Ensuring sample is representative and large enough to draw conclusions 

Protecting the identity of your sample by ensuring anonymity or confidentiality

Avoiding response bias such as acquiescence or responding in a socially desirable way 

Carefully selecting the type of questions you will ask: 

--Open-ended questions often provide full answers but can be hard to score 

--Closed-ended questions tend to provide less rich answers, but are easier to score 

Considering memory limitations and reducing cognitive load when possible 

Taking into account the order of questions and response options 

Determining the question content 

Quantitative research will end with a number – for example, heart rate in beats per minute, score on the Beck Depression Inventory, or number of steps a person took in a day. Quantitative research will differ in the amount of whatever your variable is measuring. Quantitative research tends to have variables measured on ordinal, interval or ratio scales of measurement. 

Qualitative research, on the other hand, will result in a holistic description. For example, people may be asked to describe how much exertion they felt while completing exercise, may be asked to describe how they feel, or may be asked to discuss something that happened in their life. Qualitative research is more likely to result in data on a nominal scale of measurement (that is, categorization) than quantitative research. 

Longitudinal designs and Cross-sectional designs are both interested in understanding how change progresses over time. 

Cross-sectional designs do this by essentially taking a snapshot of several different groups at the same time. For example, if you are interested in the development of number cognition in children, you might look at 2 year olds, 4 year olds, and 6 year olds today to get an idea of how number cognition changes over those 4 years. 

One problem with cross-sectional designs is the potential for a cohort effect – that is, it’s possible that children who are 6 years old today have had different educational or historical experiences than children who are 2 years old today will have. 

Longitudinal designs protect against cohort effects by following the same cohort over time. There are several different types of longitudinal designs, including trend studies, cohort studies, cohort sequential studies, and panel studies. Longitudinal studies, though they protect against cohort effects, are subject to attrition and history threats to validity.