 3269

Quote:
 Originally Posted by harlesmd Sorry but I didn't really get your point. Which concept is wrong ?
I really don't want to confuse people here, so I sent u a PM with a simulation.

 Tags Biostatistics-Epidemiology, Step-1-Questions

Message:
Options

## Register Now

In order to be able to post messages on the USMLE Forums forums, you must first register.
User Name:
Medical School
Choose "---" if you don't want to tell. AMG for US & Canadian medical schools. IMG for all other medical schools.
 AMG IMG ---
USMLE Steps History
What steps finished! Example: 1+CK+CS+3 = Passed Step 1, Step 2 CK, Step 2 CS, and Step 3.

Choose "---" if you don't want to tell.

 Not yet Step 1 Only CK Only CS Only 1 + CK 1 + CS 1+CK+CS CK+CS 1+CK+CS+3 ---
Favorite USMLE Books
 What USMLE books you really think are useful. Leave blank if you don't want to tell.
Location
 Where you live. Leave blank if you don't want to tell.

## Log-in

Human Verification

In order to verify that you are a human and not a spam bot, please enter the answer into the following box below based on the instructions contained in the graphic.