How will you maximize construct validity in an observational study to see who is more likely to hold open a door for another person, men or women?
Construct validity refers to the issue of how well your test actually measures what it is supposed to measure. In this case, your test only seems to be meant to measure who holds doors open for other people more often. It is not meant to look at whether men hold doors open for women or whether women hold doors open for other women. It is not meant to try to (as far as I can tell) determine when or why a person holds a door open for others.
What this means is that one aspect of construct validity should not be much of a problem. When you write down how many times men and women did or did not hold the door open for others, you are clearly measuring what you mean to measure. There is no serious worry that your test will measure something else. However, it will be hard to make sure that you are really taking the right steps to properly measure your variables.
I would set up my test by videotaping traffic going through a door. I would then look at the tape. This would be better than live observation because I could go back and look at the tape multiple times to be sure I was not making any mistakes. There are some issues you need to think about, though:
- You might want to record how often men and women, respectively, walk through the door alone. You do not want to end up saying, for example, that men do not open doors for others as often when what really happened is that men tended to be entering alone and did not have the chance to hold the door.
- You will need to think about how you treat groups. If three men come walking up to the door and one holds the door for the other two, will you say that one-third of the men held the door for others? Alternatively, will you look at groups and ask how often one person from a group holds a door open and how often no one from the group holds the door open.
- You will need to define what constitutes holding the door open for another person. Does the person have to open the door and let the other person pass ahead of them? Or does it count if the person passes through the door first, but reaches behind them to make sure the door stays open for the next person coming through?
- You will need to make sure that you are looking at a single door or at the first of a set of doors. Buildings often have two sets of doors in close proximity to one another, presumably to minimize loss of warm/cool air when the outermost doors open. You do not want to observe the second door because behavior at that door may be conditioned by behavior at the first door. In other words, if someone holds the first door open for me, I am likely to hold the second door open for them even if I would not otherwise have done so.
All of these are issues that you need to think about when setting up your study. You need to carefully define how you will measure this behavior so that you can be sure that you are measuring how often men and women hold doors open, not how often they travel in groups or any other variable.