Chi-squared tests and many other analyses are covered in my book, SPSS Made Easy: www.amazon.co.uk/SPSS-Made-Easy-Statistical-Researchers/dp/B0DJGR4Z5K
Thank you really for your explanation. Could you please show us how to report these results in a table and which values are needed to be included in the table. Best
Great video! I see however that there are only 4 tshirts in Scotland however. Is having at least 5 observations in each group an official assumption of chi2? I’ve been told I can’t do a chi2 when I have less than 5 in a cell and so tried to use a Fishers exact test instead (except my analyses is more than 2x2 so im not sure if Fishers exact test is appropriate). Sorry for ur stalking channel and spamming u with questions btw. Ur videos are just too good!
Yes, that's right, Fisher's exact test is often used as an alternative when there are fewer than five expected cases, and it's not unusual for it to be used with tables larger than 2 x 2.
Hi, I need to compare three groups with different number of samples. Independent variable is categorical from 1-6. Can I use percentage as input? Thank you. Best
Hi Naida, thanks for your question. Percentages wouldn’t usually be used in chi-square tests, unless you’ve converted them into an ordinal variable (e.g., high percentage vs low percentage).
@@naidababic7778 The Kruskal-Wallis test is suitable if you have (1) one independent variable with (2) three or more levels (e.g., groups, conditions) and (3) the data don’t meet the assumptions of a one-way between-participants ANOVA (e.g., the data aren’t normally distributed). It could be used with percentages.
Dear David, thank you very much for your answer, at the end I did Kruskal Wallis H, since it can meet all listed vriteria. It is my pleasure to watch your videos, they are very informative and helpful. Thank you for your time. Best
Thanks again for another great video, these are way easier to follow than others I've been watching. Trying to match these tests up with other scenarios to make sense of them, would this be an appropriate test if you were looking at say male/females buying behaviour e.g. whether men were more likely to buy a meal at subway and woman a single sandwich?
Hi Grace, thanks for the question. Yep, that's right, the chi-squared test would work in that case as you'd be looking at frequencies (i.e., how many males/females, how many meals/sandwiches).
@@DavidRobinsonPhD Thanks David, would that make the independent variable sex and the dependent the purchase behaviour? Could you have used a parametric test for either of the examples (yours/mine) or is non-parametric better in these cases? Sorry for the 20 questions, this is the closest I've got to understanding stats.
@@gracegibson5443 Hi Grace, yes, that's right. You can think of the dependent variable as the thing that varies as a result of the independent variable. Therefore, sex (a common variable in studies) won't ever be the dependent variable, as it's fixed (and can't, for example, be the result of sandwich choice!). Re. parametric tests, there's no parametric equivalent of the chi-squared test.
@@DavidRobinsonPhD Do you have any further videos on the chi-squared test? I'm looking to explore expected counts and observed counts to determine which variable levels may have the most impact on association. I read that cluster bar charts can help visualise what is going on.
Chi-squared tests and many other analyses are covered in my book, SPSS Made Easy:
www.amazon.co.uk/SPSS-Made-Easy-Statistical-Researchers/dp/B0DJGR4Z5K
Thank you really for your explanation. Could you please show us how to report these results in a table and which values are needed to be included in the table. Best
Awesome, thanks!
Thanks José, glad it was helpful!
@@DavidRobinsonPhD It would be great to have a video from you explaining in what cases to use what type of test. Are you on twitter?
@@joseluisalves5829 I'm not on twitter, unfortunately. Thanks for the suggestion - I'll add that to the list!
Thank you
Thank you so much man!
THANK YOU !
You're welcome!
How do you report the table you were able to come with in SPSS for an APA format research paper?
Great video! I see however that there are only 4 tshirts in Scotland however. Is having at least 5 observations in each group an official assumption of chi2? I’ve been told I can’t do a chi2 when I have less than 5 in a cell and so tried to use a Fishers exact test instead (except my analyses is more than 2x2 so im not sure if Fishers exact test is appropriate).
Sorry for ur stalking channel and spamming u with questions btw. Ur videos are just too good!
Yes, that's right, Fisher's exact test is often used as an alternative when there are fewer than five expected cases, and it's not unusual for it to be used with tables larger than 2 x 2.
Hi, I need to compare three groups with different number of samples. Independent variable is categorical from 1-6. Can I use percentage as input? Thank you. Best
Hi Naida, thanks for your question. Percentages wouldn’t usually be used in chi-square tests, unless you’ve converted them into an ordinal variable (e.g., high percentage vs low percentage).
@@DavidRobinsonPhD thank you very much for your answer. Would you recomand for precentage compare Kruskal Wallis H?
@@naidababic7778 The Kruskal-Wallis test is suitable if you have (1) one independent variable with (2) three or more levels (e.g., groups, conditions) and (3) the data don’t meet the assumptions of a one-way between-participants ANOVA (e.g., the data aren’t normally distributed). It could be used with percentages.
Dear David, thank you very much for your answer, at the end I did Kruskal Wallis H, since it can meet all listed vriteria. It is my pleasure to watch your videos, they are very informative and helpful. Thank you for your time. Best
Thanks again for another great video, these are way easier to follow than others I've been watching. Trying to match these tests up with other scenarios to make sense of them, would this be an appropriate test if you were looking at say male/females buying behaviour e.g. whether men were more likely to buy a meal at subway and woman a single sandwich?
Hi Grace, thanks for the question. Yep, that's right, the chi-squared test would work in that case as you'd be looking at frequencies (i.e., how many males/females, how many meals/sandwiches).
@@DavidRobinsonPhD Thanks David, would that make the independent variable sex and the dependent the purchase behaviour? Could you have used a parametric test for either of the examples (yours/mine) or is non-parametric better in these cases? Sorry for the 20 questions, this is the closest I've got to understanding stats.
@@gracegibson5443 Hi Grace, yes, that's right. You can think of the dependent variable as the thing that varies as a result of the independent variable. Therefore, sex (a common variable in studies) won't ever be the dependent variable, as it's fixed (and can't, for example, be the result of sandwich choice!). Re. parametric tests, there's no parametric equivalent of the chi-squared test.
@@DavidRobinsonPhD Do you have any further videos on the chi-squared test? I'm looking to explore expected counts and observed counts to determine which variable levels may have the most impact on association. I read that cluster bar charts can help visualise what is going on.
@@gracegibson5443 Hi Grace, thanks for your question. Unfortunately, this is the only video I have on chi-squared tests.