ON conversing with many researchers , most of the researchers clearly mention that not even chi-square test can be used
If it is so , any anyone please attach a pdf or article stating that chi-square cannot be used as a statistical test in case of studies where participants have been recruited through convenience sampling
Further , for studies using convenience sampling can we compute Odd Ratio and Confidence Interval what ever the answer may be can anyone please send a link or pdf .;
Thankyou
Sorry Ashish I do not have a paper to send you, but can say that very many published studies are based on convenience samples. You can conduct analyses on these data, according to your hypotheses, The limitations are around the generalisability of the results to broader populations. So, develop your hypotheses from the literature and this will indicate the analyse to use - assuming the data are in the correct format. Kate
Ashish,
I have never heard such a statement from other researchers that chi-squares are not appropriate when convenience sampling has been applied. These are two distinct issues. Convenience sample MAY limit possibility to generalise your findings to a wider population, however if your target sample is hard to reach it is perfectly ok to use convenience sampling. On the other hand, chi squares is just an example of statistical method to analyze your results... So perhaps talk to some other researchers or ask why it is not appropriate to use chi squares....
Best of luck
To the best of my knowledge, convenience sampling does not have any restrictions for the statistical tools to be used. As Kathleen and Witold mentioned, convenience sampling has its limitations of generalizability. However, often time we have to sample within some constraints that should justify (in that context) the use of convenience sampling. Yet, you may ask those researchers for the references if they have any regarding their suggestion of not using chi-squire or other stat tools for the data gathered from convenience smaple.
Convenience sampling: "A statistical method of drawing representative data by selecting people because of the ease of their volunteering or selecting units because of their availability or easy access. The advantages of this type of sampling are the availability and the quickness with which data can be gathered. The disadvantages are the risk that the sample might not represent the population as a whole, and it might be biased by volunteers. For example, a study to determine the average age and sex of gamblers at a casino that is conducted for three hours on a weekday afternoon might be overrepresented by elderly people who have retired and underrepresented by people of working age. Also called accidental sampling."
Read more: http://www.businessdictionary.com/definition/convenience-sampling.html#ixzz3iWH6unw9
It has nothing to do with what methods you compute.
Appreciate the answers of all researchers but i had attended a conference on Thesis writing - one of the research scholars mentioned i cannot use chi square because the assumption of chi-square is based on randomness the study has to be of random sampling
I have found the following articles which clearly mentions that chi-square follows the assumption of random selection of data but it is not uncommon to find use of chi -square test in non -probability sampling
I have shared this as it may be useful for all researchers like us
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3900058/
http://www.biochemia-medica.com/2013/23/143
I think you can apply statistical techniques for analysis, but result may not be generlized. convenience sampling is one type of technique for selection of units from the population, but those units may not represent whole population.
There are some methods not recommended for convenience sample such as regression analysis. If you use chi-square you tell how your convenience sample possible could differ from a strictly randomized sample where all on the globe had an equal chance to be drawn to your sample.
Beatrice Marianne Ewalds - Mam just check out the link i shared
2 days ago , it clearly mention chi square test is a nonparametric test and based on assumption of random sampling
\
Ashish Pundhir
There are many sites with many recommendations and many opinions on this issue
- Badges
- Science method
- Similar topics
- Mathematical Sciences
- Statistics
- Sampling Design