Hello,, i am a Lecturer at Institute of Management Studies, University of Peshawar, Pakistan. Currently i am doing my Ph.D on "Decision to choose defensive silence" The perspective of expectancy theory".
i am going to apply within subject factorial design, . i have 1 dependent and 3 independent variables, each at 2 levels (2*2*2*2= 16) so i gt 16 hypothetical situations, through which i want to measure the impct of each IV on the decision behavior of the respndents.
can u please tell me waht should be the sample size for within subject factorial design
Why are you getting into such a meta-subject. I would suggest you to do something more relevant.
It sounds like a very interesting project.
Your needed sample size depends on your expected effect size and the statistical power you want for detecting a significant effect. When you figure these out, you can use a free sample size calculator called G*Power available from http://www.gpower.hhu.de/en.html
An old fashion rule that is still sometimes used is 30 participants per cell, which would be 480 in your case.
Dear David Dunaetz. thank u for ur reply. Actually i am using within subject factorial design for the first time and don't have good command on it. i feel ur reply to be very relevant but don't understand the terms u used as its all new for me. can u please help me out in this matter.
i have selected the Banking sector of one of the district, in one of the province of pakistan. In the district of Peshawar, there are total 34 banks, so this is my population, now guide me what should me my sample, while i am using within subject factorial design
Dear Diedre Thank u for ur reply.
i have selected the Banking sector of one of the district, in one of the province of pakistan. In the district of Peshawar, there are total 34 banks, so this is my population, now guide me what should me my sample, while i am using within subject factorial design
Amen,
It sounds like you should collect data from all 34 banks. Even if all 4 of your factors are within subjects, 34 is not a very large sample size. With N =34, many effects cannot be detected at a statistically significant level.
Statistical power and effect size are important statistical concepts. It would be worth your time to understand them. Almost all statistics textbooks used at a graduate level will cover these concepts, especially if you attempt to use a complex ANOVA analysis like this one.
When designing an experiment, the main objective is to detect the effects of factors on the response variable. Poor detection of the effects depends on the sample size (power of the test), ie the size of the sample should generate a power of at least 90% of detecting the difference established.
Thank u Dr. Manuel for your reply. can u please tell me what wil be sample size in my case, because i am still not getting it.
As per Salkind 8th Addtion:
The Research Randomizer
The Research Randomizer (http://www.randomizer.org/) is a cool little tool that assists you in performing a simple random sampling and assigning participants to experimental conditions.
Doug’s Random Sampling Applet
If you don’t find the above fun, try Doug’s Random Sampling Applet at http://www.
dougshaw.com/sampling/. You enter the population size and the size of the sample,
and Doug computes which participants you need to select (by number).
How to Calculate Sample Size
At http://www.ehow.com/how 5262463 calculate-sample-size-formula.html, eHow provides a formula for calculating the appropriate sample size for your research. You will need a calculator and a z-score conversion table (the site provides a link to one).
Random Number Generator
At http://www.graphpad.com/quickcalcs/randomn1.cfm, GraphPad Software created this tool for generating random numbers. Simply enter the number of columns and rows you would like in addition to the range of possible integers, and voila!
In terms of the psychical number there is no right or wrong because no sample will represent your population perfectly.
- Badges
- Science topic
- Similar topics
- Mathematics
- Statistics
- Sampling Design
- Sample Size