Question
Asked 4th Jun, 2012
Sample Size calculation
In a study of 154 subjects (77 in test & 77 in the control group), the formula asunder was applied. Kindly help me for justifying this sample size, I am really zero in statistics. I shall remain grateful if anyone can help me in this regard.
Also, is there any pilot study for my case?
Sample Size Formula:
n = t2 pq
__________
d2
t = 1.9 (>90%)
p = .1 (10%)
q = 1-P (.9)
d = .05
n= 129
drop out at @ 20% = 25
n= 129+25= 154
Treatment arm= 77
Control arm= 77
Most recent answer
Agreed
Popular answers (1)
Oregon Health and Science University
Dr. Khan, I would like to respectfully suggest that you buy and read a statistics textbook. And perhaps your thesis committee members can be of help. It is unseemly to post a question on a public website asking for someone else to write the justification for your thesis sample size. This is your responsibility.
3 Recommendations
All Answers (16)
Hungarian University of Agriculture and Life Sciences
Actually, there is a theoretical approach for required minimal sample size:
n = ( t * S / D )^2
where t = 1.96 for infinite df (since df = n-1 by default), S is the estimated/expected deviation of data (or uncertainty of the measurement), D is the expected accuracy.
Regards, Laszlo
Dear Laszlo Baranyai, Thanks for your prompt reply and help, ok I got your explanation. Kindly help me that how I defend this sample size in my thesis or justify this sample size, is this formula is enough to write in the thesis or any other way to justify inclusion of 154 subjects in the research if, please send me that formula. Kind regards. Dr. Muhammad Khalil Khan: fjdc2000@yahoo.com
Hungarian University of Agriculture and Life Sciences
Dear Muhammad Khan, OK, if you reverse the calculation, accuracy can be computed. On the other hand, if you make 1 mistake with e.g. classification out of 77 in a class, it is 1.298%. Therefore the maximal theoretical probability is approximately 98%. Without any information about your data and the objectives, I can only recommend to browse for keywords: accuracy/error of prediction/classification, AIC, BIC, RMSECV, etc.
Regards, Laszlo
1 Recommendation
Dear Laszlo Baranyai, thanks and appreciate for your help. But I don’t understand about “to browse for keywords: accuracy/error of prediction/classification, AIC, BIC, RMSECV, etc.” because really I don’t know ABCD of Statistics. Can you explain what I have to write in section of sample size with sound justification of sample size 154 (77+77). Best regards. Dr. Muhammad Khalil Khan: fjdc2000@yahoo.com
Hungarian University of Agriculture and Life Sciences
Dear Dr. Muhammad Khan, as I still know nothing quantitative (neither qualitative) about the data, please refer online sources like: http://www.statsoft.com/textbook/
Regards, Laszlo
Dear Laszlo Baranyai, thanks for continues support. In my methodology I used 4 dental indexes are as under:
1. Measuring of periodontal pocket depths, using a standard periodontal probe with 5 mm marking.
2. GI recording, using the Gingival Index of Löe and Silness (1963) as modified by Talbott et al (1977).
3. Dental Plaque index recording. PI will be determined using the Turesky modification (1970) of the Quigley-Hein Index (1962).
4. Extrinsic tooth stains determined on the labial surface of the 12 anterior teeth, using the Lobene extrinsic tooth stain index (SI).(Lobene, 1968)
All these measurements/recordings done at 0 month, 1 month, 3 months, and 6 months in both control and test group. Kind regards. Dr. Muhammad Khalil: fjdc2000@yahoo.com
Hungarian University of Agriculture and Life Sciences
Dear Dr. Muhammad Khan,
thank you for the details about your data but for me, as an outsider of medicine, quantitative description is still missing. According to my limited knowledge in medicine, your sample size may be sufficient for scientific publication but probably more will be required for official treatment proposal. I am afraid no one (else) can give you sound justification ready to be inserted into the thesis. With my best regards, Laszlo Baranyai
Dear Laszlo Baranyai, thanks for your response in my case. Yes you are right that it is difficult to justify, can I just write the formula directly in the thesis as it is.
Thanks. Best regards.
Dr. Muhammad Khalil: fjdc2000@yahoo.con
Oregon Health and Science University
Dr. Khan, I would like to respectfully suggest that you buy and read a statistics textbook. And perhaps your thesis committee members can be of help. It is unseemly to post a question on a public website asking for someone else to write the justification for your thesis sample size. This is your responsibility.
3 Recommendations
Dear Teresa, nice to see your comments, and agreed with you, about my responsibility. I appreciate you in this regard and follow your instruction in future. Kind regards. Dr. Muhammad Khan
Univerzitet Union - Računarski Fakultet
Perhaps a sound reference from perspective of clinical research statistics would help here?
Civil Veterinary Hospital Talash, District Dir Lower, Khyber Pakhtunkhwa, Pakistan
Dear Dr Muhammad Khan. I am sending you a power point presentation of sampling size calculation but please remember that this is for known population. I have asked many persons but i could not found any solution to calculate sample size for unknown population.
Dear Muhammad Saleem, Assalmo Allikum, Thanks for sending presentation. I appreciate your concern about my problem. Best regards.
Civil Veterinary Hospital Talash, District Dir Lower, Khyber Pakhtunkhwa, Pakistan
It is my pleasure. Unfortunately it is difficult for us to find applied statistician. Those who got degree in statistics are mostly theoretical and they only give you example from books.
Similar questions and discussions
Related Publications
The type and scope of data, our attitudes towards them, and their availability have fundamentally changed in recent years. Never before has there been more data than today. Never have they been so readily available. And never have there been greater opportunities for analysis, preparation, and presentation. So-called infographics, frequently animat...
