interesting question and remarkably many have suffered from gaining weight during this period

bulimic eating is a defense mechanism against stress

thanks for sharing

Before making the comparison, keep in mind the cohort effect, or period effect, or both. You want to use the Chi-square, meaning that there will be no adjustment for covariates. Adjust the result for all possible confounders and risk factors. This will increase the strength of your study.

interesting question and remarkably many have suffered from gaining weight during this period

bulimic eating is a defense mechanism against stress

thanks for sharing

Endorse the response of Md. Belal Hossain to adjust confounders. As chi-square will not address the effect of confounders, study findings are not realistic so employee some more reliable test. I suggest to use logistic regression with Chi-square to increase study strength.

To address main part of your question, as both the study population belongs to different cohorts, ideally it is not possible to combine some factors. As household income in 2019 will be different from 2020 due to pandemic it might act as a confounder for many other factors in the study which decreases the study strength.

Hermann Gruenwald Wassan Nori , researches on this question show opposite findings. The main thing is lack of physical activity with its consequencies on social interaction, growth etc. Since school cafeterias were closed, children are forced to find other food sources, interestingly my research suggest that they starved, probably due to masks and shorten school stay. Also, the consumation of sweetened juices decreased in comparison to previosu year.

Ceriantly, pandemic has get same benefits. home cooking is one of them. Also, people restore the tradition of family meals that strengthen family bonds providing examples to children and adolescents.

Regression and adjusting for confounders

About 10% of children responeded that they had not eaten during time spending in school. May they practiced some kind of intermitent fasting.

@ Ubaid-Ur-Rehman Zia , I seriously agree with your suggestion that logistic regression be used. This will enable him to obtain COR and AOR for different confounders. For each confounder or factor, a base category is identified for the regression analysis.

Actually I am confused. Your question was to 'combine prevalence' then you have mentioned 'before and after the pandemic'. Do you mean 'compare' the prevalence of some habits before and after the pandemic? If so, in this case you can not use the Chi square test. You have to use the McNemar test.